The role of digital communication in patient-clinician communication for NHS providers of specialist clinical services for young people [the Long-term conditions Young people Networked Communication (LYNC) study]: a mixed-methods study

Aims

1. To evaluate the impacts and outcomes of DCCs for young people living with LTCs. Many young people disengage from health services between the ages of 16 and 24 years, resulting in poor health, but they are prolific users of digital communications. As such, young people using the NHS present the best current opportunity for addressing our second research aim.

2. To provide a critical analysis of the use of DCCs by NHS specialist care providers. DCCs are being widely embraced by clinicians working with young people, but the NHS is currently underserved by research evidence to prepare it for this development.

Objectives

1. To engage young people in the implementation of the research.

2. To observe and explore with young people with LTCs – and where appropriate a parent/carer, clinicians and managers – the use of DCC in the NHS for a variety of clinical conditions, how it is used and with what impact and issues related to ethics and patient safety.

3. To investigate the impact of DCCs on health outcomes for young people with LTCs and on their engagement with, and use of, health services.

4. To describe the cost of implementation and ongoing provision of DCC and how it varies across different clinical conditions, and to understand the value of this service to patients and clinicians.

5. To identify and explore the use of patient-reported outcome measures (PROMs) for future cost-effectiveness studies, which can be used across disease areas to capture the impact of DCC.

6. To evaluate and synthesise published evidence on the use of DCC by health professionals with young people with LTCs.

7. To develop and disseminate guidance for NHS providers and commissioners on the use of DCC, to provide insights for policy-makers from current NHS use of DCC and to consider the need for future research.

We chose not to specify a specific technology to study in order to future-proof the study and to avoid limiting the utility and validity of the results. The digital communication ecosystem is rapidly changing. 20 For example, the introduction of smartphones is changing the way people access information and use communication channels. The use of communication channels also evolves. For example, Twitter was initially used primarily for broadcasting, but then became used for communicating between individuals or within small groups. 58 We studied the use of DCC via whatever medium was being used by the patients and clinicians we studied, in order to draw out results that are transferable across technologies. This allowed us to learn from a small number of established systems for DCC for young people in the NHS, and from the informal, unregulated development of this means of communication between clinician and patient.

The research focused on young people (aged 16–24 years) who have LTCs such as diabetes mellitus, cancer, cystic fibrosis, sickle cell disease and other conditions that require engagement with specialist (secondary or tertiary) clinical services. We studied NHS providers of specialist clinical services. We were concerned with communication between patients and specialist (secondary or tertiary care) clinicians/clinic teams who have already been in contact with each other in the clinical setting, where communication was, or there was potential for it to be, in both directions – patient to clinician and clinician to patient. We did not include specifically the delivery of therapeutic interventions via digital communication media59 [such as cognitive–behavioural therapy (CBT)],60 nor digital communication that solely involves the delivery of information on disease prevention and health promotion. 61 However, where the use of a digitally delivered intervention or the delivery of disease prevention and health promotion information formed part of ongoing patient–clinical team communication, they were included.

Chapter 3: what works for whom, where, when and why?

• What works for whom, where, when and why?

Chapter 4: ethical implications of using digital clinical communication to support young adults with long-term conditions

• What concerns do patients and clinicians have about confidentiality in relation to DCC?

• How does this form of communication affect the patient–clinician relationship and the clinician’s duty of care?

• What regulatory framework is needed to reassure patients and clinicians regarding its use?

Chapter 5: patient safety

• What are the significant risks to patient safety associated with the use of DCC in the context of supporting young people with chronic disease?

Chapter 6: health economic analysis

• What value do patients place on DCC?

• What are the direct resource use implications for the NHS of implementing it?

• How does the direct resource use vary when used with different patient groups?

• What are the resource implications for scaling up in the NHS?

Chapter 7: information governance

This chapter analyses the data from the interviews with IG officers with reference to the research questions for Chapters 4–6.

Chapter 8: impacts on health-related outcomes

• How is the impact on the health status of patients currently evaluated?

• Using existing clinical data, what is the impact on health status of patients?

Chapter 9: generic patient-reported outcome measures

• What generic outcome measures are available to assess the impact of DCC?

• In the future, how can its effectiveness be measured across health conditions?

Chapters 10–12: scoping reviews

• What is the evidence in the literature to support, refute or add value to the case study findings?

Chapter 13: discussion

• What are the risks to patients and to NHS specialist care providers from its use?

• What policy and procedural changes are needed for gaining benefit and limiting harm?

• In which clinical areas is benefit most likely and how is benefit most likely to be achieved?

• What future evaluation is needed and how should it be undertaken?

In response to requests from NHS England and NHS Digital (formerly the Health and Social Care Information Centre) for assistance with the development of their policies and procedures, we undertook the analyses reported in Report Supplementary Material 1. The consensus conference, which also contributes to the above research questions, is also reported in Report Supplementary Material 1.

We tackled the following question through our internet search described in Chapter 3 in order to form our sampling frame.

• How, and for what purpose, is this form of communication taking place (or not) in the UK?

We tackled the following question through maintaining awareness of developments in the ethical, legal, policy and governance landscapes relevant to DCC in the UK to ensure that our research took account of this.

• What is the ethical, legal, policy and governance framework for DCC?

In Appendix 1, we provide a report on all our patient and public involvement (PPI) activity and the impact it had on the research, except for PPI activity linked to the identification of a generic outcome measure which is reported in Chapter 9. Our appendices also include copies of the quick reference guides produced in response to discussions with NHS Digital, from our data analysis. These are A4 leaflets available to freely download from www2.warwick.ac.uk/fac/med/research/hscience/sssh/research/lyncs/ (accessed 5 March 2017). The NHS Digital website carries a link to these guides.

Case site inclusion criteria

Inclusion criteria for clinical sites were as follows.

• The clinic provided specialist care for young people with LTCs.

• The LTC was currently expensive to the NHS.

• The clinical team expressed interest in the use of two-way digital communications to a minimum, moderate or large extent.

We sampled clinical teams for diversity of long-term health conditions experienced by young people.

Case site recruitment

As a relatively large number of case sites were required, multiple recruitment strategies were used. An extensive internet search of grey literature was conducted to identify technology being used for clinical communications in the UK. An internet search using the search engine Google (Google Inc., Mountain View, CA, USA) was undertaken between December 2013 and February 2014. The keywords used were e-health, telehealth, telemedicine, digital communication, young people and young persons. Combined, the searches returned 24,626,063 hits. The first 35 pages of each search were scrutinised for relevance. In the case of potentially relevant projects, detailed information was sourced from individual trust websites, documents and reports. Further details were obtained by contacting key individuals and project leads. This search identified 71 projects involving DCC, running at 38 different NHS trusts, health-care organisations and universities. We tried to contact 14 sites that seemed eligible for participation (the search was not limited to young people with LTCs). Additionally, one clinical team was identified via the internet and approached.

Snowball recruitment was also used: the research co-applicants sent out recruitment e-mails to their colleagues and asked that they pass the information on to anyone else who may be interested; in addition, specific clinical teams were suggested by the research co-applicants, based on the known use of communication technology in their usual practice. The research team contacted 51 different clinical teams using this approach. Research team members attending academic and applied health conferences advertised the study with poster presentations and recruited potential case sites through networking sessions. Seven contacts were made this way. Finally, the research team was approached by seven NHS trusts that had seen the study listed in the National Institute for Health Research’s (NIHR’s) portfolio. A further 24 contacts were made after people already in discussion with the research team snowballed the study information to others they thought would be interested. Overall, we made initial contact with a total of 104 different clinical teams.

Interview participant recruitment

We aimed to conduct brief interviews with members of the clinical team, up to 15 patients aged 16–24 years and an IG officer at each site. Where the clinical team was < 15 patients, we aimed to interview all members of staff. Parents/carers/household members of young people with a LTC were also invited to participate in an interview.

Two weeks prior to fieldwork commencing, a research nurse from the local clinical research network or NHS trust identified eligible patients attending the clinic on days when a Long-term conditions Young people Networked Communication (LYNC) study researcher was present. These patients were sent (via post, or e-mail if appropriate) a study information sheet and given an option of contacting the researcher in advance to express interest in participating. Patients were also recruited by the researcher during their time in clinic, usually while waiting for their appointment. Patients were asked to bring examples of recent DCCs to inform interview discussion. The research nurse also identified patients who frequently DNA clinic appointments from clinic records and sent a separate recruitment letter to these patients. Young people were given a choice of medium for their interviews: in the clinic, over the telephone or via other technology chosen by the participant (including e-mail and Facebook). The researcher asked for consent either during recruitment or prior to commencement of the interview. For telephone interviews, consent was given verbally and a dated note was made by the researcher. All participants were offered a £20 high-street voucher as a thank-you token. At the end of the interview, researchers asked the patient if they were willing for their parent/carer/household member to be interviewed for the study. If they agreed, they provided contact details for the research nurse to send out an invitation letter. These interviews were conducted over the telephone, at a time convenient for the participant.

Clinic staff were recruited on an ad hoc basis during fieldwork. Prior to the start of fieldwork the researchers attended a team meeting to introduce themselves and the research, and to make staff aware of their presence on site. We aimed to interview any clinic staff who used (or would potentially use) DCC, such as nurses, doctors, professions allied to medicine and administrative staff. We specifically sampled for diversity of job role and experience of, and opinion about, DCC within each case study. Within each clinic we aimed to interview as many team members as were willing to participate, up to 15 members of staff. In small clinics we interviewed all relevant team members. In large clinics we aimed to get representation across all specialties and grades of staff. Interviews took place over the telephone, or in a location chosen by the participant (clinic space, offices, cafes, etc.). Participants were reassured that the information they provided would be confidential and that it would not be possible to identify specific clinics or staff members from our research report. It was also explained that if a concern arose about unethical or unsafe clinical practice, the researcher would need to report this to a principal investigator (PI) who would decide if it was necessary to initiate action through normal professional channels. However, no such practice was reported. During data collection we were asking clinic staff to reveal activities that may breach IG policies. This was a difficult issue, as these clinicians were likely to be rich sources of data. We therefore drew up a protocol of action for serious breaches of confidentiality (see Appendix 2). However, we did not expect to take action for activity that we found was common practice.

At each case site we aimed to interview a trust-wide IG officer. We aimed to conduct this interview prior to fieldwork commencing, so that the researcher would not inadvertently disclose clinic practices that may be out of step with trust IG policies. Additionally, this enabled an understanding of the local context in which the clinical staff operated. Interviews with IG officers took place on site in trust offices or over the telephone.

Non-participant observation

Non-participant observation of the clinic and shadowing of clinic staff took place at each clinical site. Using an observation pro forma, the researchers noted how the clinic functioned, how DCC was used and monitored within the team and what evidence there was for more widespread use of digital communication within the clinic space (e.g. clinic staff using computers, patients using smartphones in waiting rooms). When shadowing members of the clinical team, the researchers reported if, and how, digital communication was used for clinical purposes, describing the context and content of the dialogue. Shadowing lasted for up to 2 hours, whereas clinic observation lasted for the duration of the clinic (up to 8 hours) and were designed to note the use of digital communications between appointments. When possible, observations preceded interviews. The researchers made patients and staff aware of their presence in the clinic by wearing a T-shirt identifying them as LYNC study researchers and putting up posters advertising the study. Observation schedules are available in Report Supplementary Material 1.

Semistructured interviews

All interview schedules focused on how digital communication was used for clinical purposes and the advantages and concerns participants had about this use, specifically identifying risk and ethical concerns. The interview schedules were a guide and we covered the part of the schedule relevant to each interviewee. Based on initial analysis of interviews we adjusted interview schedules to ensure that we obtained the richest possible data (e.g. altering question wording). During individual interviews we adjusted the words we used to refer to digital communication according to the words used by the interviewee in response to early questions.

Additional data on ethics, safety and the generic PROMs were collected from specific sites. Data from a cancer and liver conditions site were initially analysed with a focus on ethical concerns. Based on this analysis, the interview schedule was amended to focus on emerging ethical issues, such as equity in access, consent, privacy and confidentiality, risks/benefits to staff, patient–clinician relationship, patient autonomy and duty of care. This interview schedule was used to collect data from a further three sites (renal, cystic fibrosis and mental health). These sites were chosen based on their use of DCC and diversity in health condition. Additional data on safety were collected from two clinics (diabetes mellitus and cancer) which used DCC. The diabetes mellitus clinic was chosen as the self-management of diabetes mellitus involves the young person making decisions several times a day about their dose of insulin and errors in these decisions can quickly lead to adverse events. The cancer clinic was chosen as these young people are likely to be on medication with potentially serious side effects and are coping with psychological distress as a result of their life-threatening diagnosis. At these two sites, the interview schedule was amended to focus on important safety issues, such as risks, the importance of risks to patients, how risks are accepted or reduced, engagement with treatment regime, adverse events, unintended outcomes and benefits. In the later sites, the interview schedule was revised to focus on undertaking cognitive interviews to assess the two generic PROMs identified in the literature review and from PPI activity. This interview schedule was used with seven patients and seven clinicians.

