Dental therapists compared with general dental practitioners for undertaking check-ups in low-risk patients: pilot RCT with realist evaluation

Article history

The research reported in this issue of the journal was funded by the HS&DR programme or one of its preceding programmes as project number 16/01/79. The contractual start date was in October 2017. The final report began editorial review in March 2020 and was accepted for publication in November 2020. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. This report has been published following a shortened production process and, therefore, did not undergo the usual number of proof stages and opportunities for correction. 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 2021. This work was produced by Brocklehurst et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. 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.

2021 Queen’s Printer and Controller of HMSO

Role substitution in NHS dentistry in the UK

Workforce planning is an important policy objective in the NHS to ensure that ‘the right number of people with the right skills are in the right place at the right time to provide the right services to the right people’. 1 ‘Skill mix’ is a term that is used to describe a model of care in which the whole of the clinical team is utilised in delivering service activity. 2 It can be further subdivided into role substitution and role supplementation. In dentistry, the former means that different members of the dental team undertake clinical tasks instead of a general dental practitioner (GDP), while the latter means that team members augment the activity of a GDP. 3 The use of role substitution in NHS dentistry has been advocated for some time, but its implementation appears to have lagged behind that seen in medical specialties owing to a number of factors, including professional regulation, NHS regulations and financial incentives. 2 These, along with the current evidence from the literature, will be explored in this chapter.

Professional regulation of role substitution in NHS dentistry

In 1921, the Dentists Act4 created the Dental Board of the United Kingdom, which became the professional body for dentistry and oversaw its practice. In 1956, following amendments made to the Act, the Dental Board was superseded by the General Dental Council (GDC). 5 This reflected the recommendations in the Teviot Report,6 which argued that the dental profession had become sufficiently mature to self-govern. The Act also facilitated the training of ‘dental auxiliaries’ for the first time [referred to as dental care professionals (DCPs) in the remainder of this report]. The newly formed GDC subsequently developed the regulated titles of dental hygienist (DH) and dental therapist (DT). The duties of the former related to the provision of preventative and periodontal treatment, and the latter was permitted to provide a range of direct restorative procedures and extract deciduous teeth. In 1983, UK dental schools began to offer ‘dual’ integrated training over a period of 2 years. The new ‘hygiene–therapist’ qualification was able to offer the full range of clinical activities that both DHs and DTs could legally undertake, although only an individual title of DH or DT was registerable with the GDC. 7 In 2002, the Dentists Act was again amended and allowed DTs to practise in NHS dental practices. 8 Before this time, the DT’s role had been limited to the provision of care in NHS Community Dental Service settings only; the provision of care in NHS Community Dental Service settings had formerly been restricted to GDPs since 1948.

Despite this change, DCPs were not allowed to practice as ‘front-line’ clinicians in the NHS or in private practice. Instead, patients were required to see a GDP for an examination and treatment plan before being referred to DCPs for their NHS care (a condition in both the Dentists Act and the NHS regulations9). NHS regulations9 also prevented DCP services from being procured directly, with the ‘knock-on’ effect of DCPs not being able to receive a NHS pension (unlike their GDP counterparts). However, after a review of the available literature, a landmark decision by the GDC in 2013 allowed patients to have ‘direct access’ to dental care by DCPs without a referral from a GDP. 10 This was accompanied by an expansion of the DCPs’ Scope of Practice to incorporate undertaking examinations and developing treatments plans within their level of competence. 7 However, DCPs remained unable to contract directly with the NHS under NHS regulations9 and were also unable to prescribe dental radiographs or fluoride or administer local analgesia (owing to restrictions in the Ionising Radiations Regulations11 and the Medicines Act,12 respectively).

Policy recommendations on role substitution in NHS dentistry in the UK

Policy-makers have been interested in the potential of role substitution in NHS dentistry for some time. In 1993, the Nuffield report inquiry13 redefined the concept of the dental team through which dental care could be delivered and argued that the role of DCPs could be expanded. Subsequently, increasing attention has been paid to how role substitution can deliver the level of care in the NHS that is required to meet the population health need. This is now explicitly recognised in a number of policy documents that underpin NHS care. For example, both the Prudent Healthcare policy agenda in Wales14 and the NHS Long Term Plan15 call for greater use of role substitution.

In the UK, population oral health needs are changing. In the most recent epidemiological survey, 90% of young adults are expected to have > 21 teeth in 10 years’ time, and the levels of dental caries and periodontal disease have fallen dramatically. 16 By contrast, levels of dental caries in young children have remained relatively intransigent, despite the fact that dental caries, as a non-communicable disease, are entirely preventable. 17 As highlighted in the Child Dental Health Survey in 2013, ‘nearly a third (31%) of five-year olds and nearly a half (46%) of eight-year olds had obvious decay experience in their primary teeth. Untreated decay into dentine in primary teeth was found in 28% of five-year olds and 39% of 8 year olds’ (information from NHS Digital, licenced under the current version of the Open Government Licence). 17 This pattern of disease also follows a social gradient: ‘a fifth (21 per cent) of the five-year olds who were eligible for free school meals had severe or extensive tooth decay, compared to 11% of five-year olds who were not eligible for free school meals’ (information from NHS Digital, licenced under the current version of the Open Government Licence). 17

Equally, the increasing number of partially dentate older people with varying degrees of independence are giving rise to new health-care challenges. The oral health of care-home residents is much worse than their community-living peers (e.g. caries prevalence is 73% vs. 40%, respectively) and about half of all care-home residents now retain some of their natural teeth. 18,19 Poor oral health may also exacerbate a range of medical conditions, including pneumonia and delirium, which increases health-care costs and leads to poorer outcomes. 20 Despite this high level of need, dental service provision in residential care is poor, with little emphasis on prevention. 21,22 Access to domiciliary services is difficult and unscheduled care for dental problems (including hospital admissions) is common, complex to deliver and expensive. 20,23 The World Health Organization (WHO) argues that the design of long-term care systems that are fit for ageing populations should take priority,24 and the Royal College of Surgeons of England (RCS),25 Public Health England (PHE) and the National Institute for Health and Care Excellence (NICE)26 have all called for more high-quality research. The Care Quality Commission (CQC) also highlighted the paucity of dental care in care homes in its report of June 2019. 27

Along with these pressing and emerging population health needs, there is evidence that many who attend dental practices have good oral health. In the Adult Dental Health Survey,28 71% of all dentate adults reported that they attended their GDP at least once per year. A total of 6% reported that they attended once every 2 years and a further 10% attended less often than every 2 years. The remaining 13% of dentate adults reported that they attended only when they were having trouble with their teeth. Data from the Business Services Authority (the organisation responsible for paying GDPs in England and Wales) suggest that many of those who do attend do not require any further treatment. 29 In 2018/19, out of the 39.7 million courses of treatment delivered in 1 year in England, 23.3 million (58.7%) related to patients having check-ups with no further treatment. The figures for 2017/18 and 2016/17 were 58.0% and 57.0%, respectively. That many who attend do not need further treatment is further supported by data from the Assessment of Clinical Oral Risk and Needs (ACORN) in Wales, where GDPs collect ‘risk’ and ‘need’ data as part of the Welsh Dental Contract Reform programme. A total of 60.4% and 47.8% of adults in 2019 were classified as ‘green’, that is as having no active dental caries or periodontal disease, respectively, and many of these patients appear to return to their dental practice within a 9-month period (Colette Bridgman, Chief Dental Officer for Wales, 2020, personal communication). This suggests that a substantive level of the resource being invested into primary dental care in the NHS relates to managing low-risk dental patients. As a result, there are increasing calls for the development of a NHS dental workforce to meet these emerging population oral health needs, and the procurement of relevant NHS service provision, while freeing up resources to increase the capacity to provide care and reduce social inequalities in oral health. 3

Role substitution in NHS dentistry

Dental care professional utilisation by NHS dental practices appears to be heavily influenced by the financial incentives inherent in the NHS contract. 30 NHS GDPs run their practices as businesses to offset the cost of the capital risk of the premises and the equipment that they own while ensuring liquidity to cover their overheads. 31 In medicine, transaction costs can be offset by economies of scale, which enable a broader range of services to be made available. 32 This is more difficult in NHS dentistry because, historically, many practices have not been ‘purpose built’ and some remain ‘single handed’. 33 This limits the extent of role substitution and supplementation undertaken in the NHS.

Although DHs appear to be well-accepted members of the dental team in the UK,30,34 financial considerations appear to play a significant part in the decision to use a DT in the NHS. 35,36 In one of the few studies that examined the profitability of using role substitution,37 patient charges generated did not cover the cost associated with their use in the current remuneration system. As a result, many DTs have been employed in the NHS as DHs in England rather than being utilised across their full range of skills. 35 In addition, some patients expect to pay less for treatment provided by a DCP than by a GDP. 38

Dental practices operating in the NHS are acutely sensitive to the incentives in any given remuneration system. 39,40 This can influence the institutional logic of the dental practice (the culture in an organisation that shapes the collective behaviour and actions of those who work there) and whether or not role substitution is supported. 30,41,42 Retrospective payment systems, such as fee-for-service (FFS) (where a GDP submits a claim for every single item of completed treatment), have been shown to lead to overtreatment to maximise profit. 43,44 In these systems, the incentive for practices is to increase the amount of clinical activity delivered, which may not always suit the greater utilisation of role substitution. 30 In contrast, per-capita remuneration systems pay practices a fixed level of funding based on the number of registered patients. This breaks the link between treatment activity and practice income, giving practices greater autonomy on what to focus on. 45 This may lead some practices to place greater emphasis on prevention, which would favour role substitution and supplementation (to reduce staff costs and draw on the strengths of DCPs to deliver prevention). 30 However, per-capita systems can also lead to undertreatment and patient selection, that is, a preference for low-risk patients who require little treatment, given that funding for these practices is capped and unrelated to clinical activity. 46,47

Goodwin et al. 48 argue that institutional logics at any given NHS practice include not only dentistry as a business but also professional ethics and contextual factors based on where the practice is embedded. As highlighted by Watt et al. ,49 the most important factors influencing change in dentistry include concerns about financial risk, progressive practice environment, supportive organisational structure, supportive professional networks and opportunity for training. Therefore, the drive to maintain (and maximise) the viability of a NHS practice can also be tempered by a practice owner’s sense of duty to their patients and their ideas about how best to deliver care for their patients and community. 50 In our earlier study,30 the views of practice principles on the benefit of role substitution was found to be one of the most important factors, which could ameliorate concerns about the impact of using DCPs, in relation to the underlying NHS contract. Management of change is also a potential problem with role substitution because professionals seek to protect their clinical roles and maintain traditional clinical boundaries. 51,52 Managing a transition to role substitution takes time and good human resource skills. 53 McDonald et al. 54 found that the key factors that determined the acceptability of changes to role boundaries included the clarity around roles and responsibilities as well as personal relationships with colleagues, which raise issues of mutual trust and respect. Nevertheless, the influence of the underlying NHS contract remains substantive and warrants an understanding of the historical development of NHS funding and contractual reform across the UK.

The evolution of NHS dental contract in the UK

At the turn of the new millennium, Modernising NHS Dentistry55 set the agenda in England and subsequently gave primary care trusts (the organisations responsible for procuring NHS service provision at the time) powerful new commissioning tools to improve access to NHS dentistry and increase the provision of preventative services across England. This was further emphasised in Options for Change in 2002. 56

Prior to 2006, GDPs were paid on a FFS basis. This meant that GDPs claimed for every item of clinical activity that they undertook. As described in Role substitution in NHS dentistry, this payment mechanism can have a tendency to incentivise overtreatment because GDPs’ income is directly linked to the level of clinical activity undertaken on each patient. 43,45 It also had a negative impact on practices because many of the individual items of treatment were of a relatively low value and led to GDP complaints about being on a ‘treadmill’ (a high level of clinical activity on low-value items) to maintain the viability of their practices. 31 This was particularly relevant after 1990 because the Department of Health and Social Care (DHSC) had attempted to correct an overspend on the national dental budget and had placed a downwards pressure on the annual review process that controlled how individual items of treatment were costed. 57 In 2006, a new general dental services (GDS) NHS contract was introduced in England and Wales. 58 This contract collated NHS dental activity items into three broad bands in an attempt to lessen the number of individual items of treatment that GDPs would claim for on each patient:

• band 1 – examination, radiographs and a simple scale and polish

• band 2 – restorations, extractions and root canal treatments

• band 3 – crowns, bridges and dentures.

These bands of treatment attracted 1, 3 and 12 units of dental activity (UDAs), respectively. The value of a UDA varied across NHS dental practices and was based on clinical activity and payments in a ‘reference year’ that were ‘earned’ under the previous FFS NHS dental contract in 2005. As a result, GDPs who had been undertaking significant levels of clinical activity in the reference year were rewarded with high UDA values, resulting in considerable heterogeneity and inequity across practices in England. For example, an examination (generating 1 UDA) in one practice could generate £18 whereas in another practice it could generate £30. Likewise, one or multiple fillings completed in a ‘course of treatment’ (generating 3 UDAs) would attract £54 or £90, respectively, for GDPs on the same NHS dental contract.

NHS dental contracts were raised with equity-owning GDPs in England and Wales (known as ‘providers’). They then subcontracted these contracts to non-equity-owning GDPs at their practices (known as ‘performers’). DCPs were provided with an income based on the proportion of clinical activity that they undertook, or were paid a salary. The former method of remuneration for DCPs led to a potential conflict of interest between performers (also known as ‘associate GDPs’) and DCPs, because both were funded based on the number of UDAs that they delivered. 30 Another important element of the 2006 contract was cost containment, which capped providers’ annual activity to an agreed number of UDAs per year, known as an annual contract value (ACV). 30 NHS GDPs were then paid one-twelfth of their ACV on a monthly basis. As a result, NHS GDPs’ outputs under the new contract in England were constrained and they were penalised if they underperformed (< 96% of their ACV) or overperformed (> 102% of their ACV).

The effect of this change to the NHS dental contract was significant changes in the types of treatments that were offered. For example, as providers and subcontracted performers were paid the same number of UDAs for endodontic treatment and extractions (which are often the two alternative treatment options for many teeth with advanced decay), the number of the former reduced in favour of the latter, given that the latter could be undertaken in one-quarter of the time in most cases. 39,40 Equally, one or multiple fillings completed within a course of treatment would generate the same level of revenue for a practice (3 UDAs). As a result, there was downwards pressure on the number of fillings undertaken in any one course of treatment, leading to multiple courses of treatment submitted for patients who required more than one restoration. 39,40 GDPs were also incentivised to make upper and lower partial or full dentures separately to claim two sets of band 3 payments. Preventative care was not remunerated separately and any preventative activity was expected to be undertaken in an existing course of treatment. 39,40

In addition to these unintended consequences at a regional and national level, the 2006 NHS dental contract proved unpopular with GDPs. 31 This was largely because of the loss of clinical autonomy and the need for practices to be accountable to local dental commissioners (DCs), who closely scrutinised providers’ ability to deliver to their ACV targets. GDPs argued that one form of treadmill had simply been replaced by another. As a result of this criticism, an independent review was undertaken in 2008/9 in England, which recommended the greater use of preventative care and a standardised approach to patient assessment, leading to patient care pathways. 59

This led to the development of a pilot programme in England in 2010, which was predominantly based on capitation. 60 It also required GDPs to undertake an oral health assessment that categorised patients into different risk categories (red, amber or green). However, the pilots in England were beset with a number of informatic problems, and based on concerns about the capitation payment system were relaunched in 2015 as ‘prototypes’. 61 Prototype practices were paid on the basis of a blended-funded system, drawing on features of the 2006 contract (a retrospective payment mechanism based on clinical activity) with capitation. This blended together the financial incentives to ‘care’ and prevent disease (associated with capitation payment mechanisms) with the incentives to undertake clinical activity (associated with FFS mechanisms). In similarity to the earlier pilots in England, the prototypes were based on establishing risk and then referring the patient to the appropriate patient care pathway.

