Facilitating technology adoption in the NHS: negotiating the organisational and policy context a qualitative study

Article history

The research reported in this issue of the journal was funded by the HS&DR programme or one of its proceeding programmes as project number 08/1820/254. The contractual start date was in October 2009. The final report began editorial review in March 2013 and was accepted for publication in September 2013. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HS&DR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

© Queen’s Printer and Controller of HMSO 2014. This work was produced by Llewellyn et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Types of clinical technology

Clinical technologies can be thought of as encompassing all of the methods used in order to address health issues, including drugs, devices, procedures, and organisational and support systems. 14 Technologies can thus be ‘low-tech’ as well as ‘high-tech’, and are not limited to equipment. Clinical technologies include both medical and non-medical technologies, such as those for nursing care. The focus of our research is on clinical devices, procedures and associated support systems. NTAC itself defines technologies as treatments, devices and diagnostic tools.

There is such a wide variety of medical devices that even a summary definition which attempts to be comprehensive is lengthy:

A medical device is any instrument, apparatus, implement, machine, appliance, implant, in vitro reagent or calibrator, software or material whose primary intended action is not achieved solely by pharmacological, immunological or metabolic means, and that is intended for human beings for:

• the diagnosis, prevention, monitoring, treatment or alleviation of disease

• the diagnosis, prevention, monitoring, treatment, alleviation of, or compensation for an injury

• the investigation, replacement, modification, or support of the anatomy or of a physiological process

• supporting or sustaining life

• controlling conception

• disinfecting medical devices

• providing information for medical or diagnostic purposes by means of an in vitro examination of specimens derived from the human body. 15

Medical devices include medical aids, such as wound-care products; artificial body parts, such as hip prostheses; and technical equipment, such as magnetic resonance imaging (MRI) scanners. 16 Simple devices such as thermometers, scales, latex gloves, wound dressings and beds are commonly used, and so clinical procedures can involve many devices. 17

Interventional procedures for diagnosis or treatment involve an incision, puncture, entry into a body cavity or the use of electromagnetic radiation. 18 ‘New’ procedures span a spectrum of innovation from minor adaptations of existing practice, or applying an existing procedure in a new area, through to major innovations that are genuinely novel. 19

Adoption decisions regarding interventional procedures and medical devices are more complex than for pharmaceuticals for a variety of reasons:19 outcomes often depend on operator skill, with a learning curve to be negotiated; health technology assessment (HTA) processes have only recently been established; additional physical infrastructure is often required; and good quality data on cost-effectiveness is often not available. In addition, medical devices have a relatively short product life cycle (2–10 years) and prices are prone to vary over time, whereas interventional procedures are generally delivered to a heterogeneous patient population.

Innovative health-care technologies can include new models of care and ways of organising services and staff, such as nurse-led care, integrated transmural care across the primary–secondary care interface, collaborative or shared care, hospital safety procedures, clinical decision support systems, clinical guidelines, and staff communication and information sharing systems. 20 Changes to organisational systems are often needed when new devices or procedures are put in place.

Five major types of health-care-related technology have been identified, based on the purpose of the technology:21

• Diagnostic technology enables treatment or palliative care to take place, by identifying diseases and other conditions.

• Therapeutic technology is used in treating diseases.

• Enhancing technology improves human functioning over and above what is needed to cure diseases.

• Enabling technology, also known as assistive technology, mitigates the impact of disease or disability. This includes both personalised equipment such as artificial limbs or spectacles, and universal technologies that address environmental or societal issues, such as wheelchair-accessible entrances.

• Preventative technology reduces the risk or severity of accidents and other social and environmental sources of disease or injury. This encompasses a wide spectrum of technologies, from hip protectors to airbags to sewage treatment plants.

Diagnostic, therapeutic and enhancing technologies, together with some enabling technologies, such as prostheses, are an integral part of health care. Other enabling technologies, and most preventative technologies, are not closely connected with health care, but may be based on medical knowledge.

One of the key characteristics of new health technologies is the use of information and communication technologies (ICT). 22 This includes electronic health records, digital picture archiving and communication systems (PACS) and pharmaceutical prescribing/dispensing systems;20 and telemedicine, telehealth and telecare. 23 Telemedicine can be defined as electronically mediated doctor–patient interaction for the purpose of diagnosis or planning case management. Telehealth and telecare are the two main types of systems of remote care. 24 Telehealth is remote surveillance of the health status of patients by clinicians through the collection of data about symptoms or physiological parameters. Telecare monitors lifestyle changes, including potential emergencies, to enable people with social care needs to live independently.

The three new technologies investigated in this study are examples of diagnostic and therapeutic medical devices, all of which need staff to adopt new roles and change the relationship between patients and staff. The therapeutic device is operated by patients themselves in everyday settings, following receipt of suitable training. One of the diagnostic devices is part of a new clinical procedure performed in hospitals, and requires systems to be reorganised. The appropriate setting for the other diagnostic device is disputed, it has been used in various settings: specialist hospitals; general hospitals; and in the community. This last technology uses ICT to produce digital images that are easily transportable across settings.

The technology adoption process

The process that takes place when an organisation adopts a technology that is new to it (although it may not be new to all organisations) can be regarded as comprising three main stages: initiation, the adoption decision (which may be to reject use of the technology) and implementation. 25 Williams and Dickinson26 regard the process as a cycle, because initiation, or ‘setting the adoption agenda’, has as one of its inputs consideration of the existing technology in use within the organisation. They also distinguish implementation of the decision to adopt or procure the technology from use of the technology becoming routine. When the term ‘adoption’ is used in this report, it refers to the whole of this process, not just to the decision about whether or not to procure.

The adoption process may be influenced by characteristics of the technology itself, such as complexity, scale and cost, which are related to aspects such as the degree to which it may be experimented with on a limited basis (trialability); the degree to which it is perceived as being consistent with existing values, past experiences and needs of the potential adopters (compatibility); and the degree to which the products of the technology are visible to others (observability). 27,28 Characteristics of the organisational context, such as absorptive capacity (an organisation’s ability to acquire, assimilate, transform and utilise externally held knowledge),29 may also be important.

Studies of technology adoption typically take the perspective of an individual organisation. There is also a substantial body of research on the diffusion of innovations, which may be relevant. Diffusion of innovation studies take a macro perspective, looking at the incorporation of an innovation over time into common practice across a whole system or sector of organisations. 30

See Chapter 3 for a discussion of underlying theories and conceptual models of the technology adoption process.

Organisational and policy landscape for technology innovation and adoption in the NHS

Difficulties in evaluating new technologies

Decisions about whether or not to offer technologies routinely to patients require more information than is provided by the assessments of short-term safety and efficacy for small groups of patients that are required by EU regulations, or even by the more extensive processes undertaken when a technology is assessed by NICE. 44 There is typically uncertainty about associated costs, risks and benefits, and to which population groups they may apply. 45 Furthermore, such estimates can change over time, as technologies may improve in quality, become cheaper, be superseded or be used with a wider range of patients. For example, most of today’s coronary artery bypass graft surgery patients would not have been eligible to take part in the clinical trials that established its effectiveness. 46

Evaluation of radically different technologies is complex because they may be used with a very heterogeneous set of patients, outcomes may depend on practitioner skill levels, significant changes in practice or service reorganisation may be required, and devices with similar functions produced by different manufacturers may not be equally effective (see Tarricone and Drummond,47 for example). If the technology is expensive, then this means that carrying out a clinical study is risky in terms of the commitment of financial resources.

Where there is a gap in the scientific evidence, clinicians are reliant on expert consensus informed by clinical experience, which may be misleading; in some circumstances product awareness raising by manufacturers or innovators in the form of education, training or awards may then drive uptake. 48 Promoters of telehealth and telecare have used a strategy of downplaying risk and the need for evidence by stressing that the underlying services are essentially the same, with ICT acting only as a facilitator. 23

‘Hard’, quantitative evidence on outcomes, particularly from scientific research but also from audits and other data gathering, appears to be more influential than experiential knowledge, which is typically dismissed as anecdotal, with no recognition of the potential usefulness of rigorous qualitative research,49 again, potentially biasing decisions. New technologies that represent moderate changes are most likely to be problematic, as minor changes should be acceptable by definition, and NICE gives greater scrutiny to major changes. 19

Once a new technology has been adopted, there is no national system for ensuring that its effectiveness in everyday practice is evaluated. Lourenco et al. 19 suggest a framework for staged evaluation of new interventional procedures, which would include mandatory post-market data collection to check clinical effectiveness and cost-effectiveness when the technology is implemented.

Variable local decision-making processes

Even when there is strong evidence and mandatory NICE guidance has been produced (NICE guidance on interventional procedures is not mandatory), this does not ensure adoption, as the process is left up to individual trusts, who may not have put effective arrangements in place to manage implementation. 50 Furthermore, managers may selectively shape ‘evidence’ to align it with their own goals and preferred practices. 51

Some NHS organisations have been found to have more structured processes than others for making decisions about the adoption and implementation of new interventional procedures. 52 Hospital procurement and evaluation committees often play a part in decision-making, with doctors having a key role in defining desirable characteristics of the technology and accompanying support, and in reviewing evidence on costs and benefits,53 although their involvement is often ad hoc. There is generally a lack of organisational capacity and resources with regard to the purchasing of technologies, with little co-ordination of purchasing among organisations locally, regionally and nationally. 53

The nature of stakeholder involvement at different stages of the process can affect the decisions made and the outcomes. A study on the adoption of new technologies for infection prevention and control (IPC) by NHS organisations54 produced a number of pertinent findings. Stakeholder involvement at initiation impacted on which technologies and IPC areas were considered. Those involved in the adoption decision influenced how the technologies were critiqued and what was finally selected. In the organisations where wider consultation occurred early, more diverse approaches to IPC were considered. Support by senior management at the point of decision-making facilitated implementation by mobilising resources and providing increased legitimacy to the initiatives. Involving technology users such as front-line clinical staff from the start of the decision-making process increased their commitment and provided feedback to suppliers, which could be used to help ensure that the devices procured were compatible with working practices and organisational policies. Lack of wider stakeholder engagement in implementation planning was observed to have a negative impact on implementation. Late involvement of the procurement team, due to inexperience or negative perceptions of staff, extended the process.

Lettieri and Masella55 suggest that when a hospital makes a decision about technology adoption it should consider the expected contribution to value generation and the level of sustainability. Considerations related to value generation include effectiveness, patient or family satisfaction, revenue generation, cost containment and gains in image or reputation, and, in the longer term, creating knowledge by developing new services and health-care technologies and building up new communities of knowledge. 55 Considerations related to sustainability include the degree of self-funding and ratio of fixed to variable costs (economic sustainability); coherence to strategic goals, technology acceptance among physicians and uncertainty in clinical practice (organisational sustainability); technology life cycle and fit with the existing technology portfolio (technological sustainability); and training intensity and coherence of human and physical resources (resource sustainability). 55 Context sustainability issues of coherence to the legal framework and to generally accepted ethics are unlikely to be relevant where a technology has already been approved by national agencies, but will be crucial if the hospital is involved in the development and testing of an emerging technology. 55

Ong et al. 45 suggest that three types of decision-making models are used in NHS hospitals with regard to investing in new technologies: maximising profit is generally favoured by finance directors and managers; maximising competitive advantage is generally favoured by chief executives, marketing directors, research/teaching hospitals and private hospitals; and maximising utility (health outcomes) is generally favoured by clinicians and patients. 45

The types of information that have been considered in business cases prepared in support of new technologies have included efficacy, alternative treatments, training, cost, potential savings, duration of procedure, safety, benefits, numbers affected, length of stay, pre-operative assessment and cost-effectiveness. 52

It is widely believed that business cases are only likely to succeed if they show evidence of efficiency or cost savings; demonstrating quality improvements alone will not be sufficient. 45 The criteria for medical device procurement are discussed by Sorenson and Kanavos. 53 Price is the most important criterion because of cost pressures, especially for ‘standardisable’ products. Quality criteria do, however, tend to be more relevant in specialised or complex surgical areas, where weight is given to the quality of the service as a whole. Other factors often considered include reliability, production capacity/volume, delivery date, maintenance requirements, and innovative characteristics or technical merit. Therapeutic benefit and cost-effectiveness are sometimes, but less often, considered in purchasing decisions. The emphasis on price may lead to manufacturers focusing on reducing production costs rather than investing in research and development, reducing access to innovative technologies in the long term.

Three main external influences on adoption decisions have been identified in innovators’ accounts: (a) economic, such as a focus on cost containment; (b) political, such as the existence of national regulators; and (c) ideological, such as fitting with the ‘spirit of the times’. 49

More formal involvement of physicians in procurement generally might help health considerations to be more prominent in adoption decisions. 53 Centrally planned implementation could also be more efficient and less costly. 45

Misaligned financial incentives

Given the importance of costs in adoption decisions (see Variable local decision-making processes), there is a need for supportive financial climates. 56 PbR may not be well suited to supporting technology adoption, however. Introducing new technologies always has an initial downward impact on productivity. 57 Additional activities, such as training and clinical pathway redesign, may need to be undertaken, leading to reduced patient throughput, while at the same time raising providers’ costs. Yet, under PbR, funding drops whenever activity drops, and there is no allowance for the additional initial costs.

In theory, organisations can address this through the local flexibilities allowed by PbR, such as by linking funding to the actual costs incurred by a particular trust, rather than to the national average cost. There are, however, no centrally published data on the extent to which local PbR flexibilities are being implemented, and the limited evidence that is available suggests that progress is painfully slow. Setting a local price relies on communication and dialogue between providers and commissioners, which may be problematic given the complex and varied factors to be considered, the variability in organisational processes and the conflicts of interest (a higher price suits providers, a lower price suits commissioners) (see Variable local decision-making processes). Therefore, any expectations that PbR can operate in its current form without adversely affecting the adoption of new technology look overly optimistic.

Payment by Results may also impede innovation, as there is a trade-off between innovating and productivity. 58 Innovating involves exploration, which raises risk,59 and a policy climate that relies on financial incentives for activity tends to reduce providers’ propensity to take risks. 60

Diagnostic-related group (DRG) systems organise patients into groups (or categories) on the basis of diagnosis. They began in the USA and are used to reimburse hospitals for the care they provide. In the UK, HRGs are similar to DRGs but they classify patients according to intervention rather than diagnosis. There is also the likelihood that DRG- or HRG-based funding systems may incentivise purchase of technology that yields short-term cost savings, especially if tariffs are not updated sufficiently frequently. 53

Barriers within and between sectors, professional groups and organisations

Traditionally, UK health services have operated within secondary, tertiary and primary care sectors, with little interaction between primary care and the others. 61 Lack of trust in the generalised expertise of general practitioners (GPs) has been a barrier to diffusing technology from the hospital sector into primary care. 62 Specialist clinical expertise is the main ‘pull’ factor, but enthusiasm tends to be for ‘high end’ technologies rather than the mass ‘bargain basement’ technologies required for technology transfer into primary care and to support health status monitoring in the community. 63 If technology relies on patient involvement, there can be additional barriers concerning acceptability, particularly in situations where devices are transmitting data to clinicians. Yet policy expectations are that new diagnostic tools and technological treatments will become available outside of hospital settings, even entering the home as ‘near-patient’ technologies. 22

Uniprofessional communities of practice, particularly among clinicians, can provide strong social and cognitive barriers to the spread of innovations:64

. . . strong boundaries between professional groups at the micro level of practice slow innovation spread [or flow] . . . indeed ‘flow’ is a radically inappropriate image to describe what are erratic, circular, or abrupt processes, which may come to a full stop or go into reverse.

Even within hospitals, structural complexity and professional rivalries militate against diffusing technology. 65 More than 40 specialties are recognised with little co-ordination between them; indeed, the use of technology has increased rather than reduced medical specialisation. 66 Structural constraints and cultural diversity between specialities67 foster a situation where technologies do not travel.

Organisational barriers can also limit the appropriate use of technologies; and make supply and demand issues more difficult to negotiate. Fit with the organisational ethos is important, and employees with experience of the technology, or in decision-making positions, can be either a key channel for diffusion or a barrier. 49 Clinicians become risk averse and resistant if they suspect new technology will negatively affect their work with patients or if implementation is difficult,68 and this is likely to be the case with innovative new technologies, as they will alter the patient pathway and require changes to the roles of professionals involved. Furthermore, the clinical training required to master new technologies raises organisational costs.

Government policy is for technology adoption to be demand led through world-class commissioning,1 but this is impeded by power imbalances and information asymmetry between commissioners and providers. 69 Progress with developing the NHS commissioning function has been slow, with weaknesses such as passivity still remaining after 20 years, due to shortcomings such as lack of skills and lack of clinical knowledge. 70 The upcoming reorganisation of the NHS, with SHAs and primary care trusts (PCTs) to be abolished and replaced by CCGs and a national commissioning board, is unlikely to help in the short term, and will create new sets of organisational boundaries to be negotiated.

The NHS procurement system has been found to be ineffective, complex and confusing. 71 Centralising medical device procurement is perceived to produce greater cost savings, efficiency and purchasing specialisation, but it may also inhibit competition and emphasise cost rather than quality or value considerations. 53 Several of the procurement hubs have closed, merged or been privatised, while individual trusts continue to have low-volume direct contracts with suppliers, and there may be issues within trusts of ‘maverick spending’ by individual employees who ignore their organisation’s procurement team. 71

In view of the above barriers, and of the need for social interaction in order to develop tacit knowledge which may be needed to master a new technology, networks and boundary spanning roles are likely to be important in supporting the uptake of new technologies. Champions may be effective in spreading a new technology in their immediate locale, but not necessarily more widely. 72 Change agents have been recommended as a way of overcoming professional boundaries,56 and clinical networks have been suggested as a means of facilitating uptake across both primary and secondary care. 45

Conceptual models and theories relevant to technology adoption in the NHS

Chapter 2 highlights the importance to technology adoption in the NHS of financial incentives; the use and generation of evidence during adoption and implementation processes; and boundary spanning across multiple networks and stakeholders. Linear stage models have, however, been the foundation of much research on technology adoption and diffusion. Such models tend to be techno-centric, with the capabilities of the technology being the main enabler of change73 and determining the social and organisational impacts, such as on governance structures, work routines, productivity and performance. 74

One strand of the empirical research on technology adoption has focused on information technology (IT). Theories and models that have commonly been used in this research are the Diffusion of Innovation theory, the Technology Acceptance Model, the Theory of Reasoned Action, the Theory of Planned Behaviour and the Technology Organisation and Environment Model. 75 These have been integrated together to produce a conceptual model for the IT innovation adoption process in an organisation. 75 This model is pertinent to our research study because it incorporates the three main stages of technology adoption (initiation, the adoption decision and implementation) which we have used as the basis for our definition of adoption (see Chapter 1, The technology adoption process). Initiation is subdivided into awareness of the innovation, forming an attitude of adoption, and developing a proposal for adoption; the adoption decision needs to be backed up by resource allocation for implementation; and implementation requires acquisition of the innovation and user acceptance of it, for actual use to occur. The model also highlights the importance not only of the characteristics of the technology, but also of the organisational context and of technology users.

There is an increasing recognition that the adoption of innovation in health care is a complex process, affected by a variety of factors, but understanding of the interactions is limited. 76 The Normalization Process Model engages with this complexity, focusing on the social relations and processes by which innovations are made workable. 77 A conceptual model of factors affecting the diffusion of innovations in health service delivery and organisation has been developed. 78 This model further elaborates a number of the concepts identified in the model of the previous paragraph, and also maps onto many of the issues that we have highlighted (see Chapter 2). Difficulties in evaluating new technologies (see Chapter 2, Difficulties in evaluation new technologies) relate to absorptive capacity, which is a key antecedent of user system readiness for innovation. Variable local decision-making processes (see Chapter 2, Variable local decision-making processes) correspond to aspects of having a receptive context for change and user system readiness for innovation. Sectoral, organisational and professional barriers (see Chapter 2, Barriers within and between sectors, professional groups and organisations) can be addressed by appropriate linkage between users/adopters and innovators, and by suitable communication and influence strategies, of which champions, boundary spanners and change agents are relevant to formal, planned dissemination. The issue of misaligned financial incentives (see Chapter 2, Misaligned financial incentives) is less prominent in the model, but is mentioned, and a risk-taking climate is identified as part of a receptive context for change.

These models partially address the criticisms that have been made of viewing technological innovation as a linear, rational process,79 but we also need to take account of alternative approaches that have been developed. These alternative approaches emphasise the profound uncertainties surrounding technological innovation, highlighting the complexity, political context, broader social network and the social ramifications of the adoption process (i.e. non-structural determinants such as the micro-politics of the organisational setting, interests, prevalent rhetorics, fads and sociocultural context). 80–83 Our study draws on actor-network theory (ANT) (see Actor-network theory),84,85 boundary work, risk communication and the adaptation of work practices to new technologies (see Boundary work, risk and the adaption of work practices to new technologies) because these theories and concepts have been found most efficacious in a large body of previous work on technology adoption (as indicated below).

Methodology

The research questions (see Chapter 1) investigate issues that are underexplored and, often, context dependent. Therefore we relied, in the main, on a qualitative methodology, although for one of the technologies this was supplemented by a survey which we analysed quantitatively. In effect the clinical technologies were case studies. Our core interests were focused on how intraorganisational and policy implementation barriers were negotiated (or not) but we found that other aspects impinged in significant ways. For example, the extent to which the clinical efficacy and utility of the technology was accepted as proven, the nature of the interface between primary, secondary and tertiary care, and relationships with technology producers. Hence, we looked at the technologies holistically, so the case study methodology was highly relevant. 143,144

The case studies focused on three clinical technologies: insulin pump therapy (IPT), breast lymph node assay (BLNA) and ultrawide field retinal imaging (UFRI). This focused approach was agreed in consultation with NTAC. We decided that, within the time constraint of 3 years, intensive research would be more productive than a more superficial approach over all the technologies that NTAC supported (around 15 at the time the study commenced). The three technologies identified in discussions with NTAC were those which presented the most complex and puzzling problems with regard to adoption and implementation. NTAC themselves had been unable to understand the reasons why these three technologies were underadopted despite their intensive project work. Current government policy for improving technology adoption, as expressed in Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS,1 takes a selective approach by supporting only six technologies. Similarly, in focusing on three technologies, we undertook an intensive research design. We describe the adoption issues for the three technologies in detail next.

Introduction

NHS Technology Adoption Centre was established in 2007 to directly address the problem of underutilisation of clinical technologies in the NHS. At the time of this research, NTAC was funded by the Department of Health, NHS Institute and NHS South West. NTAC is a small organisation, with, as of February 2013, personnel consisting of a chief executive, three technology implementation managers (who run the projects at the trusts), a data manager, a performance and operations manager and an administrator. 155 This modest size means that NTAC cannot support all the trusts that apply for assistance with technology adoption.