A specific interview schedule was developed for use with IG officers, with input from our IG co-applicants. This focused on if, and how, technology was used within the trust for communication about clinical issues, what policies instruct the use of technology, recorded DCC-related incidents and their professional opinions on the use of digital technologies in their respective NHS trusts. When DCC was not used, the key professionals were asked whether or not there were any future ambitions to introduce these types of communication, and if there were any barriers to, and enablers of, this introduction. Interview schedules are available in Report Supplementary Material 1.

Collection of impact data

Data are broadly categorised in two groups: (1) the cross-clinic and condition outcome data and (2) clinic-specific outcomes. Cross-clinic outcomes are appointment status (attended, cancelled or not attended), A&E attendance and hospital admission. Such outcomes, from most sites, were obtained from the information technology (IT) departments of trusts, with contact initiated by the site contact (PI, consultant, etc.) or the service manager. For tertiary care case study sites, A&E attendance and hospital admission data for their clinic patients were not available, as the patients would attend their local A&E department or hospital for emergencies.

Clinic-specific outcomes are outcomes specific to the various health conditions and sites studied. These were identified in discussion with the clinical team as being key indicators of patients’ health. These outcomes were HbA_1c and glucose levels for diabetes mellitus sites; body mass index (BMI); forced expiratory volume in 1 second (FEV₁) as a percentage of predicted (FEV₁% predicted); bed-days to receive intravenous (i.v.) therapy for cystic fibrosis sites; time to discharge for a dermatology site, where patients are discharged from the service when the skin condition is successfully treated; tacrolimus level for a liver cancer site; viral load for a human immunodeficiency virus (HIV) site; urine protein and serum creatinine levels for a renal site; and self-harm incidents and self-harm ward attendances for mental health sites. Tacrolimus level data from the liver cancer site, and BMI and FEV₁% predicted data from one of the cystic fibrosis sites, were extracted by the clinic staff, whereas all the other clinic-specific outcomes were provided by IT specialists from trusts. These aggregated data were collected for all patients in the age range of our study on these cross-clinic and clinic-specific outcomes. Data collection periods were for the year prior to the escalation of DCC in each service and every year following to the point of fieldwork commencement.

Collection of economic data about the digital communication system

A structured questionnaire was developed to collect data on resource use associated with DCC (see Report Supplementary Material 1), during staff interviews. The questionnaire was based on initial case study observation and semistructured interviews with staff. All staff, at each case study site, who participated in an interview, and whose role involved any degree of contact with patients, were asked to complete the questionnaire. Where possible, the questionnaire was given during the interview, otherwise it was sent by e-mail after the interview had taken place. The questionnaire asked staff members to report the time they spent each day on DCC-related activities, broken down by the channel of communication used (e-mail, mobile phone calls, text messaging or other, such as VoIP). Staff were not asked to report a precise figure; instead they were asked to report usage for each channel by selecting from boxes corresponding to time intervals (e.g. 0–15 minutes, 15–30 minutes). The questionnaire also asked staff to report what technology they used for DCC (laptop, desktop personal computer, mobile phone or tablet) and who provided this equipment (i.e. whether it was provided by the clinic or themselves). Data were obtained from a site administrator, or suitable equivalent, at each site on the grade of each staff member and the number of days worked per week.

A further aim of the LYNC study was to explore, qualitatively and quantitatively, the value DCC provided service users, and how this value might be elicited. A number of ‘value’ questions were explored in the interview schedule, whereby patients at each site were asked how much they would hypothetically be willing to pay for a service which used DCC compared with one which did not. After interviews at the first two case sites, the question focusing on value to patients was changed from a willingness-to-accept format (as cited in the protocol) to a willingness-to-pay format, as patients struggled to understand the original question. An open-ended question was used, but where patients struggled to respond, we tried using prompts suggesting values ranging from £5 to £50 in order to understand what values respondents might consider plausible. In sites where technology was not used for clinical communication, clinic staff and patients were asked about how they currently communicated, how happy they were with the current situation and what DCC they would find useful in the future.

Documents collated

Current trust policies and procedures were collated with help from lead clinicians and IG managers. These documents were used to inform the IG interviews. IG managers were asked to tell us about any adverse events relating to the use of DCC.

What works?

Ethics

Safety

Health economics

Impact

Description of case sites

Of the 104 clinics contacted, 47 were eligible for the research and interested in participating (Figure 1). Sampling of eligible sites aimed to achieve diversity in health condition, DCC used, the frequency of the clinic, the size of the patient group cared for and geographical location, so that each case site was unique. We initiated site set-up at 25 sites. Of these, two withdrew and three were not progressed beyond initial set-up, giving a total of 20 case sites studied.

FIGURE 1.

Flow chart showing case site recruitment.

The case sites studied focused on 14 different health specialties: cancer, cystic fibrosis, dermatology, diabetes mellitus, HIV, inflammatory bowel disease (IBD), liver conditions, mental health [both Child and Adolescent Mental Health Services (CAMHS) and Early Intervention in Psychosis Teams (EIPTs)], renal conditions, rheumatology, sexual health, sickle cell disease (a condition experienced mostly by those from ethnic minorities) and a school nurse service which covered a range of different health conditions. Clinic populations varied from children and adolescent services, to transition populations, young adult services and adult services. Clinical teams were based in a range of locations around England and Wales: nine were in the south and east of England, seven were in the Midlands, three were in the north of England and one was in Wales. The types of geographical location served by the clinics included rural areas, major industrial cities, market towns, industrial towns and London. A range of DCCs was used in these clinics; some were heavy users, whereas others did not communicate digitally with patients in any way (Table 2 provides case study site descriptors).

Description of interview participants

A total of 367 participants took part in an interview. Of these, 165 were patients, 173 were clinic staff, 13 were parents/carers and 16 were IG officers (Table 3).

Patient recruitment ranged from 0 to 16 people per clinic. The largest number of patients was recruited from cancer and diabetes mellitus clinics. It was not possible to recruit any patients from the school nurse service and no patients who frequently DNA appointments volunteered to participate in the study. Only 16 patients chose to share examples of their digital communication with a clinician with us. Interviews with patients were primarily conducted over the telephone (n = 82), but also in person (n = 41), by e-mail (n = 35), via Facebook (n = 4), via Skype (n = 2) and by text messaging (n = 1). Synchronous interviews (taking place in person, or via telephone or Skype) took between 20 and 60 minutes, with most taking approximately 30 minutes. Asynchronous interviews took longer to complete; the longest, via e-mail, took 2 weeks. Overall, few parents/carers were recruited; none was recruited from 11 individual sites. Parent interviews took place primarily via telephone (n = 7), but also by e-mail (n = 4) and in person (n = 2).

Between 3 and 12 members of clinic staff were recruited from each site. The largest number of clinic staff were recruited from CAMHS teams and the fewest from dermatology. This reflects the size of the clinical teams being studied and the number of sites focusing on each health condition. Clinic staff had a range of clinical and clerical roles, including consultants, registrars, community nurses, advanced nurse practitioners, psychiatrists, psychologists, dietitians, physiotherapists, occupational therapists, pharmacists, secretaries and administrators. Interviews with staff took place in person (n = 148) or by telephone (n = 25). Interviews lasted up to 2 hours, but the majority took approximately 45 minutes. A total of 79 staff members were also shadowed across the clinical teams. Shadowing usually lasted between 1 and 2 hours, but lasted longer (the maximum was 7 hours) if the clinician felt that it was appropriate.

Information governance specialists and Caldicott Guardians were also interviewed. Two were interviewed in each of five trusts and one was interviewed at each of six trusts. It was not possible to interview an IG officer or Caldicott Guardian at nine sites. The majority of interviews were conducted in person (n = 10), with the remainder conducted by telephone. All interviews were audio-taped and lasted between 20 and 45 minutes.

In the results chapters that follow we include illustrative quotations from our qualitative data. These are each labelled with the condition the clinic manages, participant type and participant number. We present our analysis of qualitative data by thematic focus: ‘what works’, ethics and patient safety.

Patient empowerment

The majority of clinicians agreed that digital communication increased patient autonomy by giving them more control in both the management of their condition and the way in which communication between them and the clinician was conducted. Questions can be formulated and feelings expressed, which the patient might otherwise find difficult to do. Young people also valued this opportunity:

Well I guess that’s saying, able to control their own illness but they feel in control if they’ve got a question, they’re not disenfranchised. They’ve got someone to ask about it so they can feel empowered to ask those questions and control their own health rather than don’t know who to ask or whether to come forward with things.

Cystic fibrosis 1 site, secretary, 04

. . . well I’ll just say that personally I think that method of communication is one that’s very important to people. It gives people another chance to explain their issues, you know, and it can even be good for things that maybe you’re not comfortable talking about in front of the person, you know. And e-mailing to, you know, to the person who is dealing with your care I think is an invaluable resource because you know, you can ask anything, you don’t have to wait, you can get an answer.

Liver conditions site, young person, 10

However, clinicians were also concerned that, in some cases, the availability of digital communication may discourage young people from taking responsibility for their own health. Easier access to advice about what to do in any situation could remove the need to think about the problem and develop their own management strategy. Some clinicians expressed a concern that this general dependence on easy access to a clinician could develop into a more personal dependence on a specific clinician:

I think . . . personally I do think it is too . . . I think it makes patients a little too dependent on you and you only, which we don’t like that to happen.

Diabetes mellitus 1 site, specialist nurse, 03

You know, they’re not going to have me forever to text, so they need to learn to manage these situations themselves.

Mental health 3 (outreach team) site, support worker, 03

Some clinicians observed that because of the ease of access, DCC makes it easier for young people to avoid engaging with health services:

Sometimes it makes you lazy and you text because you don’t want to have a conversation, well the same thing works for the young person because they actually remotely tell you, no I can’t make it, and then not answering if you call so you can’t have that conversation. So I guess it makes it easier for them to not turn up or to cancel stuff or to act avoidant if you like.

Mental health 3 (outreach team) site, mental health practitioner, 04

Person-centred care

Respect for autonomy in the context of health care includes a positive duty on the part of the health-care professional to support and enable a patient to make decisions that are consistent with his or her own preferences, values and ideas about how they wish to live their life. The emergence of the concept of person-centred care as fundamental to good clinical practice reflects this underlying ethical imperative. 73 The ability to have more frequent contact with a specific clinician who is known to the patient helps to build a more personal relationship and means the clinician is more likely to know that particular young person, their circumstances and what is important to them. Young people in our study more frequently identified benefits to this element of patient care from the use of digital communication than specific benefits to empowerment and control:

I think in some aspects it’s made [my relationship with the clinical team] stronger because I have more contact with them and more personal contact because I only ever e-mail like one person from the like clinical nurse specialist group if that makes sense [. . .] I think that because I have more contact now outside of clinic they kind of know me a bit better and I know them a little bit better, which is quite nice. So it doesn’t feel like I’m just ‘another patient’.

Liver conditions site, young person, 18

For me, it’s only every 3 months that you get to see a nurse . . . So when somebody’s been ill a long time, your relationship with that nurse, it’ a lot easier to contact them and they can help more with your issues because they know you there and they know your condition . . . They understand like how yours is different to everybody else’s.

Diabetes mellitus 1 site, young person, 04

Digital communication was perceived as enhancing person-centred care by our participants because it is a way of communication that young people are familiar with and use in their daily life. Communicating with them in this way can help to reduce the power imbalance in the patient–clinician relationship, with clinicians fitting into the young person’s world rather than the young person being expected to fit into the clinical world:

I think every case is individual, but on the whole if you had to rely on them ringing in the landline here and talk to you sort of through reception, nine times out of ten it wouldn’t happen because, you know, for a 14-year-old lad or whatever it’s too much like hard work . . . I suppose it’s bringing . . . not bringing yourself down but levelling . . . getting on the same level, the same wavelength.

Mental health 3 (outreach team) site, mental health practitioner, 04

Well again it’s like . . . it is like the texting and stuff because then they feel like . . . you know, that’s what they do in their everyday life, texting their friends, and I think having that contact with them through text makes them feel like we’re more on their sort of level.

Mental health 3 (outreach team) site, support worker, 03

Consent

Respect for autonomy requires clinicians to obtain consent from patients for all aspects of their care, including the use of their information. Communication by digital technology has risks as well as benefits, most notably risks related to breaches of confidentiality and data protection. Some clinician participants expressed concern at the lack of a formal consent process for the use of DCCs:

I think there probably needs to be some sort of formal conversation or information about the use of e-mail given to patients when they first attend for their first appointment, just to make sure that you’ve got their consent to communicate with e-mail. Because that doesn’t happen really, there isn’t any kind of terms of the use of . . . well there is information governance policies in the hospital but it doesn’t really translate to the clinic.