An evaluation of the first year of prototyping was published in 2018. 62 Attendance marginally increased over the course of the evaluation period (3–6%). A total of 62% of prototype practices reported delivery of preventative care to adults according to evidence-based guidelines compared with 56% in the 2006 contract, with little change to the level of prevention offered to children (60% and 58%, respectively). The evaluation concluded:

. . . progress has been made in the first year of prototyping on the key issues of improving oral health, providing appropriate care and quality, and maintaining or increasing access to merit continuation of the programme.

Reproduced with permission from DHSC. 62 Contains public sector information licensed under the Open Government Licence v3.0

The evolution of contractual reform followed a similar pattern in Wales, although the process was different under their devolved government. In 2011, eight dental practices took part in a pilot trial. 63 This process required these practices to deliver weighted ‘key performance indicators’ and a capitation target based on the number of unique patients treated in the practice within a defined time period. In 2012, the Quality and Outcome Pilot64 for adults was introduced, with the introduction of a dental care assessment and a risk-based preventative care plan. By 2016, all but two practices had reverted back to the 2006 UDA contract. The remaining practices continued as prototype practices to test new models of care and formed a ‘learning network’. An early evaluation of the Welsh pilots found a reduction in the number of patients attending NHS practices (7–10%) compared with baseline.

In 2017, the Welsh Government published Taking Oral Health Improvement and Dental Services Forward in Wales,65 which placed greater emphasis on incentivising needs-led care, role substitution/supplementation and prevention. This led to the development of a needs and risk assessment tool. In 2018 the Welsh Government increased the number of ‘contract reform’ practices to 23, and by late 2019 > 100 NHS practices had entered phase 1 of the process. This reduced ACV targets by 10% and raised all providers’ UDA values to a minimum of £25 to ensure greater parity across the nation. For this 10% reduction in their ACV, practices were required to undertake an ACORN at least once per year under the mantra ‘do it once, do it well’ (Colette Bridgman, personal communication). Phase 2 of the contract reform process saw further reductions in ACV targets (reduction to 70–80% of pre-reform levels) in exchange for greater emphasis on the provision of prevention, increasing the quality of care provided and greater utilisation of role substitution/supplementation. These policy objectives were further emphasised in The Oral Health and Dental Services Response to a Healthier Wales,66 published in 2018, and are being underpinned by a formal evaluation process undertaken by researchers at Bangor University. As of spring 2020, over one-quarter of the total NHS practice population was using ACORN and involved in the Dental Contract Reform programme in Wales.

In Northern Ireland, the predominant financial mechanism used to pay GDPs still resembles the pre-2006 contract in England and Wales, that is, the majority of funding draws on FFS and items of service as part of a retrospective payment system. An Oral Health Strategy for Northern Ireland67 was published in 2007 and reiterated the main themes identified by the Northern Ireland Primary Dental Care Strategy in 2006, namely a shift away from treatment in favour of prevention of disease while maintaining access to services. In 2014, the Northern Ireland Health and Social Care Board initiated a pilot NHS dental contract across the province, which involved > 30 GDPs. Similar to the English pilots, the chosen remuneration mechanism was prospective in nature, so reduced the incentive for GDPs to engage in unnecessary treatment. The underpinning policy intention was that practitioners would receive the equivalent gross income during the pilot period that they would have received under the GDS had they maintained their activity and list as per the baseline period. As highlighted in our evaluation of the pilot NHS contract led by Bangor University, we found ‘rapid changes in the patterns of care provided by GDPs to patients (compared with the control practices) when they moved from a FFS system to a capitation-based remuneration system’ and ‘there were statistically significant reductions in the volume of all treatments in the intervention practices during the capitation period’ (contains information licensed under the Non-Commercial Government Licence v2.0). 57 This highlighted the sensitivity that GDPs have to underlying incentives in the NHS dental contract.

In Scotland, the prevailing dental contract funding mechanism was similar to Northern Ireland’s model. After a consultation exercise in 2016, the Scottish Government published its Oral Health Improvement Plan68 in 2018 with a 41-point plan that focused on prevention, an oral health risk assessment and a personalised care plan. As highlighted by Brocklehurst et al. :57

. . . in contrast to the other three home nations, there were no plans for scrapping the item-of-service system of remuneration; instead there was an intention to ‘streamline’ items of service payments, which would be progressively introduced.

Reproduced with permission from Brocklehurst et al. 57 Contains public sector information licensed under the Open Government Licence v3.0

The proposed remuneration mechanism was based on a ‘mixed economy of item of service, capitation and continuing care payments’57 (contains public sector information licensed under the Open Government Licence v3.0).

Evidence base for the use of role substitution in dentistry

Proponents of role substitution argue that it has the potential to free resources, increase the capacity to care for high-need populations, improve access and reduce oral health inequalities. 69–71 In an international review, Nash et al. 72 concluded that ‘access to basic dental care will not be available without the utilization of dental therapists in the workforce’, and Johnson73 argued for a paradigm shift ‘from treatment to prevention, wellness and self care’.

Role substitution by DCPs has been established for some time in a number of European countries. 70,71 Sweden and the Netherlands legalised the independent use of DCPs in 1964 and 1978, respectively. 74 Finland, Denmark and Norway have allowed independent practice since 1994, 1996 and 2001, respectively. Similar practices are found in Switzerland, which started in 1997, and in Italy, where DHs have been able to work as independent practitioners since 1999. 74 In the USA and Canada, DH is a growing profession, and DHs can practice with varying degrees of independence in a number of US states, including California, Colorado, Montana, Nebraska, New Mexico, Oregon and Washington, and Canadian provinces, including British Columbia, Alberta, Saskatchewan and Manitoba, although some restrictions remain in terms of settings. 75 Tasmania has a liberal regulatory model in which DHs and DTs practice independently and can own their own practices. 74 DTs are also considered to be independent in New Zealand, although they are not able to treat adults. 74 In Samoa and Singapore, DTs must work under the supervision of a GDP. Fiji has allowed DTs to assume independent responsibility for managing clinics since 1985, and DTs have been allowed to practice independently in South Africa since 1994. 74

The research evidence for the greater use of DCPs is emerging. Despite suggestions from some elements of the dental professional that role substitution is unsafe, this is not borne out by the limited literature that is available. Two studies by Calache and Hopcraft,76,77 undertaken in Australia, found no evidence to question patient safety when adult patients were managed by DTs. No differences were found between DTs and graduating GDPs in infection control, local analgesia, cavity preparation, placement of restoration, occlusion and patient communication. 77 Knowledge and clinical skills were rated as good to high, and 80% of DTs were considered safe to treat adult patients. The potential of DCPs not to identify oral cancer is another frequently cited concern in terms of patient safety. However, in an in vitro study, Brocklehurst et al. 78 found that the diagnostic test accuracies of GDPs and DTs were similar when they were presented with a judgement task that required them to distinguish between malignant and benign diseases of the oral mucosa. Values of sensitivity and specificity were 81% and 73% for GDPs, compared with 77% and 69% for DTs, respectively. DTs also missed fewer frank carcinomas in the judgement task.

Although the social acceptability of DTs appears to be positive, public awareness of DTs as a professional group is not widespread. 38,79,80 In the UK, our earlier study found that the attitudes of practice owners were critical in the process of patients accepting treatment from DTs. 30 Despite this, it does appear that adults are willing to receive treatment from DCPs under the NHS and there is evidence of increased patient satisfaction. 80 In one of the Calache and Hopcraft studies,77 patients were very satisfied with the dental treatment that was provided. In another study conducted by the same team,81 356 direct coronal restorations were placed with high patient satisfaction. In Canada, 65.8% of respondents stated that they would visit an independent DH to maintain their oral health. In the USA, 98.0% of respondents strongly agreed or agreed that their care by DCPs was satisfactory (based on RAND criteria; RAND Europe UK, Cambridge, UK). 82 Perry et al. 83 reported 98.0% patient satisfaction. For DTs, the majority of adults and children’s caregivers receiving care in Alaska had a positive patient experience. 84 In the Netherlands, one-quarter of those surveyed (n = 1500) agreed that ‘simple dental treatments’ could be performed by a DH rather than by a GDP, and 15% agreed that a DH could undertake a check-up. 85,86 In another study, 85.0% of patients who received a simple restoration for the treatment of dental caries provided by a DCP claimed to be satisfied with the treatment provided. 87

In terms of efficacy, Wang and Riordan88 evaluated a population oral health programme undertaken in Norway that allowed DHs to undertake a check-up over a 3-year period. The quality of care provided was judged not to deteriorate over this time frame. 88,89 This is supported by Kwan et al. ,90 who compared the performance of DCPs with GDPs in an epidemiological programme. Sensitivity values for GDPs ranged from 0.54 to 1.00, and those for the DCPs from 0.80 to 0.94. The direction of effect in these studies has also been replicated by Patel et al. ,91 Kwan and Prendergast92 and Hopcraft et al. 93 In terms of restorative management by DCPs, Bader et al. 94 found that 74 out of the 84 direct amalgam restorations placed by DTs were considered to be completed to an adequate standard, compared with 32 out of 41 restorations provided by GDPs. This was not dissimilar to Wetterhall et al. ’s study,84 which found that 88.0% and 78.0% of amalgam restorations were rated as ‘satisfactory’ for DTs and GDPs, respectively. Calache et al. 81 evaluated 356 restorations placed by DTs in 115 patients. After 6 months, 95.0% were judged to be successful by blinded evaluators. This concurs with studies by Battrell et al. 95 and Freed et al. 82 Dyer et al. ’s Cochrane review96 identified four randomised controlled trials and one non-randomised controlled trial. Three found no evidence of a difference in retention rates of fissure sealants placed by DCPs compared with GDPs, and the study comparing the comparative effectiveness of ‘atraumatic restorative technique’ restorations found no difference in survival rates after 12 months. 96

These results concur with the findings of a number of studies97,98 led by our team that were undertaken with NHS clinicians. In 2012, we showed that both DTs and GDPs had comparable sensitivity and specificity in the recognition of occlusal caries in vitro. 97 In 2015, we tested the diagnostic test accuracy of DTs in vivo. 98 A total of 1899 adult NHS patients were screened by both a DT (index test) and a GDP (reference standard) prior to their routine NHS check-ups in 10 busy NHS practices. Both sets of clinicians made an assessment on the presence of dental caries and periodontal disease. The sensitivity and specificity values of DTs were 0.81 [95% confidence interval (CI) 0.74 to 0.87] and 0.87 (95% CI 0.78 to 0.92) for dental caries and 0.89 (95% CI 0.86 to 0.92) and 0.75 (95% CI 0.66 to 0.82) for periodontal disease, respectively. This suggests that DTs could recognise the two most common oral diseases. Brocklehurst et al. 78 found that the diagnostic test accuracies of GDPs and DTs were similar when they were presented with a judgement task that required them to distinguish between malignant and benign diseases of the oral mucosa. Values of sensitivity and specificity were 81% and 73% for GDPs, compared with 77% and 69% for DTs, respectively. DTs also missed fewer frank carcinomas in the judgement task.

Although diagnostic test accuracy is important as a measure of efficacy, a more important consideration is an assessment of the effect on health of using DTs in front-line roles. Given that precise estimates of effect, recruitment, retention and fidelity were unknown, we undertook a study to test whether or not a definitive trial could be feasible in the NHS. 99 Over a 15-month period, adult NHS patients were randomised into one of three arms in two NHS dental practices: patients who saw only a DT (arm 1), those who saw both a DT and a GDP (arm 2) and, finally, those who saw only a GDP (arm 3). The initial recruitment rate for the study was 33.7%. This figure increased to > 82.1% when telephone calls or face-to-face recruitment were utilised. The retention rates were 60% for the DT and DT/GDP arms and 70% for the GDP arm. The proportion of sites with bleeding on probing (BOP), a measure of gum inflammation, at the end of the 15-month period was 46.7%, 14.5% and 32.1% in arms 1, 2 and 3, respectively (44.6%, 17.2% and 26.5% at baseline, respectively). Similarly, plaque levels were found in 68.2%, 43.7% and 60.9% of patients, respectively (60.5%, 44.9% and 56.3% at baseline, respectively). Because these values did not differ significantly from baseline values, it suggested that oral health can be maintained by role substitution.

The majority of the studies examining the efficiency of DCP utilisation are from the UK. 100,101 Harris and Sun102 concluded that role substitution in the UK ‘may be limited to particular situations where conditions are conducive’. Patients with high levels of disease (many restorations) were thought less suitable for referral because GDPs would undertake care in fewer visits and there was a risk of complications necessitating referral back to a GDP. 99 In our earlier study30 we found that the extent of role substitution in NHS dental practices appears to be relatively limited and largely restricted to the use of DHs for routine periodontal treatment. This appears to be directly related to the incentives and disincentives in the NHS dental contract. This meant that ‘NHS dental practices that utilised fewer non-dentist team members were associated with higher levels of technical efficiency, that is as role substitution in NHS practices increased, their relative efficiency dropped’. 30 As highlighted in the report, ‘when UDAs were used as the output measure in England, NHS dental practices operated at a mean level of efficiency of 64.0%. This changed very little when the outputs were measured in terms of number of patients seen, or the number of treatment plans generated. NHS dental practices that did not use any form of role substitution had a higher mean level of efficiency (68.0%; n = 39)’ (contains public sector information licensed under the Open Government Licence v3.0). 30 Any use of DCPs was found to be associated with statistically significantly lower efficiency scores (14.0% lower for UDAs and 11.0% lower for treatment plans or patients seen) than no use of role substitution.

Discussion

As highlighted in this chapter, the potential use of role substitution in NHS dentistry is becoming increasingly recognised by policy-makers. Changes made by the GDC now make this more possible, but there remain regulatory barriers at this point in time: NHS general dental services regulations, the Medicines Act12 and the Ionising Radiations Regulations. 11 Financial incentives in NHS dental contracts also play a key role but can be mitigated by the institutional logics in a dental practice.

All dental contract reform programmes undertaken across the UK have placed an emphasis on increasing patient access and prevention while maintaining and improving the quality of service being delivered. In many cases, this is facilitated by the development of care pathways based on the risk status of the patient. In Wales, the use of role substitution is an explicit objective in the second stage of the contract reform process, and it is implicit in many other national programmes.

This means that it is timely to draw on the existing evidence base for the greater use of DCPs. However, the quality of the literature is relatively poor and, as Dyer et al. ’s Cochrane review96 highlighted, the number of trials empirically testing role substitution is limited. Although there have been a number of diagnostic test accuracy studies to demonstrate efficacy, no trials have examined the effectiveness of using DTs, instead of the GDP, to undertake the check-up. As highlighted by our feasibility study,99 there was some reservation about recruitment and retention rates in busy NHS dental practices. There were also specific design questions that remained unanswered. For example, what is the estimated size of effect of such an intervention and should the design be based on superiority or on the basis of non-inferiority (i.e. are DTs better than or no worse than a GDP in undertaking the check-up and any subsequent treatment that is required)? The latter question is a fundamental one and significantly influences the design of a definitive trial.