The NTAC approach is to work at selected trusts with, for the technology concerned, all relevant stakeholder groups to resolve adoption barriers, understand adoption enablers and embed the technology in day-to-day working practices. This modus operandi is clearly understood by NTAC as an implementation project rather than a pilot or trial. The outcome of NTAC’s learning at selected sites (usually three or four for any particular technology) is distilled into an online HTWT guide. These guides are intended to give trusts that were not selected as implementation sites, along with others who may wish to implement the technology concerned in the future, sufficient information to enable them to undertake an implementation project without NTAC’s active involvement.

The NTAC process starts by identifying suitable technologies through putting out a call to industry. The criteria for selection are indications that the technology will result in a step-change in health-care provision; the technology should already have some, albeit suboptimal, uptake in the NHS; the technology should be in an area of national policy priority; the technology should be in a clinical area of major focus; and, preferably, the technology should have undergone some form of appraisal. Overall, an NTAC implementation project is focused on an underadopted technology that, nevertheless, has a substantial evidence base for clinical efficacy and utility, including system-wide benefits.

For all clinical technologies, with respect to ‘substantial evidence base’, NTAC’s remit is somewhat problematic. Under European directives, clinical technologies are classified as ‘devices’ (see also Chapter 1, Types of clinical technology). In the UK, EU directives are implemented by the MHRA. Under this regulatory regime, manufacturers are required only to demonstrate that the technology is safe and fit for purpose; unlike pharmaceuticals, they do not have to undertake clinical trials to demonstrate clinical efficacy (see Cohen and Billingsley156 for a comparison between the EU requirements for devices and the more stringent US regulations). Although, compared with the USA, the somewhat looser EU requirements can be criticised, the knowledge and skills to benefit fully from new technologies are often acquired only after adoption into clinical practice. 19 On the other hand, it should be noted that informal evaluation through practice does not always work well. Some technologies enter the market and become firmly embedded in routine practice, but later are found to be ineffective or harmful. 19 Thus, the clinical efficacy and utility of many technologies becomes apparent only over time, as the learning curve plays out. 157

NHS Technology Adoption Centre’s approach of working with trusts on implementation does fit with the profile of clinical technologies. As the efficacy and utility of many technologies only becomes fully apparent over time, an implementation project which takes place over an extended period of time and engages all stakeholders to maximise learning should give the best chance of revealing clinical efficacy and utility. After an NTAC technology implementation manager has completed a due diligence report to ensure that the technology meets the criteria for selection (outlined above), NTAC puts out a call to trusts to invite their interest in being an implementation site for the technology concerned. The criteria for site selection are the overall thoughtfulness and credibility of the application in terms of patient benefit; capacity of the trust to implement the technology; and perception of adoption and implementation barriers. The preferred NTAC model is to work with three trusts out of those who apply, each with differing implementation barriers in order to maximise the on-site implementation learning that feeds through to the HTWT guides.

Theoretical framework

We argue that technology implementation can be understood through ANT. The clinical efficacy (or not) of any technology is not the only determinant of successful adoption and implementation. A key aspect is whether or not the technology becomes embedded in a network and whether or not that network is powerful. An actor network does not have any a priori form, it is not confined to sectors, organisations or even projects, it traces associations as new events and interventions occur. 158 A network, according to ANT, does not only associate human agents;159 non-human agents (actants) are also mobilised and assembled. 88 In ANT terms, a clinical technology is an actant. Such actants can acquire agency in one of two ways: either actions that were carried out by humans are delegated to them or the actant displaces a human (or humans) in the network. 160 Hence, a basic tenet of ANT is that social phenomena cannot be analysed through human intentions and actions alone; non-human entities also have agency and, hence, participate in the networks that determine social outcomes. 161 These non-human actants have material form; they may be technologies but they may also be concrete practices. 162

Our analysis of NTAC, through ANT, begins by arguing that NTAC’s starting point for its implementation projects was somewhat underspecified (i.e. implementation projects were defined more by what they were not than what they were). NTAC’s first chief executive recognised that if an implementation project began as a ‘pilot study’, the agency attached to the practice of ‘pilot study’ would seriously diminish the possibility of successful implementation. So the starting point was that an implementation project was not a pilot.

The remainder of this chapter addresses the following issues: NTAC’s approach to the implementation of clinical technologies; their views on implementation barriers and enablers; and their intentions and experiences with the HTWT guides. We then cover the answers to our survey questions on the HTWT guides and participants’ responses to the video method of capturing their online guide experience and conclude with a short summary of the main issues raised in this chapter. (The interview evidence on the HTWT IPT and BLNA guides is presented in Chapters 6 and 7.) In the sections below we include data from the first NTAC chief executive, the second NTAC chief executive, the project implementation manager for the IPT and BLNA cases (the same person undertook both) and, briefly, the implementation manager for the UFRI case (this was terminated before we started the research, as described under Methodology, Chapters 3).

What is the NHS Technology Adoption Centre approach to the implementation of clinical technologies?

In ANT terms, the pilot study – as a practice – is an actant. The agency of a ‘pilot study’ in negating ‘total buy-in’ and, thus, rendering technology implementation optional in a network was understood by NTAC’s first chief executive. She was determined that the trusts accepted that their project was truly about implementation rather than a pilot which could be terminated.

My aversion to pilots was well known and very well documented; having been involved in them. If we genuinely wanted to understand the barriers we would never do that without getting total buy in from the trust we were going to work with.

First NTAC chief executive

The end point of NTAC’s implementation was also clear. It was agreed from the beginning that learning from each of the projects would be embedded in a HTWT guide for the benefit of future trusts who wished to adopt and implement the particular technology. Enrolment is a key stage in the ANT sociology of translation. 89,163 The agency of the HTWT guide in enrolling and, therefore, enabling technology implementation was assumed by NTAC.

The idea behind the guide was to showcase the success of the sites we’d worked with and that would enable other trusts . . . Here’s your business case. Here’s your costing model. Here’s your evidence. It [the HTWT guide] brought everything together, basically.

Implementation manager: IPT, BLNA

The NTAC approach was absolutely clear in one other respect – it had a ‘bottom-up’ agenda. It worked at the trusts on implementation. It was not a national organisation that issued dictates from above.

I could write how I think we should overcome the barrier of tariff, ok? And I could be absolutely correct . . . But if I can’t back it up with an example, then people will say, ‘This is an academic exercise.’. . . Some national organisation, they don’t live in the real world. That was the whole idea of working with the trusts.

Implementation manager: IPT, BLNA

Between the start (not a pilot study) and end points (the production of a HTWT guide) there was no structured process; no set ‘model’ for clinical technology implementation at the trusts. In part, this stemmed from a lack of central direction from the Department of Health.

I think, initially, I inherited a business plan which had been developed by a consultant to look at NTAC – but it was extremely outdated. And quite honestly I don’t feel at that point the Department of Health was clear itself on what it saw NTAC doing.

First NTAC chief executive

The NTAC approach laid considerable responsibility on the project implementation managers, who pursued their own operational plans within the overall remit of ‘implementation’. The first chief executive compared this to an early-stage company.

I was only one step ahead of the project manager. That’s what happens in an early stage company. And we were an early stage company to all intents and purposes. I was reasonably comfortable with that scenario.

First NTAC chief executive

In consequence, the NTAC project managers had considerable autonomy in setting the implementation agenda.

There was no model. There was no laid out plan of how to do it . . . She [the Chief executive] wanted people who could come in, who were experienced, who had enough seniority to push this change forward. To be in that environment, you needed to have some initiative of your own . . . [the Chief executive] said to me, ‘I don’t care how you do it . . . Here’s your budget. Now go away’ . . . I had a blank piece of paper.

Implementation manager: IPT, BLNA

In ANT terms, however, the ‘blank piece of paper’ perspective is somewhat naive as human agency always depends on the power of the network within which it is embedded; an actor is an effect generated by many interacting material and non-material entities. 164 A hospital trust is a world that is ‘pre-figured’. 165 NTAC staff will have to engage with and transform pre-existing networks around, for example, the patient pathway to ensure technology implementation. The ‘blank piece of paper’ is, actually, already inscribed. This inscription does not preclude a technology from transforming the pre-existing networks. Indeed, NTAC staff showed awareness of this. The actors, material artefacts and practices that make up any network are continuously being reconfigured through new events and interventions. 164,166 Indeed, clinicians and managers are not always aware of current practices.

Let me tell you something. Nine times out of ten, people do not know how patients are moving through the system now, so never mind with a piece of technology that completely changes it.

Implementation manager: IPT, BLNA

Implementation projects enabled ‘stakeholder engagement’, which, although NTAC staff did not use ANT language, proceeded through Callon’s ‘four stages of translation’. 89 The first stage is ‘problematization’, where a problem is defined as shared across disparate sets of actors but one set of actors privileges itself in stating how the problem can be solved. In the case of NTAC, the problem is how a technology can be implemented through stakeholder engagement. In the quote below, an NTAC project manager positions himself as able to begin to solve the problem as he, as opposed to the clinicians, is aware of all the relevant stakeholders:

The first thing to do is identify the team. And it’s not just the clinical personnel. I mean, clinicians often think that the whole world revolves just around their clinical colleagues, but what about the finance manager? What about the commissioner? What about the procurement? Clinicians say ‘What’s procurement?’ You say, ‘Your supply department.’ Then they say, ‘Who are they?’. Well actually, they’re the ones that are going to get this technology in, so you’d better find out who they are.

Implementation manager: IPT, BLNA

Along with stakeholder engagement, the NTAC staff declared that they possessed the requisite project management skills. Callon’s second stage is ‘interessement’,89 where privileged actors propose roles for other sets of actors and gain their commitment to certain courses of action. At this stage, NTAC staff state that clinicians are unable to ‘work through project plans’, thereby appropriating for themselves the roles of project managers. The project implementation managers also proposed certain goals such as awaydays and a ‘tight time frame’.

They’re clinicians by day, they’re clinicians by night too, they don’t have the luxury of being able to work through project plans . . . It’s up to us to do that . . . And within that, I would say we need to incorporate things like the awaydays, but plan for it at the beginning . . . [And] there needs to be a tight timeframe.

Implementation manager: IPT, BLNA

Callon’s third stage is ‘enrolment’,89 where the proposed roles and courses of action are consolidated. In the next quote the NTAC project manager consolidates the local network team through connecting them to ‘subgroups’ (or other networks). He then assigns further specific roles and responsibilities to these subgroups while positioning the role of the local clinical champion, with whom NTAC communicates, as ‘overseeing’ all the subgroups.

It’s not just about having an overarching team locally, but subgroups as well to give ownership of certain elements . . . Ok, the nursing staff, they’re the ones who’ll see these patients, ultimately. Yes, it’s a pathology test, but you’re the ones that follow them through. Go and design the pathway. Bring it back in six weeks and we can talk about it together. You go and speak to the accountant . . . you need to go and draw up the specification for procurement. Let’s have a procurement subgroup. And that’s what the clinical champion needs to oversee, all of those subgroups.

Implementation manager: IPT, BLNA

The fourth stage is the ‘mobilization of allies’, where further actors are drawn into the networks which can communicate with each other through the ‘immutable mobiles’ (e.g. procurement specifications) established at the enrolment stage. 89 After enrolment, through clinicians and managers, NTAC has produced new patient ‘pathways’, ‘specifications for procurement’ and ‘business cases’. These are the immutable mobiles that enable communication at a distance and act as ‘obligatory points of passage’ for the separate, but now interconnected, networks in pathology, physiotherapy, psychotherapy, PCT commissioning and patient representation.

You need your clinicians from all the areas the technology impacts on. So, for example, even though it’s a technology that sits in pathology, it impacts on theatres, on the wards, on nuclear medicine, on pathology, the nursing team, physiotherapy, psychotherapy. All those departments need to be included and, equally, the decision-makers, the people who will sign off the business cases, the trust accountant, the PCT commissioner, and also, where possible, a patient representative.

Implementation manager: IPT, BLNA

Overall, NTAC respondents stressed the implementation gaps that NTAC fills in its ‘downstream’ and ‘practical’ approach to innovation in the NHS.

NTAC is the only organisation in the NHS looking at the downstream delivery of technology. There’s so much about introducing technologies – you know, having a relationship between innovation and research and development. But NTAC is really looking downstream where something isn’t working; why a particular technology is not being taken up by the NHS.

Implementation manager: UFRI

I do think there is an ethos of us being a practical organisation rather than a theoretical one.

Second NTAC chief executive

Where a technology is not ‘being taken up’ into a network, NTAC works in a practical way ‘on the ground’ through the translation stages outlined above to try to ensure that networks are created and sustained around the technologies concerned.

The next section explores NTAC’s views, based on its experience, on what the generic implementation barriers are in the trusts and how these can be overcome.

NHS Technology Adoption Centre’s views on implementation barriers and enablers

The first chief executive revealed how her thinking on barriers had shifted, making reference to ‘pathways’, ‘clinical utility’ and ‘systems impact’. Technology adoption and implementation always take place within the context of pre-existing networks. The concepts of ‘clinical utility’ and ‘systems impact’ capture aspects of this issue. There will be a pathway along which patients move, and the pre-existing networks of clinicians and practices will have shaped how this patient pathway has developed. These networks cannot be easily recast to enable technology to enter, so much has to be done to make the trusts become comfortable with the new ways of working, which involves creating new networks.

My view was it must be the NHS’s fault. It’s difficult, it’s reactionary, it isn’t interested in technology. But as we started to talk to clinicians and managers we realised what issues there are in bringing devices and diagnostics through the system as compared to drugs . . . we started to think about the pathways, clinical utility, all the really serious questions that needed answering to make a trust comfortable that they should bring a new technology in.

First NTAC chief executive

We argue above that four stages of translation need to occur for a technology to be implemented into networks. At the first stage, NTAC has to convince the clinicians of the technology’s clinical utility. This means more than scientifically showing that the technology is clinically efficacious; it also implies that it has clinical utility (i.e. the technology can work in the context of clinicians’ perceptions about how the care of patients should take place). 167 Clinical experience has considerable weight in judgements on clinical utility. 168 Hence, translation may not occur, may be only partly achieved or may be delayed through successive implementation cycles. Therefore, management is crucial in enabling change.

So it’s about making the change possible, changing the pathway, changing their [clinicians’] mindsets and – and this is a word that we use a lot now – enabling.

Implementation manager: IPT, BLNA

The first NTAC chief executive emphasised how the technology can act on systems to ‘change the whole dynamic’. In doing so she expresses the central ANT insight – that any(thing) that changes the dynamics of practice is an actor. 88

The technology can suddenly change the whole dynamic of the procedure.

First NTAC chief executive

The second chief executive continued in this vein, through pointing out the necessity for new ways of working and developing new networks around the technology. Law164 uses the concept of ‘punctionalization’ to express the idea that the networks which embed technologies are often taken for granted. Clinicians become aware of the networks that generate effects (such as clinical pathways, training and the tariff) only when they need to be changed to incorporate a new technology.

You’re helping organisations to re-design their clinical pathways, training implications and tariff consequences and all of those sorts of things . . . It’s a step-wise process of [the clinicians] saying, ‘What do we need to change or how do we need to change the way we’re working to use this [technology]’?

Second NTAC chief executive

For change to work, the ANT perspective draws attention to local agency. There may be powerful elite actors who agree to change, but there also have to be agents ‘at the coal face’ who implement change. All these actors (elite and ‘coal face’) have to be connected for change to happen. 162 This NTAC project manager points out the importance of consultants and nurses and the relationships between them.

I’ve learnt a hell of a lot about how to navigate around the importance of relationships, the importance of who is engaged, ownership. Yes, you need to pamper the consultant’s ego, but it’s the nurse on the ground that ends up doing it.

Implementation manager: IPT, BLNA

Within these network relationships, PbR constitutes a possible major barrier to technology implementation because any new technology disrupts established networks, requires the formation of new networks and, therefore, at least in the short term, reduces patient throughput and, in consequence, reduces income under PbR.

NTAC staff identified PbR and associated tariffs as a major barrier to technology adoption and implementation.

I have witnessed these barriers. I have witnessed finance managers saying, ‘Well, we’re going to lose income, so we can’t bring in this test.’

Implementation manager: IPT, BLNA

The first chief executive explained that if NICE made a technology mandatory it could also create an interim HRG code. She noted, however, that this could take time to materialise.

NICE puts out a guideline for a new piece of technology which may not have a tariff. It tells the trust they have to adopt it but it doesn’t tell them how they’re going to get paid for doing that. What I was trying to push NICE to do is create an interim HRG code for the moment that the guidance is given so that a trust can start using it and develop some reference costings around it. But that’s going to take time coming through and NICE, of course, is going to have to change.

First NTAC chief executive

The second chief executive noted that the lack of a tariff for many technologies meant that the trusts incurred costs that were not reimbursed.

Lack of tariff for these technologies means that the trust are shelling out on disposables and sometimes on capital but it’s not picked up at all [in the tariff] so again it can be a real cost thing for all of them.

Second NTAC chief executive

The HTWT guides were intended to enable the trusts to overcome these barriers. The next section explores NTAC’s intentions and experiences with these guides.

NHS Technology Adoption Centre’s intentions and experiences with the How to Why to guides

As mentioned earlier, the HTWT guides were the end point of the NTAC process, as understood in its agreement with its funding organisations.

NTAC was originally funded to produce the guides as well as work with a small number of organisations.

Second NTAC chief executive

One intention was that the HTWT guides would be a ‘roadmap’. In ANT terms, NTAC intended the guides to be an ‘obligatory passage point’:87

They can follow a very clear road map of what needs to be done at each stage; [for example] If you look at the guide for breast lymph node the documentation there will include a patient consent form to have all of that done in one surgery. So they haven’t got to think of any of the things . . . all the pieces of the jigsaw – it’s all there ready to download. So the one thing they do need to do is to get the right people round the table. And we even tell them who the right people are.

First NTAC chief executive

Another intention was that the guides were ‘tools’, particularly as templates for business cases to trust management.

The ‘How to Why to’ guide gives them the tools. So if they need to make a case to their manager, there are costing models, there is the core basis of a business case there, so they’re not starting completely from scratch.

Second NTAC chief executive

A key neglected issue, however, was the need to publicise the guides so that clinicians and managers in the trusts knew they were available.

So the notion of dissemination was the production of the guides, which would be a resource to the NHS. What wasn’t given any thought was how the guides were publicised and used.

Second NTAC chief executive

Later, NTAC hired an agency to publicise the guides and, hopefully, enable knowledge diffusion about the technology concerned throughout the NHS.

Yes, publicise the guides, in the hope that there’d be an element of diffusion that way. We’ve gone through a professional communications agency.

Second NTAC chief executive

However, on reflection, it was felt that there was a certain naivety in thinking that the guides would work to promote diffusion of the technologies.

So the methodology was . . . to share learning with the NHS via the How To guide. And I suppose, looking back, it was naive, but the vision was then that people would pick up those guides and adoption would filter through.

Implementation manager: IPT, BLNA

In the next section we explore the extent to which clinicians, managers and patient representatives did use the guides.

Clinicians’ and managers’ views on and experiences with the How to Why to guides

We report here, first, on responses to the IPT survey of a network of English clinicians who were actively engaged in trying to increase the uptake of IPT at their trusts. We asked these clinicians about their awareness and use of the HTWT guides. Second, we summarise participants’ filmed comments on both content and ease of navigation when actually using the online guides.

The results of the survey are as follows. First, we asked simply if respondents (all clinicians) were aware of NTAC; 54% were. Of those who were aware of NTAC, 63% were aware of the guides. We asked next about use of the guides. There were four predefined categories of ‘use’: a business case to your trust to develop your service; a business case to your commissioners to develop your service; an information resource; and to contact other trusts with experience of developing their service. Respondents were asked to tick all that apply. Of those who had used the guides, 96% saw it as an information resource, 13% used it for developing a business case to the trust, 13% for developing a business case to commissioners and only 8% to contact other trusts. Finally, we asked about the helpfulness of the guide: 54% were neutral; 20% found the guide somewhat helpful; 14% thought it extremely helpful; and 11% somewhat unhelpful.

A summary of comments (from clinicians and managers) while navigating the HTWT online follows. First, we discuss the main issues raised on the written content. Participants said that there is lots of useful information for managers and clinicians, if they can find the site and persevere in looking through it, although it can be quite hard ‘to see the wood for the trees’. Others found the site ‘generally unengaging’ and thought that the content generally emphasises the complexity of the process and is a long read. This was thought to be a danger because the content emphasises a time-consuming technical process, but political issues might be more important for implementation success, for example getting people on side at an early stage and identifying champions. The guide provides information aimed at commissioners, which is a strong point. The language is accessible for clinicians but may not be for patients.

Second, the main issues raised were on navigation and presentation. The website may need to provide different ‘ways in’ for different stakeholders with different purposes and interests. The layout of the guide assumes that people will read the executive summary first, but this may not be the case. The links to NICE guidance, in particular, are helpful. More visual interest would help, for example images of different insulin pump models might engage the attention not just of patients but of other stakeholders too. The slowness of some pages to load could be a barrier to people’s use of the site.

Summary

Although NTAC personnel did not think they worked to a clinical implementation ‘model’, we concluded that their approach built new networks through the ANT sociology of translation. 89 Once new networks were in place they adopted project management techniques to maintain momentum and ensure an end date. Individual project managers had considerable autonomy. Although this often worked well, in the case of UFRI, the management of this project may have been an aspect of its early close-down (see this case study). In terms of key implementation barriers, NTAC identified PbR and the tariff. Implementation was enabled through actively ensuring stakeholder engagement and paying particular attention to rethinking patient pathways. Intensive work did generate results in accelerating technology implementation at particular trusts but NTAC’s reliance on the HTWT guides to ensure wider diffusion throughout the NHS was not very successful. In particular, clinicians rarely used the guides to build a business case to trust management and commissioners, despite NTAC concentrating efforts on the business case aspect. The issue of how best to ensure the wider diffusion of efficacious clinical technologies throughout the NHS is discussed at some length in Chapter 8.

Introduction

This case study focuses on a clinical technology used to detect and diagnose peripheral pathology in the retina. UFRI technology was regarded by NTAC as proven, clinically efficacious but underadopted. Our initial understanding from NTAC was that the main adoption barrier was lack of clarity over where the technology was best situated (e.g. in primary, secondary or tertiary care, or all of these). NTAC staff also commented on the possible problems engendered by implementing in secondary care when most benefit may accrue within primary care (i.e. staff in secondary care may feel somewhat exploited). These issues were evident in the case study but there were other significant barriers. There was ambiguity over not only where the technology is best deployed but also how it should be used. There was also uncertainty over what UFRI should be used for, i.e. the target medical conditions for which it is most efficacious. These ambiguities and uncertainties are still not fully resolved and remain barriers to the adoption of UFRI into clinical practice, as further discussed in relation to our empirical evidence below.