Dermatology site, consultant, 02

Because I think sometimes people might just say, oh do you have an e-mail address, and then people might not feel able to say no [. . .] so I think getting consent both to use e-mail, who can use e-mail and for what purposes, and consent for do they know about things like security of the link, so I think that’s a big one.

Liver conditions site, psychologist, 01

The practices of obtaining consent varied among clinicians in this study. A few clinicians said that they obtained written consent, some felt that it was sufficient that the patient understood the risks and others referred to implied consent to share test results:

For me times have moved on but as long as my responsibilities are professional, as long as I explain those risks to them before I use it, if they are well aware of the risks, well, the ball’s in their court.

Liver conditions site, consultant, 02

So in terms of asking them you know, about contacting them with their results the next day, they’re given an option of e-mail or telephone call . . . So I get them to e-mail me so their e-mail is coming into me rather than me e-mailing them, so that way I know that it’s their e-mail. I don’t think there’s any governance issues around it because in terms of risk or safety because the patient’s e-mailing you in and so you’ve already got implied consent at the point of the consultation we had the previous day and agreeing how would you like to get your results tomorrow [. . .] And so it’s their choice and it’s their consent.

Cystic fibrosis 1 site, specialist nurse, 03

Communication and trust

In the context of health care, trust is considered a cornerstone of doctor–patient relationships. DCCs, as discussed earlier in the chapter, can increase patient autonomy and engagement, but it does not remove the need for trust. For some participants the use of digital communication enabled a trusting relationship to develop more easily:

So I know that she’s going to text me, so then I actually get some feedback on what we’ve actually just done and how we can move forward. So she is, she’s able . . . it enables her to be able to tell me how she’s feeling and what she needs and how she needs it and her true feelings, where she can’t do that when I’m sat in the room with her. So it’s a useful tool for her.

Mental health 3 (outreach team) site, dialectical behaviour therapy co-ordinator, 10

I think it’s knowing that I can contact her easily and that she’s so nice about it all the time, she’ll always send a really friendly reply. I think I have a lot more trust in her, I feel quite confident in her care.

Liver conditions site, young person, 07

However, both young people and clinicians expressed a view that the physical co-location, both during consultation and, even more importantly, when patients need direct care, is important. Direct contact can enhance trust from both a patient and clinician perspective:

I’m trusting these people with my life; literally trusting these people with you know, with your life. This isn’t just a question of, you know, you’re asking somebody to do something for you, write a letter for you or something like that. This is your health, this is, you know, this is important, there’s nothing more important than, you know, ultimately . . . and I think also there are so many things that you could fabricate in an e-mail or a text, you could say you feel fine when actually you’re not that fine . . . so I just don’t think you can replace that personal one on one sort of. Well I suppose from their perspective it’s they need to see you and I suppose from a patient perspective, you know . . . So from a patient perspective I quite like to know who is treating me, who is calling the shots, and yes to be able to just discuss with them properly why they’re doing the things that they’re doing, or you know, what other options there are.

Liver conditions site, young person, 16

So I mean from a more clinical point of view then if you were giving them, whether it’s good news or bad news or . . . I think that needs more of a face-to-face contact, you know, to be seen by the consultant. You know, you could never text someone, oh by the way, you’re progressed, because the way that it’s dealt with is obviously incredibly important, especially the way that they take to it psychologically and that kind of thing. So I think for the more serious conversations, good or bad, they need to be done face to face. And then like I said from my point of view just giving the advice that I would, it wouldn’t be appropriate to giving it by text or that kind of thing.

Cancer 1 site, dietitian, 03

Building and maintaining trust when using DCCs may therefore require renegotiation in the context of the clinician–patient relationship and a careful consideration of individual views of patients. Trust may depend on the patients’ needs and preferences and what they think is appropriate to communicate via digital means.

Hazard

Digital clinical communication may include patient identifiers, condition-specific details (e.g. blood glucose levels), and other potentially sensitive and confidential information. Participants explained that such sensitive information might be inadvertently disclosed to parties other than the intended recipient.

In the following quotation, a patient attending a diabetes mellitus clinic emphasises the importance of considering inadvertent disclosure of sensitive information because there might be third parties seeking to benefit from unauthorised access to such data:

I’d like to think there wouldn’t be people out there that want to use people’s health information and stuff like that, but there probably are, that do want to kind of use that to their advantage. So, yeah, I think that [inadvertent disclosure] would be a big issue . . .

Diabetes mellitus 2 site, young person, 02

The following quotation, from a participant with an IG background, suggests that the scope of what should be regarded sensitive information might be broad, as it is possible to make inferences about the patient’s condition even if this is not explicitly communicated:

So, for example, we’ve probably had lots of communications that have gone on for quite a long time that could allow a third party to infer the clinical condition of a patient.

Diabetes mellitus 2 site, Caldicott Guardian, 01

In the following quotation, a clinician from a mental health clinic reveals that, in the past, sensitive information, such as patients’ names and their addresses, has sometimes been disclosed by mistake to unintended recipients:

There have been incidents and there still are incidents where clinical information gets sent to the wrong person [. . .] that can even be lists of patients names and addresses that meant to go there [i.e. were meant to be sent], but for whatever reason went to the wrong person. That does happen.

Mental health 4 (early intervention) site, case manager, 03

Consequences

Participants suggested that inadvertent disclosure of sensitive information could have negative effects on the patient’s mental health and on their well-being. It could also affect the trust between the patient and the health service.

In the following quotation, a patient attending a diabetes mellitus clinic reasons about potential adverse effects on patients resulting from the disclosure of their LTC. The quotation suggests that this might be a cause for concern, not only for conditions where known social stigmata might come into play (e.g. in relation to sexual health), but also for conditions such as diabetes mellitus, for which there might not be an obvious social stigma. Patients might perceive living with a LTC as something very personal, irrespective of the particular condition, and they might not wish to disclose this widely:

I think some people it may affect because I know like quite a few people who hide the fact that they’re diabetic.

Diabetes mellitus 1 site, young person, 13

This sentiment is echoed in the following quotation by a clinician from a different diabetes mellitus clinic. In the quotation the participant reasons about the breadth of factors as to why patients might not wish their condition to be disclosed, and about the potential impact on patients’ mental health and social relationships of disclosure:

Let’s say a patient doesn’t want anybody to know they have diabetes, you know, for whatever reason, cultural reasons, social reasons, and the world finds out they have diabetes. Well that can cause a lot of problems for them, socially acceptable, cause depression, so you know, there are those associated things.

Diabetes mellitus 2 site, consultant, 01

The patient providing the following quotation explores the potential adverse consequences on social relationships further. The patient argues that they would be most concerned about disclosure of their condition to colleagues at work. The patient expresses their concern that this might affect their relationship with colleagues or even their career:

Certain people that I work with [. . .] I suppose there would be that kind of thing, like, would that [inadvertent disclosure of the patient’s long-term condition] affect my relationship at work with them; would they think something different of me; would they always be like, ‘oh did you know this about [name], or did you blah blah blah?’ kind of thing. [. . .] I wouldn’t want it to affect my career or my job or my relationship with the people at work and stuff.

Diabetes mellitus 2 site, young person, 02

In the context of young patients, establishing a relationship between clinician and patient based on trust is particularly important. In the following quotation, a clinician from a cancer clinic argues that inadvertent disclosure of sensitive information might erode trust and jeopardise this relationship:

. . . it [inadvertent disclosure] would be a breakdown of relationship between the patient and yourself, which is probably the most dangerous part of it.

Cancer 2 site, consultant, 03

Causes

Participants suggested that potentially sensitive information might be disclosed inadvertently for a number of reasons, including hacking and interception of communication; loss or theft of mobile phones and computers; shared access to e-mail accounts and computers; forwarding or sending of e-mails to the wrong recipient; and excessive distribution of communication (e.g. copying a message to people who should not receive it).

In the following quotation, a patient expresses concern about inadvertent disclosure of sensitive information as a result of loss or theft by relating this to previous experience in other contexts:

Or if it was stored on some sort of device or something, and then as has happened for God knows how many years where people leave laptops on trains and it’s got people’s like, you know, their national insurance numbers on it and their tax codes and things like that, and their health, NHS numbers and things, that would be an issue.

Diabetes mellitus 2 site, young person, 02

Personal e-mail accounts and mobile devices can be shared easily between people. A clinician from a mental health clinic suggests that this might lead to inadvertent disclosure of information:

So for some young people it could be . . . there could be an element of risk if you’re sending stuff to phones or tablets if they leave them lying around. We don’t have many young people at all who don’t want their parents involved, for example in their care, but we do have some.

Mental health 2 (CAMHS) site, consultant, 02

Digital communication usually involves typing e-mail addresses or telephone numbers, or selecting a contact from a drop-down menu. Participants suggested that it was easy to make mistakes in these actions and to send information to the wrong recipient. A clinician from a cancer clinic describes such a scenario in the following quotation. The mistake was triggered by lack of familiarity with the mobile phone and – most likely – poor usability of the device:

I mean some of the problems that you have, and I’ve done it with another colleague’s phone, is texting . . . the hard bit is, as you say, with patient safety and confidentiality, is believing you’re texting the person when you’re not, you end up texting someone else.

Cancer 2 site, cancer nurse specialist, 11

Digital communication can be distributed easily to different recipients. There is the potential that unauthorised recipients are included by mistake or intentionally and that in this way information is disclosed to third parties. In the following quotation, a participant with an IG background suggests that such excessive distribution might breach the confidential nature of the communication between clinicians and patients:

And the other bit of that [risk of inadvertent disclosure] is the copying in of people. So suddenly this is no longer a confidential conversation between two people, there are other people, they can be forwarded. So there’s all that sort of worry from an information governance point of view.

Cancer 2 site, IG and data protection manager, 01

Mitigation

Participants described a number of strategies they adopt to prevent inadvertent disclosure of sensitive information. These included (1) limiting the use of DCCs; (2) implementing various barriers and safeguards, such as personal identification numbers (PINs), passwords and location-specific protection of devices; and (3) double-checking e-mail addresses and telephone numbers (to prevent wrong recipient errors). A further form of mitigation described was (4) explicit or implicit patient consent, where clinicians assume that the patient is willing to accept the risks when they engage in DCC.

In the following quotation, a clinician from a cancer clinic describes how they restrict the use of digital communication to sending appointment reminders. In this way, they aim to eliminate the risk of inadvertent disclosure of sensitive information:

If it was anything about say their diagnosis or deterioration in their condition, or if I had information like from our psychosocial, I would never share that on text, never, you know, it’s just . . . it’s very basic what I text and what they text me for generally. It’s more about appointments.

Cancer 2 site, social worker, 07

Some participants suggested they used passwords and PINs in order to keep the content of their digital communication secure. A patient attending a diabetes clinic describes this in the quotation below. The patient suggests that they felt reasonably well protected (in terms of security) because they were using passwords for their e-mail accounts and their phone, which were not shared:

I think I’m quite secure because I have a password that nobody knows and then I’ve got my phone encrypted as well, and I’ve also got a secure password on e-mail.

Diabetes mellitus 1 site, young person, 14

Awareness of the possibility that mistakes can be made when sending e-mails and text messages can prompt people to be particularly attentive to this action. A clinician from a cancer clinic describes, in the following quotation, how they pay particular attention when texting patients with similar or identical names in order to prevent mistakes:

I live in fear of sending the wrong text to the wrong person. So I’ve got a couple of people that are called [name] so when I have to text a [name] I make double, treble checks that I’ve got the right person.

Cancer 2 site, cancer nurse specialist, 08

Some clinicians described consenting patients as another strategy to deal with the risk of inadvertent disclosure of sensitive information. In the quotation below, a participant with an IG background highlights the importance of consenting patients, which implies that patients accept the risks associated with DCC. Although this might not prevent information from being disclosed inadvertently, the participant suggests that the organisation would be covered from a legal point of view:

. . . so we’d gain that consent from the beginning. Because once we’ve got that if the communication breaks down at least we’re covering our back as such to say well actually this was agreed with the patient.

Mental health 4 site, IG officer, 01

Hazard

Participants suggested that DCC could be subject to a broad range of communication failures, such as communication not taking place, delayed communication and misinterpretation of the communication content.

A patient attending a diabetes mellitus clinic described an example of failed communication. They described a situation where they were in need of urgent advice concerning the appropriate dose of insulin. They attempted contacting the clinic through different modalities, but were unsuccessful and did not receive any reply because the clinicians apparently did not receive any of these communication requests:

Yes, I think there was a time I was getting hypos [hypoglycaemia episodes] quite a few days in a row at the same times, and I must have rung actually, because normally if it’s a hypo they want to deal with that quite quickly, so I will try and ring. And if they . . . obviously if no one picked up so I’d text I think two of the nurses and I didn’t get a response, so I tried ringing again. And then no one picked up, so I think my mum tried texting them and ringing them and then after a few days they got back to us. But then when I said, I have tried to text and call you, they said, we haven’t got any texts or calls.