As a result, we proposed undertaking an individually randomised pilot study in the North West of England over a 15-month period based on the following population, intervention, control, outcome (PICO) format:

• population – adult, asymptomatic, low-risk, routine dentate or partially dentate NHS patients

• intervention – check-up and any subsequent treatment undertaken by a DT

• control – check-up and subsequent treatment by a GDP (‘treatment as usual’)

• outcome – proportion of sites that BOP (measured at six sites per tooth).

To facilitate the design of the definitive study, a realist-informed process evaluation was undertaken alongside the pilot study to explore the acceptability of using DTs as front-line clinicians, patient crossovers (from one arm to the other) and treatment fidelity. We also captured the contextual factors that shaped the intervention in NHS practices, mechanisms that sustained or potentiated effects, unexpected pathways and consequences and the contextual factors that shaped implementation. 103 This was framed from a realist perspective to understand ‘what is it about a programme that works for whom, in what circumstances, in what respects, over which duration’. 104 Realist methodology is becoming increasingly used in health services research because it recognises the complex and contingent nature that underpins the settings for new interventions and service delivery. This approach to process evaluation ‘supposes that regularities in the patterning of social activities are brought about by the underlying mechanism constituted by people’s reasoning and the resources they are able to summon in a particular context’. 105

Finally, we undertook a third workstream (WS) to develop and test the health economic data collection tool and rehearse the health economic analysis of the intervention compared with the current practice. We adopted the viewpoint of both the NHS and the patient, and collected resource use data that included the costs of dental consultation by GDPs and DTs, the use of primary and secondary NHS dental services, and participants’ out-of-pocket expenses relating to any dental problems during the pilot’s follow-up period.

Aims and objectives

The aim of this study was to inform the design for a definitive trial by undertaking a pilot study to determine whether or not DTs can maintain the oral health of low-risk routine NHS patients, who form the predominant proportion of the regularly attending practice population. The objectives of the research were to:

• determine the most appropriate design of a definitive trial

• determine whether or not BOP is the most appropriate primary outcome measure (POM) (and if not, determine the most appropriate measure)

• confirm the appropriateness of the non-inferiority margin of the chosen outcome measure for the definitive trial (i.e. whether or not the effect estimate lies within an appropriate margin of non-inferiority)

• further investigate recruitment, retention and fidelity rates

• confirm willingness to be randomised among study participants

• determine the potential for patient crossovers between arms (e.g. where the patient’s condition is considered too complex to be managed by the DT)

• undertake a process evaluation underpinned by a realist framework to understand what works, for whom, why and in what circumstances

• rehearse the health economic analysis and assess the health economic data collection tool, to inform the definitive trial design

• explore patients’ preferences in a focus group setting, to inform a preference elicitation exercise [e.g. discrete choice experiment (DCE)] in the definitive trial.

Structure of the report

This report is arranged in chapters, as follows. Chapter 1 provides a review of the literature and describes the historical context of role substitution in dentistry. Chapter 2 describes the methods and results of the pilot study. Chapter 3 describes the realist-informed process evaluation and Chapter 4 details the health economic element of the study. Chapters 5 and 6 present the discussion and conclusions of the study.

Introduction

In 2013, the GDC expanded the Scope of Practice for both DHs and DTs following a review of the scientific evidence. 10 This allowed both DHs and DTs to see patients for the first time as the front-line health-care worker. This meant that they could now see patients directly without a referral from a GDP, undertake an examination, formulate a treatment plan and provide clinical treatment within their competence. For DHs, this amounted to the provision of preventative and periodontal care and for DTs this also included the provision of temporary and permanent restorations.

These changes formed the basis of this study and an opportunity to test an alternative care pathway using DTs to undertake the check-up on low-risk routine patients in the NHS. Two earlier NIHR grants investigated the efficacy and efficiency of using DTs in NHS practices [NIHR/CS/010/004 and Health Services and Delivery Research (HSDR) 11/1025/04]. The former demonstrated the diagnostic test accuracy and feasibility of using these groups to undertake check-ups and the latter found that dental practices using DTs in the NHS could be better organised. Pilot studies determine ‘whether something can be done, should we proceed with it, and if so, how’. 106 In WS1, we undertook such an approach to inform the design of a definitive trial.

Aims and objectives

The aim of this WS was to inform the design for a definitive trial by undertaking a pilot study over 15 months to determine whether or not DTs could maintain the oral health of low-risk routine NHS patients, who form a significant proportion of those who regularly attend NHS practices. The objectives of the research were to:

• determine the most appropriate design of a definitive trial

• determine whether or not BOP is the most appropriate POM (and, if not, determine the most appropriate measure)

• confirm the appropriateness of the non-inferiority margin of the chosen outcome measure for the definitive trial (i.e. whether or not the effect estimate lies within an appropriate margin of non-inferiority)

• further investigate recruitment, retention and fidelity rates

• confirm willingness to be randomised among study participants

• determine the potential for patient crossovers between arms (e.g. where the patient’s condition is considered too complex to be managed by the DT).

Methods

Workstream 1 was an individually randomised pilot study undertaken across the North West of England over a 15-month period:

• population – adult, asymptomatic, low-risk, routine dentate or partially dentate NHS patients attending high-street dental practices

• intervention – check-up and any subsequent treatment undertaken by a DT

• control – check-up and subsequent treatment undertaken by a GDP (treatment as usual)

• outcome – proportion of sites that had BOP.

Pilot study setting

Eight high-street NHS dental practices were recruited across the North West of England. The geographic distribution of the practices is highlighted in Figure 1.

FIGURE 1.

Geographic distribution of participating practices.

Patient pathway

The patient pathway for the pilot study is provided in Figure 2.

FIGURE 2.

Participant pathway through the NHS practice.

On arrival for their check-up, identified patients were asked by a member of the research team if they had any questions about the study and if they wished to participate. If they agreed, the participant was then asked to sign the consent form and a unique patient identifier was provided to ensure anonymity. A unique patient identifier reference sheet was kept throughout the study, which enabled the participants to be tracked, when necessary. The participants were examined by a trained epidemiologist, who first checked if the patient was eligible for the study and then undertook the baseline measurements for the case report form (CRF) (see Appendix 1). The baseline CRF recorded details that included:

• age and gender

• exemption status and employment type

• ethnicity

• number of teeth remaining and number of sites (six per tooth)

• number of sites with BOP (to enable proportion to be calculated)

• number of sites with visible plaque present

• Oral Health Impact Profile-14 items (OHIP-14)107 score

• reported levels of dental anxiety.

The participant was randomised, using the North Wales Organisation for Randomised Trials in Health (NWORTH) Clinical Trials Unit sequential dynamic adaptive randomisation algorithm, to see a DT (intervention arm) or their GDP (control arm). 108 After randomisation, each participant received a check-up in accordance with the arm that they were allocated.

Results

The outcome of the pilot trial and suitability of a definitive trial design were assessed using the following:

• willingness to be randomised – quantified by considering the numbers approached and who subsequently consented (or refused based on not wanting to be randomised) (see Recruitment and retention)

• recruitment, retention and fidelity rates – quantified throughout following the Consolidated Standards of Reporting Trials (CONSORT) flow chart process (see Recruitment and retention)

• numbers of participant crossovers between arms (e.g. where the patient’s condition is considered too complex to be managed by the DT) – quantified from the logs kept by each dental practice (see Protocol deviations)

• appropriateness of BOP as the POM – quantified by considering the potential effect size indicated by preliminary exploratory analysis (see Baseline measurements and Missing data)

• appropriateness of the non-inferiority margin (i.e. whether or not the effect estimate lay within an appropriate margin of non-inferiority) – quantified by considering the outcome of the preliminary exploratory analysis (see Exploratory analysis of the primary outcome measure).

Recruitment and retention

A total of 546 NHS patients were invited to participate in the pilot trial. A total of 217 patients agreed to participate over a 5-month period (March 2018 to July 2018) and 328 stated that they were not interested in participating. This equated to a recruitment rate of 39.7% and a mean recruitment rate of 43.4 participants per month. The mean number of participants recruited per site was 27.1, across eight NHS dental practices over the 5-month recruitment period (Table 1). Out of the 217 patients who were randomised, 107 were randomised to the DT arm and 110 were randomised to the GDP arm. The arms appeared well balanced on stratification variables (see Table 1).

TABLE 1 - Summary of recruitment by the dental practices

Dental practice	Date first participant recruited	Date last participant recruited	DT arm (n)	GDP arm (n)	Overall, n (%)
1	5 April 2018	3 May 2018	16	21	37 (17.1)
2	17 July 2018	24 July 2018	9	9	18 (8.3)
3	24 April 2018	29 May 2018	16	17	33 (15.2)
4	2 May 2018	5 June 2018	18	18	36 (16.6)
5	13 April 2018	9 July 2018	12	9	21 (9.7)
6	15 March 2018	26 April 2018	6	4	10 (4.6)
7	11 May 2018	22 June 2018	14	13	27 (12.4)
8	12 March 2018	11 July 2018	16	9	35 (16.1)
Exempt	–	–	10	13	23
Non-exempt	–	–	97	97	194

A total of 50 participants did not complete the follow-up assessment at 15 months. Among these 50 participants (23.0% of the total sample), nine participants withdrew from the trial and 41 were lost to follow-up. These 41 participants were lost because of the logistical difficulties of organising a follow-up appointment with an epidemiologist at the practice that the participant was attending. In some cases, participants could not be contacted (by mail, e-mail, telephone or text message), and other participants were unable to attend the practice on designated days owing to work or other commitments. The attrition rate was lower than the attrition rates obtained in previous dental-practice based trials. 99,115 This is represented in the CONSORT flow diagram (Figure 3).

FIGURE 3.

The CONSORT flow diagram for the pilot trial.

The attrition rates for the DT arm and the GDP arm were comparable (n = 24, 22.4%, vs. n = 26, 23.6%, respectively) over the 15-month period. There was no statistically significant difference in attrition across the randomisation stratification variables (arm, p = 0.873; dental practice, p = 0.676; number of teeth, p = 0.421; exempt from patient’s charge, p = 0.432).

Baseline measurements

Participants in the pilot trial were predominantly female (72.4%), non-smokers (92.2%) and white (93.1%) and were not exempt from patient charges (89.4%). The mean age of the participants was 46 years, with a standard deviation (SD) of 15.8 years and a range of 18–87 years. The participants’ occupations are summarised in Table 2.

TABLE 2 - Occupation of the participants

Occupation type	Percentage
Clerical and sales (e.g. administration, salesperson)	12.0
Professional and managerial (e.g. teacher, doctor, manager)	35.5
Semi-skilled and unskilled (e.g. factory worker, labourer)	2.8
Skilled blue collar (e.g. electrician, plumber, craftsperson)	7.8
Other	40.6
Prefer not to say	1.4

The mean participant Index of Multiple Deprivation (IMD) score was 22.49, with a SD of 17.02, and was comparable with the mean IMD score of the population (Figure 4).

FIGURE 4.

Distribution of IMD scores.

Descriptive statistics of the dental-related measures at baseline are provided in Tables 3–5.

TABLE 3 - Dental-related measures at baseline: total sample

Measure	n	Mean	SD	95% CI	Median	IQR	Range
Number of teeth	217	25.51	2.89	25.12 to 25.90	26	24–28	11–28
Number of sites	217	153.04	17.32	150.72 to 155.36	156	144–168	66–168
Number sites with BOP	217	5.08	5.97	4.28 to 5.88	3	1–8	0–36
Number sites with dental plaque	217	3.30	4.95	2.64 to 3.96	2	0–5	0–30
Number of dental caries lesions	217	0.01	0.10	0.00 to 0.02	0	0–0	0–1
Percentage of sites with BOP	217	3.39	3.98	2.86 to 3.92	2	0.60–5.13	0–23
Percentage of sites with dental plaque	217	2.25	3.32	1.81 to 2.69	1.19	0.00–3.03	0–19
OHIP-14 score (0–56 points)	216	0.85	2.03	0.58 to 1.12	0	0–1	0–19
Anxiety of GDP rating (1–10 points)	217	2.10	1.77	1.86 to 2.34	1	1–3	1–10
Anxiety of using DT rating (1–10 points)	217	2.12	1.72	1.89 to 2.35	1	1–3	1–10

TABLE 4 - Dental-related measures at baseline: DT

Measure	n	Mean	SD	95% CI	Median	IQR	Range
Number of teeth	107	25.48	3.02	24.9 to 26.06	26	24–28	11–28
Number of sites	107	152.86	18.12	149.39 to 156.33	156	144–168	66–168
Number sites with BOP	107	5.60	6.20	4.41 to 6.79	3	1–9	0–28
Number sites with dental plaque	107	3.43	5.28	2.42 to 4.44	2	0–4	0–30
Number of dental caries lesions	107	0.02	0.14	0.01 to 0.05	0	0–0	0–1
Percentage of sites with BOP	107	3.79	4.24	2.98 to 4.60	2.08	0.60–5.95	0–17.95
Percentage of sites with dental plaque	107	2.36	3.61	1.67 to 3.05	1.19	0–2.88	0–19.33
OHIP-14 score (0–56 points)	107	1.07	2.43	0.60 to 1.54	0	0–2	0–19
Anxiety of GDP rating (1–10 points)	107	1.95	1.73	1.62 to 2.28	1	1–2	1–10
Anxiety of using DT rating (1–10 points)	107	1.98	1.7	1.65 to 2.31	1	1–2	1–10

TABLE 5 - Dental-related measures at baseline: GDP

Measure	n	Mean	SD	95% CI	Median	IQR	Range
Number of teeth	110	25.54	2.77	25.02 to 26.06	26	24–28	14–28
Number of sites	110	153.22	16.59	150.08 to 156.36	156	144–168	84–168
Number sites with BOP	110	4.58	5.73	3.50 to 5.66	2.5	1–6	0–36
Number sites with dental plaque	110	3.17	4.63	2.30 to 4.04	1.5	0–5	0–25
Number of dental caries lesions	110	0.00	0.00	0.00 to 0.00	0	0–0	0–0
Percentage of sites with BOP	110	3.00	3.70	2.30 to 3.70	1.76	0.60–4.28	0–23
Percentage of sites with dental plaque	110	2.13	3.02	1.56 to 2.70	0.94	0–3.07	0–19
OHIP-14 score (0–56 points)	109	0.64	1.52	0.35 to 0.93	0	0–1	0–11
Anxiety of GDP rating (1–10 points)	110	2.24	1.80	1.90 to 2.58	1	1–3	1–8
Anxiety of using DT rating (1–10 points)	110	2.26	1.74	1.93 to 2.59	1.5	1–3	1–8

Discussion

The findings in this pilot trial appear to show that DTs were non-inferior to GDPs (within the non-inferiority boundary set a priori) over the 15-month time period. The point estimates were similar between the two arms in terms of the POM (BOP) and for the remaining clinical measures used the: proportion of sites with dental plaque and the number of new dental caries lesions. There were also no differences (within the non-inferiority boundary) between the two arms in terms of oral health-related quality of life (measured using the OHIP-14) or prospective anxiety about receiving treatment from a GDP or DT. For the process outcomes, there were also no differences between arms in any of the clinical measures. This would suggest that DTs are non-inferior to GDPs in their clinical management of low-risk NHS patients over the chosen timescale and using the selected POM. However, finding a ‘significant’ result in a pilot study should not be overinterpreted. 118,119 The estimates that we used may have still lacked precision, particularly the variability on our proposed primary outcome, and this may have influenced our ability to determine non-inferiority (owing to the reliance on the CI estimates for the parameters). In Results, we provided both ITT and per-protocol analyses, because the latter in this context is more conservative. However, it is also possible that the measures used for this study were exhibiting floor effects because of the low-risk patient population that we drew from (under-representing the variability).