The UFRI case study was undertaken first. At the time this research was funded, one trust had successfully applied for NTAC’s support to implement UFRI. However, by the time the research began, this trust had withdrawn from the NTAC project and was not using UFRI. Brief details of this site are given in the following section.

Ultrawide field retinal imaging

The technology is a scanning laser device which provides wide-field (up to 200°) retinal images;170,171 this wide field view captures approximately 82.5% of the surface of the retina. 172 Colour images can be obtained without dilating the pupils, corneal contact or high degrees of illumination. 173 These digital images can then be viewed on screen. 174 Once in the computer, images ‘can be magnified, enhanced, annotated, printer, stored or emailed’. 173 The peripheral retina can indicate evidence of pathology in many ocular conditions. 172 Hence, Eyemap images enable identification of peripheral eye abnormalities and should aid the earlier detection of retinal pathologies.

The patient pathway for the detection and diagnosis of eye disease is rather complex and somewhat uncertain. In the UK, eye disease is often first detected by ophthalmic opticians (or optometrists) on the high street. Dispensing opticians are only usually licensed to provide glasses based in current prescriptions. In contrast, optometrists are qualified to carry out eye examinations in primary care and, hence, are in a position to assess the health of the eye. If an optometrist detects eye disease they may refer the patient to an ophthalmologist in secondary care; sometimes this referral will be via the patients’ GPs. Ophthalmologists are medically qualified, specialising in the diagnosis and treatment of eye pathologies and practising eye surgery. Many eye conditions that result in sight loss are painless (e.g. cataract, open-angle glaucoma, age-related macular degeneration and retinal detachment). 175 Thus, for many serious eye conditions, pain will not prompt individuals to seek medical attention; 39% of the UK population do not have their eyes examined on a regular basis. 176 It is calculated that 50% of sight loss is avoidable through early detection and intervention. 176 Detection of disease in the retinal periphery has been problematic: ‘Until recently, the retinal periphery went largely unimaged, mainly because until now there existed no easy way to image it’. 177

In current practice the gold standard for examination of the retinal periphery, ‘includes a dilated retinal examination by an ophthalmologist with indirect ophthalmosocopy and often requires scleral depression’. 172 [Scleral depression is a technique which depresses the wall of the eye inwards (see Albert et al. 178).] Dilation of the eye is time-consuming, uncomfortable and results in blurring of vision, which lasts for several hours; it is inconvenient for patients and, as compared with undilated examinations, results in reduced patient throughput for practitioners. In the UK, the gold standard examination is somewhat rare in practice, usually in primary care an eye examination by an optometrist is done without dilation; dilation of the pupil is normally only performed by an ophthalmologist within secondary care. 179 Even when performed in secondary care, only 10–12% of the retina can be viewed at any one time,180 and no digital image is available for later review. These issues suggest that an Eyemap digital image (which can be done without dilation) may replace, duplicate or add to traditional methods. Among studies to date, there is currently a lack of consensus over these possibilities. One study evaluating the clinical efficacy of UFRI suggests that it ‘does not replace the dilated fundus [interior surface of the eye including the retina] examination, but is a powerful tool in patient education, photo-documentation, and retinal and systemic disease detection and prevention in conjunction with our current instrumentation’. 173 On the other hand, Nath et al. 181 found that UFRI detected some lesions missed during a dilated examination, and concluded that UFRI ‘appears to be additive (as well as duplicative) to the dilated exam’. Friberg et al. 182 tested UFRI as a screening device for specific eye diseases and concluded that its use is viable and merits further research. Mayers173 suggests that the use of UFRI ‘aids in the early detection of a variety of eye conditions [including] diabetic retinopathy, various forms of macular degeneration, posterior vitreous detachment, retinal holes and tears, hypertension, some types of leukaemia, and retinal detachment’. Silva et al. 183 concluded that non-dilated UFRI images compare favourably with both traditional dilated photography and a dilated clinical examination in determining diabetic retinopathy and diabetic macular oedema (watery fluid collection in the macula, the area of the retina where vision is keenest); they noted, however, that the UFRI images were obtained more rapidly, thus enhancing the efficiency of the screening process. Witmer and Kiss172 suggest that the full utility of UFRI will only be revealed over time.

The issue of whether, and in what circumstances, UFRI could replace, or add to, the gold standard examination was a matter of some debate among the participants in this study, as discussed further below.

Is Eyemap a proven technology?

The ways that ‘proof’ is best obtained is contested in clinical practice. In the light of this situation, along with the issue that regulators do not require clinical technologies to be subject to clinical trials, the question of whether or not Eyemap is ‘proven’ is considered in broad terms in this section. Debates around evidence in scientific journals are discussed along with those of awareness and ‘take-up’. Aside from proof, clinicians’ perceptions of Eyemap’s utility in clinical practice is discussed in the next section.

Specialists in TCs are aware of publications and evidence for UFRI, but, as highlighted later, this knowledge tends to remain confined to specialist centres.

There are countless publications out there about the device and its usability.

Consultant ophthalmologist, TC2

In the UK, these specialists are instrumental as clinical ‘champions’ and in raising awareness of the technology.

I got interested in the use of Eyemap at least five years ago. I’m always looking for new technologies that can help me understand better retinal diseases . . . And that was, for me, the main drive to try to get an Eyemap into the X Eye Hospital. I succeeded with it and we had the first Eyemap of the current generation in the NHS.

Consultant ophthalmologist, TC4

One specialist indicated that, in the UK, advertising and awareness of the technology was still not high.

I think the whole advertisement and awareness of the Eyemap and what it’s capable of has not been very high. As I say, I only came across it by chance. Now, that may be due to resources in the company or it may be just the way they’ve targeted telling people about it, but a lot of how we find out about things in reality is going to meetings or speaking to colleagues who’ve got something and you hear about it.

Consultant ophthalmologist, TC1

Therefore, as may be expected, some respondents stress that interest is more intense in other countries where take-up is higher than in the UK, especially among optometrists.

Germany uses Eyeco a lot, Italy uses Eyeco a lot, there are over 4,000 machines in the States . . . So, the UK is lagging very, very far behind . . . I would say about 90 per cent of the machines are optometrist, most optometrists in the States have one.

Consultant ophthalmologist, TC2

I’m off to Norway to lecture in October. I’m going to be lecturing to more optometrists in one room than I can lecture to in 24 road shows in this country. Because they’re so keen to embrace that modern technology and they can see the benefit for the patient. And you know, in the States, it’s huge.

Optometrist, HSO

When assessing whether or not clinicians accept that there is supporting evidence for a technology, one relevant issue is that some sites follow the lead set by others. In the implementation site that withdrew, the next respondent comments that his trust adopts technologies after they have been assessed elsewhere.

We get technologies that are already known to most people. We know that it’s useful, we know where it will be used, and we know about some of the research that has been done.

Consultant ophthalmologist 4, WIS

Late adopters of technology are less likely to have unrealistic expectations over what a technology can deliver. Knowledge about UFRI and what it can deliver is being generated in tertiary care. Therefore, outside of this somewhat rarefied realm, expectations were sometimes unrealistic and, therefore, not met. The business development manager in WIS articulates a discourse of unfulfilled expectations around ‘undilated pupils’.

The key technical aspect, which the equipment had promised, was that examinations could be done with an undilated pupil. Basically [under current procedures] there is a 45 minute wait for drops to be activated [to dilate the pupil] and not being able to drive, you’ve ruined your day . . . But what we found was that all this technology could do is add another test to the current patient pathway instead of replacing it with another pathway.

Business development manager, WIS

In addition, a tertiary specialist confirms that the expectations of ‘some people’ exceeded what the technology could accomplish.

Some people expected never to have to do a clinical exam again, the gold standard is clinical examination, I’m sorry but there’s no way of getting away from actually looking at the patient. There is no imaging out there that will give you all the answers without looking and talking to the patient. So I think the hype around it rightly or wrongly was much more than it could deliver.

Consultant ophthalmologist, TC2

So if the technology could not replace the ‘gold standard’ clinical examination, what was its clinical utility?

Clinical utility: what are the benefits for practice?

The first point is that, based on his experience, one tertiary specialist thought that Eyemap could sometimes substitute for clinical examination.

It does actually help the diagnosis more than I thought . . . we’re beginning to find that just looking at the photograph is often better than the doctor looking in. Or at least, it’s just as reliable and, in many ways, more reliable . . . it does lead, I think, to safer practice. You’re missing less pathology just due to taking a wide field.

Consultant ophthalmologist, TC1

In contrast, in WIS, based on arguments that retinal pathology usually occurs in the centre, one ophthalmologist declared that Eyemap was ‘just not useful’ in clinical practice:

It’s just not useful in clinical practice. The majority of problems do not occur in the far periphery, the majority of problems occur in the centre. The majority of meaningful problems, of problems that are going to affect your vision, are easily imaged by our current systems, OCT [optical coherence tomography] and other technologies.

Consultant ophthalmologist 7, WIS

This difference of opinion may, at least partially, be explained by the differential type of conditions that are treated in secondary and tertiary care. Obviously, tertiary care specialists treat rarer and more complex pathologies. Therefore, the benefits of the technology may accrue differently across the two settings. On the other hand, the tertiary specialist quoted below identifies ‘efficiency’ as his first priority; this seems to be a generic advantage. Indeed, as explained above, expectations of ‘time savings’ through being able to carry out undilated examinations was a prime reason for WIS considering Eyemap in the first instance. In the view of this tertiary specialist, efficiency is increased due to the rapidity of image collection with the Eyemap.

I realised if I wanted to record a diabetic, I could do an [Eyemap] photograph in thirty seconds, if I sent them for a colour photograph on the standard fundus [interior surface of the eye] cameras it’d take fifteen minutes/half an hour for the same result . . . in terms of why we have the Eyemap: firstly, it’s efficiency . . . it’s much quicker.

Consultant ophthalmologist, TC1

Despite the comment above from the secondary care ophthalmologist who maintains that since most pathology is situated centrally there is little benefit in imaging the periphery, there is some consensus among the tertiary specialists that the fundamental benefit of Eyemap is the capacity to image the periphery of the retina.

If you do a very, very thorough imaging of the back of the eye peripherally you cover up to about 70 to 80 degrees and Eyeco covers to 200 degrees. So a lot of the stuff that goes on, on the very periphery of the eye you won’t see with any other imaging techniques.

Consultant ophthalmologist, TC2

The main benefit of this camera, over anything else, is wide field imaging.

Consultant ophthalmologist, TC1

I think there’s no better system on the market at present to visualise the peripheral retinal vasculature, to the point that, if I could, I would have all my patients who need an angiogram imaged with an Eyeco. Obviously, we can’t because we’ve got only one system. But if I could, that is the only system I would use.

Consultant ophthalmologist, TC4

The optometrist conveyed a specific narrative on patient benefit.

I couldn’t work without an Eyemap as there’s two thirds of the retina that I’ve not looked at. I’ve got enough patients now who are asymptomatic, one’s a contact lens patient, very high prescription [and, therefore, at increased risk for retinal detachment]. And there right in the periphery was a huge tear . . . had she [just] gone to bed that night, she would have probably woken up the next morning and had very little vision in that eye.

Optometrist, HSO

As mentioned earlier, there is uncertainty over where and how Eyemap is best used and for what target medical conditions. With the presence of uncertainty in all of these key areas, assessment in specialist TCs over time appears to be required. Indeed, this specialist queried the validity of any assessment of Eyemap in secondary care.

You know, if you go to [WIS] and they don’t like it, well frankly, who’s heard of them from an eye department point of view? It’s not like a big teaching hospital, is it?’

Consultant ophthalmologist, TC1

One aspect where there was some consensus over clinical utility was for patient education.

Where should ultrawide field retinal imaging be situated and what should it be used for?

As discussed above, WIS decided against going ahead with the technology. The technology was demonstrated by Eyeco at the trust but there was no subsequent in situ trial to determine clinical utility. The lead clinician made the decision not to continue with the NTAC implementation project after the demonstration. Although he declined to be interviewed, he summarised his objections in an e-mail. These were as follows: the equipment is bulky; maintenance is expensive; it is more suitable for optometrists; the demonstration did not impress; no evidence that the image resolution is better than their current digital photography; they were led to believe that pupil dilation was not necessary but the images are better with dilation (e-mail communication, 24 March 2010). At interview, the seven other consultants who had been present at the demonstration also commented. As the demonstration preceded the interview by more than 1 year, their recollection of events was quite poor. Their remarks are briefly summarised as follows. One consultant did not remember the demonstration at all and was unsure whether or not he was involved in the decision. Three consultants thought that the main view was that the technology was not needed at a district general hospital. Of these three, one was of the opinion that it was better situated in primary care and one that it was of most benefit in tertiary care. Two felt that the resolution was not good enough. The seventh (who declined to be recorded) thought that the decision had actually been made by the lead clinician (who declined to be interviewed).

Yet, as discussed in the previous section, after working with the technology, the tertiary specialists thought the technology conferred several benefits. This first consultant thought that a case could be made on the grounds of efficiency. He commented further on its potential use in secondary care:

In terms of a hospital [secondary care], I think it has a big role in the eye department, but it hasn’t been taken up widely, and why is that? I think it’s partly people not considering the full benefit of it. It’s partly the change in need in ophthalmology. Everybody’s looking at ways of being more efficient. I think this is one way.

Consultant ophthalmologist, TC1

And he was specific about its use ‘in the clinic’, as opposed to research.

What do we use it for in the clinic . . . Increasingly, we would take photographs of diabetics, people with any retinal vascular problem . . . Inflammatory conditions, again, it’s very useful to take an Eyemap picture. Choroidal nevus [pigmented area] or, you know, suspicious tumours or something, again, an Eyemap will probably go out further.

Consultant ophthalmologist, TC1

But another tertiary specialist commented that the secondary care clinic in his hospital had not adopted Eyemap.

If I don’t have the Eyemap the other cameras are very inferior . . . but in the rest of the hospital people seem to be quite happy with the cameras they have.

Consultant ophthalmologist/ocular oncologist, TC1

At stake here is the issue that the research findings on Eyemap and clinical experience of Eyemap in tertiary care are not being disseminated to secondary care. There is no formal conduit for this to happen. Without a national policy to ensure diffusion of clinical technologies to appropriate sectoral sites, it seems, at the moment, that the clinical utility of Eyemap will not be exploited outside of tertiary care.

How does Payment by Results impact on the adoption of ultrawide field retinal imaging?

Under PbR, trusts are funded on the basis of activity. In secondary care, non-urgent ophthalmology referrals from primary care are a significant source of income to trusts.

The trusts are making money, so there is a disincentive [to transfer patients to primary care]. They’re actually happy with the false positives because they get money for old rope.

Optometrist, HSO

Even if referrals turn out not to have been appropriate (‘false positives’), under PbR there is no incentive for the trusts to work with commissioners on ‘demand management’.

From a PCT point of view, clearly we’re looking at the appropriateness of referrals to the trust. Everybody’s [all PCTs] trying to do demand management . . . We’ve picked five areas to look at, one of which is ophthalmology . . . orthopaedics, ENT and ophthalmology are our biggest referring specialities . . . In ophthalmology we have 16,000 new appointments each year . . . we [in the PCT] know that there are lots of referrals made that could be managed in primary care, or by opticians.

Lead commissioner for WIS, PCT

Yet, as discussed above, a proposal to open an eye clinic in primary care staffed by optometrists was ‘stamped upon’ by one of the secondary care ophthalmologists. The reaction of one powerful hospital ophthalmologist may be seen, in part, as professional hostility towards optometrists expanding their skills but it may also have been triggered by the possibility of eventual loss of income to ophthalmology departments in secondary care. Although, at first, the introduction of Eyemaps into primary care would be likely to increase referrals as more disease was detected:

So the consequence of having Eyemaps in every optician would be more referrals. So it’ll be good for business for the hospital.

Consultant ophthalmologist, TC1

Over time, as optometrists improved their abilities to interpret the Eyemap images, referrals should become more appropriate, in terms of both false positives and false negatives. This process could be greatly facilitated by education. As mentioned earlier, the Eyemap producer has hired a specialist in the USA to whom images can be sent for comment.

If you have a support mechanism . . . Eyemap use a very good guy in America, and rather than refer the patient to hospital I refer to him first . . . Tell me what it is, what you think it is. And you get a response back.

Optometrist, HSO

However, clearly this is a somewhat ad hoc solution that, although helpful at the level of only one specialist, could not support all optometrists if the adoption of Eyemap became widespread in primary care.

Discussion and summary

As highlighted at the beginning of this chapter, where, how and for what issues are not yet fully decided with respect to UFRI. Domestication is a term used in the social studies of technology, a metaphor implying that a technology has to be ‘tamed’ before it can find a place and fit into practices and routines (see, for example, Berker et al. 185). UFRI is a technology that is not yet domesticated. A place had been secured in tertiary care but, unlike IPT and BLNA, where the place for adoption was uncontested, it was unclear and subject to professional dispute if UFRI could be successfully used in secondary and primary care. Moreover, unlike IPT and BLNA, there were differing professional opinions over, first, the extent to which UFRI was proven; second, the clinical utility of UFRI; and, third, the target conditions for which the technology is best deployed. There was some consensus over these three in tertiary care, but this knowledge was not being diffused out to secondary and primary care.

We have argued in a previous chapter that the NTAC processes are based on ANT principles. In ANT terms, the agency of any human or non-human actant always depends on the power of the network within which it is embedded. 88,159 The possibility of embedding UFRI in a network, and thus ensuring its agency, is problematic. One attraction of ANT, for this research, is that networks transcend organisational boundaries as they extend themselves spatially. 186 One of the key characteristics of UFRI is that it could benefit patients and the health-care economy as a whole through enabling cost-effective earlier diagnosis of serious eye conditions, which could result in loss of vision. However, as evidenced here, the potential for embedding UFRI in an educative network that brought primary, secondary and tertiary care clinicians together to enable earlier diagnosis seems, currently, remote. The division between primary and secondary–tertiary care, the financial incentives for activity in secondary care and professional jurisdictions currently all conspire in the UK to keep UFRI confined to specialist use in tertiary care, along with a few entrepreneurial optometrists. Even sharing knowledge across specialists in tertiary care faces obstacles. One specialist required us to sign a confidentiality agreement before he agreed to be interviewed, on the grounds that his research using UFRI gave him a competitive advantage over his peers. Yet the potential for UFRI to improve patient care remains. Aside from detection and diagnosis, the tertiary specialists remarked on the possibilities for Eyemap to advance their knowledge of ophthalmological conditions.

The UFRI technology did not go forward as a project, but, even if it had, from the evidence in this chapter it seems unlikely that it would have succeeded, as the barriers to implementation of this technology were so complex and difficult to overcome. Only measures to achieve some integration between primary, secondary and tertiary care, so that knowledge can be transferred upstream from tertiary, through secondary and onwards to primary care, would improve the adoption of this technology. In Chapter 8 we reflect on whether or not current government policy (as exemplified in Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS1) can improve adoption and implementation rates for ‘hard to domesticate’ technologies.

Introduction

This case study focuses on IPT, used in the treatment of type 1 diabetes. Type 1 diabetes is a long-term condition that affects around 250,000 people in the UK and requires lifelong treatment with insulin. 187 Rates of type 1 diabetes have been increasing over time, with the greatest increase in children younger than 5 years of age. People with type 1 diabetes are unable to produce the natural hormone insulin, which is needed to control and use glucose. Most people with the condition control their diabetes through multiple daily injections (MDIs) of insulin. IPT, also known as continuous subcutaneous insulin infusion (CSII), is an alternative method of treatment to insulin injections by syringes or insulin pens. The pump provides a CSII, thus replacing the need for MDIs and typically producing better control of blood glucose levels.

Insulin pump therapy was first introduced in the UK in the 1970s, initially within the context of research studies on type 1 diabetes. However, mainstream adoption of the technology was limited, linked to concerns about efficacy and safety and the potential financial burden on the NHS. 188,189 In the past decade, technological advances have resulted in a new generation of smaller, more portable, more efficient and user-friendly pumps with additional safety features. 189–191 Several studies in recent years have demonstrated the benefits of IPT in type 1 diabetes compared with MDIs. 190,192–199 In particular, it is suggested that the continuous infusion of insulin provides not only improvement in metabolic control, but also increased physiological and psychological well-being. 189,200 Other benefits include improved patient outcomes [e.g. lower glycated haemoglobin (HbA_1c); HbA_1c is a measure of the average plasma glucose concentration over a period of time, in people with diabetes, a higher plasma glucose concentration indicates poorer control of blood glucose levels], reduction in all grades of hypoglycaemia (hypoglycaemia is a low blood glucose level, which is too low to provide sufficient energy for the normal body functions), a reduction in blood glucose concentrations, fewer blood glucose swings, and a lower daily insulin dosage compared with insulin injection therapy. 201 It is also suggested that quality of life for patients and their family and treatment satisfaction are likely to be better on pump treatment than on MDIs,145 particularly for people who have experienced significant and continued control problems on MDIs. 201 However, comprehensive patient education, such as carbohydrate counting, and frequent self-monitoring of blood glucose or continuous glucose monitoring are necessary components of successful IPT. 190

The IPT case is the only one of the three technologies studied that had been the subject of a national technology appraisal process; however, this did not guarantee that the evidence base for the technology was universally accepted, as the findings illustrate. NICE issued technology appraisal guidance on IPT in 2003, which was further updated in 2008, and recommended it as a clinically effective and cost-effective treatment option for people with type 1 diabetes, whether adult or child, for whom MDIs have failed, and for children aged < 12 years if MDIs are not deemed practical or appropriate. 202 Alongside the technology appraisal, NICE produced a commissioning guide to help health service commissioners plan and deliver services in line with the guidance. This suggests that the standard benchmark rate for the uptake of IPT should be 12% of people with type 1 diabetes, and 33% for children younger than 12 years old.

A national working group on insulin pump services203 used a variety of sources, including national registers, manufacturers’ records and published reports of pump practice in various countries, to estimate the uptake of IPT at an international level. These data suggested that some countries (USA, Israel and Germany) were using pumps for about 15–20% of people with type 1 diabetes. A typical figure for Europe (e.g. France, Sweden and the Netherlands) was around 10% of people with type 1 diabetes using insulin pumps for routine management. In contrast, overall UK pump usage was estimated at around 1% of people with type 1 diabetes using insulin pumps for routine management.

A subsequent review of IPT in England was undertaken by the Medical Technology Group in 2010. 204 They carried out a survey of all PCTs (n = 152) in England to ascertain levels of insulin pump provision. Of these, 87.5% responded to the survey and the data indicated that the average rate of pump use was 3.7%, with rates across the country ranging from 0.25% to 13%. 204 Thus, although the rates were higher than those estimated by the Department of Health Working Group in 2007, they were still some way from the NICE recommended benchmark of 12% and considerably lower than in most other countries of comparable economic standing and level of health-care provision.