Diabetes mellitus 1 site, young person, 13

A clinician with an IG background raised similar concerns with respect to delayed communication. The clinician suggests that when patients are communicating urgent requests with clinicians using the latter’s designated contact details, then there is the risk that the response might be delayed if the clinician is away and the patient’s needs would potentially be left unaddressed:

One of the obvious risks is if for example we give a patient an e-mail address to use and it is mine as a clinician and they e-mail me and they’re in crisis and they choose to e-mail me whilst they go into crisis, and I’ve just gone on 2 weeks’ leave, who is going to pick that e-mail up?

Mental health 4 site, IG officer, 01

Some participants also expressed concerns about the possibility that digital communication can be ambiguous, or that it can be interpreted in ways different from what the sender intended. In the following quotation, a clinician from a diabetes mellitus clinic expresses such a concern about the possibility of patients misinterpreting e-mail and text message communications:

I do worry sometimes that information can be misinterpreted depending on how you phrase it in a text or an e-mail.

Diabetes mellitus 1 site, consultant, 07

A participant from a sickle cell disease clinic provided a more concrete example of possible misinterpretation of the content of digital communication. In the following quotation, the participant explains that there might be a failure to detect a critical detail when the examination and diagnosis are based on photos and images that were e-mailed by the patient:

There is always a risk that you will [miss] something that you would be able to pick up when you physically examine because you’re only as good as the photos that have been taken. And they’re two-dimensional rather than obviously in real life you can see far more perspective on it.

Sickle cell disease site, specialist registrar, 09

Consequences

Participants identified a large number of potential consequences of communication failures. These included, for example, delays in, or complete failure of, providing relevant clinical information and advice. Participants suggested that this might potentially discourage patients from seeking relevant advice. In the worst case, communication failures might require escalation to emergency services or even delay patients from seeking necessary help from emergency services. Participants suggested that communication failures could also contribute to anxiety and uncertainty for patients while they are waiting for a response from their clinician. Finally, participants pointed out that patients might react to communication failures in an abusive way, thereby negatively affecting the well-being of clinicians.

In the following quotation, a clinician from a diabetes mellitus clinic suggests that a critical consequence of the failure to establish communication with the relevant clinicians was that patients might give up seeking help even though they are in need of advice:

I guess the worst possible thing is that a patient doesn’t get the contact response that they want and give up, but really needs to speak to somebody. So you know, whichever method they use, they use, don’t get a response from us and just think, oh I don’t want to bother anybody; I won’t try again.

Diabetes mellitus 2 site, specialist nurse, 03

Some participants suggested that in the absence of advice as a result of communication failures patients might deteriorate. The following quotation is a continuation of a previous quotation given above. A patient attending a diabetes mellitus clinic described above that they had been unable to establish communication with their clinic when they were in need of advice about their insulin dose. When prompted by the interviewer about how the patient dealt with this situation, the patient suggested that they attempted to manage the critical situation themselves, but did so inadequately and, as a result, suffered a temporary deterioration in their well-being:

So how did you cope with them [hypoglycaemia episodes] while you were waiting for a response?

I just altered my insulin dose myself to how I thought it would be. Which actually turned out to be wrong, so it did leave an impact because obviously it took a good week to sort itself out.

A patient attending a sickle cell disease clinic describes, in the following quotation, an instance of communication failure that left them feeling very poorly and upset during an emergency admission to hospital. The patient describes that there had been a problem with medications and the patient was unable to contact the clinic to rectify the situation:

I went to hospital, when I was in the hospital I was taken to a ward and in that ward like I got like bad treatment, like I didn’t get the medication that I needed and stuff. And I was in pain all night, and they were prescribing me like paracetamol. Obviously I tried to call them on their mobile to say like what’s going on, but obviously their mobile phones were off because it was out of working hours. [. . .] I could have died. I could have died. I could have died. I could have died or I could have passed out or I could something. I could have even got angry and just started swearing or something, causing a disruption.

Sickle cell disease site, young person, 10

A clinician from a mental health clinic describes, in the following quotation, the importance of trust for the therapeutic relationship. When patients do not receive a prompt reply to their messages, they might feel ignored and this could have adverse consequences for the relationship between clinicians and patients:

Like if you did get to the stage where you use instant messenger kind of thing, that people can see when you’ve read that message, and I think that can be really harmful for like a therapeutic relationship if like a service user could see that a clinician had ignored . . . well not ignored, just hadn’t replied to their message but had read it and that kind of thing.

Mental health 4 (early intervention) site, occupational therapy student, 05

Finally, there might be a potentially negative impact on the well-being of staff. A clinician from a mental health clinic suggests, in the following quotation, that there have been instances where patients reacted to communication failures in an abusive way, which can put stress on staff:

My other biggest element is we work with families, where there are difficulties, and quite often lots of difficulties, more than often lots of difficulties. But also quite often within the parental side of things, there can be personality type difficulties, or just people who are not happy with the service. And I’ve seen, very early when I started here, I was on the receiving end, I think it had been forwarded to me rather than sent directly, but a very negative, very abusive complaint, basically unjustified I have to say. But from a parent who didn’t have the ability to reason or think or wait for a response, so it can become very problematic and can be very damaging to clinicians potentially.

Mental health 2 (CAMHS) site, psychologist, 14

Causes

Participants identified a multitude of potential causes for communication failures, including distraction (e.g. during Skype consultations); not answering/accepting communication from unknown numbers; not being able to establish the patient’s identity; delay in picking up/responding to messages (e.g. people working part-time, out-of-hours query, patient feeling too unwell to respond); inability to access the internet on mobile phones as a result of a lack of signal or credit; poor usability of devices; difficulty expressing clearly information requests in text messages; patients downplaying the severity of their condition in text messages; learning difficulties; and inability to check correct understanding of communication content in asynchronous communication.

Some participants suggested that their condition, for example cancer, could leave them feeling unwell for extended periods of time. During these periods they might prefer to avoid all social contact, and they might perceive e-mails and text messages as unwelcome disturbances and disregard these. In the following quotation, a patient attending a cancer clinic describes such situations, which put them in a different state of mind, and where they ignore messages sometimes for weeks while they are feeling unwell:

Like when I first came in I thought, I want to use my phone, but it’s like actually well I don’t want to use it [. . .] you end up leaving your phone for days. I mean the longest I’ve spent was I didn’t look at my phone for 2 weeks, and obviously when you look you get loads of messages and whatnot. And so using it when you want is one thing, you’ve just got to be able to be well enough to use it. Because sometimes your head is in another world and you’re just not sure when you want to use it or how to use it sometimes.

Cancer 2 site, young person, 03

In the following quotation, a clinician from a diabetes mellitus clinic describes how digital communication can create delays in getting advice across to patients as a result of its asynchronous nature. The clinician suggests that this was because a conversation sometimes requires additional information, and then there might be delays until a problem is resolved while messages are being read and replied to:

Whereas if it’s e-mail it’s then them responding to my message with another e-mail and then waiting for me to pick that up and respond again. So there’s a little bit of delay there whereas telephone, it’s quicker.

Diabetes mellitus 1 site, consultant, 07

Some participants contrasted textual DCC with in-person and telephone consultations. They concluded that the absence of additional cues about posture and tone might be a cause for communication failures, as it was easier to misinterpret information and to miss subtle details. In the following quotation, a clinician from a mental health clinic describes how this could adversely affect the therapeutic relationship because they might be less aware if a patient was struggling:

I suppose text messages it’s difficult to gauge, like I’ve said before, kind of subtleties in somebody’s tone of voice or, you can often from a person’s voice you can pick up on what might, you know . . . if there’s anything going on for them. And you can ask them questions about it, whereas you just don’t get that in a text message. So if that person is struggling then perhaps we might not be aware of that issue if and when we’re texting them.

Mental health 4 (early intervention) site, case manager, 06

Mitigation

Participants suggested that the risk of communication failures might be mitigated by (1) limiting the use of DCCs (e.g. using it for only simple things, such as appointment reminders), (2) providing training in asynchronous DCCs, (3) ensuring that enough time is set aside to write and read DCCs carefully (see Inadvertent disclosure of sensitive information), (4) using alternative means of seeking advice in an emergency and (5) following up DCCs by other means.

In the following quotation, a clinician from a cancer clinic describes that they restrict the use of text messaging, for example, to simple, unambiguous communication, such as arranging an appointment:

No, text is for the movement of small . . . it’s usually for the movement of a small fact, a small low-risk fact. ‘I know I said 9.30, but can it be 10.30?’ Or, ‘can you ring me this afternoon?’, or something like that. Text is fine for that.

Cancer 2 site, consultant, 03

The provision of training and guidance in the use of digital communication might increase the confidence of clinicians in managing and avoiding potential risks. A clinician from a mental health clinic explains in the following quotation that they required guidance in order to deal with new hazards that are introduced through the use of digital communication:

But then I think professionally to be given advice around what happens then [when digital clinical communication is used]; about sort of patients, what happens if somebody says at 12 a.m., maybe on our trust . . . for example if I’ve got a Facebook for mental health and I’m using it, what if they put on my Facebook or Twitter that I’m going to kill myself, what then, where do I go with that? It’s huge information.

Mental health 2 (CAMHS) site, consultant, 05

Some participants suggested that in the case of urgent queries, patients would not sit back and wait for a delayed response, but proactively seek out alternative means of communication in order to get advice. A clinician from a diabetes mellitus clinic describes, in the following quotation, a situation where they had been contacted by e-mail, but had not responded immediately. The patient’s parents then made contact by telephone fairly quickly:

Yes, because I had that when I received an e-mail while [colleague] was on annual leave, I received an e-mail from a parent. But before I’d actually seen the e-mail the parents rang me, so, you know, because I hadn’t responded the parents rang, so . . .

Diabetes mellitus 1 site, health-care assistant, 01

In the following quotation, a patient attending a diabetes mellitus clinic describes that they would follow up previous e-mail communication by a visit in person to ensure that there had been no communication failures:

Well I’d e-mailed her on the Saturday and then I popped in on the Monday or the Tuesday I think, to just kind of follow it up and say, ‘I’ve had a bit of nightmare kind of thing, did you get my e-mail?’

Diabetes mellitus 2 site, young person, 05

Hazard

The clinicians at the different clinics used several information systems and they communicated with patients through a number of different media (e.g. telephone and various forms of DCC). Participants (clinicians) suggested that there was the potential that some of the content of these various conversations might not be recorded, or that it might be recorded in different places, and as a result some information might not be available to clinicians or patients, or they might not be aware of its existence.

In the following quotation, a clinician from a cancer clinic describes that they communicate digitally with patients while they are out of the clinic. The participant suggests that subsequently recording this information in the patient’s notes poses challenges:

I think the challenge is of actually ensuring that that communication gets in to the patient record. So I can be remote, and I can be responding to patient queries, but how that actually gets recorded into the patient record is a challenge.

Cancer 2 site, IG and data protection manager, 01

A participant from a mental health clinic describes the difference between guidelines for recording communication content and actual practice. The participant explains that although the guidelines require a record of every communication to be kept, in practice this is often not followed through:

Well technically we are supposed to write in the paper notes word for word every text message, every e-mail that we have. And I have to confess I’ve not always done that, in fact I frequently haven’t done.

Mental health 4 (early intervention) site, psychologist, 07

Consequences

Participants explained that failure to record that communication between a clinician and a patient had taken place, or to record what was said during the communication, could result in other clinicians working with the patient being unaware of this prior communication. Some participants suggested that, in consequence, this failure might lead to unnecessary duplication of questions and advice given to patients. Some participants also pointed out that failure to record what advice was given could leave gaps in the audit trail and that there was also the potential for subsequent disagreement between patients and clinicians about what was actually said.

In the following quotation, a clinician from a diabetes mellitus clinic describes how patients might interact with more than one clinician. The participant suggests that, if these interactions are not recorded, then there is no way of knowing what advice has already been provided and advice previously provided by another clinician might be repeated unnecessarily:

Say, for example, this patient saw a dietitian and I wouldn’t know unless somebody writes in there [electronic system]. So I will be able to read it if the patient has seen a dietitian or even a diabetic nurse before, so I know what kind of information they will have had from the diabetic nurse, so I can talk to my patients when I see them accordingly because otherwise I will just repeat the same thing what the diabetic nurses have said.

Diabetes mellitus 2 site, registrar, 06

A clinician from a mental health clinic recalls, in the quotation below, a recent incident where failure to record communication with a patient by text message over an extended period of time led to gaps in the documentation:

So the example that I can think of was a fairly recent death with a young person who was communicating with – a clinician was communicating with a young person a lot by text, but it wasn’t being entered into the paper records, because we don’t have electronic records throughout our system. And after the event there was lack of information readily available that was documented about that interaction.