The interpretation of these findings poses some interesting design questions. The purpose of this pilot trial was to determine the most appropriate design for a definitive trial, determine the appropriateness of the POM and the non-inferiority margin, and test recruitment, retention and fidelity rates. It would appear that any future definitive trial to determine the clinical effectiveness of role substitution in NHS dentistry should be based on non-inferiority. No differences were found in the oral health status of patients over the 15 months of the pilot. In addition, despite high retention and fidelity rates, recruitment rates were moderately low (39.7%) when based on the number of patients approached (mean recruitment rate of 43.4 participants per month). These two findings may have substantial implications for a definitive trial in this low-risk population group given the larger number of participants that would be required for a non-inferiority design. This could extend the duration of the definitive study to the point at which the value of information generated becomes questionable in this population group. Equally, there are challenges from a policy perspective, where the recommendations are to see low-risk patients every 2 years. 120

A further challenge is the type of POM that should be used. In this study, we considered BOP, which is routinely assessed in clinical practice to measure the degree of gingival and periodontal inflammation. Earlier studies suggest that the levels of BOP are a very good indicator of periodontal health and tissue stability. 112–114 As a pragmatic primary end point that is relevant to both patients and clinicians, BOP would also appear to be sensible and it was among the outcome measures used in our earlier feasibility study. 99 However, based on evidence from this 15-month pilot study, it could be argued that there are floor effects (i.e. the signal/event rate is so limited among low-risk patients that it makes differentiation between the two arms difficult). When compared with the Adult Dental Health Survey16 (a cross-sectional sample of the general population), the levels of both periodontal inflammation and plaque in this sample were much lower. Given that we used experienced epidemiologists in this pilot trial, this would suggest that NHS patients exhibit particularly low levels of need when only low-risk participants are included, which this only increases the potential for floor effects.

This is even more relevant for clinical measures, such as dental caries. Dental caries was chosen as a secondary outcome measure rather than a potential POM in this study because it has a low prevalence and longer time to expression in routine low-risk NHS adult patients. 16 This was supported by the data from this 15-month pilot, in which no new lesions developed, and by our feasibility study, in which the proportion of sites with active caries was between 1.19% and 2.60%. 99 Systems developed for the detection of small incremental changes in dental caries status, such as the International Caries Detection and Assessment System (ICDAS), have been advocated, but some argue that these suffer from measurement bias. 121,122 According to Frencken et al. ,122 ‘caries prevalence should not be based on the dmf/DMF [primary dentition decayed-missing-filled/permanent dentition decayed-missing-filled] index but on cavitated dentine carious lesions (d/D-component) as the M- and F- component do not refer to a disease stage’. Although it is acknowledged that these tools are used more in epidemiological surveys, signal/event definition, low event rate and slow progression of dental caries in low-risk patients remain problematic when marginal differences are being assessed in a clinical trial.

These issues are pertinent to the design of a definitive trial. In the recent NHS contract reform process undertaken in Wales, where need and risk where measured (using the ACORN), over two-thirds of NHS patients were judged to be low risk (Colette Bridgman, personal communication). As a result, even if we were to include a higher-risk patient population in the study sample for a definitive study, it is likely that any signal of effect would be drowned out by the proportion of low-risk NHS patients. It may be possible to consider analysing a ‘complete’ sample (including high- and low-risk patients) using an appropriate zero-inflated model, but this may not facilitate a full interpretation of effect or non-inferiority. The literature indicates that there are a number of ways of considering the development of various thresholds for a pilot study, but these should all be considered a priori. 118,119,123

It would appear that the types of clinical outcomes that would be used for a definitive trial on role substitution in dentistry may not be appropriate to test for non-inferiority. So, what would be an appropriate outcome measure? Patient-related outcome measures and patient-related experience measures may be two options, but their sensitivity and signal-to-noise ratios may, again, be problematic and appear to be further down the causal pathway to assess clinical effectiveness and/or non-inferiority. These issues are explored further in Chapter 6, Recommendations for further research.

Summary

The policy-relevant question that underpinned this pilot trial was whether or not role substitution is clinically effective in NHS dentistry. By virtue of the nature of the policy context for role substitution, it did appear that low-risk populations were the most relevant population to test a priori. The results from this study suggest that relaxing the eligibility criteria to include NHS patients with greater risk and greater clinical need may be more appropriate from a design perspective. However, the ratio of low to high risk remains problematic in a practice context, given the potential of the former diluting any effect seen in the latter. This does not mean that the results of this pilot lack any significance from a policy perspective. The results suggest that DTs could provide care for low-risk regular attenders in their own right or between longer recall frequencies with GDPs, although caution needs to be exercised in overinterpreting the findings from a statistical perspective. Given that the oral health of regular attenders is improving (see Chapter 1), the results of this study will help inform ongoing and future NHS dental reform programmes across the UK.

Introduction

Workstream 2 was a theoretically-driven process evaluation that focused on the micro, meso and macro factors that influence implementation of role substitution in general dental practice. WS2 ran in parallel to the pilot trial using a realist-informed approach to understand ‘what is it about a programme that works for whom, in what circumstances, in what respects, over which duration’. 104 This built on the qualitative work undertaken alongside WS1, which explored operational factors that influenced the delivery of the pilot trial. Realist methodology is used extensively in health services research; the approach recognises the complex and contingent nature that underpins the settings for new interventions and service delivery. This approach alongside process evaluations ‘supposes that regularities in the patterning of social activities are brought about by the underlying mechanism constituted by people’s reasoning and the resources they are able to summon in a particular context’. 124

Aims and objectives

The aims and objectives of this WS were to undertake a process evaluation informed by realist methodology to understand what works, for whom, why and in what circumstances.

Methods

Moore et al. 103 argue that process evaluations should capture the contextual factors that shape the intervention, mechanisms that sustain or potentiate effects, unexpected pathways and consequences and the contextual factors that shape implementation. To aid the process, realist principles were used to develop an understanding of ‘what works about the intervention, why, how and under which circumstances’. 105,125 As Rycroft-Malone et al. 126 articulate, theories can capture how complex interventions and systems can interact, so a realist approach tends to combine the elements of both substantive theory and stakeholders’ own theories about how an intervention might work:

. . . rather than identifying linear cause and effect relationships through successionist logic (x causes y: often illustrated through logic models), realist enquiry is concerned with identifying the underlying generative mechanisms about how interventions work [or not].

Rycroft-Malone et al. 126

Although logic models are helpful for specifying intervention components as inputs and outputs, they can be less useful for developing contingent explanations between them. 126 As highlighted by Brand et al. ,127 ‘to build implementation and delivery knowledge it is helpful to understand how context (individual, social, cultural, organisational) interacts with intervention components and underpinning mechanisms to bring about desired outcomes’. In this sense, a mechanism is both the resource that an intervention provides and the recipients’ reasoning and response to it. 128 Dalkin et al. 128 argue that these mechanisms are not dichotomous in their action (on–off) but, rather, operate across a continuum, interacting in a context to produce a specific outcome. Realist methodology defines mechanisms as a reaction to ‘a resource that the intervention provides and the recipients’ reasoning about and response to that resource, and the context in which that mechanism will be activated’. 125 As highlighted in Chapter 1, this form of approach adds real value to our overarching research question about the use of role substitution, given the range of factors that can influence its utilisation in NHS dental practices.

In this chapter, we also use the idea of theories to refer to key stakeholders’ ideas about how an intervention works, which is derived from their experience. 129 Kislov et al. 130 argue for a ‘shift of focus from ‘theory’ as a relatively isolated, static, reified source guiding implementation, towards embracing ‘theorising’ as a set of processes that aim to use empirical data actively in developing, validating, modifying, and advancing conceptual knowledge in the field’. Theories in this sense are defined as ‘an ordered set of assertions about a generic behaviour or structure assumed to hold throughout a significantly broad range of specific instances’. 131 Kislov et al. 130 argue that mid-range theories (MRTs) are particularly helpful in understanding the factors that influence the implementation of new working practices. Here, the evaluation is focused on a narrow range of factors that could potentially influence adoption, as opposed to grand theories, which aim to ‘construct all-encompassing meta-narratives that span space and time’. 131 Realist methodologies commonly present these theories as a function of contexts (C) and mechanisms (M) that lead to a particular outcome or set of outcomes (O) [f(C,M,O)]. This can necessitate the apportioning of a specific factor to either a context or a mechanism, which can sometimes appear to stakeholders as a contorted process, particularly at the outset where initial programme theories (IPTs) are being developed. As a result, a number of realist researchers are now simplifying this process and using IF-THEN propositions capture a combination of f(C,M) that lead to any given outcome. 132,133

The process for this study is outlined in Figure 5. Initially, this involved working with a stakeholder group to develop IPTs, framed as IF-THEN statements. This was followed by a process of testing and refining the IPTs using semistructured interviews to develop ‘mid-level’ programme theories, to provide an explanatory framework for role substitution in NHS dentistry. 105 It then culminated in a final stakeholder group to ‘sense-check’ these theories and place them in order of priority. Ethics approval for the study was granted by the East Midlands NHS Research Ethics Committee (#17/EM/0365), as part of the larger study.

FIGURE 5.

Overview of a realist-informed process.

Results

The results of the different stages of the realist-informed process evaluation are provided in Initial programme theory development, Theory testing, Refinement of the initial programme theories and Feedback on the final programme theory.

Refinement of the initial programme theories

As a result of our theory-testing phase, we prioritised and modified our IPTs (Table 16). These were then taken for further testing with a different set of stakeholders, as outlined in this section.

TABLE 16 - Modified IF-THEN statements

Theme	Number	IF-THEN statement
Contractual	1	IF the NHS contract allowed DTs to be on the National Performers List THEN they could open and complete a NHS course of treatment, increasing the relative efficiency of role substitution (through UDA generation) and the patient’s journey through the practice
Institutional logics	2	IF DTs are seen to be endorsed by the GDPs in the NHS practice THEN patients are more likely to trust being treated and managed by DTs
	3	IF DTs are used to their full Scope of Practice THEN they can pose a threat to some GDPs (e.g. associate dentists), leading to a reluctance to delegate clinical tasks
	4	IF DTs are used in NHS practices THEN they are commonly underutilised as DHs
Regulatory	5	IF DTs could prescribe medicines and radiographs THEN they could undertake NHS treatment without the need for the GDP to see the patient, which would improve the relative efficiency of role substitution and the patient’s journey through the practice
Patients’ experience	6	IF patients were better educated about the DT role THEN they would accept treatment by DTs more readily
	7	IF patients trust the DT THEN they are just as happy to be treated by a DT as they are a GDP
Logistics	8	IF NHS dental practices had sufficient surgery space and/or time THEN role substitution is more likely to be utilised

During this phase, attention was paid by the research team to the specific contexts and mechanisms that influenced the use of role substitution in NHS practices. In this section, we present the main findings in the different theory areas (transcripts are presented in Report Supplementary Material 2).

Contractual

General dental practitioner, DT and DC groups agreed that, under the current NHS contractual model, the use of role substitution can be difficult to implement. DTs work for NHS dental practice owners (providers) that are run as small businesses that seek to maximise profit while discharging their professional responsibility to their patients (context). NHS GDS regulations9 do not allow DTs to open a course of treatment (mechanism). This means that DTs cannot act as the front-line clinician, reducing practice efficiency (outcome) and causing disruption to the patient’s journey (outcome). All of the stakeholders agreed that this creates difficulties managing their ‘books’ if the GDP or DT have not synchronised their diaries, because any referrals from GDPs to DTs then require a separate appointment. Therefore, GDPs argued that the NHS dental contract does not support the use of a team-based approach (outcome). NHS GDS regulations9 require the claims for the number of UDAs undertaken to be split between those that opened a course of treatment and those that undertook the treatment. This creates an ‘internal market’ (mechanism) leading to competition in the practice between associate dentists and DTs and a disincentive for associate dentists to refer to DTs (outcome). In addition, GDPs argued that the target-driven nature of the current NHS contract to deliver to their ACVs (mechanism) does not allow the freedom or flexibility for GDPs to experiment with different models of care. Instead, GDPs argued that they have to be innovative, in spite of the current contract, to utilise role substitution.

One model that was suggested by the DTs was based on ‘upwards referral’ (akin to medicine), whereby a DT could triage patients and refer only if patients are seen to require treatment that is beyond their Scope of Practice. Although many DTs felt confident in recognising situations in which they were likely to encounter tasks that were beyond their Scope of Practice, some felt that additional training would be beneficial. As a professional group, DTs were highly frustrated by the time taken to refer a patient back to a GDP. At worst, the patient would need to be recalled to see the GDP for a check-up and then referred back to the DT, often taking months. This only served to confuse the patient and led to inefficient processes in NHS practice.

The DC group argued that there could be an option of using an intermediate performer level that would allow DTs to open preventative courses of treatment in the NHS, in a broader recall structure. This broader recall structure may include extended recall periods, DT fluoride application and GDP oversight for clinical safeguarding. The DC group also felt that the current NHS contract, based on UDAs, is not fit for purpose in contemporary dentistry. NHS GDS dental contract reform and other innovative initiatives were cited as opportunities to realign the remuneration system for DTs away from the reliance on UDAs. More broadly, it was suggested that the current contract did not sufficiently incentivise the level of access and the quality of care required by local populations.

Figure 6 presents contractual CMO configurations.

FIGURE 6.

Contractual CMO configurations.

Institutional logics

All stakeholders were in agreement that the use of DTs was strongly influenced by the prevailing culture in NHS practices and the equity-owner’s view(s) about role substitution and how best to deliver care for their patients and their community (context). Endorsement of ‘team dentistry’ and the provision of support and clinical leadership by the equity owner (provider) in an NHS practice (mechanisms) were seen as critical to help overcome some of the difficulties associated with role substitution in the current NHS contract. Examples provided by the stakeholders included their influence on securing additional surgery space (see Logistics), exploring different care pathways or internal payment systems for staff (performers) that work around current NHS service regulations. At its simplest, endorsement also related to the very practical task of providing reassurance for patients who are being referred from GDPs for treatment with DTs. There was unanimous agreement that a supportive team approach was required for effective use of role substitution. Endorsement of DTs through outward displays of trust, positive feedback and support were recognised as essential for effective implementation. There was also recognition that some GDPs require some experience of working with DTs. Overall, equity owners were seen to be supportive of role substitution in practice because ultimately it represented a mechanism for increasing profitability: meeting the needs of the local population using a less expensive workforce (outcome).

The GDP, DT and DC groups also recognised that DTs working to their full Scope of Practice can also be seen as a professional challenge (mechanism) to associate dentists and that this could influence the institutional logics in a NHS practice. This was for three reasons. The first mechanism, immediate conflicts in the ‘internal market’, was referred to above. The second related mechanism was the associate dentists’ perception that this added only complexity to intrapractice referrals. The third mechanism related to more medium-term concerns about DTs ‘taking over’ their future role.

Where role substitution was not working, a number of the stakeholders mentioned that DTs would often be undertaking DH roles (outcome). This was highlighted particularly by those in the DT stakeholder group, who stated that they’d often be asked to perform as DHs (outcome). Many DTs reported that there was a feeling of resignation among many in the profession and that they would be fortunate to work to their full Scope of Practice (outcome) in a typical NHS practice. Ultimately, this results in deskilling of the DTs workforce, leading to a loss of clinical confidence (outcome) and many DTs not returning to their full Scope of Practice (outcome).