The NHS Technology Adoption Centre project on insulin pump therapy

NHS Technology Adoption Centre selected IPT to go forward as a technology implementation project so that it could identify the challenges associated with implementing IPT and suggest ways to overcome them. Three NHS trusts were selected to work as implementation sites; these were organisations that responded to the NTAC call for implementation sites and were selected because they wanted to develop their IPT service in line with NICE guidance (see Chapter 4 for a more detailed description of the NTAC selection process). Three MSs were identified to work with the implementation sites, these were organisations that were already using insulin pumps with many of their patients and could provide clinical mentorship to the teams in the implementation sites. The experiences of the sites involved in the IPT project were collated into a HTWT guide. 205 This was intended as an online resource that could be used to help the adoption of IPT throughout the NHS.

The case study

The main data collection took place in four NHS organisations: two NTAC implementation sites for IPT; one MS; and one organisation that had initially applied to be an implementation site, but was not selected by NTAC. This latter organisation was keen to increase the use of pump therapy and, as such, would be typical of the organisations that NTAC was aiming to target with the HTWT guide. In all sites, interviews were conducted with a range of individuals involved in implementing IPT, including commissioners, clinicians, diabetes nurse specialists and business/procurement managers. A total of 23 interviews were conducted across the four sites; details of the interview sample by site are provided in Table 3.

Alongside these qualitative interview data from the case study sites, supplementary sources of data included a survey of clinicians about IPT use in the NHS in England and documentary analysis of e-mail correspondence received by the patient support group INPUT between September and November 2011.

Before presenting the main findings relating to the IPT case, some brief background information on each of the four NHS organisations studies is outlined, along with a summary of the results of the survey we conducted on IPT uptake in England.

Uptake of insulin pump therapy: survey findings

As outlined in Chapter 3, Research methods and Data analysis and Appendix 3, we conducted an online survey of a network of UK clinicians actively engaged in trying to increase IPT uptake to assess the current level of uptake of IPT. Table 4 summarises the key findings from this survey, comparing estimated uptake of IPT in May 2012 with uptake levels 3 years previously.

These data suggest an increased use of IPT, in line with the findings of the Medical Technology Group survey of 2010. However, they also indicate that, in summer 2012, 65% of respondents reported that pump use was lower than 10% in their NHS trust (in other words, below the 12% target recommended in the NICE guidance). Moreover, of the 62 respondents working in trusts where the uptake rate of IPT was < 5% 3 years previously, only 29 (47%) had managed to raise this level to > 5%.

Our qualitative data collection attempted to explore in more depth the reasons for the lower than recommended uptake of IPT in the NHS in England, and the difficulties encountered in attempting to increase the uptake rate. Analysis of the findings reveals a number of key themes in terms of factors that appear to facilitate or hinder the adoption of IPT, ranging from those on the individual level, including patient- and clinician-focused factors, through to more organisational- and system-level issues relating to past history and experience of IPT including resourcing, financing and commissioning issues. Each of the key themes is discussed in more detail in the following sections.

The patient pull for insulin pump therapy

A number of patient level factors appear to be important in terms of the adoption (or otherwise) of IPT. These include patient-driven demand for the service, acceptability of the technology and the importance of patient self-management of the technology. In relation to the first issue, views differed in terms of the extent to which patient demand was important in driving the introduction of the technology. In some cases, patient requests for IPT were seen to be a major driver:

We are very good at being ahead of our patients in what they know about either their condition or their treatment plan or their medications. But with insulin pumps, there was a feeling that we were only just one step ahead of our patients because they came in so suddenly really that the patients were asking for them and we were saying, ‘Oh, hang on a minute. We’re not so sure that we have the skills and knowledge to facilitate this for you.’ So there was a definite perceived, you know, we weren’t skilled enough to run these kind of programmes.

Diabetic specialist nurse, IS2

By contrast, other interviewees did not perceive a significant patient demand for pumps; rather they felt that it tended to be the clinicians who raised the possibility of using a pump as an alternative to MDIs.

One of the things that I have noticed when I’ve been to meetings and so on elsewhere that some of the people say patients ask for pump therapy. It doesn’t happen in my experience very much . . . Yes it tends to be us that says, well actually I think you might benefit, would you consider a pump? There is also a significant turning down of going on pump therapy.

Diabetic consultant, IS1

Linked to the above point, there was a fairly consistent view among many of the clinical staff interviewed that there were issues relating to patient acceptability of pumps that accounted for some of the lack of uptake of IPT. For some patients, this was about how the pump would fit in with their lifestyle or how they felt about having to be attached to a pump at all times; for others it was the requirements that went hand in hand with using a pump, for example the need for active self-management and regular blood glucose monitoring.

There are a significant number who think about it and then just don’t want even to take anything any further. They’re clear in their own minds that they don’t want a pump. And I think the minority are the ones who would be prepared to consider a pump and they then go on to have an assessment, a formal assessment. And of those who go forward, not everyone in the end decides that they want a pump. Having seen one, talked about it in more detail, quite a number of them have come to the conclusion that it’s not for them . . .

Diabetic consultant 3, NIS

However, for others the pump was seen to provide them with a greater freedom to manage their diabetes, despite some initial concerns about the pump itself:

Well there’s the body image, which probably is more of a concern for females, understandably . . . There’s a classic quote where the partner calls them the Bionic Man or Bionic Woman because they are linked up to the machine. Certainly in my study . . . I think the patient experience is honestly amazingly positive. Even people who have apprehensions about going on the pump, but still probably want to do it because their control is not where they want to be or they’ve got severe hypos, generally speaking a few months into pump they would not go back to injections.

Specialist medical trainee, NIS

The comments above represent a rather divergent set of opinions on patients’ views of IPT from the perspective of professionals delivering the service. It also appears that professionals’ views do not always correspond with the views of patients themselves. This is illustrated by some of the e-mail correspondence received by the patient support group, INPUT.

[My] consultant said: ‘If you use a pump you will be a failure to yourself.’ ‘It’s not suitable for you because you work.’ ‘If you have to miss a meal you’ll be in trouble because the insulin is going in all the time.’

Patient A with type 1 diabetes

My daughter has had diabetes for 4 years. The clinic team are saying it’s not the right time for a pump. She is aged 9, will be 10 in April. [Her] control is very poor. The diabetes clinic said ‘how would we work out her pump ratios when she’s so up and down?’

Parent of child B with diabetes

The influence of clinicians on the implementation of insulin pump therapy

Clinician-related factors appeared particularly influential in the adoption of IPT; this was particularly the case at a consultant level. Where consultants providing a diabetic service were motivated and enthusiastic about IPT, they played an important role in leading the introduction of the technology.

. . . we have always been a forward looking trust I think, although we are only a district general hospital I do think that we have a philosophy of, the philosophy that we have in this department is that our patients should not miss out on any treatment that could be of benefit to them and that is available in the UK and that’s how it has always been.

Diabetic consultant, IS1

Equally, where consultants were sceptical or wary of the technology, they represented a significant barrier to its implementation. For some clinicians, this appeared to relate to bad experiences of using pump in the earlier stages of developing the technology, alongside some more recent negative incidents.

And you have to remember that I did actually use pumps in 1982, so I’ve been using pumps for a very long time . . . So I’m not sceptical because I’ve had nothing to do with them. I’m sceptical because I’ve seen the good things and I’ve also seen the bad things about pumps. Like the girl we admitted last weekend, in very severe diabetic ketoacidosis [ketoacidosis is a dangerous complication of diabetes caused by a lack of insulin in the body], with a pH of 6.8, who nearly died, and I’m not egging this up because she’d . . . we hadn’t put her on a pump, she was put on a pump elsewhere and she didn’t self-manage properly. When she felt sick, she took her pump off and she was that close to dying.

Diabetic consultant 3, NIS

In other cases, resistance appeared to be tied up with local politics and personalities:

And that’s one of the sort of barriers . . . internal barriers to implementation and it just kind of reflects the bigger picture I think and problems within the sort of structure of the team that was highlighted by this project. So I think some of the barriers to this project aren’t really barriers to this project, they’re more to do with identifying barriers that are within . . . our Diabetes Team . . . one individual in particular who was asked to be involved but because of other work commitments he wasn’t, couldn’t get involved but if it isn’t his project he doesn’t want to know and he has done quite a few things to try and sabotage the . . . and he’s still trying to do things to sabotage it. It’s not just in this, it’s in other things as well. And those are the sort of kind of things that actually aren’t necessarily surmountable.

Project manager, IS1

The strength of evidence for insulin pump therapy

One particular issue that emerged, and may partly account for the differences of opinion at a clinical level, was the strength of the evidence supporting the introduction of IPT. Some clinicians perceived the evidence supporting the use of IPT as rather ‘shaky’.

I don’t think we have the evidence yet and that is a really important question. I think it was an assumption when we started our journey with pumps really so hopefully if you can get people good at this then their needs for support get less. I don’t know if that’s born true and I think this comes back to your patient selection that you need to be very careful about the reasons that the patient is selecting a pump. The one thing that it does not do is take away your diabetes it makes it even more in the front and you know whether they are able to manage it themselves and again I think we haven’t to my knowledge got very good data on long-term outcomes from our pump clinic.

Diabetic consultant 4, NIS

I’m not even a hundred per cent sure about the proven clinical benefits. If you have . . . do the proper randomised controlled trial, pump versus intensive therapy with an equivalent amount of input from health care professionals, you don’t get a great difference in Hb_1c or in anything else. You might get a difference in patient satisfaction in favour of pump, but that’s in folk who want to go onto the pump. So I think sometimes the benefits of pumps are a bit overstated.

Diabetic consultant 3, NIS

This second quote raised an issue that a number of interviewees at site NIS referred to: whether it was the additional educational input and support that pump patients received that made the most difference, as opposed to the technology itself. However, at other sites, the evidence for pumps was much less disputed, with IPT being described as a ‘proven technology’. With the endorsement from the NICE technology appraisal that IPT was a clinically effective and cost-effective treatment, this has enabled organisations to develop a case for their commissioners for increased funding of IPT.

I suppose, on the surface, it looks expensive but there’s then the evidence to show that . . . that’s why NTAC have taken it on . . . because the improvement of the blood result, HbA_1c , in preventing complications . . . is a proven technology for improving progress with [diabetic] complications.

Diabetic specialist nurse 2, IS1

PCTs have to follow NICE, have to be seen to be following NICE guidance, so I know it’s a time of austerity and the PCTs have other priorities, but a trust should always play the NICE guidance card and say, look you know, that’s why we’ve got NICE guidance because it’s the best outcomes for this group of patients.

Diabetic specialist nurse, MS

However, there were some who were critical of the wording of the NICE guidance, which indicates that IPT should be used where MDIs have not been effective in controlling blood glucose levels. This led some interviewees to suggest that IPT was perceived as a ‘treatment of failure’:

And at the moment, very much NICE guidance is geared to pump use being used as a treatment of failure. So if you fail with your multiple daily injections, i.e. you’ve still got poor control, or you’ve still got low blood glucose. It’s a negative process . . . it’s sort of like a downward staircase on diabetes. Start on maybe twice daily injections or your control is not very good – let’s put you onto the next treatment, which might be structured education and basal bolus regime [a basal-bolus regimen, which includes an injection at each meal, attempts to roughly emulate how a non-diabetic person’s body delivers insulin]. Oh you’ve not done well with that, you’ve failed with that. Right, let’s go onto something else. And eventually you get to pump therapy. By that time, the question is, is the patient demoralised, is the clinician demoralised? Why is pump therapy used as the treatment of failure, if it’s supposedly the best treatment? Compared to cancer treatment, you wouldn’t . . . We’ll start with something but if you don’t get better then we’ll use the next one. Whereas in diabetes, it is like that.

Specialist medical trainee, NIS

This could lead to a dilemma as to how to proceed with patients who might benefit from a pump, but did not meet the NICE criteria.

But there are definitely people whose diabetes control is okay but actually they probably would do better if they were on a pump, but that’s not covered by NICE guidance, so they’re currently excluded and you can’t really encourage them to let their HbA_1cs go up, which you know some of the conversations that not just us have had but we’ve you know . . . well do you let them increase their [blood glucose] . . . so they meet the criteria?

Project manager, IS1

The initial investment costs of insulin pump therapy

A key issue relating to the introduction of IPT was the service investment required, including the availability of specialist pump trained nurses, educational preparation of patients to use a pump, and the cost of purchasing pumps and related consumables. The upfront investment costs in terms of staff and patient education were seen to be significant. Furthermore, several interviewees raised questions about the extent to which investing in service development for the introduction of IPT took services away from non-pump users and could result in inequity.

And I think the other barrier is the fact that, certainly in [this organisation], we expend a great deal of health care time with the small number of people who are on pumps and that is to the detriment of people who are not on pumps. We have a limited number of physicians, nurses, dieticians and if they’re working with pumps, it means that they’re not working with our other non-pump population. And I think we have a tension there; that the more resources we devote to pumps, the less we’ve got for somebody else. So it’s not just the cost of the pump and the pump consumables, it’s the health care professional time as well.

Diabetic consultant 3, NIS

Others raised concerns about the costs of the technology itself, suggesting that the cost was kept unnecessarily high by the producers, which, in turn, acted to limit the wider scale uptake of IPT.

So there’s been a lot of input from the pump manufacturers, but then there ought to be because prices of the pumps is ridiculous. And if you want to ask about the constraints, I’ve said it to everybody so I may as well say it to you, if an iPad can be 600 quid, why is an insulin pump £3,500? . . . And if – it strikes me that it’s not any technical – I don’t know anything about technology, but I think they’re milking it. And if you could even – that’s your – that’s your biggest constraint.

Diabetic consultant 2, NIS

Funding and commissioning insulin pump therapy

Each of the four sites had different arrangements in terms of the commissioning of IPT services. In the first implementation site (IS1), the trust had a close working relationship with commissioners. The PCT had been instrumental in driving the adoption of IPT and agreed to provide funding for the service without a formal business case.

So that’s when I got involved, I suppose that’s why they’re saying it’s been driven primarily by the PCT because once I became involved, I did actually start trying to structure it and say, you know, let’s look at it in a more comprehensive way; rather than just clinicians thinking this is a great thing to do, without any thought about how we’re going to pay for it.

Commissioner, IS1

. . . unlike other places we don’t actually have anything signed between the PCTs and us, we didn’t have to do a business case or anything. We sat around the table. . ..

Diabetic consultant, IS1

The second implementation site worked with a large number of commissioners, which could cause difficulties as different commissioning bodies had different ways of working.

I think what’s difficult is they all have a different way of working. For example, the main PCT that we work with . . . are quite happy to be invoiced directly by the insulin pump companies. I order a pump and the consumables from the pump company; they invoice the PCT; the PCT pay the bill. The patient then starts ordering consumables from the company and the PCT get the bill. That’s brilliant. I don’t have to have a cupboard full of stock. I don’t have to sign off invoices. I don’t have to get involved. So that works really well. Then you have another PCT, who shall remain nameless – there are two or three of them – where they insist on recharging back to the children’s hospital, who then have to pay the company, and then the company will supply. And then the bill will come back to the children’s hospital, we pay the bill, then we recharge the PCT . . . So you add another step. An unnecessary step in some respects, but it’s in there . . . And the other problem that that then causes is the PCTs do not get any bulk buying discount and they are missing a trick . . .

Diabetic specialist nurse 1, IS2

In the MS, the acute trust had a commissioning manager who acted as a point of liaison and negotiation with the PCTs, which was seen to have played an important supporting role in the introduction of IPT.

. . . not every trust has got like my post and my team that can be the conduit so sometimes the individual departments struggle with, who do they speak to within the PCT or we can help with that. We facilitate it, we package it in a way we know the PCTs would be amenable and so I think that’s quite important . . . We call it commissioning, which is a bit of a daft name, . . . so we agree the contracts, we agree the quality in the case of the PCT, we’re always mindful of whatever we agree is, I’ve got an inward looking role and an outward looking role and we’ve got to keep that, that communication going both ways from my department basically.

Commissioning manager, MS

In the fourth (non-implementation site; NIS), the organisation had been used to working on a block contract arrangement and had discretion over the use of their resources, which meant they could decide the level of investment they made in pumps. This was initiated when a clinical academic consultant came to work in the service and was interested in establishing a pump service. However, the consequence was that the service had largely grown around the practice of this particular clinician without being more widely adopted by or embedded in the rest of the organisation.

[The consultant] . . . was told in no uncertain terms that he could not develop a service bigger than that because he was an academic . . . It was not clear what his long-term shelf life would be . . . so he could then leave me with a huge number of people on pumps . . . We have to be quite careful that if he did move we’re then not left with a service and nobody and no budget to cover it.

Clinical manager, NIS

In summary, the implementation and MSs were able to get the PCTs to fund IPT because of the NICE guidance and, as a consequence, they were not particularly worried about the lack of a national tariff for IPT. Indeed, having the PCT pay directly for the costs was seen to be an advantage, as a tariff only provides the average national cost, which may be less than the actual cost.

I guess the danger with having a specific HRG [for the pump service] is making sure it reflects the cost and responds to changes in cost. Whereas the advantage to us at the moment is we don’t have to worry about it. If someone’s having a pump, the PCT pays for it direct.

Diabetic consultant, IS2

However, a persistent concern raised related to the infrastructure required to establish an IPT service, in terms of nurse specialist provision and patient education (as discussed in the previous section); this was a cost that the organisation typically had to absorb as part of setting up the IPT service.

The role of NHS Technology Adoption Centre in supporting implementation

NHS Technology Adoption Centre clearly played a key role in supporting the implementation of IPT for the two implementation sites. This included acting as an initial catalyst to ‘do something’ about IPT, introducing a formal project management structure, providing a map of how to get started and implement IPT and developing the systems that were needed to support the use of the technology (see Chapter 4 for a description of the NTAC process of working with implementation sites).

I mean I think it was if you like the catalyst, participating in this project was the catalyst that we needed to get us to focus on this.

Diabetic consultant, IS1

They [NTAC] were phenomenally helpful. We knew where we wanted to be, but weren’t sure of the map to use to get us there . . . NTAC were really good in helping us to get the people in the room who needed to be in the room, to have the right conversations. Project management – I think that’s what we really lack and what they did really well.

Diabetic specialist nurse, IS2

That’s the difference with NTAC. It’s a project. One of the things that’s come out of it, I would say, is that there’s been joint working with commissioners and PCTs and a recognised procurement process . . . The whole problem in the past is that you had patients waiting for pumps but it was either applying for funding individually, . . . or finding an ad hoc way of doing it.

Diabetic specialist nurse, IS1

The NTAC method of supporting technology adoption seemed particularly suited to the sites attempting to introduce IPT, providing a co-ordinating structure and project management that had been lacking, despite prior enthusiasm or intent to support IPT. The lack of a co-ordinated approach to IPT provision is equally apparent in some of the patient feedback received by the support group INPUT:

My consultant supports my belief that an insulin pump should now be tried as soon as possible and referred me to a Dr X at the X Diabetes Centre. Dr X has agreed that a pump should be tried soon, however he has warned me that there may be a delay of possibly 5 months before the local PCT gets around to agreeing to this proposal.

Patient C with type 1 diabetes

I would like some guidance if possible on how much choice I have on the insulin pump that is provided for my 5 year old. We have had funding approved for a pump, although the pump of our choice – we are told is too expensive by the PCT. Our hospital have said they will support us in whatever pump we choose if we can obtain funding, but have also said that if we don’t make up our minds soon the funding will be retracted.

Parent of child D with type 1 diabetes

Although acknowledging the important input that NTAC provided in structuring and signposting the implementation of IPT, there was also a view from some interviewees that the hard work only started once the NTAC project was coming to an end and they had to actually implement all the things they had planned on paper. In one of the two implementation sites, this responsibility sat with an internal project manager for IPT.

Well, it’s like anything; the pump project is in some way a theory. It was setting a structure, having guidelines, having a structure to every stage of pump therapy, but then, whatever we’ve decided now as part of this pump project – you’ve got to put it into practice! . . . That’s when the real hard work starts, doesn’t it? Up to now it’s been on paper.

Diabetic specialist nurse 2, IS1

Discussion and summary

Overall, the NTAC process seemed relatively well matched to the needs of IPT implementation sites; these were sites that were generally receptive to IPT and were already working with the technology, but often in an ad hoc way. NTAC brought a structured project management approach that enabled them to formalise the implementation of IPT. From an ANT perspective, the network of human and non-human actors was less complex in the IPT case, compared with the other two technologies studied in the research. Certainly, in the implementation sites, work had already been undertaken to address issues related to problematisation and interessement. However, the sites were lacking the guidance and ‘know-how’ to move further along the translation path in a systematic and structured way; the input from NTAC provided the support that these organisations needed. Through applying a project management approach and enlisting the involvement of key stakeholders, the NTAC model enabled and supported implementation sites to move through the translation path from problematisation/interessement to enrolment and mobilisation. Introducing IPT requires changes to the patient pathway and the design and delivery of diabetes services, for example in relation to patient and staff education and dedicated specialist nursing support. In turn, this requires a different set of networks and relationships to be developed. However, the fact that the evidence base for the technology was relatively well established and accepted (at least in the implementation sites) meant that the NTAC model of working had a good fit with the organisation’s needs in terms of developing this new network of actors, agreeing and co-ordinating roles, and embedding new structures and processes required to sustain the innovation.

The data reveal a marked contrast between the experiences and views of the two implementation sites and the NIS, where the receptivity to the technology appeared more mixed, particularly in relation to the benefits of IPT, the evidence supporting its adoption and the feasibility of establishing a wider adoption programme. From an ANT perspective, it could be hypothesised that the NIS was still at the problematisation stage, with human actors (particularly the clinicians) not agreeing on the nature of the problem and whether or how it needed to be addressed. This site did not receive the type of external input and support that NTAC provided to address the translational stages from problematisation through to interessement and beyond. The existence of an online resource in the form of a HTWT guide did not appear to provide a comparable substitute to the more ‘hands-on’ role provided by NTAC to actively facilitate the translational process. However, the question remains as to whether or not the NTAC process would be sufficient to address the barriers in the NIS. The issues at this site appeared to be more to do with differences of professional opinion over the value and benefit of the technology, rather than more practical problems of how to ensure implementation.

Introduction

This case study focuses on the adoption of BLNA, a diagnostic technology innovation21 for the treatment of breast cancer patients. Its adoption brought changes to the patient care pathway – which were generally perceived by health-care staff to be beneficial – but also required changes in work practices that were not always easy to implement. It reveals the importance of local factors on innovation processes and the impact of PbR12 on how stakeholders assessed the business case for adoption.