Mental health 4 (early intervention) site, consultant, 01

Causes

Participants described a number of causes for the failure to record digital communication in patients’ notes. Participants suggested that mobile communication was not logged automatically, and the content of text messages was not as easily transferable to the patient’s notes as other forms of communication (e.g. e-mail). Some participants suggested that documenting e-mails also posed problems as these might not integrate easily with the patient’s notes and it could be a time-consuming activity. Devices have limited storage, and text messages and e-mails might have to be deleted even when they are not documented elsewhere. In addition, participants suggested that they needed to determine for themselves what kind of DCCs they deem reasonably relevant to document. This could lead to significant variation in practice.

A clinician from a cancer clinic describes, in the following quotation, that they are engaging in frequent digital communication with patients. The participant explains that they are struggling to record these interactions because there is no easy way to record text messages:

Because you’re doing it [texting] all the time during the day, you might sometimes text people four, five times in an hour, if not more. And you might text about 20 different patients a day, I think what I do find hard is having that ability to back it up, what I’ve done, and what my contact’s been with that patient.

Cancer 2 site, cancer nurse specialist, 11

A participant from a diabetes mellitus clinic indicated, in the following quotation, that they deleted all e-mails that they did not deem relevant. As this is a subjective assessment, there might be significant variation about what gets deleted and what is stored:

I wouldn’t necessarily keep hold of e-mails if they weren’t relevant to keep hold of.

Diabetes mellitus 2 site, dietitian, 04

Mitigation

Some participants suggested that treating every communication equivalent to an in-person consultation might mitigate the risk of not recording communication content. Risk could also be mitigated by limiting the use of DCC to forms that are more readily transferred to the patient’s notes (e.g. preferring e-mail over text messaging), or generally by restricting the use of digital communication to simple interactions with little or no clinical content.

In the following quotation, a clinician from a diabetes mellitus clinic explains that they recorded the clinically relevant content of digital communication to the same extent that they would if it had been a regular consultation:

Anything that’s clinically relevant I will cut and paste the conversation and put it in our electronic notes so that the rest of team are aware that I’ve done XY or Z in the same way that I would if there was a consultation, I do a summary of what that consultation is.

Diabetes mellitus 2 site, specialist nurse, 03

A different clinician from the same clinic describes in the following quotation that they preferred to use e-mails over text messages because recording of e-mails was easier:

I also like e-mails because of one simple reason; you’ve got a very nice little record to keep, and texts are not always easy to put back in the clinical notes. But e-mail you can clearly print off and put in the clinical notes, or scan it in. Texts, possible, there’s no reason you can’t, but I think technically it’s probably easier to do it this way.

Diabetes mellitus 2 site, consultant, 01

Restricting or avoiding the use of digital communication is described, in the following quotation by a clinician from a mental health clinic, as another strategy for reducing the risk of not recording clinically relevant information. The participant argues that there is a lack of clarity in the organisation’s guidelines about how digital communication should be recorded and for this reason the participant limited their use of these forms of communication:

So limited use actually of digital communication, and I guess mindfully really because some clinical incidents have happened when there’s been lack of clarity about how e-mails for example or texts are used and how they link with clinical recording of information. There isn’t clarity in our policies or processes at the moment, hence why personally I’m not using as much as I potentially could. Because of those issues.

Mental health 4 (early intervention) site, consultant, 01

Hazard

Patients can initiate digital communication at any time and they might have requests for urgent advice. Some participants (clinicians) reported that they sometimes respond to patients’ queries without consulting the patient’s notes.

In the following quotation, a clinician from a cancer clinic explains that they thought they did not need to refer to their patients’ notes because they were familiar with the patients and their history:

I might have 12, 15 service users that I work with at any one time, so I know their circumstances. I don’t need to refer to clinical notes, I retain everything.

Cancer 2 site, social worker, 07

A clinician from a diabetes mellitus clinic reflects on the nature of calls that patients are likely to make. The participant argues that mostly these would be requests for urgent advice and that, in such cases, consulting the patient’s history was not necessary as they were focusing on the immediate situation:

Those calls are, ‘I’m in an acute situation what should I do about this?’ And it doesn’t matter then what’s gone on before, it’s about ‘what do we do now to keep you safe?’ So I don’t need to access any notes at that stage.

Diabetes mellitus 2 site, specialist nurse, 03

Consequences

Some participants suggested that failure to consult the patient’s notes prior to engaging in DCC can result in the clinician’s reliance on an incomplete patient history, and they might duplicate or contradict advice given previously by a different clinician (see Failure to record the content of digital communication), or they might miss relevant pieces of information in their decision-making.

In the following quotation, a participant from a diabetes mellitus clinic points out that there might be gaps in their knowledge of the patient’s history if they do not consult the patient’s notes because other clinicians might have worked with the patient:

So I might have seen them, but you don’t always know what’s happened kind of in between.

Diabetes mellitus 2 site, dietitian, 04

Another clinician from the same clinic echoes this by explaining that they could not simply rely on what the patient relates at a particular moment in time and failure to check the notes might lead to critical pieces of information being overlooked:

It’s very, very common for patients not to report everything, because they know it, they’ve lived it, they tell you the bits they think are relevant but miss out something that’s quite critical.

Diabetes mellitus 2 site, specialist nurse, 03

Causes

Participants mostly identified perceived familiarity with the patient and frequency of contact with the patient as a reason for not consulting the patient’s notes. Some participants also identified the non-medical nature of many digital communications as a further reason for not consulting the patient’s notes. In addition, one participant suggested that, if a patient has an urgent query, it could be addressed independently of the patient’s history.

In the following quotation, a clinician from a diabetes mellitus clinic explains that whether or not they consulted the notes depended on how well they thought they knew the patient:

Generally speaking, if it’s a new person then I would, you know, definitely would like to refer back to the notes just to see who the patient is, because I wouldn’t just do it on the cuff. But most of the patients I would know who they are, so that sort of helps.

Diabetes mellitus 2 site, consultant, 01

A participant from a cancer clinic describes, in the following quotation, that they do not look up the patient’s notes in cases where they are communicating with a patient’s parent about a logistical matter:

And then another example for not having to look [up the notes] is trying to arrange a mum to pick up some tablets, so I know that the tablets will sort it, so it’s just a matter of texting her back and telling her to where to pick stuff up from.

Cancer 2 site, cancer nurse specialist, 08

Mitigation

Some participants perceived that they had a sufficiently high level of familiarity with the patient to mitigate patient safety being compromised by not referring to the patient’s notes. Some participants also described double-checking after the communication has taken place as mitigation.

In the following quotation, a clinician from a diabetes mellitus clinic explains that, even though they might engage in digital communication without having consulted the patient’s notes, they would afterwards double-check the notes in case they were worried or felt they did not have an up-to-date knowledge of the patient’s circumstances:

Yes, particularly if it’s someone I haven’t heard from for a while, I would go to my clinical notes to think, ‘have they had any contact with anyone else about this, and what have they said to them?’ Just to make sure that I didn’t go and contradict something that somebody else had previously said to them, or missed.

Diabetes mellitus 2 site, specialist nurse, 03

Limitations of the health economic component of the Long-term conditions Young people Networked Communication study

The LYNC study was an in-depth qualitative assessment of the use and impact of digital communications, defined broadly and evaluated across a diverse collection of case studies. Although this context yields numerous insights into current practice, it does impose significant limitations on what could be achieved in terms of formal economic evaluation within the LYNC study. We therefore begin with an acknowledgement of these limitations, and a clarification of the goals of the health economic data collection and analysis that was performed. As we explain in Chapter 2, the LYNC study was not designed to support formal economic evaluation of digital communication. Instead, our work might best be viewed as a feasibility study to motivate and inform future economic evaluations of specific interventions involving digital communication. We discuss the relevance of our findings to such evaluations below, but acknowledge fully that our results (qualitative and quantitative) should be seen as indicative rather than conclusive. Our collection of data on willingness to pay, for example, was never intended to provide precise and valid estimates of the value young people attach to the use by services of digital communication. Given that participants were recruited from diverse sites, with only a few participants per site, this would never have been possible. Instead, our aim was to use the opportunity afforded by the collection of qualitative data from service users to understand how they might interpret such questions, posed in different forms, and gain an indication of the range of values we might encounter. We discuss this experience in further detail below.

The challenge for any economic evaluation of DCC is that the term covers a broad range of practices, and the costs and benefits of these practices will vary according to aims of the clinic and the specific approaches involved when clinics adopt DCC. It is, therefore, not helpful to consider the cost-effectiveness of DCC in general terms, as any particular decision around the adoption of practices involving DCC would involve a specific economic evaluation, which was beyond the scope of this study. It may be helpful, however, to distinguish between two types of specific decision problem involving DCC. The first type would involve a well-defined intervention or practice for which the use of DCC is an integral component. There were some examples of this among our case studies, such as the delivery of dialectical behaviour therapy via text message. In such examples, it is more appropriate to consider an economic evaluation of the intervention as a whole, rather than of DCC in isolation. Such an evaluation would require a robust assessment of the incremental impact of the intervention relative to a suitable comparator, in terms of costs and benefits; however, it was not possible to obtain the data required for this within the study.

The second type involves the uptake and expansion of DCC as part of routine activities of a service. Our findings relate primarily to this second type of decision problem, although many of our findings are relevant to any economic evaluation of an intervention with a DCC component. However, the requirement for robust data on the incremental impact of the expansion of DCCs remains. The LYNC study was cross-sectional, rather than a before-and-after study, and did not include formally matched controls. It was therefore not possible to carry out any incremental analysis of the costs or benefits of changing practice regarding digital communication. The data on costs reflect total time spent on digital communication activities, which will include activities that would previously have been conducted using conventional communication channels, rather than the net impact of the adoption of digital communication on workload. Although staff commonly expressed the view that digital communication was more likely to have increased, rather than decreased, their workload, they were unable to quantify the net impact with any accuracy. This is understandable given that the adoption and impact of digital communication is a gradual process. It was, therefore, not possible to collect valid data on the incremental cost of digital communication within the LYNC study. Similarly, we did not collect primary data on outcomes and it was not possible to discern the incremental impact of digital communication on outcomes from the clinic data that were available, as would be possible in a study that was explicitly designed to evaluate the impact of a specific change in practice in a well-defined setting which involved collection of intervention and comparator data on costs and outcomes. However, the qualitative and quantitative data we collected provide substantial insights for such evaluations, which are discussed in the next section.

Insights for future economic evaluations

The first question to consider is whether or not an economic evaluation of the ongoing adoption of DCC in routine clinic practice is required. It might be thought that this is an inevitable evolution of clinical practice and that the cost implications of this change are likely to be immaterial, or even positive (cost saving). Our survey of staff shows significant variation in DCC use across sites and among respondents, and suggests that sites are making active choices about the adoption of DCC, with substantial cost implications. In particular, the key driver of cost is staff time, which typically accounts for > 90% of the direct costs associated with DCC use. Furthermore, we found a strong association between the use of DCC and the seniority of staff involved, with consultant-grade staff reporting substantially less time spent using DCC. This influences the total cost impact of DCC use, and may largely be a result of the different roles staff have in the clinic and the different relationships they have with service users. However, it is also possible that this reflects, at least in part, generational effects. As current young doctors become consultants, they may be more willing to engage with DCCs, and this may lead to an increase in DCC-related costs.

As we explain in Limitations of the health economic component of the Long-term conditions Young people Networked Communication study, the direct costs reported may well involve activities that were previously carried out using conventional channels of communication, rather than additional activities. When combined with efficiencies generated through the use of DCC, it is possible that the net impact of DCC might be to reduce staff burden and the cost associated with it. Examples of such efficiencies relate mainly to managing appointments and reducing missed appointments. There is the potential to replace in-person contacts for routine reviews. However, it is notable that no staff reported that DCC had reduced their workload overall, and many reported that it had in fact increased it. The importance of this finding is reinforced by those staff who perceived clinic managers to hold expectations that DCC would increase efficiency and reduce pressure on staff time, expectations which did not match reality for staff directly working with young people. It is the benefits of DCCs that cause the pressure. Respondents reported that DCC improved engagement with young people. There were examples provided of specific groups that responded favourably to DCCs, but there was also a general effect as young people preferred to use digital media. DCCs also improved relationships between young people and the clinicians, and made it easier for young people to contact staff between appointments. These are positive impacts, but dealing with this broadened and enriched engagement with young people took up considerable amounts of staff time. Furthermore, DCCs were in some cases reported to break down the professional/personal distinction in the relationship between young people and clinic staff, so that increased communication occurred that was not always directly related to clinical need.