Figure 7 presents institutional logics CMO configurations.

FIGURE 7.

Institutional logics CMO configurations.

Regulatory

NHS dental practices and DTs are required to work to the Ionising Radiation Regulations Act, Ionising Radiation (Medical Exposure) Regulations11 and Medicines Act12 and the GDC’s Scope of Practice (context). The DTs inability to prescribe radiographs, fluoride varnish and local analgesia were all cited by the stakeholder groups as creating substantive inefficiencies (outcome) and practice disruption (outcome) when managing NHS patients. The DT group argued that this could manifest itself as the need to rebook the patient’s appointment, which had significant financial implications under the current NHS contract. Some DTs identified mechanisms for working around the inefficiencies of obtaining a prescription; these typically involved leaving some ambiguity in the treatment plan or GDPs issuing prescriptions post hoc, based on a trusting clinical relationship with the DT. The inability to prescribe for even basic treatments reduced the DTs’ sense of professional autonomy and was an addition burden for the referring GDP (outcomes).

All stakeholder groups agreed that the current Scope of Practice for DTs could present challenges in a NHS practice environment. Examples were given of how certain clinical presentations would initially appear to be within the DTs’ scope but later require a GDP to intervene (outcome). In these instances, the inefficiencies (outcome) caused by the need for a GDP intervention outweighed the efficiencies gained by referring to a DT. If DTs had a wider Scope of Practice (mechanism), it was thought that these instances would be less common. There was unanimous opinion that, when operating within their Scope of Practice, DTs could perform clinical tasks as well as GDPs.

Figure 8 presents regulatory CMO configurations.

FIGURE 8.

Regulatory CMO configurations.

Patients’ experience

Patient acceptability was considered to be important at the theory-testing phase, but most patients who were interviewed in this phase argued that they would accept new ways of working. Given the importance of capturing the patient voice, key themes are supported by quotations from the patient group:

I probably wouldn’t be bothered who I saw in the first place, as long as you knew there was a structured approach.

22.33

But if I went somewhere else, I would accept the way they worked straightaway, almost certainly, and probably without question.

22.636

I would have faith in the profession, that they would do that. It’s not as if you just drag somebody off the street and say, fill this bloke’s teeth. They have been trained.

25.132

I would go there in expectation that I was going to be seen by the most appropriate person for my condition . . . I wouldn’t want to know, I’d just expect, it’s like if the car mechanic said, which of our mechanics would you like to do your car.

25.227

I would make the assumption that once you’d asked the question, they were qualified to deal with what was wrong, and off you go, give it a try.

27.459

Put our faith in the system. So, I would, unless the person that was treating me was very indecisive or didn’t look the part, or looked a bit rough or something, which gave me doubt as to their proficiency, then otherwise I’d accept that the dental practice knew what they were doing.

25.172

I’ve absolutely no preference who I see. I know some people are really adamant that they only want to see the person that’s the most specialised, but I think that’s just ignorance as to what that person can do really.

26.355

I don’t think it’s the job title or . . . it’s just the situation. You know, it’s fine, I don’t think it matters who’s doing it, as long as they can meet the needs, and that’s what’s happening . . . you know, they want to know they’ve seen the right person for it to be resolved.

27.268

Yes, so, yes I think in that situation. And also, if you’ve got a persistent problem, you really would want it fixed . . . So, I think that, you know, if it was really bothering you, you’d see whoever could solve it, you wouldn’t mind.

27.449

Some of the patient stakeholders also raised the issue of complexity as a factor:

You’re in a more complex situation, maybe having bridges . . . which, touch wood, you know, we’re not in, then I can understand why somebody might feel that they needed a particular person. But people like us, we wouldn’t mind at all, we wouldn’t be bothered. We’d be quite happy.

27.131

Others argued about the importance of consistency, in terms of the clinicians that they see in practice:

That’s probably one thing that will probably affect the changes more than anything else is that somebody my age probably for simplicity expects the same person because it’s stability, it’s a sense of comfort.

22.289

Perhaps older people who are more used to seeing, the dentist, you know, the authority figure, man, usually, traditionally, might be a bit wary, perhaps the other people, not so much.

25.423

They also emphasised the role of the whole of the dental team:

It’s the receptionist which is the first face that you meet, which is probably the person that can make the difference.

22.25

A key element was the need for the practice to communicate any process of change:

But as long as I was explained the benefits I would receive, and I think they’ve got to be, if that was projected that way, then I would probably see that.

22.785

Our user researcher provided her thoughts on the key context, mechanisms and outcomes that govern patients’ response to role substitution in NHS dentistry. Given that the public have no understanding of the current NHS system (context), a number of key mechanisms were identified:

• A low level of patient awareness of what DTs can do (mechanism) means that patients are receptive to DT utilisation (outcome) only if they are explained the benefits (mechanism).

• Patients want minimum disruption to their daily lives (mechanism).

• Continuity and consistency are important (mechanism) to build trust.

• Trust in the ‘system’ of NHS dentistry (mechanism) also reduces patient anxiety (outcomes).

Figure 9 presents patients’ experience CMO configurations.

FIGURE 9.

Patients’ experience CMO configurations.

Logistics

Both the GDP and the DC groups argued that there were significant logistical barriers to the successful implementation of role substitution. Role substitution in NHS dentistry requires additional surgery space and/or surgery time to enable DTs to undertake patient management (context). The GDP and the DC groups argued that NHS dental practices that had only one or two chair practices would find it difficult to make role substitution work under the current NHS contract. Role substitution is predicated on NHS practices having sufficient clinical space to employ a DT. This was often difficult, because most high-street practices were not purpose built for team dentistry. Instead, their design had been strongly influenced by the incentives in previous NHS contracts that rewarded high levels of clinical activity undertaken by GDPs to maximise profitability (mechanisms). This ‘small business’ economic model was exacerbated further by the fact that NHS practices were essentially independent contractors with no structural investment from the NHS (mechanism). This meant that practice expansion was difficult (outcome) and that members of the dental team that could maximise profit (DHs) were preferred over DTs (outcome).

Some suggestions were made by the DCs about how best to counter this, including extended opening hours, building extensions and forming hubs for local health services in primary care, where capital resources would be pooled between providers to enable patients to access NHS dentistry, ophthalmology and pharmacy.

Figure 10 presents logistics CMO configurations.

FIGURE 10.

Logistics CMO configurations.

A common theme to these five areas of programme theory relates to the NHS dental contract. These five areas have been collated into one model (Figure 11) to highlight the different influences on role substitution in NHS dentistry. It highlights how contractual issues affect the business model in NHS dentistry and the institutional logic of the individual NHS dental practice. In turn, these can affect the logistics within the practice in terms of decisions around the numbers of surgeries needed and how much of the available clinical time is given to role substitution. It also highlights how regulatory issues affect the business model of the NHS dental practice while concomitantly directly affecting utilisation (given that DTs have a duty in their own right to discharge their Scope of Practice in accordance with the GDC regulations). Regulatory and contractual issues are seen to operate at a macro level and influence the business model and institutional logic at the micro level. These factors operate independently to the meso level, where there is little that DCs can do to facilitate role substitution (given the contractual and regulatory underpinning of the NHS system at the macro level), bar structural investment, which is difficult given the independent contractor status of NHS providers.

FIGURE 11.

Overarching model.

Discussion

The realist-informed process evaluation added value to the study and captured many of the themes that were discussed in Chapter 1. Following a detailed and iterative process, it elicited five principle areas of programme theory: contractual, institutional logics, regulatory, patients’ experience and logistics. This enabled the research team to understand and capture the contextual factors that shaped the intervention, the mechanisms that sustained or potentiated effects, the unintended consequences and the contextual factors that shape the use of role substitution in NHS dentistry. 103 As highlighted by Rycroft-Malone et al. ,126 these theory areas were complex and to a large extent interdependent. The underlying generative mechanisms were largely at the macro and the micro level, with the latter being driven by the former and related to individual, social, cultural and organisational factors. 127 They were also seen to operate across a continuum and provided the research team with a rich explanatory narrative of the factors that potentially influence adoption. 128,130

The two most important elements that appear to shape the practice of NHS dentistry in the UK are contractual and regulatory factors. Despite the relaxation of the Scope of Practice from the regulator’s perspective, the requirements under the GDS contract for dentists to open a course of treatment appears to significantly limit the practice of role substitution in NHS dentistry. 9,10 This is further potentiated by the fact that DTs cannot contract directly with DCs, which appears to undermine strategic objectives at the level of the policy-maker to promote the use of the whole dental team. 9,14,15 The results from this WS also highlight how this appears to fundamentally affect the institutional logics of NHS practices. As highlighted by Goodwin et al. 48 and Watt et al. ,49 the institutional logics at any given NHS practice include not only dentistry as a business but also professional ethics and contextual factors, based on where the practice is embedded. The drive to maintain (and maximise) the viability of a NHS practice does appear to be tempered by a practice owner’s view about their sense of duty to their patients and their ideas about how best to deliver care for their patients and community. 50 However, contractual incentives remain a fundamental driver. Therefore, the results in this study concur with our earlier work, where the views of practice principals were found to be highly influential. 30 This is important, as promoting role substitution and any potentially disruptive transition caused by changes to local working practices takes time and good human resource skills. 53 In this study, most examples of successful implementation of role substitution in NHS dentistry appeared to be where the practice owners had found workarounds to the contractual and regulatory barriers.

The level of social acceptability of role substitution appeared to be high and patients appeared to trust the ‘system’ of NHS dentistry. Most suggested that they’d accept new ways of working but would want NHS practices to articulate the benefits of being managed by DTs. Equally, many highlighted the importance of continuity and consistency in their care and the need to keep disruption at a minimum. This appears key to the success of role substitution in NHS dentistry and was also raised by many dentists and DTs, who were frustrated at the difficulty of ensuring a smooth transition for patients, given the restrictions in the GDS contract.

From a dental public health perspective, these factors are critically important given the trend in future population health needs. 16,29 Oral health is improving overall and where there remains existing intransigent problems or new oral health challenges there will become a pressing need to better target the available resources in NHS dentistry. 17–19 All of the dental contract reform programmes undertaken across the UK place an emphasis on increasing patient access and prevention while maintaining and improving the quality of service being delivered. 59–68 It would appear that a key element in any reform programme would be to change the underlying contractual and regulatory incentives.

Summary

Substantial barriers exist to the practice of role substitution in NHS dentistry. The two most dominant factors were contractual and regulatory barriers. These appeared to drive the institutional logics at the micro level, with little influence exerted at the meso level by DCs. Where role substitution was successful, practice owners had found workarounds to the macro factors that dominated any potential implementation.

Introduction

This chapter reports the methods and results of the within-trial economic evaluation (see Methods, Within-trial economic evaluation, and Results, Within-trial economic evaluation, respectively). This analysis developed and tested the health economic data collection tool. It also rehearsed the economic evaluation that would be conducted in the definitive trial of the intervention compared with current practice. Both pieces of work were designed to inform the definitive trial design. With regard to effects, we used the clinical measure BOP and oral health-related quality of life as measured using the OHIP-14 when estimating cost-effectiveness. The economic evaluation was conducted following best practice guidelines conforming to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS). 137

The clinical measure BOP and the OHIP-14 are not ideal for use in an economic evaluation because it is unclear how important decision-makers consider a unit change in either measure to be. This means that any calculated incremental cost-effectiveness ratio (ICER) is difficult to interpret except in the sense that higher values for an ICER are less likely to be considered worthwhile than lower values. For this reason, preparatory work undertaking an investigation of patients’ preferences for being seen by DTs over being seen by GDPs is undertaken. Participant views were explored in the setting of focus groups and one-to-one interviews. The results of the qualitative work informed the development of a preference elicitation exercise (e.g. a DCE) that would be administered as part of the future definitive trial. The preference elicitation exercise in a definitive trial would be able to provide a metric to value the benefits of the intervention in monetary terms. This would extend the cost-effectiveness analysis into a cost–benefit analysis (in which both costs and benefits are measured in commensurate, normally monetary, units). The methods and results of this work are presented in this chapter (see Results, Discrete choice experiment: preparatory work, and Results, Discrete choice experiment: preparatory work, respectively).

This chapter finishes with a brief summary of the principal findings of the two elements of the health economics work (see Summary).

Aims and objectives

The aims of WS3 were to:

• pilot the health economic data collection tool and rehearse the health economic analysis to inform the definitive trial design

• explore patients’ preferences through focus groups to inform a DCE in a definitive trial.

Methods

Results

Within-trial economic evaluation

Completeness of data

A total of 167 participants were eligible for inclusion in the primary ITT sample and were included in the analysis of costs. At 6 months, 133 respondents returned the PCQ; however, owing to missing ID numbers, this could not be matched to individuals and is not included in this analysis. A total of 12 respondents (95 included in the ITT sample comprising 47 and 48 in the DT and GDP arms, respectively; 28 were subsequently lost to follow-up) returned the PCQ at 12 months, describing health-care appointments and out-of-pocket costs in the past 6 months. Among these respondents, 44 provided complete time and travel details of a health-care appointment in the past 6 months; these were used to generate a time and travel unit cost.

Estimates of resource use and cost

Table 17 describes resource use. No formal testing was conducted to explore difference between groups. Similar mean numbers of visits to dental practices were observed, with the average appointment duration being longer in the DT group (see Table 10). The number of band 1, 2 and 3 treatments were similar in both arms, as were the number of emergency appointments. The mean amount of time spent with a GDP in the DT group appeared to be large, although this is understandable, because the GDP would be undertaking more complex and time-consuming procedures given the nature of the study (where these were likely to be treatments that could not be undertaken by the DT). Notably, no participants reported use of other primary care services and this suggests that data collection might be simplified for the definitive trial. The number of people reporting additional DH visits in the PCQ was small (n = 2), as was the number of people referred to secondary care (n = 4) and the number of those reporting a visit to an emergency or out-of-hours dental service (n = 1).

TABLE 17 - Resource use per patient over the trial period in the ITT analysis

A&E, accident and emergency; N/A, not applicable.

In the DT arm some participants received band 3 treatments, which were provided by a GDP and costed accordingly.

Resource use	Data source	DT arm (n = 83), mean (SD)	GDP arm (n = 84), mean (SD)
Number of visits to dental practice	CRF	3.3 (2.3)	3.0 (2.2)
Average recall appointment duration (minutes)	CRF	18.4 (5.9)	14.0 (6.2)
GDP time (minutes) in DT arm	CRF	14.5 (39.0)	N/A
DT time (minutes) in GDP arm	CRF	N/A	0.2 (2.2)
Number of visits to other primary care, mean (SD)	PCQ	0 (0)	0 (0)
Number of band 1 courses of treatments	CRF	1.9 (0.9)	1.8 (0.9)
Number of band 2 courses of treatments	CRF	0.3 (0.7)	0.2 (0.5)
Number of band 3 courses of treatmentsa	CRF	0.1 (0.3)	0.1 (0.4)
Number of emergency band treatments	CRF	0.3 (0.8)	0.2 (0.5)
Number of dental specialist visits	PCQ and CRF	0.02 (0.15)	0.02 (0.15)
Number of emergency visits (e.g. to A&E or dental emergency clinic)	PCQ	0.1 (0.5)	0 (0)
Number of additional hygienist appointments	PCQ	0.09 (0.47)	0 (0)

Table 18 describes the average (mean) cost per patient for each cost category in each arm. As would be expected, the similar UDA payments reflect the similar number of bands described in Table 17. Prescription costs were low for both groups owing to a small number of prescriptions being reported in the CRF (5 in the DT arm and 4 in the GDP arm) and the relatively low unit cost of these medications. Given the minimum impact on cost estimates, this also suggests that data collection might be simplified for the definitive trial.