NHS Technology Adoption Centre selected BLNA for its HTWT guide programme in 2008. The evidence base for BLNA was considered strong and its adoption was consistent with the NICE guidance for breast cancer care:206

The National Institute for Health and Clinical Excellence (NICE) guidance for Early and Locally Advanced Breast Cancer (2009)207 advocates that minimal surgery, rather than lymph node clearance, should be performed to stage the axilla for patients with early invasive breast cancer. The intra-operative analysis of sentinel lymph nodes offers the opportunity to streamline the management of breast cancer patients as part of a cohesive and comprehensive service, and according to a review in the Histopathology Journal (July 2009), this test is accepted as a reliable technique.

NHS Technology Adoption Centre’s own assessment of BLNA identified six key benefits:206

1. reduction in acute hospital admissions

2. reduction in overall length of stay

3. improved efficiency for the NHS

4. improved quality of life

5. higher quality services and support for patients

6. long-term savings for the NHS.

Hospital trusts across England applied directly to NTAC to take part in the implementation project in the spring of 2008 and, following a process of site visits and telephone interviews, six sites nationally were selected. Owing to unforeseen circumstances, two subsequently dropped out. A total of 34 interviews were carried out between March 2010 and September 2011 (Table 5). We also conducted participant observation of the final BLNA project awayday in January 2010, which was attended by representatives from each trust.

Breast lymph node assay

Breast cancer affects more than 45,000 women each year in the UK. Under the established breast cancer clinical pathway, a patient diagnosed with breast cancer will undergo a mastectomy, during which the so-called ‘sentinel lymph’ node is removed and tissue sent for analysis. If this biopsy shows that there has been a metastasis (i.e. the cancer has spread to the sentinel lymph node), the patient will then be readmitted for a second operation to remove the remaining (axillary) lymph nodes. Apart from the additional psychological stress for the patient of having to endure a second hospital admission and operation, the current pathway can pose a clinical risk for patients as the cancer has further time to metastasise between operations.

Breast lymph node assay is a new test, the results of which can be ready within 30–45 minutes and so can be completed intraoperatively. If metastases in the lymph nodes are identified, surgery is extended to allow the remaining lymph nodes to be removed. Hence, the adoption of BLNA represents an improvement to the current pathway, in terms of both reducing hospital length of stay for patients and improving survival rates. 2

Getting started

At the project’s initiation in 2009, a Strategic Steering Group was set up by NTAC to provide long-term strategic direction. This was followed by a series of awaydays where representatives of each participating trust (clinicians, histopathologists, nurses and procurement officers) gathered to be updated on the current status of the project and to discuss any issues related to its execution. At the final awayday in January 2010, particular emphasis was given to the current situation of the technology (and the implications of the withdrawal of the product being used, scheduled for August 2010; see Dealing with technological uncertainty), the HTWT guide, long-term strategic visions for the project and building up success stories related to BLNA adoption.

The afternoon sessions focused on the validation of the HTWT guide. Four working groups discussed the content of the guide and possible changes. Copies of the current version were distributed and participants given the opportunity to review its core components: data plan (to populate the business model); standard operating procedure for pathology; clinical evidence; procurement; and benefits versus barriers. Discussions focused on a number of areas, including the creation of new roles/responsibilities for histopathologists; the impact of BLNA on patient experience; and the standardisation of clinical pathways compared with local protocols. There was a general concern that the short time available to conduct the test would make the histopathologists’ job more difficult. For some trusts, ensuring timely results would incur additional staff costs. Concerns were also raised about the psychological implications for patients of the ‘one-step’ rapid diagnostic test and instant clinical intervention.

For NTAC, this review exercise was essential in order to ensure that the guide was correct and in a format that was useful, practical and effective for the implementation sites. NTAC’s goal was to progress towards the guide’s ‘sign-off’. In the event, the main obstacle to this was the guide’s assumption of a standardised clinical pathway and, although the guide was eventually published, on the evidence of our subsequent interviews, this issue was never satisfactorily resolved.

Understanding the benefits

Our evidence points clearly to the BLNA case study being an example of an innovation championed by clinicians. However, all the surgeons in all the participating trusts stressed that their support for BLNA was on account of the important benefits it would have for patients. In this respect, its adoption was a ‘no-brainer’, as one surgeon described it:

Well to be honest it was a no brainer, you’ve got reduced length of stay, reduced hospital visits, reduced anaesthetics, theatres, reduced surgeon’s time . . . it was a no-brainer that if you could do this as a one-hit operation then that’s what patients would want.

Surgeon 2, IS3

Because clearly that’s good for patients because it reduces the number of operations. It’s good for clinicians because they get a result quickly, and it’s good for the people who are paying our bills because they see a saving . . . So the clinical argument was won inside a matter of an hour when I presented the business case.

Director of strategy and planning, IS2

The key thing that is the advantage to the patients is avoiding a second surgery, a second anaesthetic. That is a key advantage of this procedure. So, even though we were concerned about time and theatre space and number, because it has reduced the number of patients we do per list . . . I think we, as a team, have decided that the patient comes first . . . before we started intra-operative testing, up to six weeks to find out what is the ancillary staging would have taken up to six weeks.

Surgeon 1, IS3

However, clinical staff did have some concerns about the new consenting process that patients would have to taken through. A surgeon observed:

The patient’s consent has to be careful because the patient is consenting for a sort of either/or option and therefore consent can be arguably a little bit difficult, because you’re saying to a patient, we’re going to do this, if it’s positive, we’ll do your axilla and if it’s not, we won’t and therefore there’s an uncertainty. The patient’s going to wake up with an uncertainty. The first thing they’re gonna say to you is, was my axilla clear?

Surgeon 2, IS3

Although surgeons accepted that the consenting process would need to be amended, they did not anticipate it would be difficult for the breast care nurses to implement; the latter, however, had a different view:

I think what we haven’t thought through with this technique is how we support patients is we’ve seen it as great that we get earlier results, great that we don’t have to re-admit patients to come back and have ancillary surgery but we haven’t thought about the impact for the patient . . . at the moment there’s no information leaflets, there’s no plans or processes to support it . . . The patient will go to sleep with uncertainty about what they will wake up to. So for some they’d have had their lymph nodes removed and some they won’t . . . So that nobody really had a lot of consultation with her about that dread that she felt when she woke up.

Consultant nurse, MS

A breast care nurse emphasised the need to consider the psychological impact of patients, reinforcing the sense of divergence in understanding of the challenges among surgeons and nurses:

Yes as you say, it is really for them, the uncertainty of, you know, will the test come back showing that I don’t need to have any lymph glands removed, therefore, you know, my cancer is not as bad as it could be. Or if they come back positive then, oh dear my cancer is, you know, as bad as it can be in their minds. And it’s the psychological impact of that. And they just knew that if they had that drain there, that’s what had happened.

Breast care nurse, IS3

On this evidence, it became apparent that patient acceptability might be a significant problem. 68 Acknowledging this, the clinical lead at implementation site 2 (IS2) proposed that it could be resolved through providing patient information:

The patients would be happy to be offered a one-stop operation rather than the possibility of two. What we don’t know is how our patients will take this because they know that when they wake up and they have a drain you know that their lymph nodes are positive . . . so what sort of psychology it will throw them into we don’t know so if two people come for the same operation . . . So what we’re going to do is offer some pre-operative information giving that the important thing is to have a one stop operation rather than a two stop operation and if it’s positive of course it can be treated. At least we know what it is. So that’s the only tricky side but we have patient information for that.

Clinical lead, IS2

However, further investigation by breast care nurses convinced them that leaflets would not be enough by themselves, a conclusion that ran contrary to NTAC’s expectations and would have implications for the business case:

I wanted to know the specific psychological needs of the women who’d had the assay and woke up to find that they’d had their lymph nodes removed as well because the assay was positive, to make sure that we had the correct service with the resources to offer the care that the women needed. And previously, when I’d been to an NTAC training day in Manchester, there all of the staff said that there was no additional resources needed, breast care nursing wise . . . but when I actually went to [Mentor site] it was very different. The ward nurses told me that they’d all worked on a dedicated breast ward for many years, were very experienced, and they found it very hard to deal with the women, the first day post-operatively and often had to call the breast care nurses over to support the women because they didn’t feel that they could do it adequately. And then speaking to the breast care nurses, they could demonstrate that they were regularly called to the ward, just out of the blue, for extra support that they needed to give the women. So I then, you know, realised that I needed to put in a business case for some more breast care nursing hours.

Breast care nurse, IS3

It should also be noted that, as even the strongest advocates of BLNA adoption acknowledged, the intraoperative test is not as reliable as the ‘gold standard’ post-operative test, in terms of numbers of both false positives and false negatives. Hence, the post-operative testing would have to be continued and some patients might find themselves undergoing a second operation despite the initial, intraoperative ‘all clear’:

The problem is of course histology is the gold standard, you always have to do that, you always have to use the gold standard, as what passed before. So it’s called a false positive or a false negative, where it’s probably, and it has happened of course when histology has found a micro amount that we didn’t have because it was in the slice sent to them, and not been put in our slice . . . Obviously if histology do find a positive then the patient can still come back and have the further surgery . . . and it has happened, there’s been about four cases I think where, I don’t think they necessarily brought the patient back each time, because they were micro amounts, I think in some cases they just treated with radiotherapy.

Biomedical scientist, histology, MS

Again, the costs of running two tests would have to be accounted for in the business case.

Making the business case

In any organisation, management decisions on the adoption of a new practice will be influenced by its financial implications. Hence, the first hurdle to be overcome by the trusts involved in the BLNA project was to gather the evidence and make a business case for its adoption. Once compiled, typically this would then have to be prioritised against competing business cases:

So we started to work up the business case for the inter-operative analysis alongside other business cases for breast care. And that other big business case was around the reconfiguration of the outpatient department to make the service much more efficient, to actually look at how we could do parallel clinics . . . we actually were successful with the business case for the reconfiguration of the unit because we could see that that would make some financial benefit, but also it would actually improve patient experience because we could run parallel clinics. The one that didn’t go forward was the intra-operative assessment analysis, and that’s the one that we’re putting forward this year . . . Because there’s only a finite amount of money . . . and that scheme was not given priority . . . It would be quite surprising to get support for more than one big business case actually in one directorate.

Operational manager for breast care, IS1

One reason why it might not get prioritised was because, financially, BLNA presented considerable risk to trusts that chose to adopt it. Most obviously, this risk arose from the apparently inequitable distribution of the overall reduction in costs that BLNA was capable of delivering: put simply, under the PbR regime, the commissioners (i.e. the PCTs) stood to gain from the reduction in bed-days and theatre time, whereas the providers, the trusts, stood to lose income for the same reasons. The difficulties of making a business case in such circumstances were acknowledged by clinicians:

For all it’s said to be a cost neutral exercise, in that one isn’t more expensive than the other, the trust actually loses money in that they don’t get the income from the second operation anymore . . . it does mean that from a financial point of view and a service provision point of view . . . there may be some reticence from a hospital to lose income, even if it’s also less work.

Consultant breast surgeon, IS1

This is what this is about; it’s a two-way process here. It’s about some of the issues that I have in presenting a robust business case to get this forward. So there are lots of things that we all recognise is best for patients – there’s no doubt about it – but it’s trying to actually recognise the revenue consequences of this and the potential loss of income to the service as well . . . so . . . I have to actually argue it from a patient experience point of view, you know, but also from a financial point of view. So it’s trying to join those up.

Operational manager for breast, IS1

One strategy was to look for savings elsewhere, but this was not an easy task:

But actually, as a pragmatic business manager, you have to say that if this is going to cost you more and you’re losing income, which you are with this case – there’s no doubt about it because these patients would be coming back in or some of them would be coming back in for a second procedure – the cost of it is more because the assay alone is £250 a time, plus all the lab costs. Everything is going to cost you more to run this. So the only way we could take that hit is either to get more money from the commissioners or to make savings within our own department to offset the cost. And that’s the difficulty we have; how can we identify savings to support this?

Operational manager for breast, IS1

The lack of a national tariff for BLNA was identified as the underlying obstacle to trusts and PCTs being able to agree to adopt it:

What we did have to do was put a business case together for the Checklog system and convince the trust’s Planning Committee that that was a good idea. Clinically and financially. In fact I think they got their sums wrong and we probably shouldn’t have been doing it! What I hadn’t realised was that saving the NHS money did not necessarily save the trust money. It saves the PCTs money but the PCTs don’t pay that money to the trust so the trust actually loses money. It is much better for the trust to do two operations rather than one because they get two tariffs. Which is one of the crazinesses.

Pathologist, MS

Not surprisingly, the potential for a financial conflict of interest between service provider and commissioner was seen as being contrary to patient interests:

Throughout the Health Service there’s no global view. Every section has to worry about its budget so that if one section saves money for somebody else they may be penalised but the other section benefits, whereas what actually matters is that the patients are getting a better service and potentially saving money as well.

Lead pathologist for breast pathology, MS

So, unless some sort of arrangement could be reached with the PCTs, the trusts might find the implementation of BLNA unsustainable:

From what I’ve seen a lot of the barriers though are purely financial, the arguing between PCTs and who’s going to fund this and yes you’re saving theatre slots but what are you going to fill them with and we’ll have to pay for more operations.

Biomedical scientist, IS2

This problem could only be resolved through negotiation between the trusts and their PCTs of a local tariff or pass-through payment. Much would then depend on the relationship between the trust and its PCTs. At IS3, this did not prove to be an obstacle:

And we came together with well what on earth, there must be a mechanism under the tariff in some way of introducing an incentive to put this in place . . . a pass-through payment. . . What happened was our finance person contacted the strategic health authority who contacted the Department of Health to find out how we could work this, because I think initially for us in the first year it was going to be more expensive, but less expensive year two, three, four and five. So we had to justify that. And it is called a pass-through payment. And it is to do with innovation. I can’t remember how it worked.

Assistant director of commissioning, IS3

Implementation site 1 (IS1), although it had a good relationship with its PCT, perceived that negotiating a pass-through payment would be difficult:

I would say we have a good relationship with our commissioners, but we all recognise that we’re not in a time of plenty . . . So it’s about trying to do the best you can with the limit you have . . . The best outcome would be if the commissioners said, ’Yeah, we recognise it’s best for patients. We accept it’s a cost pressure to you. We will give you a local tariff,’ which is over and above what we get for PbR, then we’d be fine . . . But it’s not going to be that easy because the commissioners have an efficiency target to meet as well and, you know, they’ve got lots of other services that are coming to them in the same way with innovation.

Operational manager for breast, IS1

The unpredictability of the outcome of local negotiations between trust and PCT was an obvious obstacle to trust managers, who questioned the lack of push for innovation at a national level:

You know, so why isn’t there a dialogue at a national level between the PbR national guidance around tariff and innovation? So if you could have done that in the very early stages to say this is a new innovation, you know, we recognise it costs more than the current one but the patient experience is so much better, if it would actually come with a new tariff, then, actually, you’d be halfway there . . . I think if it were more joined up and if the innovation came along with guidance around how you could maximise your income through an appropriately agreed national tariff recognising that innovation, then it would save any issue and any dialogue with the commissioners locally because it would be very clear that this was a worked up clear cost of this procedure and the benefits would be recognised as well. So definitely . . . it would reduce the barriers.

Operational manager for breast care, IS1

Loss of income was not the only financial challenge faced by the trusts. All trusts faced having to make a significant capital investment in assay equipment (for which the favoured solution seemed to be breast cancer charity support), training for pathologists and, in some instances, refurbishment of accommodation for the equipment and pathologist. Recurrent costs included the purchase of reagents for the test. In addition, there could be additional staff costs. As noted earlier, more breast care nurse time was needed for counselling patients and, in some trusts, the recruitment of additional pathology staff would be needed to cope with the increased workload:

And staff as well, I mean we are without a doubt having to struggle badly for staff when this gets implemented, 3 days a week with a qualified member of staff out to satellite lab all day, it’s going to hit us hard.

Biomedical scientist, IS2

These various financial factors combined to make the creation of a credible business case challenging for some of the trusts involved. For the MS, however, it seemed that a different culture prevailed. Here, the adoption of BLNA was spoken about as being consistent with the trust’s vision to become ‘a world class hospital’, whose pursuit required an active approach towards innovation:

I think one of my roles as General Manager is to kind of unblock that inertia and actually encourage people to come forward with ideas and innovation.

General manager, MS

In the view of the MS clinical lead, this research culture was under threat and this would have repercussions for innovation in the NHS:

The history of innovation is that it tends to happen within very large hospitals with lots of academics . . . If you lose your academics . . . you lose the ability to innovate because you’ve lost the people who will do it. And the people that are left are not paid to innovate, they’re paid to push patients through the door. Many of them are highly intelligent individuals who would like to innovate but the moment they come up against the hurdles that are put in their way in terms of committees and requirements to go and sort out things and, the treacle is such that you just don’t start.

Clinical lead, MS

Dealing with technological uncertainty

Breast lymph node assay adoption was surrounded by uncertainty around the supply of the necessary technology. Initially, two different commercially available systems were selected for improved detection and analysis of breast lymph nodes: Orglog, a technology which uses a Japanese technique known as one-step nucleic acid amplification (OSNA), and Checklog, a technology (US-based company) using the technique of quantitative reverse transcription-polymerase chain reaction. Both rely on an existing procedure known as a sentinel lymph node biopsy, a selective but accurate method where tissue is surgically removed and sent to a pathology laboratory for tests. The MS chose Checklog. The implementation trusts were testing both technologies in order to decide which to adopt.

At IS1, the decision over which technology to choose was taken by the lead clinician who would be responsible for the BLNA service in the trust. As the laboratory manager for cellular pathology explained:

It was based purely on what was recommended as he’s the surgeon. So he would have to be happy with the results and it was his decision which one we went on . . . I don’t know, that would have to be up to our procurement department. Again, which is – I don’t have much input into it. Because my role in it if it had happened would be to provide the technical support. The interpretation of the results were going down to the surgeon, so he had to be happy with whatever process was in place. And therefore I left it very much up to them to decide what they wanted to do, and my job would have been to literally do the tests and make sure we have the money in the budget to cover it.

Biomedical scientist, IS1

However, while the trusts were, in some cases, still deliberating on which system to choose, Checklog withdrew its product (DNAseek) from the market in August 2010, citing financial reasons, leaving trusts favouring this system with a major problem. As no intellectual property rights were in place, trusts favouring the Checklog system would have the option of developing equivalent systems in-house. IS1 was interested in adopting the Checklog system as it had already concluded that it was superior to Orglog.

We knew that [xxx] had been doing the single probe technique . . . So we knew that was available. But we thought that the two-probe technique, the Checklog one, was superior. In fact, we had some meetings with the Checklog representative. We’d had those meetings before the bid from the HT [NTAC] came in because we’d heard from the research side that this was a new technique, and so we’d engaged with the Checklog representative. So we were aware of it, but we’d built the business case around the Checklog rather than the OSNA one, which was, in retrospect, probably not such a wise idea because we didn’t know [large company X] were going to pull out.

Consultant breast surgeon, IS1

For those trusts not in a position to provide their own version of the now unavailable test, the option was to go with what (for some) was the second choice technology:

. . . [Mentor site] is unusual, it’s very well staffed from a pathology point of view and also it has its own molecular unit, you know, which is fairly unusual, and we have nothing like that. So . . . they are able to continue with their existing expertise and their existing machinery. And one or two other places, a couple of other places, have linked up with them to keep what I think essentially is the DNAseek system going, even though that’s now withdrawn and the consumables are withdrawn, they’ve been able to produce their own consumables. We are unable to do that. That’s because we don’t have the technical expertise and certainly our laboratory manager says they’re not going to do it and it’s a technical function rather than a medical function anyway, so, I mean, I wouldn’t actually push them to do that. I don’t think they’re capable of doing it.

Consultant histopathologist, IS1

Implementation site 3 (IS3) had also initially selected the Checklog system, having rejected Orglog because of its perceived failings. However, unlike IS1, it had the skills needed to develop its own in-house assay and, at the time of interview, this was in the process of being validated:

. . . mainly because the commercial assay has been withdrawn, there isn’t an assay available at the moment apart from the OSNA system. The OSNA system has major failings, major failings on the basis that it’s a single-marker assay at the moment, there is no internal quality control.

Pathologist, IS3

Reconfiguring clinical workflow and practice

Successful adoption of BLNA would only be feasible with adaptations being agreed in the clinical workflow and in the working practices of trust staff. How significant these adaptations would have to be depended on circumstances at each implementation site. As a consequence, for some sites, this necessitated a lengthy series of negotiations between the different staff groups involved – surgical teams, pathology and nurses64 – and this impacted on the making of the business case:

So I guess the complication with this is that you’re working with the labs and you’re trying to pull the business case together for lots of different factors, whereas generally, as a business case, you’d write it for yourself and your own department. You’re trying to sort of pull it together from all different areas.

Operational manager for breast, IS1

For pathology laboratory staff, these changes included having to learn a new procedure against a background of increasing complexity of their work:

Yes, complexity of work is increasing constantly. When I started, a typical mastectomy for breast cancer might have five paraffin blocks of tissue taken with an H&E on each one, now well [xxx] will be able to tell you more exact figures about the kind of blocks and complexity on the breast work, but it’s what, 20 plus blocks?

Biomedical scientist, IS2

In the case of Checklog system (before it was withdrawn), pathologists at the MS received training from the company, a factor that helped smooth its adoption.

Most notably, pathology staff would now have to be available to do test tissue samples on demand, a change that increased the pressure on them in two ways. First, the test had to be completed within a short time frame, which meant that a pathologist had to be available but also meant that they needed to get the sample to pathology with minimum delay:

. . . the big thing for us is we knew it was going to be a high pressure test with a lot of pressure on a biomedical scientist to provide that result accurately and quickly with a surgeon banging on your door.

Histopathologist, IS2

Where pathology was in the same building as theatre, this was not a problem but this was not the situation at some trusts. For the latter, one solution was having runners available in the pathology laboratory to collect samples from theatre as soon as they were ready. Another was to have the pathologist assigned to BLNA testing located close to the operating theatre suite.

At the MS, this was not a problem as space was available next to the theatres:

We had space to set up the laboratory next to theatres there, so one of the things that has been different about us is that we’ve always had this done in theatre.

Pathologist, MS

At IS2, this meant adapting a room close by for pathology use, Of course, this had cost implications:

. . . there’s a room identified . . . that can be used yes so that’s why we are going to build, well they have to remove this and make it into a proper lab yes and that is going to cost about £14,000.

Clinical lead, IS2

A consequence of physically relocating the assigned pathologist close to the theatre was that the pathologist would be deprived of the support of colleagues that they would normally be able to call on when faced with a difficult diagnosis, thereby increasing the risk of error:

. . . there will be a single band 6 or band 7 biomedical scientist isolated away from their colleagues in a room with a surgeon wanting a result. They’ve got no, here, if you’re struggling with a frozen section because it’s difficult to cut you can call on a colleague, maybe somebody more experienced, can you come and have a look at this and help me out. There, you won’t be able to do that, you will be isolated and that’s, to me is the real difference . . .