When assigning costs to resource use implications in an economic evaluation, standard methodology requires that unit costs reflect the opportunity costs of the resource from the perspective of the decision-maker. 77 This has implications for costing the impact of DCCs in terms of reductions in appointments missed. This opportunity cost is likely to be high where staff are travelling to meet young people for appointments as considerable staff time will be wasted if the young person fails to turn up. In the case of clinic visits, however, the opportunity cost depends on how services respond to missed appointments. If the time is used to catch up on a backlog of attendees, or paperwork, then the opportunity cost of the missed appointment will be lower. The opportunity cost of reducing DNAs is, therefore, highly variable depending how a clinic operates and can be extremely difficult to quantify. It is clear, however, that missed appointments impose a cost because they restrict the ability of staff to use their time productively, and require resources to chase attendees and rearrange appointments. NHS England has estimated that the average cost of missed appointment is £103,78 although this is likely to be a considerable overestimate, as it assumes that the time that was allocated to the missed appointment is completely wasted. Nevertheless, there are likely to be considerable costs associated with missed appointments. A further implication of the standard costing approach relates to DCC carried out by staff in their own time, using their own equipment. From a health service perspective, this technically involves zero cost. Services need to be aware of the impact this is having on staff to avoid issues of burnout, absenteeism and staff turnover. It may be that services need to be reorganised to ensure that sufficient time is allowed for DCC, and investments need to be made in dedicated out-of-hours DCC services, which would have clear cost implications.

Although decision-makers need to be aware of these cost implications, and not expect DCC to be cost saving, they should take into account the potential benefits of effective use of DCCs. Although it was beyond the scope of this research to quantify such benefits, we have identified clear, consistent and credible pathways through which DCCs can lead to improved short- and long-term outcomes. Services cannot benefit young people unless they engage with them and DCC was widely perceived to improve this engagement. Although this does imply an increased clinic workload, this needs to be set against the benefits to young people who might otherwise not have engaged with services. The impact of DCCs across sites could be described as enabling a move from intermittent to continuous care, with improved monitoring and therapeutic interaction between appointments. A number of staff reported how this led to improved short- and long-term health outcomes. DCC also improved the richness of information available to staff, which was reported by several respondents to enable more personalised and responsive care, which would again lead to improved health outcomes.

The findings of our study highlight the importance of considering the benefits to young people of DCCs beyond the direct impact on their health. A recurring theme reported by respondents was that living with chronic health conditions negatively affected young people’s lives in ways that went beyond the clinical consequences of the condition itself. Young people reported a loss of control and independence as a result of the need to manage their condition. There was considerable anxiety reported as a result of the uncertainty around their health and the potential for it to deteriorate. In many conditions, young people are faced with challenges around learning to manage their health and their treatments. This can involve learning a complex task in a short period under considerable pressure. DCCs were reported to help young people with these challenges to their well-being. This has important implications for any economic evaluation because these benefits may be difficult to capture using standard preference-based health utility measures, such as the EuroQol-5 Dimensions, but are clearly highly valued by young people.

Our willingness-to-pay exercise was an exploratory attempt to estimate the value of DCC for the young people, although the results must be interpreted with considerable caution, given the limitations to the study mentioned above. Our main aim was to explore how we might elicit this value and assess how young people might respond. We initially carried out this stated preference elicitation as a willingness-to-accept exercise to avoid the results being constrained by income. However, respondents found this perspective hard to relate to conceptually. We therefore revised our approach and asked a willingness-to-pay question. This was a more comfortable approach for respondents, as it was consistent with how they purchase the technology used for DCCs, but it did mean that responses reflect budgetary constraints. Where respondents still struggled to quantify the value they placed on digital communication, the interviewer was allowed to prompt the interviewee by offering suggested values. A range of £5–50 was used as prompt (although the most common value was £10). In addition, although it was explained that this was a hypothetical question, several respondents objected to having to pay for a service they considered themselves entitled to and their zero responses did not reflect their perceived value from DCCs. Such protest responses were, however, uncommon. The one site where protest votes were common was a mental health site, where the protests reflected a perception that this was another example of ongoing discrimination against mental health services.

Given the lack of consistency around approaches to the willingness-to-pay question, the use of prompts (which can lead to anchoring bias), the obvious difficulty many respondents had in answering the question, the diversity of contexts in which responses were collected and the small number of responses per site, the results presented can be considered only illustrative at best. However, they are consistent with many users placing a high value on the use of digital communication. When combined with the qualitative findings about the value users place on the non-health benefits associated with digital communication, our results suggest that a more robust estimation of this value would be an essential component of future formal economic evaluations of service delivery interventions involving digital communication. It may be that using an alternative approach, such as discrete choice experiments,79 in such an evaluation would give a more accurate estimate of the value young people experience from DCC.

Although we were unable to quantify the potential cost-effectiveness of DCCs across the study sites, cost-effectiveness evaluations of interventions in the conditions managed at the clinics illustrate the potential benefits that DCCs might achieve. In diabetes mellitus, for example, the potential benefit of reducing HbA_1c levels from 7.5% to 6.5%, as recommended by NICE,80 via a structured education programme, was estimated to be around £3500 per patient. 81 The UK Prospective Diabetes Study 41 reported that improved blood glucose control reduced the mean cost of managing complications by £771 per patient. 82 The costs associated with a mental health crisis requiring inpatient stay have been estimated at £2900 per day. 83 These examples illustrate the potential magnitude of cost savings achievable, to be set alongside the gains in patient health. There is significant potential for DCCs to be cost saving overall when these costs are taken into account, although it must be noted that the savings do not necessarily accrue to the service incurring the DCC-related costs.

Digital clinical communication as patients’ main requested method of communication

There was consensus among the IG officers that current society is technologically advanced and young patients have now adopted digital communication platforms as their main methods of communication. The IG officers were strongly in favour of the health-care sector using mainstream communication methods in order to deliver a high quality of service:

So, you know, we’re listening to what our service users are saying about how they, particularly the younger age group, how could this be part and parcel of how they communicate with each other on a day-to-day basis. And more and more, as individuals in our personal lives, we use digital communication, social media, so much more and actually we need to be able to respond to that within health.

School nurse service, IG lead, 01

A majority of the IG officers remarked that patients were increasingly making requests to engage with clinicians through digital communication and were resistant to receiving information relating to their health via conventional post:

But my personal view would be that particularly if you’re talking about young people, you know, their means of communication is not by letter and telephone, it is increasingly interactive with various forms of social media. And, therefore, to engage and to have relevance to those service users the organisation needs to use the tools and means best suited.

Mental health 2 (CAMHS) site, IG head, 01

Another incentive identified for establishing digital communication within the NHS was the recent Caldicott Review in 2013,33 which asserted that patients’ preferred methods of communication should be respected. The review was quoted by some of the IG officers:

But now we have it in the form of the Caldicott Review which you may be familiar with, and Dame Fiona Caldicott said it’s ridiculous that we’re not bending to patients’ demands for use of modern technology; the NHS needs to move into the 21st century and they concluded that it’s OK to communicate with patients by e-mail so long as the patient consented and they are aware of the risks.

Cystic fibrosis 1 site, IG lead, 01

Ten of the 16 IG officers interviewed indicated that responding to patients’ requests was essential and viewed DCCs as effective methods of communication. They gave examples of where it was better than using the postal service:

. . . It’s just kind of engaging with people in the way they want to be engaged with. We have a lot of people who are no fixed abode and they’re sofa surfing somewhere. Their life is on their smartphone.

Mental health 4 site, IG officer, 01

As a direct response to patient requests, a number of IG officers reported that their NHS trust was considering introducing text messaging and e-mailing facilities:

So, we’re beginning to allow e-mail communications with patients. It’s still in its infancy, policy change only happened a few months ago . . .

Diabetes mellitus 2 site, Caldicott Guardian, 01

We again, another development, we are looking at, at the request of patients really, an ability to use text messaging as a part of a keeping in contact. So, you know, a sort of checking up service, how are you today, you know, how are you feeling?

School nurse service, IG lead, 01

Digital clinical communication promotes patients’ access to their health records

The IG officers acknowledged policies to enable patients to have access to their own health records:

. . . The Caldicott Review is very supportive of new technologies. And there are lots of demands being made on trusts’ intent on the health service generally in terms of patients being able to access and manage their own care across the internet.

Cystic fibrosis 1 site, IG lead, 01

Eight IG officers said that their trusts were implementing, or had plans to implement, developments to achieve this goal:

. . . There’s obviously a requirement of any new provider of our clinical systems that they will have integral to them clinical interactive systems in that clinicians will be able to interact with the online clinical information out and about when they’re agile working. And that also there would be a patient portal so that patients could also be having access to certain aspects of their record and be able to input to their record . . .

Mental health 2 (CAMHS) site, IG head, 01

I mean there are isolated systems out in the trust and the one I can think of is renal patient use. So that’s online and it’s, you know, patients can access information about themselves and also clinicians can input into that system. So that’s an example of an online system that is accessed by both patient and clinicians managing their condition.

Sickle cell disease site, IG officer, 01

The popularity of e-mail communication as the principal option for meeting this goal emerged from the interviews:

. . . We would like to think we’d ultimately be able to use e-mails more securely so that we can provide patients with information about their care, electronic patient records, all these things will be not far away.

Diabetes mellitus 2 site, IG officer, 01

Yeah, ultimately I would like to think that we could e-mail patients copies of the letters and things like that from their outpatients. So at the moment we can send discharge summaries and stuff to GPs [general practitioners] but it would be nice one day if we could copy patients into the correspondence.

Diabetes mellitus 2 site, Caldicott Guardian, 01

Digital clinical communication improves patients’ service engagement

Six IG officers reported an improvement in young patients’ service engagement as a result of receiving either personalised or automated text messages:

We do phone patients to talk to them about appointments coming up and make sure that we improve DNA rates.

Liver conditions site, IG officer, 01

Sending text messages to patients was considered normal irrespective of its purposes across many of the NHS trusts where our participants worked:

Well we text a lot of patients, because we text them all prior to outpatients. We text them all prior to theatre to make sure they’re going to turn up.

Diabetes mellitus 2 site, IG officer, 01

We do use text services. Outpatients and orthopaedics use text services . . . We have our own in-house text servicing, so you can use it or not. I think some of the busier clinics do use text reminders.

Arthritis site, Caldicott Guardian, 01

Eight NHS trusts where our interviewees worked were implementing other digital communication systems in efforts to improve appointment attendance:

We are shortly due to implement a system called Lorenzo which enables us to collect e-mail addresses and things like that, and it works in a way which will enable e-mailing appointment letters to patients.

IBD 1 site, IG officer, 01

The use of social media would be both to promote the service and to engage the service user.

Mental health 2 (CAMHS) site, IG head, 01

Innovative service proposals

The development of Skype clinics was reported from across more than half of participating trusts, although implementation was at varying stages in different NHS trusts:

So in the video world I’m aware there are people that [name] mentioned like speech therapy, paediatrics, who are doing some sort of Skype stuff. But mostly they’re people waiting for us to give them a solution that they can use it.

Sickle cell disease site, IG officer, 02

One IG officer reported that Skype clinics had been piloted in the wrong clinic, but the trust still remained interested in trying it out in a sexual health clinic:

They did pilot Skype clinics, in I think it was dermatology, but found that actually it was the wrong kind of clinic to run via Skype because dermatologists still need to be able to physically see and examine the conditions. So those were very quickly stopped.

Liver conditions site, IG officer, 01

A key motivation for initiating Skype clinics across different health conditions was to promote patient engagement with health services:

We‘re about to pilot two areas where we are going to Skype patients from clinic . . . Because we’re a specialist hospital, we have very specialist services, we have patients who travel over from [area], which is, you know, 100 miles. And so they want to try and do consultations by Skype.

Arthritis site, Caldicott Guardian, 01

One IG officer commented that sexual health clinics were leading innovation as they mostly served a younger population:

The trailblazers in this regard have been sexual health services, and I think that’s largely because of the particular cohort of patient they‘re dealing with, it will be teenagers very largely and under 25s . . .

Cystic fibrosis 1 site, IG lead, 01

As for consultations, obviously I mean we’ve all had a good laugh about the idea of the Skype consultation in sexual health and what that might consist of [laughs], you know, obviously the clinicians there are going to have to draw the line about how far they can take a consultation.

Cystic fibrosis 1 site, IG lead, 01

Being able to deliver care remotely and its use in triage were talked about as key advantages of video communication:

. . . Be able to put video by every intensive care bed in the trust, so that’s 60-something beds will be remotely supported by intensive care professionals sitting behind a computer. And we’re doing some other things in terms of being able to send diagnostic images from one location to an expert somewhere else.

Sickle cell disease site, IG officer, 02

You might not be able to get to see the dermatologist in person, but maybe you’ll be able to get to the point whereby you meet a specialist nurse in a video clinic who could press the button to record it at the right moment to escalate it through the process.

Sickle cell disease site, IG officer, 02

Perceived benefits to health-care service delivery

The IG officers identified a range of benefits from using DCCs.