TABLE 18 - Average cost per patient for each cost category in each arm

All costs are in 2019 GBP.

Cost type	DT arm (n = 83), mean (SD)	GDP arm (n = 84), mean (SD)
NHS (£)
UDA payments	126.50 (127.74)	114.93 (130.10)
Prescription medicines cost	0.15 (0.65)	0.13 (0.63)
Secondary care costs	4.66 (24.22)	2.95 (19.03)
Total NHS costs	131.31 (132.57)	118.01 (131.26)
NHS costs minus patient contributions (prescription and dental charges)	44.30 (70.29)	36.75 (64.03)
Time/materials (£)
Reusable and consumables costs	22.08 (30.41)	24.53 (53.18)
Staff costs	39.78 (42.80)	44.70 (60.52)
Total	61.86 (71.07)	69.22 (111.67)
Patient (£)
NHS dental charges	86.57 (93.40)	80.73 (104.68)
Prescription charges	0.43 (1.94)	0.54 (2.14)
Over-the-counter medication charges	0.16 (1.37)	0.21 (1.45)
Private treatment	21.00 (153.48)	5.02 (40.38)
Time and travel costs	12.06 (8.05)	10.37 (7.48)
Total	120.23 (176.62)	96.86 (122.82)

Although the cost of private care appears higher in the DT arm, it is worth noting that private treatment was recorded for a small number of patients (n = 6) and the private costs for one of these patients in the DT arm was nearly £1400. Thus, although infrequent, the use of private care would be important to capture in a definitive trial given its potentially high costs.

Figure 12 illustrates the distribution of costs in each of the two arms. This shows that the data are highly skewed (as expected), with a small number of participants incurring substantially higher costs and > 60% of participants in both arms having total NHS costs of < £50.

FIGURE 12.

Total NHS cost by arm: (a) GDP arm and (b) DT arm.

Clinical effectiveness data

The evidence on the impact of the intervention on both BOP and OHIP-14 scores are described in detail in Chapter 2. Figures 13 and 14 describe the distribution of scores in the study population. As Figure 13 illustrates, more participants in the GDP arm than in the DT arm had close to zero sites (where there was BOP). However, in both arms a small number of participants had BOP at 10% of sites. The low proportion of sites with BOP is to be expected given the selection criteria for the study, but it is possible that the distributions would change in a larger sample (should a definitive trial become available).

FIGURE 13.

Proportion of sites with BOP by arm: (a) GDP arm and (b) DT arm.

FIGURE 14.

Oral Health Impact Profile-14 items score by arm: (a) GDP arm and (b) DT arm.

With respect to OHIP-14 scores, the majority of participants had a very low score (where low scores indicate better oral health) (see Figure 14). However, a small number of participants in both arms had higher scores. As with the other outcomes, it is possible that the distribution of scores would change if data from more a sufficiently large sample were available (e.g. from a definitive randomised controlled trial).

Incremental cost-effectiveness analysis

The adjusted cost-effectiveness analysis investigating BOP showed that the DT strategy was, on average, less effective (incremental proportion of sites with BOP 0.0006, 95% CI –0.005 to 0.006) and more costly (incremental cost £14.85, 95% CI –£8.50 to £38.20) than the GDP strategy. Therefore, in terms of BOP, the DT strategy is dominated by the GDP strategy and, thus, no ICER is calculated (Table 19).

TABLE 19 - Summary of the cost-effectiveness analysis

WTP, willingness to pay.

There were missing clinical outcome data for one participant, who is not included in this analysis.

A positive incremental proportion of site with BOP means that DT is more effective than GDP.

Intervention strategy	Mean NHS cost (£) (95% CI)	Mean proportion of sites with BOP (95% CI)	Incremental cost (£) (95% CI)	Incremental proportion of sites with BOP (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP per proportion of site with BOP avoided
						£0	£100	£500
Incremental cost per additional site with BOP avoided
DT (n = 82)a	55.66 (34.15 to 77.17)	0.018 (0.013 to 0.023)	14.85 (–8.5 to 38.20)	0.0006 (–0.005 to 0.006)	Dominated	9.3	9.0	9.2
GDP (n = 84)	36.75 (22.86 to 50.65)	0.013 (0.009 to 0.017)
Intervention strategy	Mean NHS cost (£) (95% CI)	Mean OHIP-14 score, points (95% CI)	Incremental cost (£) (95% CI)	Incremental OHIP-14 point (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP for each OHIP-14 point reduced
						£0	£100	£500
Incremental cost per OHIP-14 point reducedb
DT (n = 82)a	55.66 (34.15 to 77.17)	0.54 (0.09 to 0.97)	14.85 (–8.5 to 38.20)	–0.64 (–1.51 to 0.23)	23.20 per OHIP-14 point reduced	9.3	84.4	92.3
GDP (n = 84)	36.75 (22.86 to 50.65)	1.01 (0.26 to 1.76)

When benefits were considered in terms of the OHIP-14 score, the DT strategy was, on average, more effective (incremental OHIP-14 score –0.64 points, 95% CI –1.51 to 0.23 points) but more costly (incremental cost £14.85, 95% CI –£8.50 to £38.20) than the GDP strategy (see Table 19). The incremental cost per unit reduction in OHIP-14 score was £23.20.

The cost-effectiveness plane (Figure 15) illustrates that, for the majority of the bootstrap iterations, the DT strategy is more likely to have more sites with BOP than the GDP strategy. Contrary to this, in terms of OHIP-14 score the majority of the bootstrap estimates show that the DT strategy provides better OHIP-14 scores than the GDP strategy. The DT strategy is also likely to be more costly than GDP (Figures 16 and 17).

FIGURE 15.

Cost-effectiveness plane showing the incremental proportion of sites with BOP and incremental cost (DT vs. GDP). Because the expected difference in BOP between arms is very small (over-the-threshold ranges of WTP considered), the decision is that the GDP arm is more likely to be cost-effective, primarily as a result of lower cost.

FIGURE 16.

Cost-effectiveness acceptability curve showing probability cost-effectiveness for a given WTP threshold per proportion of site with BOP avoided. A positive incremental OHIP-14 score means that the DT arm is associated with a higher quality of life than the GDP arm. WTP, willingness to pay.

FIGURE 17.

Cost-effectiveness acceptability curve showing probability cost-effectiveness for a given WTP threshold per OHIP-14 point reduced. WTP, willingness to pay.

With respect to BOP, Figure 16 shows that there is a 91% chance that the current practice with GDPs would be a more cost-effective strategy than a DT strategy should society be willing to pay £100 per proportion of site with BOP avoided. With respect to OHIP-14, there is an 84.4% chance that the DT strategy would be a more cost-effective strategy than the current practice with GDPs, should society be willing to pay £100 for a 1-point improvement in OHIP-14 score (Figure 18).

FIGURE 18.

Cost-effectiveness plane showing incremental OHIP-14 score and incremental cost (DT vs. GDP). The GDP strategy is expected to be less costly, as it is for BOP, but for the OHIP-14 score, it is also expected to be less effective (although the CIs for a difference include ‘no difference’).

The evidence on costs and cost-effectiveness should be treated cautiously because the results reported are not conclusive given that they were derived from pilot data. Figures 12–14 illustrate that there was marked variability in both cost and clinical effectiveness parameters and it is likely that these results may change substantially, should data from more participants become available in a definitive trial.

Sensitivity analysis

The results of the sensitivity analyses for which high (Table 20) or low (Table 21) NHS costs are used are very similar to the base-case analysis, indicating that the results are not sensitive to the differing methods for costs that have been considered in this WS. Furthermore, the cost-effectiveness results that take into account the patient and carer perspective (derived from PCQ data) do not change the main findings either (Table 22).

TABLE 20 - Sensitivity analysis using high NHS costs

WTP, willingness to pay.

There was missing clinical outcome data for one participant, who is not included in this analysis.

Note

Incremental costs, incremental proportion of sites with BOP and incremental OHIP-14 point estimates derived using sureg in Stata, adjusted on covariates the age, sex, baseline proportion of sites with BOP, baseline OHIP-14 score, charge exemption, number of teeth and dental practice.

Intervention strategy	Mean NHS cost (£) (95% CI)	Mean proportion of sites with BOP (95% CI)	Incremental cost (£) (95% CI)	Incremental proportion of sites with BOP (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP per proportion of site with BOP avoided
						£0	£100	£500
Incremental cost per proportion of sites with BOP reduced
DT (n = 82)a	81.15 (55.77 to 106.53)	0.018 (0.013 to 0.023)	15.41 (–13.13 to 43.95)	0.0006 (–0.005 to 0.006)	Dominated	15.7	15.7	15.6
GDP (n = 84)	59.73 (41.94 to 77.53)	0.013 (0.009 to 0.017)
Intervention strategy	Mean NHS cost (£) (95% CI)	Mean OHIP-14 score, points (95% CI)	Incremental cost (£) (95% CI)	Incremental OHIP-14 point (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP for each OHIP-14 point reduced
						£0	£100	£500
Incremental cost per OHIP-14 point reduced
DT (n = 82)a	81.15 (55.77 to 106.53)	0.54 (0.1 to 0.97)	15.41 (–13.13 to 43.95)	–0.64 (–1.51 to 0.23)	24.08 per OHIP-14 point reduced	15.7	84.0	92.3
GDP (n = 84)	59.73 (41.94 to 77.53)	1.01 (0.26 to 1.76)

TABLE 21 - Sensitivity analysis using low NHS costs

WTP, willingness to pay.

There was missing clinical outcome data for one participant, who is not included in this analysis.

Note

Incremental costs, incremental proportion of sites with BOP and incremental OHIP-14 point estimates derived using sureg in Stata, adjusted on covariates the age, sex, baseline proportion of sites with BOP, baseline OHIP-14 score, charge exemption, number of teeth and dental practice.

Intervention strategy	Mean NHS cost (£) (95% CI)	Mean proportion of sites with BOP (95% CI)	Incremental cost (£) (95% CI)	Incremental proportion of sites with BOP (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP per proportion of site with BOP avoided
						£0	£100	£500
Incremental cost per proportion of sites with BOP avoided
DT (n = 82)a	30.18 (11.60 to 48.75)	0.018 (0.013 to 0.023)	14.30 (–5.42 to 34.02)	0.0006 (–0.005 to 0.006)	Dominated	5.1	5.0	4.6
GDP (n = 84)	13.76 (2.22 to 25.31)	0.013 (0.009 to 0.017)
Intervention strategy	Mean NHS cost (£) (95% CI)	Mean OHIP-14 score, points (95% CI)	Incremental cost (£) (95% CI)	Incremental OHIP-14 point (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP for each OHIP-14 point reduced
						£0	£100	£500
Incremental cost per OHIP-14 point reduced
DT (n = 82)a	30.18 (11.60 to 48.75)	0.54 (0.1 to 0.97)	14.30 (–5.42 to 34.02)	–0.64 (–1.51 to 0.23)	22.34 per OHIP-14 point reduced	5.1	85.3	92.5
GDP (n = 84)	13.76 (2.22 to 25.31)	1.01 (0.26 to 1.76)

TABLE 22 - Sensitivity analysis using patient costs

WTP, willingness to pay.

There was missing clinical outcome data for one participant, who is not included in this analysis.

Note

Incremental costs, incremental proportion of sites with BOP and incremental OHIP-14 point estimates derived using sureg in Stata, adjusted on covariates the age, sex, baseline proportion of sites with BOP, baseline OHIP-14 score, charge exemption, number of teeth and dental practice.

Intervention strategy	Mean patient cost (£) (95% CI)	Mean proportion of sites with BOP (95% CI)	Incremental cost (£) (95% CI)	Incremental proportion of sites with BOP (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP per proportion of site with BOP avoided
						£0	£100	£500
Incremental cost per proportion of sites with BOP avoided
DT (n = 82)a	121.06 (82.05 to 160.10)	0.018 (0.013 to 0.023)	17.00 (–29.15 to 63.14)	0.0006 (–0.005 to 0.006)	Dominated	28.9	29.1	28.3
GDP (n = 84)	96.86 (70.21 to 123.52)	0.013 (0.009 to 0.017)
Intervention strategy	Mean patient cost (£) (95% CI)	Mean OHIP-14 score, points (95% CI)	Incremental cost (£) (95% CI)	Incremental OHIP-14 point (95% CI)	ICER (£)	Probability (%) of the DT arm being considered cost-effective at different threshold values for society’s WTP for each OHIP-14 point reduced
						£0	£100	£500
Incremental cost per OHIP-14 point reduced
DT (n = 82)a	121.06 (82.05 to 160.1)	0.54 (0.1 to 0.97)	17.00 (–29.15 to 63.14)	–0.64 (–1.51 to 0.23)	25.60 per OHIP-14 point reduced	28.9	79.6	91.6
GDP (n = 84)	96.86 (70.21 to 123.52)	1.01 (0.26 to 1.76)

Discussion

This WS has demonstrated that we can collect the data required for an economic evaluation using costs based on either UDAs or microcosting of dental care. The economic evaluation has been rehearsed and consistent results have been produced across the range of analyses conducted. On average, the DT strategy was more costly and less clinically effective in terms of BOP. Over the range of values for society’s WTP to get one less site of BOP, it is debatable whether or not DTs are cost-effective (although this may have been influenced by the floor effects of the BOP measure). In contrast, in terms of OHIP-14 scores, the DT strategy would be increasingly likely to be considered cost-effective as society’s WTP threshold for a 1-point reduction in OHIP-14 score. GDP care is estimated to be less costly but also less effective (although the CIs for a difference include ‘no difference’). Given that the measure of cost-effectiveness is incremental cost per 1-point improvement in OHIP score and the mean difference in OHIP score is likely to be approximately 0.5 points, the ICER is likely to be less than the WTP thresholds considered, especially when the WTP threshold increases.

Critical here is ‘what is society’s WTP for a unit change in each outcome measure?’, which is uncertain. It is for this reason that we undertook the preparatory work for a DCE to be developed. An alternative to a DCE would be a contingent valuation study. However, valuing a unit reduction in BOP or OHIP score would exclude people’s preferences over the way that care is provided. However, it is important to note that, although we have been able to use the data to estimate cost-effectiveness results, these may be unreliable because we do not have an accurate representation of costs and possible outcomes. For example, private costs are uncommon but very high. As the incidence of incurring private costs is quite low, we may not have an accurate estimate of the impact on total mean of private costs.

The interviews demonstrated that there are many issues that patients consider in relation to the use of DTs to conduct check-ups. The rationale for including a wider sample was to elicit a broad range of perspectives, but this also meant that we gathered data from a broader range of participants with disparate views. These data suggest that the social acceptability of using DTs is mixed. They also highlight the need for patient education in terms of how a new system might work, alleviating concerns about training and competence, and the referral process. This suggests that understanding patients’ information needs ahead of service change is vital, particularly because concerns about continuity of care and loss of long-held personal relationships with GDPs were raised a number of times.

The potential attributes and levels collected in the study primarily relate to process and organisational factors about the service. One of the proposed purposes of the DCE is to develop a tool that could be useful to gather information for NHS decision-makers. A potentially important element of this is the inclusion of a cost attribute, which appears appropriate, given the responses provided in the interviews.