Histopathologist, IS2

Finally, dealing with the new procedure raised concerns that pathology staff at some trusts would be unable to carry out their normal workload while they are testing the assay, and hence would demand recruitment of additional staff, further escalating the cost implications.

The other significant area where adaptations in working practices were called for was, naturally enough, in theatre. Extra time would be needed when the BLNA assay was positive in order to do the axillary clearance, making the management of theatre lists more complicated. However, although this was widely anticipated by surgeons and theatre staff, how these changes should be best managed was a matter of continuing debate and, to some extent, trial and error, and would not be fully understood until the new procedure was introduced. The problem essentially was one of theatre list management and, specifically, accommodating the extra procedures (and their effect on operation time):

Theatre planning, in terms of advance planning we don’t know how much of that theatre capacity will be needed. Like any organisation we’re trying to reduce our costs, which means we’re trying to get a better throughput in theatres, this conflicts with that slightly, in terms of having to allow the flexibility to, to have an extra 15 or 20 minutes on the session for each of these procedures, not all of which would be used . . . the figures I’ve got is 28 to 30 per cent would need the further surgery.

General manager of surgery, MS

He went on to explain the additional complexities of theatre planning introduced by BLNA:

So there are going to be some patients that will need the clearance, and others won’t, so what are we going to do about that? How do we plan theatre time to make it as efficient as we are now because we expect to be running at, at least, a minimum of 90 per cent efficiency? . . . How do you plan a list when you don’t know if a woman’s going to be on the table for another forty minutes? And I know the consultants say they can do some other work, but it’s still not going to make efficient use of the theatre if they’re carrying on doing other things.

Operational manager for breast care, MS

Given the concerns over the implications of BLNA for efficient theatre utilisation, getting theatre staff on board was perceived as the biggest barrier for some of the participants. As a consultant pathologist at the MS commented:

It became very difficult to get theatres to agree to go live, to actually start acting on our results . . . because if the result is positive, then further surgery will ensue, this could go on for another half hour to an hour, if you have three cases on the list, and we do, in fact we’ve had four, if they’re all positive then that list is going to overrun . . . That was my biggest barrier, was getting the theatre staff on board. Once they were, that was it, plain sailing, you know, we’ve not really looked back.

Biomedical scientist, histology, MS

Various adaptations to theatre procedures were being experimented with as a way of mitigating the knock-on effects of additional surgery:

It’s 40 minutes longer anaesthetic and it’s 40 minutes for the surgeon to twiddle their thumbs and surgeons don’t like twiddling thumbs. So we . . . we work that round is that the sentinel node is the first thing that’s done, so you go in, you do your sentinel, you send it off and then you tackle whatever breast pathology you’re dealing with from the breast, whether that be mastectomy.

Surgeon 2, IS3

One approach was to assume that test results would be negative and to ‘close up’ patients while waiting for the result:

So we’re actually already closed up and then we’re just waiting and usually it’s another five minutes once you’ve done that, just waiting for the phone call. The phone call comes, negative, wake the patient up. The drip is already on, the stitches . . . the wound’s already closed, bang, gone. If the phone call says it’s positive, then basically just get gowned and gloved, the registrar’s there, takes the dressings all down, cuts the skin and stitches back, we’re straight back in. You do the clearance, you’re adding another 20 minutes, half an hour for a clearance on top.

Surgeon 2, IS3

However, such ‘in surgery’ workarounds could not be guaranteed to be foolproof:

You get the list and there may be two sentinel lymph nodes on there, [say the] first and second cases and you know you think: Excellent, send the person over and then they can come back when they’re finished. But something happens and for some reason, either the patient’s cancelled or they get pushed down the list and then you’ve got the person waiting over there to do the assay not doing anything else and of course if they get pushed down the list then it’s basically wasted time.

Consultant histopathologist, MS

[If] it finishes late then you’ve over utilised your list and conversely if you’ve got three patients on and you’re expecting one of them to be positive and none of them are positive you may underutilise your list so there are you know its swings and roundabouts.

Consultant breast surgeon, MS

Hence, a question that preoccupied the surgical teams at each trust was how best to organise the theatre list to accommodate the possibility of a patient requiring a second procedure while limiting the disruption to the list as a whole. The exact solutions they came to varied between the trusts, though most were using a similar technique, based on trying to predict which patients would be more likely to test positive for spread of cancer and using this to order the operating list.

At IS2, the surgical team had developed heuristics that they believed would enable them to anticipate – and, hence, mitigate – the impact:

If somebody has a high risk tumour, a grade three tumour which is large it is more likely that they are likely to be positive than somebody with a smaller tumour which is grade one or grade two. So you will want to leave the grade three ones towards the end of your list rather than, because if you do the grade three ones in the beginning and then it’s positive then you’re going to spend more time doing your list by about an hour.

Clinical lead, IS2

A similar procedure had been discussed at the MS but not, as yet, adopted:

In Cardiff it was being used . . . there was talk about stratifying patients as to in their risk groups as to how likely the patients were to be positive or negative and then planning this [theatre list] appropriately.

Consultant breast surgeon, MS

Other factors to be taken into account in theatre list management emphasised how the success of BLNA depended on a closer co-ordination between theatre and pathology teams, in that the hours in which pathology staff would be available could not be assumed to extend over the whole of the time it might take to complete a theatre list:

And also you want to do it within time because the pathologist maybe has to leave at five o’clock so you don’t want to leave your last sentinel lymph node later than four o’clock. So you want to do that within time and this procedure has such a crucial working relationship with the pathology department actually.

Clinical lead, IS2

Hence, the solution adopted by one trust was to schedule patients judged likely to require a second operation at the start of the list:

It does, it just means trying to get to a new optimum way of working and trying to use a proper sensible basis to try and predict what is likely to happen and tailor your list accordingly.

Clinical lead, IS2

However, theatre staff would always have to be ready in case things did not go according to plan. In extreme cases, there might even be a need to cancel an operation if time was running out, but this risked disruption to wider theatre list planning, quite apart from the stress and anxiety inflicted on the patient.

The final element of the impact of BLNA adoption on work practices concerned the role of the breast care nurses in dealing with patients and also had cost implications:

But the breast care nurses would like to be there for the patient when they’re waking up so that, you know, they’re waking up and they know the prognosis is not so good, but they’ve got someone who they know and who has some expertise. And so that’s time and money for the breast care nurses as well.

Consultant breast surgeon, IS1

Not surprisingly, given the above, those championing the adoption of BLNA found making a business case a difficult proposition and one that, in the case of the implementation trusts, the HTWT guide did not necessarily prove to be of much assistance.

NHS Technology Adoption Centre and the How to Why to guide

Several interviewees at the implementation sites remarked that the HTWT guide had been of little help in tackling the hurdles that they faced in getting approval for BLNA adoption and they did not expect it to be of much assistance elsewhere:

Well, the business case is normally written by our manager with advice from ourselves. So I provided the information from the How To Guide and the empty boxes that you have with it suggesting that they fill in the numbers around our case. But they felt that that didn’t mesh with how business cases are submitted within this trust. You couldn’t just paste the How To Guide . . . And so that sort of didn’t progress.

Consultant breast surgeon, IS1

Discussing the value of the HTWT guides, the operational manager for breast care at IS1 commented:

And that’s the problem, you see. All the arguments that they give for bringing this in except, you know, the patient experience – are all the issues I have and it doesn’t tell you how to overcome those issues . . . It was just a reminder. I mean, actually, most managers, if you’re used to writing cases, would be able to know how to be able to write a business case. And obviously each organisation does things differently, but it actually just reminds you of what you should be considering. But if you know your service, probably the information – It’s just a confirmation really. I didn’t find it useful. I didn’t learn anything from it because it actually didn’t give me the answers to my own problems locally. And the things that they suggested, I already knew. So it was more difficult.

Operational manager for breast care, IS1

One specific issue raised by the participants in the implementation trusts was the collection of evidence that the guide stipulated would be needed to underpin the business case. This was a problem because the data required were not routinely collected and doing so would demand a significant effort which the trusts could not resource:

[xxx] wanted some data, the breast surgeons told him they didn’t have access to it and he sent an email to me saying, surely you must be able to get this data, it’s easy and essentially it wasn’t easily because we don’t routinely keep that data. It was how many . . . positive nodes wasn’t it, they wanted a number putting on that, so that would mean like reading through 300 patients.

Biomedical scientist 1, IS2

Reports for every mastectomy, wide local incision, so I said, well if you’ve got the staff, send them, come and look through them, feel free, we haven’t . . . And there was a move afoot at that meeting where they want to collect as much data as they possibly could and set up centralised data collection. One trust, one site would be responsible for collating all this data.

Biomedical scientist 2, IS2

Crucially, NTAC was criticised by some participants for not having involved PCTs in the development of the BLNA guide:

And I thought right from the outset, when we put in our application, it was absolutely key to have your commissioners on board with you on the project right from the start. And I think that, again, we were fairly unique within the group when we went – when we had our first NTAC awayday, the others all sat up when they saw [xxx] there. They said god, why didn’t we think of having a PCT lead on our project straight away, because I mean let’s face it, they’re the people that are paying for it. So, you know, you’ve got to have them on board, engaged in what you’re trying to achieve right from the start.

Clinical pathologist, IS3

Overall, participants acknowledged that being involved in the BLNA project had brought advantages. It had created the opportunity to work with staff from different trusts, had enabled comparing of practices and exchange of ideas about solutions to problems:

I think the premise of NTAC is very good for getting different areas within a trust, the clinicians, pathologists, pathology staff together whereas otherwise they might have never of had a meeting, oh no I can’t do that, you know how these things work but because they were being pushed into it externally it’s more likely to actually happen and then also you’d got those groups of people that were then forced into meeting and working together within the trust, being forced into meeting and working with similar groups of people from other trusts as well.

Biomedical scientist 2, IS2

However, there was scepticism that this mutual learning would have any transferable value for trusts in dealing with what many perceived as the major stumbling block for BLNA adoption – the lengthy and difficult negotiation process between trust and PCT:

I think the aim if I’ve understood it correctly was that by the four pilot sites doing what they did they would end up assessing a new technology which for anybody else that wanted to take it up and run with it all they had to do is to pick this up, take it along, get it agreed and you could roll the technology out very quickly; and I agree that that is a laudable aim . . . I just don’t see having been through the process on several occasions you know with business cases, it always seems to come down to those local negotiations that are going to go on between the PCT and the trust and that in reality is what’s taken the time to get sorted out.

Directorate manager of pathology, IS2

I think we could have achieved what we needed to achieve anyway. What did come out of it from NTAC was hopefully a guide to others as to what they were going to have to go through which I think will be a useful guide but I’m not sure that that is going to stop the same problems that we’ve experienced because you’re still going to end up with the same negotiations, the same discussions with a different set of PCTs . . . it isn’t going to stop the interactions that have to be gone through with the PCTs and the organisation.

Directorate manager of pathology, IS2

With respect to the HTWT guide specifically, concerns were expressed that the low level of awareness of NTAC and its activities would blunt its value:

I don’t think a lot of people have heard of NTAC. People have enquired about the assay, I’ve pointed out the guide, and it doesn’t seem to be that anyone’s heard of it or known they can access it, so there’s definitely a problem it’s all been very worthy stuff . . . and then hardly anyone in the NHS even seems to know it exists. They need to get a higher profile really.

Biomedical scientist, histology, MS

Discussion and summary

The adoption of BLNA was arguably the most complex of the three case studies we undertook and this clearly impacted on the speed with which the implementing trusts were able to complete their projects and, because adoption was contingent on local factors, it also impacted on the way in which they went about it. Yet, at the start, the prospects must have seemed good, the basic case for adoption was strong and, when looked at from the perspective of the NHS as a whole, unambiguous and adoption was consistent with NICE guidelines for breast care. As we also saw from the interviews, clinicians were unanimous about the patient benefit and their role as organisational champions clearly made a difference. 72 However, staff who had direct contact with patients (i.e. breast care nurses) were not so confident that patients would be able to grasp this when they were already likely to be experiencing anxiety and pressure.

The complexity of the BLNA case study arises from a number of organisational factors78 that came into play at the trust level (i.e. financial implications and risks;58 adaptation of clinical workflow and reconfiguration of surgical and non-surgical work practices,73,119,121 some of which required negotiation across professional groups;64 and technological uncertainty). Of course, it is these trust-level challenges that HTWT guides are designed to tackle, but the effort of translating a generic template for innovation so that local factors could be taken into account was time-consuming. Making a case for adoption demanded significant effort if trust management – and other stakeholders – were to be convinced that this could be done within acceptable levels of risk – financially and operationally. Trust management also had to weigh up the relative benefits of different service development options and possibly prioritise them.

From a financial perspective, PbR meant that trusts might not themselves experience the benefits identified by NTAC, at least in the short term. To add to the challenge of making a business case, there would also be costs of equipment and staff training. To help with the presentation of the business case, HTWT guides include a business model template that set out in a standardised, tabular format the key elements to be addressed. Reflecting the complexity of the BLNA business case, its HTWT guide also included a costing model to help trusts estimate the financial implications of implementation. We found no evidence, however, that it was used by the trusts participating in the BLNA project.

At the clinical pathway level, reorganisation of theatre lists was necessary as the intraoperative testing has the potential to cause both under-runs and over-runs of theatre lists. How this reorganisation would be done would vary from trust to trust. In addition, for some trusts, these changes could result in additional costs, for example the recruitment of new staff or the provision of laboratory space close to the theatre. Service level adaptation was required not only for breast cancer surgery itself, but also for outpatients, follow-up clinics, bed use on main wards and other services which shared the same facilities.

A key area of non-surgical work practice reconfiguration was in pathology. The need to train pathology staff in the skills required to perform the assay was relatively easily satisfied. Trusts found it less easy, however, to meet a core requirement – rapid testing and reporting back of results. Ideally, the assay would be done in theatre, in an adjoining room or, at least, in the same building. However, not all of the trusts could satisfy this requirement without making significant changes to pathology work practices (e.g. creating a new ‘satellite’ pathology laboratory next door to theatre; using ‘runners’ to take the biopsy to the pathology laboratory and return with the results, which added to costs of implementation). Moreover, as a histopathologist would now have to be ‘on-call’ whenever a breast cancer surgery list was scheduled, there were concerns that pathology staff would be unable to carry out their other duties at these times. Pathology staff themselves expressed reservations about the impact of time pressures and, in some cases, isolation from other laboratory members on the reliability of results. The workplace studies research literature on the importance of colocation for informal collaboration and the role of the latter in the routine achievement of dependable work in medical and other work settings (see, for example, Luff et al. 208 and Buscher et al. 209) substantiates their concerns about the risks involved.

Finally, the BLNA project was affected by uncertainty around the supply of the technology. When the project started, supply of the technology appeared secure, hence NTAC would have had no reason to see this as a risk factor. By the time one of the suppliers withdrew, the project was well advanced and the trusts were able to adapt. However, this contributed to delays in implementation as the pathology labs re-evaluated their options.

In summary, the barriers faced by the trusts participating in the BLNA project can be attributed to a range of factors: funding issues (i.e. financial reimbursement mechanisms, paying for additional pathology and nursing staff); the reconfiguration of services in order to fit with new clinical procedures (i.e. surgical lists, bed use, reorganisation of pathology work to ensure good access to testing facilities from operating theatres); user training and acquisition of new skills (i.e. for pathology staff); and uncertainties about the availability of the technology itself. In ANT terms, they can be understood as outcomes of complex interactions between heterogeneous networks of human and non-human ‘actors’. 84,88,89 In the BLNA case study, we can identify two key groups of human actors: trust staff (clinicians, pathologists, nurses, business managers) whose alignment is needed to adopt the innovation, and PCT commissioners, with whom the trust must negotiate how the adoption of BLNA will be paid for. For BLNA, the evidence base furnished by NTAC was effective in achieving problematisation, the first stage of ‘translation’, but subsequent stages of interessement and enrolment proved more difficult and contingent on local factors, so that each trust came up with its own solution. First, supporters had to negotiate the prioritising of competing business cases. Second, uncertainties over the availability of the technology had to be resolved and changes in working practices demanded by its adoption agreed upon.

Financial risks mean that business managers might not be convinced that adopting BLNA was in the trust’s best interests. However, as we saw in the case of the mentor trust, where a trust has a strong research culture and reputation to maintain, other factors come into play and tip the balance in favour of adoption. 49 However, there was concern that this research culture was under threat from increasing bureaucracy. Of course, without some trusts being in a position to offer mentoring in innovation, the NTAC model cannot work.

For the participating trusts, the BLNA project had had real benefits. Attending the awaydays enabled staff members to compare practices, to extend their professional networks and to engage in ‘social learning’. 73 However, what was unclear was whether or not there had been useful learning that could be ‘packaged up’ in the HTWT guide for reuse by others in the future, that is to enable other trusts to overcome the barriers and follow the translation process more easily through to successful completion. 45 In other words, the HTWT guide was seen as providing a template for innovation at only the most generic level. In ANT terms, the guide was a useful tool for problematisation, but of more limited value for interessement, enrolment and mobilisation. In the absence of a central push for innovation that would drive this translation process, its value would inevitably be limited. No information is available to us about the subsequent use of the guide by other trusts.

Even if there were a stronger push on trusts from the centre to innovate in the NHS, the BLNA study demonstrates that financial instruments such as PbR would still be a major impediment to securing the interest of all key actors and persuading them to co-operate. 19,52 The impact of PbR makes itself felt when the costs of innovation cannot be contained within the boundaries of the health-care provider (trust) or commissioner (PCT) and a national tariff for the innovation does not yet exist. In the absence of a national tariff, adoption of an innovation such as BLNA by the NHS will depend on local factors, in particular a willingness on the part of service commissioners to find an equitable solution to sharing the costs and benefits. However, when this does not happen, for an innovation like BLNA, it creates a major obstacle to the achievement of mobilisation, the final stage of translation and its adoption across the NHS.

What are the main organisational and decision-making processes and challenges specific to the adoption of the project technologies? What are the barriers and enabling factors?

When any trust is considering the adoption and implementation of a clinical technology they will be concerned about clinical efficacy and utility. Madden48 points out that there are sometimes disputes over how medical devices (including clinical technologies) are best evaluated; she reports that practitioners argue that greater reliance should be placed on the ‘real world’ of clinical experience over and above the ‘proof’ provided by systematic reviews and statistical evidence from clinical trials. The connections between evidence in peer-reviewed clinical journals and clinical practice have long been described as ‘loose’210 and this persists even in the present context of ‘evidence-based medicine’. 211 Even if research-based scientific knowledge is taken up, it is interpreted and often reframed in clinical practice where power dynamics and political agendas between professional groups hold sway. 64,212 So, even when the scientific evidence is clear (and often it is not; rather there are competing bodies of evidence), disputes about efficacy may persist. 212 These complexities are inevitable because evaluation of technologies occurs mostly in practice and, therefore, evidence on clinical efficacy and utility only emerges over time. 19 Clinicians are often more convinced by the reported experiences of their colleagues than by forms of scientific evidence. 48,167 Fitzgerald et al. 212 outline four contingencies that influence the decisions of clinicians when deciding whether or not to use an innovation: evidence of adverse outcomes without the innovation; financial incentives for use or, at least, a neutral position on costs; evidence that other professionals use; and patient compliance and favourable reactions. However, ‘use’ cannot immediately be taken as clear evidence for ‘usefulness’, as some technologies are rapidly taken into practice but then later prove to be less than efficacious. 213,214 Another pertinent evaluation issue is that there may not be a lack of evidence per se, but hard-pressed staff in NHS organisations often do not have the time or skills to identify and evaluate it. 20 Indeed, it is argued that the explosion of medical knowledge greatly exceeds clinicians’ capacities to understand, assimilate and translate into their practice. 211 Expectations over what an innovation can deliver vary according to how close the professionals concerned are to processes of knowledge production. 215

Our case studies indicate that when the clinical utility and efficacy of innovations is still in question, academic clinicians in tertiary care are best positioned to evaluate the innovation concerned as they are most closely involved in knowledge production. As argued above, their expectations over what the innovation can deliver are likely to be more realistic. IPT is a mature technology. Although there were still some issues being raised over efficacy, at the NTAC implementation sites in particular, doubts over efficacy were muted. Indeed, for IPT the main focus was on increasing uptake. In contrast, for UFRI, debates over clinical efficacy and utility were central to both adoption and implementation. In particular, three questions were not fully resolved: where the technology should be situated (in tertiary, secondary or primary care, or all three); how the technology should be used; and what should be the target clinical conditions. Some consensus between the consultant ophthalmologists on the how and what issues was emerging, however, from the four tertiary sites. Academic clinicians in tertiary care were highly instrumental in leading the adoption and implementation of BLNA. Their collaboration was sought by NTAC in gathering evidence of cost savings that adoption of BLNA would generate, and findings from a study conducted by them were subsequently published in a peer-reviewed journal. 216

However, Stevens et al. 217 point out that ‘What is important is that any system for identifying and evaluating new technologies is linked to the system for knowledge dissemination and implementation. It must not stand in splendid isolation’. The UFRI and BLNA case studies indicate that the evaluation of these technologies is tending to stand in splendid isolation. The knowledge generated through use of these technologies within the sites concerned is underexploited because there is no link into any formal system for dissemination. The increasing development of joint higher education and health service collaborations, for example AHSCs, CLAHRCs and, most recently, the new AHSNs, should be a means to rectify this, although there is only limited evaluation of their roles to date. 218,219 However, recent evidence does support the concepts on which CLAHRCs and AHSNs are based. For example, across 12 case studies, Kyratsis et al. 220 found that local, practice-based, peer-mediated networks were seen as important for knowledge diffusion on technology implementation; in contrast, the centralised dissemination of evidence had only a minimal to moderate impact.

This theme of senior academic specialist clinicians driving and enabling technology implementation was echoed in both the UFRI and BLNA case studies. For UFRI, a few consultant ophthalmologists working in specialist centres (e.g. ophthalmic oncology) were working with Eyemap both for research and in the clinic. The issue for wider implementation was that their developing skills and knowledge with the technology remained trapped within their units. There were no formal channels for this expertise to be more widely shared. There was not even an informal network of these tertiary experts. Indeed, one of the consultant ophthalmologists expressed some concern that we might spread his research results to the others, and thus reducing their publishing impact. In the case of BLNA, apart from the role of academic clinicians noted above, clinical champions were identified as essential to success; being able to demonstrate their involvement was a condition for trusts being accepted into the NTAC project. NHS improvement networks (e.g. QIPP, cancer network) were singled out as important dissemination routes. The value of more informal professional networks was noted by our interviewees (e.g. pathology, procurement) for sharing knowledge of techniques and collaborating on equipment purchasing, but clinical networks did not play a significant role.