Saving time and money for the patient, particularly those living in remote areas, and for the NHS was identified as a key benefit:

Telephone saves the patient a hell of a lot of bother. We also have . . . we provide services for patients from quite a wide geographical area, including the [place], so if it saves patients coming across the [place] then it’s a significant benefit . . . So you know we’re absolutely sold on modern technology.

Diabetes mellitus 2 site, Caldicott Guardian, 01

And, you know, it’s using these technologies to be able to provide that quality care to service users with maybe reduced service, if you know what I mean, reduced number of staff. So you might be able to see more patients, but your team might actually be smaller, because what you’re doing is you’re using other types of engagement methods with them.

School nurse service, IG lead, 01

Being able to provide services at the right time was also identified as a benefit:

So another benefit of e-mail . . . if a result comes through that means the patient needs to be reviewed, it’s much more instantaneous if you can phone them on a mobile or you can e-mail them, than sending a letter to be typed, getting that letter typed, putting in the mail. Mail is mainly second class now, so you know, there is inherent time delays.

Liver conditions site, IG officer, 01

Current digital clinical communication-related policies

Thirteen of the 16 IG officers reported the existence of policies in their organisations that were devised specifically to cover either text messaging, e-mailing, social media and/or the use of handheld mobile devices:

We have an e-mail policy, we have a mobile device policy, specifically for phones, and we have a social media policy for staff.

Arthritis site, Caldicott Guardian, 01

We‘ve got a large policy and protocol set and guidance set as part of IG, so there’s individual protocol for use of internet, use of e-mails, . . . and social networking, so you know, as these technologies come on board we create policies for them and sort of review the existing ones.

Sickle cell disease site, IG officer, 01

There was recognition that policies need to evolve as digital communication services evolve:

And if anything, my role as a Caldicott Guardian has been about saying to our information governance team, we can’t keep on with the same policies as we’ve had because we need to assure that there are these modes of communication for patients.

Cancer 2 site, Caldicott Guardian, 01

So I think our policies are beginning to acknowledge that where communication has come directly from the patient using e-mail, or where the patient has requested an e-mail correspondence, you know, these are the standards that we expect you to use.

Cancer 2 site, Caldicott Guardian, 01

One IG officer described how compliance with policies was monitored:

In terms of the Fair Processing Notice for individual projects we spot check them from time to time, and we don’t just spot check the new projects, we spot check elements within the trust. We will just walk into one department and say ‘hello we are from IG and let’s see what you’re doing’, which doesn’t make us very popular but it gives the organisation assurance and it gives our managers assurance.

Liver conditions site, IG officer, 01

Training courses during staff induction and mandatory refresher online courses were described among measures for enforcing policies related to DCCs.

Policies in development

A majority of IG officers reported having plans to develop digital specific policy for DCC:

They are currently working on guidance regarding e-mailing patients but this is very much in draft stage.

Liver conditions site, IG officer, 01

But I know the trust doesn’t currently have a policy for e-mailing patients because it’s a bit of a taboo subject within the NHS, there’s no national guidance on it, so . . . But it is something that the trust’s going to be embarking on anyway.

IBD 1 site, IG officer, 01

No existing policies

A few trusts did not permit DCCs, and there were no active policies in existence to cover such activities, but the IG officers in these trusts were aware that it was taking place. There were also no policies reported that covered clinicians giving out their contact details to patients, although most trusts were developing communication systems that enabled patients to e-mail clinicians:

Patients can obtain clinicians’ e-mail addresses from e-mail signatures and there is no policy that currently states that they cannot do it.

Liver conditions site, IG officer, 02

But anybody who has got any logical thought could work out an individual’s e-mail anyway in the organisation because it’s got corporate format.

Mental health team 3 (outreach team) site, IG head, 01

Generic policies that cover digital clinical communication activity

Eight IG officers referred to generic IG policies when discussing guidelines for DCC. The most commonly quoted were data protection and confidentiality policies:

Well under the information governance framework we have to have an overarching information governance policy which stipulates accountability, communications and overarching policies.

Cystic fibrosis 1 site, IG lead, 01

We also have a code of confidentiality for IT users . . . I’m not sure if that specifically talks about contact details, but it does talk about protecting confidential information, and certainly also the e-mail one talks about sending confidential information out. So there’s bits covered in different places. As I said before, it’s not necessarily patient centric at the moment, but certainly we do have four policies in place.

Liver conditions site, IG officer, 01

Protocols for digital clinical communication operationalisation

Consent had to be sought from patients to use DCCs, irrespective of whether the patient had initiated the communication or the clinician:

Some service users may request that communication is done via e-mail. If this is the case the following must happen: health professionals involved must inform the service user that e-mail is not a secure method of transferring information; ensure the service user is happy to accept the risk involved with using e-mail as a method of communication. This must be formally recorded in the service user’s record.

Mental health 2 (CAMHS) site, IG head, 01

All interviewees mentioned that patients had to be made aware of the risks involved with using private e-mails that were not secure:

So we tell clinicians that if the patient is happy to communicate in that way and they’ve been made aware of the risks it’s fine to do.

Sickle cell disease site, IG officer, 01

When the implementation of a new service was being considered, they considered the appropriateness of the patient group and undertook a privacy impact assessment:

There would be a clinical risk assessment of the appropriateness of providing care or service in that way.

Mental health team 3 (outreach team) site, IG head, 01

We do privacy impact assessments now. You know, we’re required to, you know, it’s part of the Data Protection Act we have to do . . . actually it’s not a legal requirement, but the Information Commission recommends that we do privacy impact assessments.

IBD 1 site, IG officer, 01

In NHS trusts where DCCs had not been formally implemented, exceptions were made to the rule to facilitate it and the IG officers detailed the procedures they followed:

We have a parent who has learning difficulties and she’s requested that we only communicate with her via e-mail because she’s got some special software on her computer that allowed her to be able to understand, read and listen to . . . We’ve responded to her and we’re putting in a particular measure to be able to communicate with her. The service have done a risk assessment to be able to do that . . .

School nurse service, IG lead, 01

A key issue that arose in discussion was about making an entry in a patient’s clinical notes after a patient–clinician digital interaction and payment for the clinical encounter:

So if we conduct an outpatient appointment with somebody on the telephone we will record that, so we get paid for it. So we record that it happens, but not necessarily the content of it.

Diabetes mellitus 2 site, Caldicott Guardian, 01

Payment, that’s probably the main one in an acute hospital. And of course we log the outcome of a discussion because that’s the patient safety record of what happened. So it’s imperative that something is recorded.

Diabetes mellitus 2 site, Caldicott Guardian, 01

The IG officers talked about the difficulties of logging patient consent:

Well I think one of the issues both for e-mail and for Skype is how do we effectively record someone’s consent and particularly for e-mail where it might be used a little bit more widely.

Diabetes mellitus 2 site, Caldicott Guardian, 01

E-mails . . . would tend to be printed out and stuck in the paper record and in the new system we have I guess the same will happen, it will be scanned in: printed out and scanned in like we’ve done with the e-mails. Texts are harder because you can’t print them out the same, and you have to write them down.

Mental health 4 site, consultant, 02

Protocol for incident management

Information governance specialists reported that the infrastructure for DCCs has to be adequate and secure. They wanted clinicians to use the software and hardware provided by the trust in order to minimise the occurrence of adverse incidents. Reinforcing the idea that clinicians had to apply encryption to e-mails sent to e-mail addresses that were considered not be secure was a priority across all NHS trusts:

But as soon as you start e-mailing patients, because I can’t, if I had to e-mail you a list of patients, I would have to apply encryption to it, so I’d have to do it as an attachment encrypt it, you know.

IBD 1 site, IG officer, 01

We’ve got a means of encryption e-mail to any e-mail address as it goes out. And so we’ve been using that, and so we can encrypt whatever we want to any organisation.

Mental health 4 site, IG officer, 01

Our interviewees also talked about management of mobile devices:

So we have all our mobile devices are managed through this mobile device management solution and it tracks where the devices are, if it goes out of a certain radius within the hospital it blocks itself and encrypts, so you can’t get into it . . . it becomes a brick, and we’ve adopted that.

Cancer 2 site, IG and data protection manager, 01

Essentially to access your smartphone with e-mail you must have a passcode that locks the screen. The screen locks after so many minutes . . . we’ve got setting in our e-mail servers that force that onto the device that you’re using.

Mental health 4 site, IG officer, 01

The majority of IG officers reported dealing with DCC incidents that involved sending text messages to the wrong number or more than one recipient and mistyping an e-mail address:

We had one incident where we had text reminders going to the wrong number, because the wrong number could have been typed in, in the first place, or the patient doesn’t use that phone anymore and they’ve given it to someone else.

IBD 1 site, IG officer, 01

Certainly we do have e-mails going to the wrong address. It’s usually human error, and that would come in through the information governance team in terms of incident reporting.

Mental health 2 (CAMHS) site, IG head, 01

Another example of an incident was the storage of patient data on personal mobile phones:

The only thing that we were aware of is that sometimes staff may take photos of patients using their personal mobile phones. So we tell people not to do this because it’s against trust policy, because obviously, you know you don’t want personal information belonging to a patient to be on a personal device belonging to a member of staff.

Sickle cell disease site, IG officer, 01

Clinical concerns

Fears were expressed that clinicians would not check that their e-mail had been received and acted on, and that patients would expect responses when clinicians were not available:

With e-mail, I think the big danger with e-mail is that things get assumed. So you’ve sent an e-mail, you don’t get a response straight away, you assume that something’s happening and I think e-mail is probably the least useful tool between patient and a clinician, personally.

Arthritis site, Caldicott Guardian, 01

One of the obvious risks is if for example we give a patient an e-mail address to use and it is mine as a clinician and they e-mail me and they’re in crisis and they choose to e-mail me whilst they go into crisis and I’ve just gone on 2 weeks’ leave, who is going to pick that e-mail up?

Cancer 2 site, IG and data protection manager, 01

The IG officers pointed out the challenge they face to maintain the integrity of patient records:

There is . . . the integrity of the record . . . if information about a patient was going to be dissipated so that it’s no longer accessible in the way that we’ve worked hard to achieve, there will be a loss rather than a gain.

Mental health 4 site, IG officer, 01

I can be remote, and I can be responding to patient queries, but how that actually gets recorded into the patient record is a challenge, because we’re not working through an electronic patient record. So if the e-mail was coming in via the patient record, and you were responding, and that correspondence was all in the record, there would be an advantage to that. But there isn’t usually.

Cancer 2 site, Caldicott Guardian, 01

One interviewee was concerned about the quality of written communication:

. . . In days gone by, somebody would have written to you, you might have dictated a response which a secretary would have typed. They’d have picked up most of the typographical errors, some of the grammatical errors. And there’d have been some sort of quality assurance of the response. Even down to your secretary coming back to you and saying, did you really mean to say that, it sounds a bit, you know, rude or flippant, or whatever. Or even, perhaps the patient will be upset by that. Whereas that’s lost and you’ve got people, perhaps, entering into a dialogue where they’ve forgotten that they are having a formal communication.

Cancer 2 site, Caldicott Guardian, 01

Technological concerns

Information governance specialists were concerned that the technology could hinder effective health-care delivery:

Disadvantage of course is the internet is not, regardless of how good it is within an area, you know, is not a stable platform in the sense that you get the good strong continuous signal.

Mental health team 3 (outreach team) site, IG head, 01

You don’t have guarantee on delivery, because people can make mistakes with e-mail addresses.

Cancer 2 site, Caldicott Guardian, 01

There was some acknowledgement that clinicians required technological training in order to maximise the benefits of using these new digital communication platforms:

There’s a massive variation in IT literacy in the . . . staff base. I think we’ve probably got a few gaps in training, and I suppose, you know, the much younger generation is going to be by default more IT literate, but that’s generalising, that’s not always the case. So I think, you know, there’s going to be a large training requirement if we have to make these technologies usable so it’s going to have a design aspect as well.

Sickle cell disease site, IG officer, 01

Technology is, always we need proper training within the staff because if they are not trained properly it won’t do anything.

Liver conditions site, IG officer, 02

Reference was made to the difficulty of maintaining up-to-date digital contact details:

The predominant one for that is changing e-mail addresses. Certainly people are happy to change their e-mail addresses every time they change their internet service provider, they will have a new e-mail address because they don’t want to keep their old one. It’s similar if they don’t tell us they moved, but people are more akin to telling people, oh I’ve moved house, but e-mail addresses kind of, oh I’m with BT [British Telecom] now I used to be with Sky, or whatever it might be.

Mental health 4 site, IG officer, 01

Organisational concerns

Interviewees reported that funding allocated to new technologies was scarce and this impeded plans for technological advancement:

I think for me as well the other challenge on this is the cost. So funding in the NHS is tight. The funding goes as it should do, the priority goes to clinical care. The technology side of clinical care and especially the development side is something that can be extremely difficult to obtain the funding for.