Although we have demonstrated that we can collect data on the use of services, in the pilot study sample only low levels of service use were reported. Some of these data, such as on referrals to secondary care, were duplicated in the CRF. This raises questions about the value of a detailed PCQ. An alternative solution might be to ask participants about the use of services at their appointment at the same time as OHIP-14 data are collected. For the majority of participants a simple yes/no question would only be needed.

The estimation of the use of private treatment was also unclear. Some data were collected in the CRF but in the PCQ the question focused on ‘problems with face and mouth’. This blurred the line between dental services for maintaining or improving oral health and dental services for cosmetic services. Furthermore, although there were no significant differences in any cost category (NHS, patient and time/materials) observed between those who had not returned PCQs and those who had returned PCQs (p > 0.05), the assumptions of missing data in the PCQ meant that where patients did not return questionnaires describing additional costs these costs were potentially underestimated.

The data collection via the CRF facilitated a microcosting exercise. For the definitive trial, consideration needs to be given to the value of this exercise. The microcosting arguably identifies the ‘economic’ cost of care as opposed to the charges for care. Differences between the two illustrate potential inefficiencies caused by charges that do not reflect the economic costs. However, such data might not be considered of sufficient importance to NHS decision-makers given the costs of collecting such data.

With respect to the UDA costing, further information about treatment bands and patient charges could have been included. The precision of current estimates of cost may be improved by revisions to the CRF, which might be completed at the time of each appointment. However, this would be balanced against the increased time for busy staff to complete the CRF.

With respect to measures of cost-effectiveness, it is unclear what the value is of a unit change in either of the two measures used in the pilot. This limitation could be overcome by seeking to formally estimate the value of a unit change. However, the findings of the qualitative work reported in the second half of this chapter suggest that a broader range of impacts is worth consideration.

Summary

This WS has demonstrated that a meaningful economic evaluation could be conducted and has provided valuable information to inform a definitive trial. The qualitative component of this WS has elicited views from patients regarding the acceptability of management of low-risk patients by DTs and allowed us to compile a list of potential attributes and levels for inclusion in a preference elicitation exercise in a definitive trial.

Summary of main findings

The aim of this study was to inform the design for a definitive trial by undertaking a pilot study to determine whether or not DTs can maintain the oral health of low-risk routine NHS patients. WS1 was an individually randomised pilot study conducted over a 15-month period. Eight NHS practices from the North West of England recruited 217 NHS patients, who were randomised into one of two groups. One involved DTs undertaking the routine check-up and any treatment during the intervention period and the other was ‘treatment as usual’ (using GDPs to undertake the check-up and treatment). WS2 was a theoretically driven process evaluation that ran in parallel to the pilot to understand ‘what worked, for whom and under what circumstances’. 104 In WS3, we developed and tested a health economic data collection tool and rehearsed the health economic analysis of the proposed intervention. We also explored patients’ preferences to inform a preference elicitation exercise to inform the design of the definitive trial.

In WS1, we approached 546 patients and 217 were recruited over the 15-month period, equating to a mean recruitment rate of 43.4 participants per month. The attrition rates were comparable with our feasibility study. 99 These were 22.4% and 23.6% for the DT group and the GDP group, respectively, and there were no statistically significant differences in attrition across the randomisation stratification variables.

A total of 13 out of the 14 protocol deviations occurred because participants were allocated to the DT group but were seen once by the GDP in error. The remaining protocol deviation occurred because a participant was allocated to a group before being randomised. The recruited participants in WS1 were predominantly female (72.4%), non-smokers (92.2%), white (93.1%) and were not exempt from patient charges (89.4%). Deprivation scores across the two groups did not show any statistically significant differences.

We found no difference between the two groups in BOP over the 15-month period in the ITT analyses. Therefore, the performance of the DT group in this pilot study was non-inferior to the GDP group. This lack of difference between the two groups was also found across all of the secondary clinical, patient-related and process outcomes. The only substantial difference to note was that 20% more participants in the DT group changed their brushing behaviour. These results were also supported by the per-protocol analyses. Therefore, the findings appear to suggest that DTs were non-inferior to GDPs in the context of a pilot trial run over a 15-month period. However, it is important to be cautious and not overinterpret these findings. 118,119 As we highlighted in Chapter 2, it is possible that the estimates that we used may have still lacked precision, particularly the variability on our proposed primary outcome, and this may have influenced our ability to determine non-inferiority (because of the reliance on the CI estimates for the parameters). We explore these issues further in Chapter 6.

Despite high retention and fidelity rates, recruitment rates were moderately low (39.7%), which was lower than our earlier feasibility study. 99 This may have been caused by the broader range of practices recruited in the pilot study compared with the two practices recruited in our earlier study. To reduce self-selection bias, we deliberately targeted a wide range of NHS practices across the North West of England, which would need to be expanded further for a definitive study. The mean recruitment rate of 43.4 participants per month, along with the non-inferiority design, may have substantial implications for a definitive trial in a low-risk population group. The duration of the recruitment period could become extensive and this could be compounded by the time needed in the intervention period to detect enough of a signal to make comparison between the two groups meaningful.

In WS2, we elicited five areas of programme theory: contractual, institutional logics, regulatory, patients’ experience and logistics. These theory areas were largely interdependent and operated at both a macro and a micro level. It appears that the two most important elements that shape the use of role substitution in NHS dentistry are contractual and regulatory factors, which in turn drive institutional logics at a micro level. All stakeholders were in agreement that the current NHS contractual model makes it difficult to implement role substitution. Despite the relaxation of the Scope of Practice from the regulator’s perspective, the requirements under the GDS contract for GDPs to open a course of treatment appeared to limit the extent of role substitution in NHS dentistry. 9,10 DTs work for NHS dental practice owners (providers) that are largely run as small businesses. The inability of DTs to open courses of treatment means that DTs cannot easily act as the front-line clinician. This has knock-on effects in terms of reducing the efficiency of the practice and in many cases causes disruptions to the patient’s journey. The ‘provider’ and ‘performer’ split also tends to create an ‘internal market’ in NHS practices, leading to competition for work between associate GDPs and DTs. We also found that it acted as a disincentive for associate GDPs to refer to DTs.

Given the need for NHS dental practices to work to the Ionising Radiation Regulations Act, Ionising Radiation (Medical Exposure) Regulations and the Medicines Act, DTs were not allowed to prescribe radiographs or fluoride. These restrictions added to practice inefficiency and patient disruption.

Similar to our earlier research, there was also agreement that the views of the equity owner were highly influential. 30 They were seen as critical in setting the tone in the institutional logics of the NHS practice and overcoming the barriers associated with role substitution at the macro level (contractual and regulatory factors). As highlighted in Chapter 3, a positive attitude towards role substitution saw the equity owners actively seeking workarounds to facilitate different care pathways or internal payment systems for staff (performers). The extent to which their estate was partitioned (e.g. securing additional surgery space) could also influence the promotion of role substitution. This was often difficult in NHS dentistry because most practices were not purpose built for team dentistry.

Endorsement of DTs through outward displays of trust, positive feedback and support were recognised as essential for effective implementation. Although the social acceptability of role substitution was positive, patients did express the need for better communication for them to understand what a DT could do clinically. Most of the patients who were interviewed argued that they would accept new ways of working but that they were concerned about the potential disruption to their care and the need to potentially rebook their appointments.

Because NHS practices were essentially independent contractors, structural investment from the NHS was more difficult. This highlighted the limited number of mechanisms that were available at a meso level for DCs to facilitate the implementation of role substitution.

In WS3, we demonstrated that the data required for an economic evaluation can be reliably collected in terms of costs based on UDAs and/or microcosting of dental care. The pilot study has identified areas where data collection can be refined and simplified for a definitive trial (e.g. simplifying the data collection of use additional health services because these were rarely reported). The economic evaluation was rehearsed and estimates of cost-effectiveness made.

The findings of the qualitative work in preparation for a DCE identified a long list of potential attributes and levels that can form the basis of a future preference elicitation exercise. This could be used to combine the different clinical and process outcomes into a single metric so that the results of the study would be more useful to decision-makers. The attributes and levels identified relate primarily to process and organisational factors about the service (e.g. practice premises).

Qualitative data also highlighted how social acceptability of role substitution can be complex, with strongly held views on either side, and that these views were affected by factors including previous oral health and experience of dental care. This suggests that there may be a role for patient information and/or education ahead of any proposed service changes.

What have we added to the evidence base?

As highlighted by Barnes et al. ,152 ‘there is considerable scope to delegate routine examinations and restorations to DTs’. It is estimated that approaching one-third of restorative interventions could be undertaken by DTs, which would produce a knock-on reduction in the number of dentists required to undertake these routine tasks. 153,154 By conducting a pilot trial and rehearsing a health economic evaluation, we have thoroughly tested the feasibility of a future definitive trial to evaluate role substitution in NHS dentistry in a highly policy-relevant area of research.

The Cochrane Effective Practice and Organisation of Care review96 found a lack of experimental or quasi-experimental evidence for the expanded role of DTs, which concurred with the study by Brocklehurst et al. 155 In this study, we have confirmed recruitment and retention rates and found high levels of clinical fidelity to the intervention in NHS practices. We have also confirmed that a non-inferiority design is the most appropriate, but highlight further challenges to empirically testing role substitution using experimental studies. Of particular note is the challenge of determining the most appropriate primary and secondary outcome measures. As detailed in Chapter 2, even if the eligibility criteria were relaxed to include high-risk patients, there is still a high probability that the strength of any effect may remain undetected, given the large numbers of low-risk patients in NHS practices. To detect clinically meaningful changes, it is also possible that we would need to extend the duration of the trial (at least doubling the duration of the pilot study) and, even then, type II errors may still be possible because any potential signal may remain undetected.

Despite this, the direction of effect in this study (i.e. non-inferiority) is not dissimilar to the non-experimental studies we highlighted in Chapter 1. A number of studies have shown that GDPs and DTs perform to similar standards in terms of their efficacy and clinical effectiveness. The results presented in Chapter 2 concur with Wang,89 Wang and Riordan,88 Kwan et al. ,90 Patel et al. ,91 Kwan and Prendergast92 and Hopcraft. 93 DTs appeared to be able to diagnose and manage low-risk patients within their competency. They were also able to perform routine restorative tasks over the study period, which supports earlier studies by Calache et al. ,81 Freed et al. ,82 Wetherhall et al. ,84 Bader et al. ,94 Battrell et al. 95 and Dyer et al. 96

The realist-informed process evaluation has reinforced findings from our earlier research. 30

Barnes et al. 152 identified the following key mechanisms, which map well onto the results of Chapter 3:

• a payment system that supports appropriate employment of DTs

• appropriate referral systems to promote DTs utilisation

• workplace values and culture*

• good communication within the practice*

• experience of working with DTs*

• team training.*

[List items marked with an asterisk (*) were described in this study as institutional logics.]

Similar to our study, Barnes et al. 152 also contrast the position of the GDC (who allow direct access) with the contractual legislation in the GDS,9 which allows only a dentist to open a course of treatment. The dominance of the contractual theory area in our study is supported by our earlier Cochrane review57 on the effects of financial incentives and also the work of Robinson et al. ,156 who argue that contractual incentives appear to ‘determine dentists’ and patients’ perceptions of need, their behaviours, health outcomes and patient satisfaction’. It is also supported by the work of Harris and Sun,101,102 who explored the efficiency of NHS practices that employ DTs. Barnes et al. 152 contrasts this with other countries, where direct access has been facilitated by favourable regulatory frameworks. 157,158

As highlighted in our study, Barnes et al. 152 also found institutional logics, which are described as intrapersonal and interpersonal factors, to be critical, alongside a lack of knowledge of DTs’ role and their Scope of Practice. Again, this concurs with earlier studies. 158,159 Attitudes towards DTs were also a key factor in influencing the amount and type of work referred to them in practice. 160–162 Barnes et al. 152 also identified the conflict in the internal market between DTs and associate GDPs (because of the competition for UDAs). 152

Our findings also concur with Goodwin et al. 48 and the work of Harris et al. 42 in two large NIHR HSDR projects undertaken in NHS practices. Goodwin et al. 48 found that ‘general dental practitioners are affected by a number of competing interests and incentives and that this can influence their behaviour in a variety of ways’. As Harris et al. 42 argue, there appears to be differences in institutional logic between NHS dental practices and NHS medical practices; they describe the differences between these two groups of NHS practitioners in terms of ‘population health managerialism’ and ‘public health good’. These relate to the extent to which clinicians feel responsible and accountable to their immediate community and the extent to which practices operate to improve the health of their local community. Among general medical practitioners this appears to be prominent and explicitly stated. This contrasts with the prevailing institutional logic in dental practices, which relates to ‘ownership responsibility, professional clinical values and entrepreneurialism’. 42 This may have implications for role substitution in NHS practices, for which the decision to support DTs and/or increase surgery space appears to be driven by a ‘business model’ (linking to contractual incentives) rather than a logic driven by population health managerialism or public health good. This appears to be exacerbated in NHS dentistry by the relative lack of levers that DCs can apply at a meso level to promote role substitution.

As highlighted in Chapter 3, the institutional logic at NHS dental practices also affects our final theory area of logistics. Logistics was also found to be a substantive barrier to role substitution in the NHS by Barnes et al. (i.e. the need for NHS practice owners to invest in additional surgery space to promote role substitution). 152 International evidence also supports these findings. Where NHS dentists are working in larger multisurgery practices, they are more likely to support role substitution because of their capacity to optimise surgery space. 163,164

The social acceptability of role substitution found in this study concurs with earlier work referred to in Chapter 1. 30 It also supports the findings from a survey undertaken in 2018 by Barnes et al. ,165 who found that both patient satisfaction and patient confidence in dentists’ or DTs’ ability was uniformly high (97% and 99%, respectively). This concurs with the studies cited in Chapter 1 in which social acceptability was positive once patients had been educated about the role of the DT. 77,80–86,97 No questions around patient safety were raised in this study, in similarity to other studies cited in Chapter 1. 76–78

The study directly builds on a 5-year Clinician Scientist Award (NIHR/CS/010/004) and a HSDR grant (11/1025/04). 30 As highlighted in Chapter 1, to date, the research team has demonstrated that DTs can identify occlusal caries in vitro, can screen for dental caries and periodontal disease in vivo and are safe as front-line health-care workers. 78,97,98 In addition, the feasibility of undertaking a full trial has been tested as part of NIHR/CS/010/004 over a 15-month period. 99

Strengths and limitations of the research

This study has provided a wealth of information to design a definitive randomised controlled trial. WS1 was conducted in NHS dental practices and recruited a relatively large number of NHS patients. This means that the findings are grounded in the experience of NHS dentists and conducted in a manner that helps research teams understand the challenges of working in a NHS environment. The pilot study has confirmed that the recruitment and retention rates of NHS patients are sufficient to test role substitution in a definitive trial. Fidelity rates were also shown to be high. It also provided estimates of effect for a future definitive study and highlighted the challenges of determining the most appropriate primary and secondary outcome measures. In addition, it revealed the impact that study eligibility might have on the strength of the signal and the impact that including low-risk patients has on a study of this nature.