In the IPT case study, patients were significant in enabling the further implementation of the technology through their demand for the service; indeed, a patient support group (INPUT) was established to advocate for access to insulin pumps in the NHS. Our data illustrate that in some cases it was the push from patients and carers that mobilised NHS provider organisations to address the issue of IPT and develop the skills, knowledge and infrastructure to develop a pump service. Although patient pressure was not an enabler for UFRI, clinicians’ desire in specialist oncology to use the Eyemap images to better communicate with patients was a driver for implementation. Problems in patient communication are very common in medical practice, adversely affecting patient management and clinical outcomes. 221 Failed communication is the most common cause of litigation. 222 Patients appreciate honest information about the costs and benefits of treatment. 223 The Eyemap images enabled the consultant in specialist oncology to give patients realistic information about treatment outcomes which he thought avoided possible litigation.

In the case of BLNA, patient advocacy was mobilised through a breast cancer charity. NICE guidelines for the care of early and locally advanced breast cancer were included in the evidence base assembled by NTAC. The NICE guidance for BLNA is non-specific in that it only advocates minimal surgery. NTAC interprets the adoption of BLNA as compatible with this (see www.ntac.nhs.uk/HowToWhyToGuides/BreastLymphNodeAssay/Breast-Lymph-Node-Assay-Evidence-Base.aspx).

In the case of IPT, in particular, the NICE technology appraisal of IPT clearly acted as an enabling factor as it provided clinicians and trusts that were keen to implement the technology with a nationally recognised (and mandated) source of evidence that they could use to support their case, for example when negotiating with commissioners. Despite the fact that NICE technology appraisal guidance is expected to be implemented in practice/service delivery within a given time frame, the organisations we studied were not meeting the expected levels of pump provision for adults and children when they started working with NTAC. However, their willingness to develop the service (as evidenced by the fact that they applied to be an NTAC implementation site), coupled with the endorsed NICE evidence and the project management support provided by NTAC, created a set of enabling factors that facilitated the implementation of the technology in practice.

A classification of the case study technologies with regard to the enabling factors for adoption and implementation that we found is given in Table 7.

In a systematic review, Robert et al. 76 found that, first, professionalism and, second, power and politics impede the adoption of non-pharmaceutical technologies in health care.

Over the course of successive UK governments, policy decisions aimed at providing treatment for people with life-threatening conditions have resulted in acute hospital care assuming increasing prominence in the overall health economy. 224 Since the inception of the NHS the power of clinicians in tertiary and secondary care has greatly exceeded that of primary care, leading to a profound division between the two. 61,62 This division also occurs in health-care systems in other countries. 225 Accordingly, the ability to command and direct resources lies for the most part with tertiary and secondary care. 58,62 This is unfortunate as it is becoming clear that further increases in the cost-effectiveness of health care can only be delivered if more patient care, particularly for those with non-life-threatening chronic conditions, is undertaken in primary care. 224 Collaboration across tertiary, secondary and primary care is crucial for the successful adoption and implementation of UFRI. In this study, this collaboration was patchy at best. In general, it was not in evidence. Power, status differentials and politics are barriers to more collaboration between primary and secondary care. The most powerful differentiation in professional jurisdictions is between routine and non-routine work: technology can alter the differentiation between routine and non-routine, hence technology can reshape professional work. 126

This potential reshaping was evident for Eyeco. Currently, HSOs can only detect eye pathologies, they cannot diagnose them. Detection is more in the nature of routine work than diagnosis. The greater education of optometrists (and GPs) in diagnosis would erode the power and status differentials between community-based optometrists and secondary care-based ophthalmologists. The current power and status differentials between primary and secondary–tertiary care created particular difficulties for UFRI and the realisation of the diagnostic potential of Eyemap.

Khandhadia et al. 174 comment that the development of Eyemap has created the possibility for community-based retinal assessment. Trials of community-based assessment (as compared with assessment in secondary care) have been shown to increase patient satisfaction, reduce numbers of patients with booked appointments who do not attend the clinic and save resources. 184 Efficacious assessment in community care could also aid in the earlier detection of eye pathologies. The tertiary specialist in ocular oncology who participated in our study commented on this with respect to Eyemap. Use of Eyemap in the community is supported by research studies. For example, Mayers173 states that the Eyeco technology ‘aids in the early detection of a variety of eye conditions including diabetic retinopathy, various forms of macular degeneration, posterior vitreous detachment, retinal holes and tears, hypertension, some types of leukaemia, and retinal detachment’. Education and knowledge dissemination from tertiary–secondary care are necessary for the successful deployment of Eyemap in primary/community care, however. The backdrop for the increased use of this diagnostic technology in primary/community care would be government policy for greater integration between primary and secondary care (see The King’s Fund224 for the issues associated with such integration). Greater integration, involving cross-boundary working, would enable the transfer of some work from secondary to primary/community care. For example, Field226 comments:

For the NHS to survive and prosper and develop, it needs to work very differently and shift care from the hospital to the community. For that to work, we have to build capacity in primary care and that means we need to innovate, not just by using technology but by working differently.

However, currently, as the UFRI case study demonstrates, power and professional politics at the primary–secondary care interface can be significant barriers to shifting care from the hospital to the community.

In the BLNA case study, political issues surfaced most clearly in the relationship between providers (the trusts) and commissioners (PCTs) of services, which, for provider business managers, were defined by PCT unwillingness to share the financial risks of adoption. In their view, this had a powerful, negative impact on the NHS’s capacity to innovate. Instead of using their power as commissioners to drive innovation for patient benefit, with one exception, PCT negotiators were seen by trust business managers as being fixated on contractual details and risk avoidance. Hence, with that one exception, it was the providers who ended up taking on the role of drivers for BLNA adoption and taking on the risks. Delays in proceeding were an inevitable consequence. Whether or not PCT-led commissioning is the right model for all services, especially those of a more complex and specialised nature, has been questioned. 227 On the basis of international evidence, Ham228 argues that the ‘best performing [health care] systems are characterised by integration of commissioning and provision with alignment of incentives for clinicians’.

In relation to IPT, the most apparent barriers to implementation were observed in the NIS, where views about the strength or rigour of the evidence underpinning pump therapy were widely different among the clinicians interviewed; some believed there was a strong case for providing IPT, whereas others questioned its safety and efficacy. In this latter group, objections were also raised in terms of the additional education and nurse support required to put patients on pumps, as this was seen to create inequity by taking resources away from non-pump users. Without external input, such as that provided by NTAC to the implementation sites, stakeholder discussions and meetings to explore differences of clinical opinion did not appear to have happened and the pump service was restricted to a clinical academic who was providing a limited service within an agreed budget.

The policy intent behind PbR is to augment activity and enable cost efficiency. 12,33,58 As yet, there is little systemic research on the impact of PbR on clinical technology adoption in the UK. Work (to date) indicates that PbR is slowing the rate of technology adoption. For example, FitzGibbon et al. 229 identify PbR as a significant barrier to the adoption of a point-of-care diagnostic technology, as the technology increased costs in the short term but this was not included in the tariff. This is clearly the case in relation to IPT, where initial investments to establish a pump service include both staff (specialist pump nurses) and resources (e.g. educational provision for staff and patients). This disincentive to adopt new technologies has been recognised for some time. In the USA, for example, Kane and Manoukian230 showed that DRG payments impacted negatively on clinical technology adoption as the reimbursement did not generally cover the cost of the technology. They used the example of cochlear implantation, arguing that many years after FDA approval, DRG payment was still well below the average cost of the implant; in consequence, many hospitals rationed the device and, eventually, the manufacturers discontinued production. Policy guidance, from the UK Department of Health, allows for pass-through payments (to cover additional costs) to be made for new technologies for a period of 2 years if agreed in advance with the PCT. 231 Pass-through payments are made over and above the relevant tariff reimbursement. 33 In this study, however, few clinicians or managers seemed aware of this possibility (but for one exception see the BLNA case, in Chapter 7).

As noted above, the policy intent behind PbR is to augment activity and reduce waiting times. 12 However, it could also have the perverse outcome of encouraging trusts to maintain (or increase) inappropriate activity, particularly when, if activity decreases substantially, the consequence may be a financial deficit. 232 The possibility of loss of income from reduced referrals may have been a factor in the lack of enthusiasm for UFRI in secondary care. As noted earlier, meeting notes that we obtained indicated that at the site that eventually withdrew there was, initially, a declared intent to reduce inappropriate referrals from primary care and, therefore, focus clinicians’ time on patients who genuinely needed to see a consultant in secondary care. However, concerns about loss of income to the trust were raised. In consequence, for UFRI, the professional power and politics issues that impeded the shifting of care from secondary to community settings (discussed above) may have been further hampered by the possibility of loss of income to secondary care.

In the case of BLNA one of the implementation sites had succeeded in negotiating a pass-through payment. There was, however, no clear process for the trust to follow. The assistant director of commissioning at the trust had contacted the SHA who, in turn, contacted the Department of Health. As a result of these contacts a pass-through payment had been forthcoming. At the other implementation sites, negotiations over how the financial impacts of BLNA would be resolved had not gone so smoothly and were seemingly contingent on the quality of the working relationship between the trust and PCT. Under PbR, providers are paid by commissioners at a fixed tariff according to the volume of activity,233 hence a change to a care pathway that reduces activity (e.g. reduced bed-days) by the provider will lead to it losing financially, whereas the commissioner will gain. In principle, as it would now have spare resources it could deploy elsewhere, the provider could make up this loss by increasing volume in another service. However, as service volumes are set out in contracts, increases in other services would, themselves, have to be renegotiated. A trust director of strategy and planning observed that the PCT was suspicious that if second operations were prevented the trust would use these ‘released’ theatre sessions to do more work that would, in turn, mean that they overperformed on the contract with the PCT.

A classification of the case study technologies with regard to the barriers to implementation that we found is given in Table 8.

Are processes for adoption generic or do the different types of technology require their own processes?

In this study we found certain generic issues, but some problems were specific to the technologies concerned. In terms of generic factors, the following are relevant: negotiating and implementing changes to the patient pathway; identifying training needs and, often, finding the resources to fund these; changing work practices; locating the resources to fund the purchase and use of the technology concerned, usually in the context of a decrease in patient throughput during the early stages of implementation and consequent loss of income under PbR; and, lastly, summoning the motivation, energy and commitment to overcome these barriers in the context of little time to devote to implementation while coping with ongoing clinical work. These generic factors demonstrate that clinical technologies are ‘disruptive’ in two senses. First, they have the potential to bring about step change improvements in patient care through disrupting traditional and sometimes out-of-date ways of working but, second, because they are disruptive to work practices they can elicit resistance. Theories of disruptive innovations in health care have been proposed, arguing that innovations are disruptive in health because they do not translate easily into ‘value-added’ business models. 234,235Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS1 suggests that ease of implementation influences the rate of implementation, so very disruptive technologies may take much longer to implement.

Current government policy as expressed in Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS1 refers to ‘. . . game-changing innovations that change patient pathways and traditional delivery systems and . . . strip(s) out the processes that no longer add value’. This rhetoric neglects the issue that technologies have to be adopted and implemented by NHS clinicians (and managers) in the face of all the generic barriers identified above. Another rallying call is: ‘We need to make innovation everybody’s job from the top to the bottom of the NHS’. 1 At the moment, innovation is in nobody’s job description: this is a generic barrier. In the case of new clinical technologies, the responsibility to adopt and implement is assumed by intrinsically motivated technology champions, usually clinicians but sometimes managers. Indeed, if technology implementation was ‘everybody’s job’, it would be no one’s responsibility as specific decision-making rights have to be delegated (or assumed) for technology adoption and implementation to progress (cf. Brunsson;236 Ahrne and Brunsson237).

Our findings indicate that the NTAC approach to technology adoption and implementation was effective at addressing generic issues around changes to the patient pathway, training needs, rethinking work practices, harnessing energy and commitment for implementation and, to some extent, funding the purchase and use of the technology concerned. To overcome these generic issues, NTAC employed the processes of project management and stakeholder engagement. As discussed earlier, NTAC established from the beginning that their work at the trusts was an implementation project rather than a trial. This did appear to ensure commitment and focus energies on a timetable with an end date and, for the MSs, the production of a HTWT guide. Also in relation to the issue that technology adoption and implementation is in nobody’s job description, the project implementation manager role supplied a person around whom energies for implementation could coalesce. In a similar manner to a management consultant, this person was an ‘objective’ outsider removed from the trust’s internal power and politics agenda and so, possibly, was more able to press ahead with implementation. Attention to stakeholder engagement ensured that all relevant parties were ‘at the table’; apparently without NTAC’s input, commissioners had been neglected, resulting in problems with funding. NTAC introduced a structured implementation schedule including meetings, presentations and awaydays, the formality of these processes imposed a dynamic that gave implementation an additional impetus alongside the day-to-day clinical use of the technology concerned. Participants also had the opportunity to share experiences, compare practices and learn from one another.

For IPT, we found that, in the two implementation sites studied, the processes underlying project management, stakeholder engagement and structured implementation schedules were largely sufficient to overcome the more generic barriers the sites were dealing with. In both sites, there was a general receptivity to IPT as a technology and a willingness to increase its use. However, the main barriers were around how to get started and how to manage the process of implementation, including the various stakeholders that needed to be involved. This does, however, raise the question of whether or not the more generic support provided by NTAC would work in organisations that would not fit the criteria of ‘early adopters’,27 such as the NIS where there was a greater contestation of the evidence and less receptiveness to the technology of IPT.

As indicated earlier, the NTAC project implementation for UFRI failed as all the interested trusts withdrew. Based on our findings, however, we judge that even if the trusts had remained interested the project would have been unlikely to succeed. The technology was too disruptive. The processes of project management, stakeholder engagement and structured scheduling (to address generic issues) would have been unlikely to solve the complex issues of where the technology is best situated (primary, secondary and tertiary care or all three), how the technology is best used and what target conditions the technology is best used for. The emerging scientific evidence in journals and the ongoing research in tertiary care (including use in the clinic) is resolving the ‘how’ and ‘what’ questions, but the ‘where’ question (apart from in tertiary care itself) is not being addressed. As argued earlier, to reap its potential UFRI requires a close working relationship between partners in primary and secondary care and willingness on the part of tertiary (or secondary) care to educate primary care. Education is required for the ‘how’ and ‘what’ issues; in addition, our research found that there is a long learning curve in interpreting the digital images. Without some integration between primary and secondary care, education on this scale seems unlikely to happen, given that there are, currently, no financial incentives for ophthalmologists in secondary and tertiary care to engage. Indeed, if the use of the technology in primary care reduced the number of referrals to secondary care (which seems likely over time), there is a financial disincentive for the trusts, although there should be savings for the health economy as a whole.

Breast lymph node assay adoption called for a number of significant changes in working practices of surgical and pathology teams. First, it made the management of theatre lists more complex as provision had to be made for additional surgical time, but which patients would need it could not be predicted with complete confidence in advance. Second, it demanded much closer co-ordination between surgical teams and pathologists, as test results had to be available within a very short time period. We saw different ways of achieving this at each trust, some of which were acknowledged to introduce some potential risk to the reliability of testing procedures. In general, the implementation of these changes was contingent on local factors, making it difficult, if not impossible, to propose a ‘standard’ for implementation that could be straightforwardly applied at any trust.

What role does the wider commissioning process play?

Against a background of perceived poor commissioning performance, a programme for ‘world class commissioning’ began in the English NHS in 2007 with the intention of developing and strengthening PCT commissioning of health services. 238 The advent of the Coalition government in 2010 led to the abolition of PCTs and their replacement by the NHS Commissioning Board and CCGs. 239 Despite these structural changes, commissioning continues to face the same challenges. Foremost among these is the ability of commissioners to negotiate on a ‘level playing field’ with the providers of care,240 which, in turn, is dependent on equal access to appropriate information about the cost and quality of health-care services. Moreover, even in a more marker-orientated regime, there are indications that a cultural climate that enables relational (encompassing trust, common values and networks) rather than adversarial (or legalistic) commissioning is more effective. 241 Effective ‘commissioning’ is more than purchasing, purchasers merely buy what is already available or reimburse providers according to their use of the services providers offer. In contrast, commissioning should involve ‘a proactive strategic role in planning, devising and implementing the range of services required’. 242

Commissioning, therefore, should be a driver for technology adoption and implementation, particularly where there are potential savings for the health economy as a whole. Equally, commissioning may involve decommissioning where an activity no longer adds value or has been shown to be outmoded. 1 To better guide the implementation of clinical technologies, Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS1 is introducing a National Institute for Health and Care Excellence Implementation Collaborative (NIC) comprising NICE, the NHS Commissioning Board, the chief pharmaceutical officer, the main industry bodies, the NHS Confederation, the Clinical Commissioning Coalition and the Royal Colleges.

The findings from our case studies indicate that a body such as NIC will be valuable once a technology has been evaluated in clinical practice. The learning curve on clinical efficacy and utility for technologies is such that the NIC will be unable to drive technology adoption until providers have evidence of added value. So we think that the NIC will be important in ensuring diffusion once clinical efficacy and utility are evaluated and the technology is domesticated, i.e. it fits into practices and routines (see, for example, Berker et al. 185).

Overall, our evidence on the role of commissioning in relation to UFRI, IPT and BLNA indicates that commissioners play a purchasing, rather than the broader strategic ‘commissioning’, role in relation to technology adoption and implementation. The lead commissioner for the WIS for UFRI indicated that he had identified ophthalmology (along with orthopaedics and ear, nose and throat) as the biggest referring specialties. He went on to say that he knew that there were ‘lots of’ ophthalmic referrals that could be managed in primary care or by opticians. Yet, when asked if commissioning should drive technology adoption, particularly where the technology concerned could enable the management of more referrals in primary care, he responded that commissioners do not purchase technology, they buy patient outcomes. This approach clearly locates the responsibility for technology adoption and implementation with the trusts. He added that commissioners are not ‘against technology if it improves efficiencies’. Such a line hardly points to a strategic role for commissioning in technology adoption and implementation. It should be pointed out, however, that without a NICE technology appraisal, commissioners are reliant on evaluations undertaken at the trusts for their evidence on clinical efficacy and utility. None of our respondents in the UFRI case study mentioned any current commissioning plans to shift some ophthalmic care out of the trusts into community settings. The only example of such a possibility was given by an optometrist where optometrists and a GP had built up a plan to open a community ophthalmic unit, but the move was ‘stamped upon’ by one of the consultant ophthalmologists at the local trust. The business and development manager at the withdrawn UFRI implementation site had initially contacted the PCT and persuaded them to fund the technology, but when the trust decided not to go ahead there is no evidence that the commissioners had responded with any objections. However, as pointed out earlier, we were unable to obtain the minutes of the meetings between the trust and the PCT over this withdrawal, as both sides stated that they had lost them. Subsequently, we put in a Freedom of Information request but still no documents were forthcoming.

Similar observations of commissioners’ roles were made by trust business managers in the BLNA case study in the sense that the acute trusts approached the PCT to tell them they should be commissioning the BLNA, although the trusts were aware that policy expressed the view that commissioners should drive innovations.

What is the role of the technology producer in supporting adoption in health-care organisations?

Faulkner243 points out that the medical device industry and the health-care state are closely interwoven. Yet, Wanless2 identified the NHS as a ‘slow and late adopter’ of medical technologies and predicted that this would, increasingly, become a performance issue. Sir Bruce Keogh, the current NHS Medical Director, commented in 2012, ‘Even with hard evidence of superior efficacy it generally takes around 15 years, in any health-care system, for widespread adoption of a new intervention’. 244 The role of the producer in supporting adoption is, therefore, potentially important in facilitating the use of efficacious clinical technologies which can leverage NHS performance. In the UK, there are around 2000 companies that produce medical devices; about 85% of these are small firms, so the industry is unlike pharmaceuticals (where companies are large and globalised). 243 As most UK medical devices companies are SMEs, their future in the UK market is rather precarious until their products are fairly widely adopted within the NHS, as private health care is used by < 8% of the UK population. This gives clinical technology producers a clear incentive to engage in strategies to increase the uptake of their devices. Yet, through the present ‘bottom-up’ (i.e. at trust level) implementation process this is not an easy or straightforward task. As there are around 250 trusts in England alone, at the most basic level this could mean 250 contracts to be negotiated and established. One of the former project implementation managers at NTAC plans to start a consultancy advising UK technology producers on how to navigate the complexities of the UK NHS when they are seeking to find a market for their new devices.

Producers have strategies to increase the uptake of their products. Currently, it is common practice for technology producers to provide demonstrations of their clinical technologies at NHS trusts in an attempt to sell the ‘competence’ of their innovations. They also sometimes offer free trials to allow potential users to evaluate and test the capability of clinical technologies. Demonstrations of technology are organised events where the capabilities and features of technology are revealed to potential adopters and users. 245 Such presentations can more resemble ‘infomercials’48 than be driven by evidence-based reasoning. Nevertheless, they represent a defining moment in the adoption pathway, designed to provide evidence to various stakeholders about the efficacy of technologies and systems, either under development or ready for use. 246 However, demonstrations are not simply about evaluating technologies (i.e. providing evidence for or against adoption). Nor do they straightforwardly evaluate suppliers (i.e. assessing the ‘status’ of producer and suppliers’ ‘track record’). They play a crucial role in the ‘negotiability’ of uncertainty and risk of adoption. Indeed, as ‘infomercials’,48 much of the promise of new clinical technologies depends on masking such uncertainty and mobilising a range of claims and expectations about the capabilities and future trajectory of innovation. In this context, it is suggested that expectations mobilised in the representation of technological innovation do not simply describe these future technologies but also help bring them into being. 245,247,248

Clinical technology adoption and implementation both happen in the context of uncertainty. As pointed out above, evidence about the clinical efficacy and utility of technologies is negotiated and evaluated over time as medical devices are used in practice. Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS1 notes that the UK is strong in biosciences (see also Faulkner,243 who states the UK is a net exporter of medical devices) and recommends ‘partnership’ with industry producers, emphasising that the NHS should adopt and implement technologies that improve productivity and add value but not cost. To this end, the following is suggested to ‘work with industry to develop a better “value proposition” for the NHS, relying less on upfront capital or revenue investment, and more on taking income from downstream revenue savings’. 1 Recognising the problems producers have in terms of accessing individual NHS trusts (see above), the government has instituted AHSNs to help with access. AHSNs would function as ‘lead customers’ working jointly with industry to scope barriers and develop solutions to problems with technology adoption and implementation. 1 Better metrics on innovation uptake is also recommended, as is making the uptake of ‘high impact’ innovations mandatory. 1 Overall, the commendation is that AHSNs would work with industry on the ‘evaluation, commercialisation and rapid adoption of health technologies . . . [whilst] creating wealth for UK PLC’. 40

The evidence from the three case studies indicates that the industry producer was actively involved only in negotiating the uptake of the technology for UFRI. Unlike IPT and BLNA, the UFRI technology had not undergone a NICE appraisal – resulting in a guideline – so this ‘push’ factor was lacking. As mentioned earlier, the clinical efficacy and utility of this technology was disputed. For example, the possibility of using non-dilated digital Eyemap images rather dilated clinical examination was debated. In consequence, Eyeco had grant-funded several clinical trials. One of these, the ‘Joslin Trial’ in the USA, recently reported:

Nonmydriatic [non-dilated] ultra wide field images compare favorably with . . . dilated fundus examination [clinical examination of the retina] in determining DR and DME [diabetic retinopathy and diabetic macular edema – both eye conditions associated with diabetes] severity; however, they are acquired more rapidly. If confirmed in broader diabetic populations, nonmydriatic ultrawide field imaging may prove to be beneficial in DR evaluation in research and clinical settings. 183

The tertiary specialists who participated in our research were aware of this study but the relevance of non-dilated vis-à-vis dilated eye examination is more relevant in secondary and primary care, as dilation is the norm in tertiary care. For the industry producer of Eyemap this is the type of research finding for which access to an AHSN, to enable rapid dissemination in secondary and primary care, would be particularly valuable. Eyeco had also funded a US ophthalmologist who was skilled both in the use of UFRI and in the interpretation of Eyemap images to advise any clinicians who had adopted UFRI, but who were on a learning curve in terms of diagnosis from the images. This person gave advice over e-mail after receiving the digital images. Again, within the context of partnership between the NHS and the industry producer, the expertise of this funded US ophthalmologist could be mobilised more widely. As mentioned above, Innovation, Health and Wealth: Accelerating Adoption and Diffusion in the NHS1 envisages industry producers taking revenues ‘downstream’ after productivity gains from the technology have become apparent. The problems with this, from the point of view of the producers (SMEs in the main), are that cost savings are difficult to estimate; such savings may accrue only after many years of using the technology (e.g. for diagnostic technologies, such as UFRI, earlier detection and prevention of disease will reduce health-care costs over the longer term); and, financially, many SMEs are dependent on positive cash flows to maintain the viability of their businesses.