Four limitations of WS1 related to the exclusion of high-risk patients, the number of practices included in the study, the duration of the pilot and the number of patients lost to follow-up. The eligibility criteria of the study were amended to low-risk patients on the advice of the funding board. Budgetary constraints also required the research team to limit the number of practices and duration of the pilot, but we felt that the design provided enough data to test recruitment, retention and fidelity rates across the practices. The data also showed relatively limited variance, suggestive of a broadly representative sample of practices and patients. In addition, although we tried to be as flexible as possible with regards to the availability of epidemiologist, the fact that we could not schedule appointments with some participants owing to their prior commitments contributed to a number of participants being lost to follow-up. The information we have about the 41 participants who were lost to follow-up suggests that participants were lost primarily because of logistical difficulties. It is important to acknowledge the limitations of this assumption (i.e. that these data are missing at random and these participants did not represent a group with different clinical outcomes to those included in the ITT analysis).

Workstream 2 has provided a novel approach to conducting the process evaluation by using the principles of the realist approach to evaluation of complex social programmes to build on the qualitative component of the study. Realist approaches can help build an explanatory account of what works in any given situation, and places emphasis on stakeholders’ experiences and perspectives to identify and refine the underlying (unseen) mechanism of action. In this study, using the principles of realist methodology were important to articulate and explain the factors that underpin intervention success or failure.

As expected, there are limitations to using realist approaches in this way. Owing to the time frame, pragmatic decisions were made not to conduct a full systematic realist synthesis of evidence. In this instance, the team chose to rely on data derived from experts. As a result, it could be argued that other information is lost. Purposive and convenience sampling could have missed other opportunities to gain insight into theories driven by a wider catchment of different stakeholders. Finally, the theories articulated are at a particular level of abstraction. However, further additional cycles of testing and refining could have developed the theories towards more practice-level specification.

Workstream 3 has demonstrated the feasibility of collecting data required for an economic evaluation from multiple perspectives (health service and patient). We have successfully rehearsed the economic analysis, which produced consistent results across the range of analyses conducted in terms of cost-effectiveness. In addition, the qualitative component of this work has provided rich data regarding the acceptability of the use of DTs in the management of low-risk patients and provided us with a number of attributes and levels to take forward in a future preference elicitation exercise. We were able to recruit both low-risk and non-low-risk participants for qualitative work, which has provided us with a wide spectrum of views on DTs and role substitution.

As discussed in Chapter 4, there are some limitations to the health economics work. With regard to the data collection processes, the collection of service use data from both clinicians and patients meant that there was potential overlap or conflict between the two data sources. The importance of accurately capturing data, such as secondary care use, and patient costs, such as private treatment, should be taken into account in a definitive trial because these costs have the potential to be high and skew the data, and alternative methods to capturing these data should be investigated; for example, capturing these data during clinic visits or through follow-up telephone calls would be a quick way of collecting data, with the potential to reduce the level of missing data. The 6-month PCQ questionnaire data were missing by design and this meant that additional assumptions were made regarding participants having consistent costs over both of the 6-month periods covered by the questionnaires. In addition, our assumption that patients with missing PCQ data had no additional costs had the potential to underestimate cost categories covered in this questionnaire in both arms. Although our microcosting work was based on a previous study, the accuracy of our unit costs could be potentially improved by further exploring materials and methods used by participating practices in a future study.

In the qualitative work, although the inclusion of non-low-risk patients provided us with a range of perspectives, it could also be argued that prominent views were included by those who would be unlikely to be affected by changes in service provision.

Possible considerations for future policy

The results of this study are important for policy, given the predicted trends in future population health needs. 16,29 There remain challenges in improving the oral health of young children, and providing adequate care for older people is an emerging public health challenge. As a result, there will be a pressing need to better target the available resources in NHS dentistry, which appear to be increasingly spent on low-risk NHS patients. 17–19

As highlighted in this report, the dental contract reform programmes undertaken across the UK place an emphasis on increasing patient access and prevention while maintaining and improving the quality of service being delivered. 59–68 Increasing the use of role substitution is an important consideration for policy-makers and explicitly stated as a strategic objective in Wales. 14,66

Although our pilot study confirmed the potential of using role substitution in NHS dentistry, our realist-informed process evaluation highlighted the barriers to its use. Contractual and regulatory factors were dominant and concurred with our earlier work30 and the international literature. 144 Given that the supply of NHS dentistry is provided by NHS practices operating to a ‘business model’, there is a need to align the financial incentives in NHS GDS contracts with the policy objective of greater use of role substitution. This is concomitant with recommendations from the WHO, which argues that there is an opportunity to think again about how we educate our future health workers. 166 Although amending the regulations to the use of medicines (local analgesia and fluoride) and radiographs may be more problematic, changing the GDS contract is within the scope of many dental policy-makers.

Implications for practice

Because this study was a pilot study, care needs to be exercised in the interpretation of the findings. As highlighted in Chapter 5, the direction of effect concurs with earlier non-experimental research and the same barriers and enablers were identified to the use of role substitution in NHS practices. Institutional logics at a micro level were found to be important but were largely secondary and dependent on the contractual and regulatory factors at a macro level. It appears that the social acceptability of role substitution is high if the DT role and its benefits are clearly articulated to NHS patients. Therefore, the implications for practice would appear secondary to the dominant factors that could be influenced at a policy-level. Because NHS practices largely operate as small businesses, financial incentives that promote role substitution appear key in shaping the response at the level of the practice.

From a patient perspective, the qualitative data collected as part of this study have highlighted that the acceptability of role substitution may vary greatly between patients, influenced by a number of factors including perceived risk, history of oral health, prior dental treatment and dental anxiety. A further understanding of these factors is of vital importance to understand how any changes to services may affect patients.

Recommendations for further research

As highlighted in Chapter 2, the interpretation of the findings of the pilot study poses some interesting design questions. The purpose of the pilot was to determine the most appropriate design for a definitive trial and determine the appropriateness of the primary and secondary outcome measures and the non-inferiority margin. It also tested recruitment, retention and fidelity rates.

It would appear that a future definitive trial, should one be warranted, should be based on non-inferiority. This would align with the findings in this study and the international literature but would require a large number of participants depending on the size of the non-inferiority margin. The relatively low recruitment rates also have practical implications for a future trial and may prolong the recruitment period to reach the desired number of participants in each arm. This could be compounded further by the length of time needed in the intervention period to detect enough of a signal (on the chosen POM) to make comparison between the two arms meaningful.

The nature of the POM is also problematic. It would appear that the inclusion of low-risk NHS patients in a future study would pose a number of design challenges in relation to the POM, yet this is the group of patients to whom role substitution would be the most relevant from a policy perspective. In this study, we used BOP as the POM. Earlier studies suggest that BOP is a good indicator of periodontal health and tissue stability and is well known to NHS practitioners. 112–114 However, it would appear that there are floor effects with this measure in low-risk patients. We anticipate the same kind of effects in clinical measures, such as dental caries, owing to low prevalence and longer times to expression. 16 As highlighted in Chapter 2, there are systems that measure incremental changes in dental caries status, but marginal changes in low-risk patients may create issues with measurement error. 121,122 As we argue, it may be possible to consider analysing a ‘complete’ sample (including high- and low-risk patients) using an appropriate zero-inflated model, but this may not facilitate a full interpretation of effect or non-inferiority. As a result, it would appear that the types of clinical outcomes that would be used for a definitive trial on role substitution in dentistry may not be appropriate to test for non-inferiority. Patient-related outcome measures and patient-related experience measures may be two options, but their signal-to-noise ratio may be problematic. They are also further down the causal pathway to assess clinical effectiveness and/or non-inferiority.

These issues highlight the challenge of designing a definitive randomised controlled trial to test role substitution in NHS dentistry. It would appear that measuring changes to health and/or care pose substantial design problems. This is further exacerbated by the length of time that may be needed to capture the effect of the intervention and the knock-on effect on the cost of the trial. This raises questions about the role of definitive studies in answering these types of important policy questions when they are undertaken on largely healthy participants.

Alternative methods to experimental designs may be advocated here. There is a developing argument about the need to embrace a broader range of research methods for policy-related and population health issues, rather than restricting the approach to randomised controlled trials. As argued by Brocklehurst et al. :167

We need to consider a broader range of conceptual and methodological approaches to increase the value of information generated. This should be undertaken either in parallel with empirical and experimental designs, or in some cases, instead of. This is important if we are going to understand the complexity and contextual knowledge of the ‘system’, within which interventions are implemented.

Reproduced with permission from Brocklehurst et al. 167 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text

What is evident from Chapter 3 is that this ‘system’ is complex, with multiple interdependent elements that operate at different macro, meso and micro levels. As Braithwaite et al. 168 argue, accounting for this complexity is key; there are ‘no simple solutions, no single ingredient’. 169

There are alternatives to experimental designs and these may be warranted here to determine whether or not role substitution is both clinically effective and cost-effective. As highlighted by Brocklehurst et al. 170 and Listl et al. ,171 ‘ex-post techniques typically evaluate the impact of an already implemented health policy programme and include a range of quasi-experimental methods including Instrumental Variables, Difference-in-Difference, panel data analyses using fixed or random effects, and Regression-Discontinuity-Designs’170 and ‘ex-ante methods include structural modelling, agent-based modelling and micro-simulation’171 to help model the short-, mid- and long-term health effects of an intervention. For example, difference-in-difference designs have been used in other dental public health studies before. 57,172 In Northern Ireland, difference-in-difference design was used to assess the implementation of a NHS pilot based on capitation. 57 This enabled the research team to measure changes in activity levels during three time periods, and the coefficients were triangulated by an interrupted time series approach undertaken in parallel. Such methodologies can ‘provide very helpful information on the evaluation of policies and interventions that would otherwise not be rigorously evaluated as the standard [randomised controlled trial-]related methodologies are neither feasible nor suitable’. 170 They might also prove eminently more affordable when compared with a lengthy definitive trial based on non-inferiority. They would also allow multiple outcome measures to be followed over time, rather than selecting one single POM.

In addition to the empirical aspect of any future study, it would be important to undertake a parallel process evaluation to further explore the factors that influence the pathway to impact for role substitution. In this study, WS2 provided helpful insight to the barriers to and enablers of role substitution in NHS dentistry. This element of the study captured how complex interventions and systems can interact and identified the underlying generative mechanisms about how role substitution works. 126 This gave us a good understanding of ‘how context (individual, social, cultural, organisational) interacts with intervention components and underpinning mechanisms to bring about desired outcomes’. 127 This allowed the research team to develop a set of focused MRTs that enabled us to determine the range of factors that could potentially influence adoption. 130 As a result, we remained focused on the barriers and enablers that could make a difference at both a policy and a practice level as opposed to constructing ‘all-encompassing meta-narratives that span space and time’. 131 As a result, we commend the use of these types of approaches alongside future experimental and quasi-experimental study designs. Future teams could make use of the different CMO configurations that were generated in this study as a starting point for undertaking a further realist cycles.

Following this pilot study, there is an important role for economic evaluation work in any future study to establish cost-effectiveness of role substitution in the management of low-risk dental patients (WS3). As discussed in Chapter 4, recommendations from this WS include potentially revising data collection methods with regard to patient-reported data, possibly integrating these into the collection of clinical data at dental appointments, given that the routine nature of appointments has the potential to achieve higher levels of accuracy and completion. Telephone interviews have been used successfully in past studies to collect details of health service use and this may also offer a method of supplementing data collection. In addition, completion of clinical data has the potential to be simplified, therefore reducing burden on participating health-care professionals and standardising the data collected. In a future study there is potential to improve the accuracy of resource-based microcosting through further investigation into materials and methods. These methods could be incorporated into either an experimental or a quasi-experimental design.

The preparation work undertaken in this pilot demonstrated the importance of including a DCE to help further understand the relative importance of different attributes. These could provide valuable data for future service design. In addition, given the need to understand the impact of service changes on people using these services, there is an important role for future qualitative work to explore some of the questions to arise out of the qualitative component of the health economics work, such as patient information needs or exploring the experience of receiving care from different health-care professionals.

Contributions of authors

Paul Brocklehurst (https://orcid.org/0000-0003-1878-9030) wrote the application to secure the research funds from the NIHR, led the project, provided dental public health input, undertook the analysis and oversaw the writing of the report.

Zoe Hoare (https://orcid.org/0000-0003-1803-5482) wrote the application to secure the research funds from the NIHR, provided statistical input throughout the study and made a substantive contribution to the design of the pilot study and the writing-up of Chapter 2.

Chris Woods (https://orcid.org/0000-0003-4546-3028) managed the pilot trial, conducted the interviews for the realist-informed process evaluation, undertook the analysis and co-wrote Chapter 3.

Lynne Williams (https://orcid.org/0000-0003-2575-9710) wrote the application to secure the research funds from the NIHR, undertook the analysis and co-wrote Chapter 3.

Andrew Brand (https://orcid.org/0000-0002-6942-0404) provided statistical input throughout the study and made a substantive contribution to the design of the pilot study and the writing-up of Chapter 2.

Jing Shen (https://orcid.org/0000-0001-8450-5003) wrote the application to secure the research funds from the NIHR, planned and conducted the DCE development work, planned and conducted the pilot health economic evaluation and co-wrote Chapter 4.

Matthew Breckons (https://orcid.org/0000-0003-3057-6767) planned and conducted the DCE development work, planned and conducted the pilot health economic evaluation and co-wrote Chapter 4.

James Ashley (https://orcid.org/0000-0003-0058-4889) wrote the application to secure the research funds from the NIHR and provided input to the study design from the perspective of a NHS GDP.

Alison Jenkins (https://orcid.org/0000-0003-1963-4903) managed the pilot trial.

Lesley Gough wrote the application to secure the research funds from NIHR and provided input to the study design from the perspective of a consultant in dental public health.

Philip Preshaw (https://orcid.org/0000-0002-7153-4865) wrote the application to secure the research funds from NIHR, provided valuable input at the design stage and proofread the iterations of the final report.

Christopher Burton (https://orcid.org/0000-0003-1159-1494) wrote the application to secure the research funds from NIHR and provided input in the writing of Chapter 3.

Karen Shepherd acted as our user researcher in WS2, co-wrote the Plain English summary and contributed to our dissemination strategy.

Nawaraj Bhattarai (https://orcid.org/0000-0002-1894-2499) planned and conducted the pilot health economic evaluation and co-wrote Chapter 4.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to available anonymised data may be granted following review.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

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 and Social Care. 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 and Social Care.

Case report form (baseline)

(PDF download)

Case report form (outcome assessment)

(PDF download)

PARC-Bangor

PARC-Bangor is the public involvement and engagement group at the NWORTH Clinical Trials Unit at Bangor University (URL: https://nworth-ctu.bangor.ac.uk/parc-bangor.php.en; accessed 9 December 2020). Members of the public across North Wales can sign up to the PARC-Bangor database if they are interested in taking part in research or taking part in developing research ideas at NWORTH Clinical Trials Unit.

Members of the PARC-Bangor network meet four times per year to discuss new and ongoing studies. Some members of the PARC-Bangor network sit on Trial Steering Committees and others engage with projects as and when their input is required. The functions of PARC-Bangor are to:

• provide insight and a public perspective to the development of research questions

• ensure that documents and information associated with a study are easy to understand for patients and the public taking part in the research

• discuss and comment on the research process described in funding applications to ensure patient perspectives are accounted for

• provide a ‘patient voice’ to NWORTH’s executive.

Being a member of the PARC-Bangor network means that members of the public can:

• learn about and influence health research in their area

• use their experience as patients, service users, carers or family members to help others

• influence study designs to make funding applications more successful

• discuss the work with research teams via e-mail, in person, or over the telephone

• access free training about health and social care research.