There was no producer push in the BLNA case study, their role was notable mainly for Checklog withdrawing their product from the market just as the project was getting under way. This caused some disquiet initially. However, in the end, the impact was short-lived and negligible. First, there had been no consensus among the biomedical scientists as to which of the two assays was best. Factors that had to be considered included material costs (reagents), precision of the assay (false-positive and -negative rates), complexity of the process, staffing and availability of training, and these weighed up differently in the individual trusts. Two had opted for the Checklog assay and two for Orglog. Second, the impact on the former by the Checklog withdrawal was resolved when Checklog consented to granting them a royalty-free licence to develop in-house a similar assay and to make it available to other trusts. The only problem was that validating the in-house assay would take additional time and so created a delay before it could be used by other trusts.

In the case of IPT, there were a number of different producers in the market for trusts to choose from and procurement of pumps was one of the issues addressed within the NTAC implementation project. Pump companies often provided practical help to trusts, for example through the company-employed pump nurses providing educational support to staff in the initial stages of setting up a pump service.

How does the presence of and intervention by NHS Technology Adoption Centre impact on the process of adoption within the institution?

We have already commented at some length on the role of NTAC in Chapter 4. Above we identify their work practices as project management, stakeholder engagement and structured scheduling. It is worth emphasising that NTAC is not focused on ‘upstream’ technological innovation; it works ‘downstream’ with trusts that have adopted a technology but are experiencing problems and want support to implement that technology. We argued that, conceptually, the NTAC approach can be understood through Callon’s ‘four stages of translation’, where the first stage is ‘problematization’, where an objective (technology implementation) is seen as shared across different sets of actors (clinicians, managers and commissioners), but one set of actors (in this case NTAC) position themselves as key in terms of accomplishing this objective. The second stage is ‘interessement’, where the key actors propose roles for the other sets of actors and gain their commitment to particular courses of action with respect to the objectives. Callon’s third stage is ‘enrolment’, where the proposed roles and courses of action are operationalised and consolidated. Finally, the fourth is the stage of ‘mobilization of allies’ where further actors are drawn into the established networks, and communication among this enlarged group is enabled through ‘immutable mobiles’. Analysis of how NTAC was perceived in the different IPT and BLNA sites does tend to support these concepts. NTAC was regarded as a ‘catalyst’ for action. As an ‘outsider’, it was viewed as being able to ‘force’ different players to come to the table – where, previously, key actors had sometimes refused to participate in discussions on technology implementation. Some participants acknowledged that their own approach to implementation had been ‘ad hoc’; in contrast, NTAC brought a co-ordinated ‘road map’. In addition, some trusts had not thought about involving their commissioners, whereas NTAC identified them as crucial, albeit that some BLNA sites did criticise NTAC for not involving commissioners. In terms of the ‘mobilization of allies’, respondents commented that once their implementation project was under way they were able to link into networks with staff at other trusts (i.e. other NTAC implementation or MSs). This networking enabled constructive exchanges on comparing practices and solutions to common problems. There were some comments, however, that the NTAC work was ‘on paper’ and the ‘real work’ began after their departure. Nevertheless, there was considerable positive feedback from the sites on NTAC’s presence at the trusts and the interventions they made to improve the implementation process.

On the other hand, there were areas where the NTAC approach could be critiqued. First, the success of the implementation projects at the sites was very dependent on the skills and enthusiasm of the individual project manager. There did not appear to be much managerial oversight of the work of these implementation managers. The trust that withdrew from the UFRI project seemed to make this decision on the basis of one particular meeting where the producer demonstrated the technology. There did not appear to have been much work to engage stakeholders in meetings before this demonstration to clarify expectations and work through difficulties. As outlined in Chapter 5, the complexities around adoption and implementation were considerable and the NTAC approach of working within secondary care (rather than across organisational boundaries between primary, secondary and tertiary care) was not really suited to this technology, so the implementation project may have failed anyway, but the way that the project was managed at the start did not seem optimal.

Second, expectations that the HTWT guides could support trusts through technology implementation without NTAC’s active intervention ‘on the ground’ look overly optimistic. In ANT terms the HTWT guide was conceived by NTAC as an ‘immutable mobile’ in the sense that the guide could ‘move around’ from trust to trust and also could achieve ‘universality’ (i.e. immutability) because it embedded the learning from several implementation sites (cf. Law and Singleton249). The responses of trusts who had used the guides, however, indicate that they felt that the guides did not have universal applicability. Respondents expressed the views that the guide did not ‘mesh’ with their particular circumstances, was only a ‘reminder’ of what they already knew and did not give them the answers to the specific problems they were experiencing.

Does the involvement of NHS Technology Adoption Centre have an impact on the sustainability of adoption? Does the technology remain embedded after NHS Technology Adoption Centre withdraws? Can the issues and processes that cause it to continue or fail to remain embedded be identified?

As this research was limited to a 3-year time scale and the IPT and BLNA case studies were undertaken after the research on UFRI, our findings cannot be definitive on the sustainability of IPT and BLNA at the sites. We did not observe any indications that IPT or BLNA would not remain embedded at the sites. The issues with regard to the sustainability of the two technologies were rather different, however. For IPT, the trusts were attempting to increase the proportion of patients on pumps who met the NICE criteria, so sustainability would equate to continuing to increase uptake until the NICE guideline was met or, indeed, overtaken. For BLNA, all women undergoing breast surgery for cancer were offered the assay, so sustainability could not be measured by increasing proportions of women avoiding the need for two operations.

Lettieri and Masella55 drew on 13 studies of technology adoption and implementation to define five dimensions that determine the sustainability of any technology (see also Chapter 2, Variable local decision-making processes):

• financial sustainability (e.g. the ability to sustain cash needs)

• organisational sustainability (e.g. the inertia or resistance to change)

• technological sustainability (e.g. the evolution of technology)

• resource sustainability (e.g. training needs)

• context sustainability (e.g. the demand of health services).

With regard to financial sustainability, none of the sites had negotiated a local tariff with the PCT (or PCTs) for either IPT or BLNA. At the IPT sites, on account of the NICE guideline, most participants reported that the PCT was willing to fund the equipment costs for the purchase of the pump and the consumables on an ongoing basis for patients who met the criteria. However, respondents were uncertain, once the NICE guideline of 15% of patients with type 1 diabetes was met, if the PCT would continue to meet the costs. For one of the sites, a specialist children’s centre, this ad hoc negotiation with the PCTs was a concern as they had 17 PCTs to deal with. Given that the PCTs would fund IPT, none of the sites had submitted a business case to trust management. Most IPT sites argued that they needed a specialist pump nurse to meet the additional patient need for education and advice to use the pumps safely. The PCT would not fund these additional staff costs; the trust had to meet them. Therefore, in terms of financial sustainability, IPT is dependent on the continuing support of the PCT and, to a lesser extent, trust management. For BLNA, financial sustainability, at the level of the trusts, was less clear. NTAC estimated that if BLNA was implemented across the entire NHS, on the basis of ward costs alone, there would be savings of £4M per annum and the total savings would be around £5.1M (see www.ntac.nhs.uk). Nevertheless, under PbR, the trusts lose income as the introduction of the assay means that 25–30% of patients undergoing breast surgery for cancer will not need a second operation. One of the BLNA sites had negotiated a pass-through payment (a ‘top-up’ to the tariff) with the local PCT but at the other sites negotiations with the PCT were ongoing and unpredictable. Financial sustainability for BLNA was somewhat uncertain, despite the potential savings for the whole NHS economy.

Organisational sustainability seems strong for BLNA. The technology inspired enthusiasm and commitment from the clinicians; it was described as a ‘no-brainer’. Also, once joint working between histopathologists, breast surgeons and theatre staff was instituted, the new patient pathway looked embedded. In contrast, the commitment to IPT looked more variable between different clinicians, with some demonstrating high motivation but others showing some resistance, even describing pumps as ‘elitist’ and a ‘lifestyle’ choice for patients. At one of the sites the continuing implementation of IPT had been delegated to the nursing staff.

As a mature technology, IPT had low technological uncertainty. There were several pump manufacturers so availability was assured. In contrast, as a newer technology there were some questions over the availability of BLNA. There were two different assays commercially available, but the manufacturer of the one generally deemed as superior had pulled out, citing financial reasons. Two of the BLNA implementation sites had the expertise to develop their own assay based on their knowledge of the withdrawn system (there were no intellectual property issues). However, the other sites did not have the expertise to do this and the second system was considered somewhat inferior. Hence, there were some unresolved questions over the technological sustainability of BLNA.

As training needs had been met at the BLNA and IPT sites, in this regard, resource sustainability was not a problem. However, the BLNA technology put pressure on human resources within pathology and introduced uncertainty into the management of theatre lists, and some sites argued that additional breast care nurses were needed to cope with the increased patient anxiety. In a similar manner, the resource sustainability of IPT depended on the infrastructure provided by a specialist pump nurse and access to educational resources for patients.

Insulin pump therapy and BLNA had associated NICE guidance, so context sustainability was, to some extent, assured. Context sustainability for IPT was further augmented by patient demand. Clinician enthusiasm and commitment provided additional context sustainability for BLNA.

NHS Technology Adoption Centre interventions focused mainly on organisational, resource and context sustainability. Although they tried to ensure financial sustainability, through involving the relevant PCTs in the implementation projects, our findings show that presence of NTAC was not always successful in ensuring formal arrangements for financial sustainability. Short-term solutions, for example the PCT funding the actual costs of IPT on a patient by patient basis, predominated. Technology sustainability was outside NTAC’s control.

What information can be gathered from the NHS Technology Adoption Centre project to assess the wider impact on how implementation is managed?

Our findings indicate that the NTAC approach is most suited to a technology like IPT, where the problems are generic ones around managing changes to the patient pathway, identifying and meeting training needs, negotiating new work practices, being a catalyst to generate energy and commitment for implementation and, to some extent, funding the purchase and use of the technology concerned. We judge that the resources that can be mobilised by a small organisation such as NTAC are insufficient to tackle a technology like UFRI, which, along with these generic issues, had a contested evidence base and necessitated large-scale intervention to span the boundaries of primary, secondary and tertiary care to overcome resistance generated by professional power and politics. Although the expertise of NTAC was clearly helpful for the BLNA project, their intervention seemed less necessary where there was clear clinical commitment and enthusiasm for implementation. We judge that technology implementation should be finessed dependent on the extent and nature of the barriers encountered, which, in turn, vary with the type of technology. The NTAC approach of working within secondary care on project management, stakeholder engagement and structured scheduling meets many needs but, on the basis of our experience with three different technologies, is somewhat limited and inflexible for some clinical technologies. This reinforces the evidence from the literature, which strongly supports the view that there is not a ‘one size fits all’ approach to implementing evidence and innovation. 250

Meeting aims and objectives

We next discuss how far we were able to meet our aims and objectives. For clarity we set them out again as follows:

The aims of this research were:

• to identify the root causes of risk perceptions over technology adoption in the ‘sponsoring’ trusts

• to map out the network of actors required for successful technology adoption

• to understand the policy, organisational and cognitive barriers and resolve cross-boundary issues

• to assess the extent to which PbR is creating barriers to technology adoption and implementation.

Our specific objectives were:

• to produce recommendations on ‘what needs to change’ for successful technology adoption [such recommendations will seek to change any misguided perceptions of risk through communication and alleviate real risk (e.g. of income loss) through cross-boundary negotiations]

• if PbR funding mechanisms are too rigid, to make recommendations of how local flexibilities can be enhanced

• to identify any new boundary-spanning roles required to facilitate technology transfer along the adoption pathway

• to work closely with NTAC to ensure that this research dovetails with its agenda (the technologies we will investigate have been chosen in consultation with NTAC).

In terms of meeting our aims, technology adoption and implementation always involves changes to the patient pathway and existing work practices, necessitates training and usually involves new infrastructure. These changes result in risk and uncertainties; these have to be negotiated along the lines that we have outlined above. In local terms it was not necessary for us to map out the network of actors needed for technology adoption and implementation because NTAC had already accomplished this. However, these local trust-based networks need to be linked into wider top-down initiatives to ensure that technologies are adopted into the most optimal settings and diffusion takes place to all appropriate organisations. We have provided our understanding of the main policy, organisational and cognitive barriers in the above discussion. The major cross-boundary issue was the lack of integration between primary and secondary–tertiary care. This prevented secondary care providing the necessary education for primary care to adopt UFRI. This issue is intractable to the efforts of a small research team. We were, therefore, unable to resolve it. As discussed above, PbR is a major barrier to the adoption and implementation of new clinical technologies. However, as pointed out, the flexibilities provided by pass-through payments and local tariffs were not taken advantage of. As we discuss, lack of trust between providers and commissioners tended to prevent this. There is a potential role for trust managers in negotiating tariff flexibilities but, currently, few managers occupy these roles. We point to ‘what needs to change’ for successful technology adoption and implementation in our recommendations on best practice below. As mentioned above, new boundary-spanning roles between secondary–tertiary and primary care were necessary for the adoption of UFRI, but PbR prevented consultants from assuming these roles, as there is no payment for them under the tariff. We were able to collaborate with NTAC but, over the life of the research, policy changes restructured the way that NTAC works, rendering its input more directly tied into specific Department of Health initiatives. In consequence, NTAC’s methodology shifted and it was less able to work ‘on the ground’ at the trusts. Our research with NTAC was, therefore, more retrospective than planned at the beginning of the research.

Conclusions

Finally, we note the limitations of our study, outline possibilities for further research and consider if it is possible to define best practice for technology adoption.

Contributions of authors

Sue Llewellyn (Professor, Healthcare Management) contributed to designing the study, data collection, data analysis and report writing.

Rob Procter (Professor, Technology Studies) contributed to designing the study, data collection, data analysis and report writing.

Gill Harvey (Reader, Healthcare Management) contributed to designing the study, data collection, data analysis and report writing.

Gregory Maniatopoulos (Research Associate, Technology Studies) contributed to data collection, data analysis and report writing.

Alan Boyd (Research Associate, Healthcare Management) contributed to data collection, data analysis and report writing.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health.

Background

Evidence indicates that patient safety and proven clinical effectiveness is insufficient to ensure the adoption and implementation of new technologies, and that the organisational and policy context may present barriers which slow or even prevent uptake. Despite government policy, innovative technologies are not yet demand-led through commissioning. Introducing new technologies initially raises providers’ costs as they necessitate training, can encounter clinical resistance, may alter the patient pathway and, therefore, patient management, and can lead to reduced patient throughput. The current funding regime of Payment by Results (PbR) rewards activity and as such it is not surprising that providers see new technologies as risky. Moreover, clinical utility has to be built and demonstrated as technologies are embedded in concrete interdisciplinary work practices along the adoption pathway.

The aims of this research are

• To identify the root causes of risk perceptions over technology adoption in the ‘sponsoring’ trusts.

• To map out the network of actors required for successful technology adoption.

• To understand the policy, organisational and cognitive barriers and resolve cross-boundary issues.

• To assess the extent to which PbR is creating barriers to technology adoption and implementation.

The technologies we are researching

Our focus is on technologies which give rise to the greatest perceptions of risk along the lines outlined above. Discussions with the NTAC led to the identification of three technologies presenting the most complex problems of adoption and implementation. These are:

1. an insulin pump with remote patient management

2. a lymph node metastases diagnostic for breast cancer

3. a system for advanced retinal imaging.

Data collection

We are collecting primary data on adoption and implementation difficulties with the three technologies over nine sites. Structured comparisons will be made between the project sites and the network dynamics within them. There will also be a focus on the specific areas of difficulty at each of the sites.

Your role in the research

To achieve these ends we are conducting semistructured interviews with clinicians and managers. These interviews (we anticipate that at each trust there will be 8–12 key players) take about 1 hour and, if participants are agreeable, we would like to tape and transcribe them. We also hope to ‘sit-in’ on two to three formal meetings if technology adoption is still a substantive agenda item and spend time (1 day) tracking the informal interactions of one to two key players in technology adoption as they go about their normal work or if this is not feasible to ask participants at interview to sketch their personal network maps. Agreeing to participate is voluntary and you (at trust or individual level) can opt out at any stage if you wish.

The benefit to you

Participation in this research should crystallise adoption and implementation problems. If your trust is still experiencing difficulties the researchers act as ‘cross fertilisers’ so that you can learn from the experiences of other trusts. You will have access to the research report and, if you wish, participate in the publications that flow from the research.

We hope that you will be able to participate in this research. Please contact us if you have any questions.

General questions

Q1. Please will you outline your role in working towards the successful adoption and implementation of (name of specific clinical technology).

Q2. What are the main clinical benefits of (name of specific clinical technology)?

Q3. What do you think are the main organisational barriers to adoption and implementation?

Prompt: raises costs; funding issues; organisational inertia; clinical resistance; managerial resistance; difficulty in forming the business case; commissioning issues.

Q4. What do you think are the main policy barriers to adoption and implementation?

Prompt: PbR; organisational targets (e.g. waiting lists); competing policy objectives; rapid policy reform.

Q5. Are there any unresolved clinical barriers to adoption and implementation?

Q6. Do you believe that adopting and implementing (name of specific clinical technology) is risky? If yes, how is this risk perceived? How is this risk alleviated?

Building a support network

Q7. Was there a key individual who championed the adoption of (name of specific clinical technology)? If yes, why was this person so crucial and would the technology have been adopted without them?

Q8. Who are the key individuals in the trust whose support was required for successful adoption?

Q9. Who are the key individuals outside of the trust whose support was required for successful adoption?

Q10. What were the main issues in building this network of support?

Q11. Has this network been sustained? If yes, how? If no, why?

Q12. Were there any individuals who were resistant to the adoption of (name of specific clinical technology)? If yes, why were they resistant? How was their co-operation secured?

Q13. Did successful adoption necessitate the creation of new organisational roles? If yes, what were these roles? Did any of these roles cross boundaries?

Prompt: between the trust and commissioners; between the trust and suppliers; across different specialisms within the trust.

Payment by Results and funding issues

Q14. Do you think that PbR created particular difficulties for the introduction of (name of specific clinical technology)? If yes, what were these?

Q15. Is the trust engaged in strategies to try to resolve these difficulties? If yes, what are these strategies?

Q16. Do you believe that (name of specific clinical technology) will eventually raise productivity? If yes, why? If no, why not?

Q17. Did the introduction of HRG4 impact on (name of specific clinical technology)? If yes, what was the impact?

Q18. Is the trust engaged in any new costing initiatives as a result of the introduction of (name of specific clinical technology)?

User acceptability issues

Q19. Were there clinical user acceptability issues in relation to (name of specific clinical technology)? If yes, what were these and how were these overcome?

Q20. Were work processes disrupted due to the introduction of (name of specific clinical technology)? If yes, what form did this disruption take and how was it overcome?

Q21. Will the benefits of the introduction of (name of specific clinical technology) accrue solely within the trust or will there be benefits outside?

Prompt: primary care; home care.

Patient acceptability issues

Q22. Were there any patient acceptability issues in relation to (name of specific clinical technology)? If yes, what were these and how were these overcome?

Q23. Does the effective use of (name of specific clinical technology) require patient/carer compliance/involvement/knowledge? If yes, what work with patients/carers was required to ensure the effective use of the technology?

Q24. Were patients involved in either the adoption or implementation processes? If yes, what form did this involvement take?

Q25. Has patient feedback on (name of specific clinical technology) been sought? If yes, are you aware of any issues that this feedback raised?

Supplier/industry issues

Q26. Was the trust encouraged by the producers of (name of specific clinical technology) to adopt? If yes, how was this manifested?

Q27. Was the trust well supported by the producers of (name of specific clinical technology)? If yes, what form did this support take? If no, what problems did this create?

Engagement with the NHS Technology Adoption Centre

Q28. How has your engagement with the NTAC assisted in the adoption and implementation of (name of specific clinical technology)?

Q29. Do you believe that there would have been adoption without the NTAC’s support?

Q30. Is your engagement with the NTAC still ongoing? If yes, what form does this take?

Q31. Has the NTAC produced a ‘How to Why to’ guide for (name of specific clinical technology)?

Q32. Will this ‘How to Why to’ guide be of any relevance to you? Or do you think that your adoption process can now be considered complete?

Q33. Do you know if the NTAC has produced any other ‘How to Why to’ guide for another technology which you are considering adopting? If yes, have you downloaded this guide? How has the guide helped you?

Closing questions

Q34. Do you believe that the difficulties the trust experienced in adopting (name of specific clinical technology) are specific or would they apply to other technologies?

Q35. Do you think that best practice can be prescribed for technology adoption? If yes, can you indicate what form this would take?

Q36. Reflecting on the adoption and implementation process for (name of specific clinical technology), if you were to undertake this process again what would you do differently and why?

Thank you for agreeing to participate in this research.