Liver conditions site, IG officer, 01

There appeared to be a tug-of-war between IG officers and IT managers in relation to the development of digital communication systems. IG officers argued that digital innovations in the NHS needed to shift from being organisation centric to patient centric in order to provide best care:

So if, you know, I don’t go to Apple and say, look you know, we just want something for young people that Apple offers in their App Store, can we just do it with you? Why don’t they have apps that just do this, because they seem to work? Whereas when you ask our IT department to design something it suits the organisation not its population.

Mental health 4 site, IG officer, 01

One interviewee reported a lack of communication between the IT developers and IG officers and that new ideas were driven forward without seeking IG clearance:

I do think that developers of systems, you know, I think the NHS as a whole, and patients at the end of it, would benefit more and technology would move quicker in the NHS, if IG was involved at the very beginning. I think quite often IG isn’t involved quick enough, so you know, you’d generally get involved at development stage, you’ll get senior people within the NHS, Department of Health, and all these people, oh yeah, that’s good, yeah, yeah. A bit like Patient Knows Best, you know, it was . . . I can’t remember if it was Department of Health or NHS England, but, you know, really rated it, and they’d done their IG Toolkit and everything, but unfortunately it doesn’t tick all the boxes for IG; it doesn’t, you know.

IBD 1 site, IG officer, 01

Ethical concerns

Information governance was primarily concerned with patient confidentiality:

Also there’s risk of who is getting that e-mail; how is it being accessed? If it’s just on the computer and that’s their computer and they’ve got password protection on it, it’s probably quite safe to assume that only they’ve got access to it. But if it’s on your mobile phone or your tablet and you can just flick open the screen and your e-mail is presented, or if it’s an iPhone and it just pops up anyway on screen first time and you can see the first chunk of it, there are risks around that.

Mental health 4 site, IG officer, 01

And the other bit of that, is the copying in of people. So suddenly this is no longer a confidential conversation between two people, there are other people, they can be forwarded. So there’s all that sort of worry from an information governance point of view.

Cancer 2 site, IG and data protection manager, 01

Some interviewees talked about the impact of security breaches and the consequences these potentially posed on the lives of individuals:

You know, if you know that somebody’s got a long-term condition, they’ve got diabetes and they’ve got a respiratory problem, if that information’s sold to, you know, obtained and sold to insurance companies, you know, they could be exploited.

School nurse service, IG lead, 01

Human immunodeficiency virus site (does not use digital clinical communication)

The clinic-specific outcome for HIV infection is viral load. The threshold for a good/desired level of viral load is < 39 copies/ml. The viral load data provided are from 1851 young people collected between 1 November 2010 and 16 May 2016.

The profile for the percentage of young people with good viral loads is given in Figure 7a. In general, over time, the percentage of viral loads < 39 copies/ml has been increasing. For example, after January 2015, the percentage of good viral load measurements for most months is at least 80%. Thus, there seems to be improvements on viral load measurements over time.

FIGURE 7.

Profiles for percentage of viral load < 39 copies/ml for HIV-infected patients. (a) Profile of percentage of good viral load rates; and (b) DNA rate against good viral load rate for HIV-infected patients.

The results of assessments of where DNA rates influence viral load measurements are summarised in Figure 7b. The higher the DNA rate, the lower the percentage of good viral load rate, with the weighted correlation being –0.76. This indicates that, when DNA rates are low, young people’s viral loads are at desired levels.

Mental health 2 (CAMHS) site (does not use digital clinical communication)

Clinic-specific outcomes for the mental health 2 (CAMHS) site are self-harm incidents, self-harm ward attendances, Strengths and Difficulties Questionnaire (SDQ) scores and Health of the Nation Outcome Scale for Children and Adolescents (HoNOSCA) scores. SDQ scores have been recorded at enrolment only. Similarly, except for 11 young people, HoNOSCA scores available are for enrolment visits only. Therefore, SDQ scores and HoNOSCA scores cannot be used to assess trend over time.

From 2010 to 2015, the number of unique young people seen in each year was 659, 726, 633, 615, 472 and 244, respectively. These numbers are used to compute self-harm incident rates and self-harm ward attendance rates. From 2010 to 2015, the number (%) of young people with at least one self-harm incidence is 11 (1.7%), 9 (1.2%), 21 (3.3%), 65 (10.6%), 31 (6.6%) and 23 (9.4%), respectively, whereas the number (%) of young people with at least one self-harm ward attendance is 11 (1.7%), 9 (1.2%), 17 (2.7%), 39 (6.3%), 17 (3.6%) and 11 (4.5%), respectively.

Figure 8 shows results assessing if DNA rate influences self-harm rates and self-harm ward attendance rates. There are strong positive associations indicating that, when DNA rate increases, both the self-harm incidence and self-harm ward attendance rates increase.

FIGURE 8.

Did-not-attend rates against self-harm rates and self-harm ward attendance rates. (a) DNA rate against self-harm rate for the mental health 2 (CAMHS) site; and (b) DNA rate against self-harm ward attendance rate for mental health 2 (CAMHS) site.

Diabetes mellitus 1 site (uses digital clinical communication)

Clinic-specific outcomes provided for diabetes mellitus 1 are HbA_1c and glucose levels. However, there are no cut-off points to determine good or bad glucose levels and so these data were not analysed.

HbA_1c levels are monitored over time, with the clinic staff using the average of the last three measurements taken to determine how good a young person’s HbA_1c level is. Most young people have three measurements taken each year. For the analysis, to assess a young person’s HbA_1c level, the average of up to the three most recent measurements in each year was calculated. Using the clinic’s definition, the ranges < 50 mmol/mol, 50–55 mmol/mol, 56–74 mmol/mol, 75–80 mmol/mol and > 80 mmol/mol correspond to low, good, borderline, at-risk and high-risk HbA_1c levels, respectively.

Figure 9a shows the annual rates of young people with different dangerous HbA_1c levels (low, at risk and high risk). The percentage of young people with low levels of HbA_1c is small. The percentage of young people with high-risk HbA_1c levels has been decreasing. Thus, there seems to be improvement of HbA_1c levels over time.

FIGURE 9.

Profiles for different HbA_1c levels and relationship between DNA rates and HbA_1c levels for the diabetes mellitus 1 site. (a) Profiles of HbA_1c levels and (b) DNA rates against percentage of high-risk HbA_1c levels for the diabetes mellitus 1 site.

Figure 9b shows the relationship between DNA rates and the percentage of young people with high-risk HbA_1c levels. There is weak positive relationship between the two rates, with a weighted correlation of 0.28. This indicates that when DNA rates increase, HbA_1c rises to high-risk levels, but this association is weak.

Diabetes mellitus 2 (uses digital clinical communication)

The clinic-specific outcome for diabetes mellitus 2 was HbA_1c level. Data were provided from April 2013 to March 2016. For most patients, in each year, one or two HbA_1c measurements were taken.

The site did not provide us with cut-off points. The ranges defined by the diabetes mellitus 1 site were used to assess HbA_1c levels over time. Annual rates of young people with low levels of HbA_1c (< 50 mmol/mol), at-risk HbA_1c levels (75–80 mmol/mol) and with high-risk HbA_1c levels (> 80 mmol/mol) are shown in Figure 10a. The percentages of young people with low-risk and at-risk HbA_1c levels are approximately the same across the 4 years. Over time, there is a decrease in the percentage of young people with high-risk HbA_1c levels. Thus, in general, the percentage of young people with risky HbA_1c levels has been decreasing.

FIGURE 10.

Profiles for different HbA_1c levels and relationship between DNA rates and HbA_1c levels for the diabetes mellitus 2 site. (a) Profiles of percentages of HbA_1c levels in different ranges; and (b) DNA rates against percentage of high-risk HbA_1c levels for the diabetes mellitus 2 site.

Results assessing if DNA rates affect the rates of young people with high-risk HbA_1c levels are summarised in Figure 10b. There is a positive association, with the weighted correlation being 0.78. This correlation value indicates that high DNA rates lead to more patients with high-risk HbA_1c levels.

Dermatology site (uses digital clinical communication)

Two clinical outcomes were provided by the dermatology site. These outcomes were whether or not a young person had been given a wig referral and time to discharge from the clinic (when young people are well, they are discharged). The clinic had not been established for long enough to provide enough information to analyse wig referral data.

To assess the effect of time, the young people in this site were categorised by three periods (December 2011–December 2013, January 2014–June 2014 and July 2014–February 2015) that were based on when they were referred/registered to the clinic. These periods were chosen based on the volume of referrals into the clinic during these times. The number of young people referred to the clinic in these periods was 76, 169 and 84, respectively.

Figure 11a shows the estimated probabilities of not being discharged from the clinic. Figure 11b provides more details for the time frame 0–15 months. The black, blue and green lines correspond to young people referred in periods December 2011–December 2013, January 2014–June 2014 and July 2014–February 2015, respectively. The short vertical ticks on the three lines correspond to young people who have not been discharged from the clinic (censored observations), with locations (on the x-axis) of the ticks giving durations the young people have been registered in the clinic so far. In general, the earlier the period of referral, the higher the line, meaning larger probabilities of not being discharged. This finding indicates that, over time, the time to discharge has been decreasing. In contrast, DNA rates remained relatively unchanged over time at this site. The difference between profiles for DNA rates and probabilities may be explained by qualitative research that has established that young people can e-mail images to get advice on their skin condition. Owing to the form of data, it was not possible to use scatterplots and Pearson’s r correlation to assess the association between DNA rates and clinic-specific outcomes in this clinic.

FIGURE 11.

Kaplan–Meier estimate for the probability of not being discharged. (a) Plot including all time points; and (b) a slice for the time between 0 and 15 months.

Cystic fibrosis 2 site (uses digital clinical communication)

Clinic-specific outcomes provided by the cystic fibrosis 2 site are days for i.v. therapy, annual review of BMI and annual review of FEV₁% predicted.

The proportions of young people who required at least 1 day for i.v. therapy in 2014 and 2015 are 44.9% (35 in 78) and 52.6% (41 in 78), respectively. The difference in proportions is large (7.7%). However, the sample is relatively small (78 young people), which may be one explanation for the large observed difference. From 2014 to 2015, the mean BMI increased by 0.23 kg/m² only (21.55 kg/m² vs. 21.78 kg/m²), whereas mean FEV₁ decreased by 0.40 litres only (74.81 vs. 74.41).

In general, comparing 2014 and 2015, there are only small differences in clinical outcomes for young people in the cystic fibrosis 2 site. This was always true for DNA rates (see Appendix 3). As we had data for only 2 years, it was not possible to use scatterplots and Pearson’s r correlation to assess association between DNA rates and clinic-specific outcomes in this clinic.

Sexual health site (uses digital clinical communication)

The clinic-specific outcome for the sexual health site is return to the clinic after a positive test, where positive means presence of a disease. The number of positive tests from 2011 to 2016 was 2135, 1887, 1841, 1789, 2386 and 1028, respectively. For 2016, the data were up to 31 March.

Figure 12 shows the profile for the percentages of young people who did not return to the clinic after a positive result. There was a decrease from 2011 to 2012, but there were increases in 2015 and 2016. In summary, outcomes have deteriorated over time.

FIGURE 12.

Profile for rates of not returning to the clinic after a positive test.

As appointments are not booked, it was not possible to calculate DNA rates. Consequently, it was not possible to assess the link between DNA rates and clinic-specific outcomes in this clinic.

Liver conditions site (uses digital clinical communication)

The clinic-specific outcome for the liver conditions site was tacrolimus levels in young people who have had a liver transplant. There is no threshold for good tacrolimus levels. The clinical objective is that tacrolimus measurements do not vary geatly over time. The consultant informed the project team that one way of summarising a young person’s tacrolimus measurements is calculating standard deviation. However, most young people had few measurements and so standard deviations were not used to summarise data. Annual mean profiles were used.

Figure 13 shows the mean annual profiles for young people. Young people have been categorised by the number of days post liver transplant. For young people up to 1000 days post liver transplant, there is variation up to year 2008. After year 2008, although there are increases in mean tacrolimus values over time, there is little variation. There is also less variation in the mean tacrolimus values after 2008 for young people 1001–2000 days post liver transplant. For young people 2001–3000 days post liver transplant, it is not clear whether or not there is less variation after 2008 because there had been increases in the mean levels over time and then a decrease for 2011.

FIGURE 13.

Tacrolimus mean profiles for patients post liver transplant.

In summary, there is less variation in tacrolimus measurements from 2008. Noting that there was a change of DCC format, the stabilising of tacrolimus measurements from 2008 may be attributed to this change.

Owing to the form of data, it was not possible to use scatterplots and Pearson’s r correlation to assess association between DNA rates and clinic-specific outcomes in this clinic. However, it can be observed that DNA rates in years 2008 to 2011 were lower than in years 2004 to 2009 (see Appendix 3, Table 21). Therefore, it seems that lower DNA rates are associated with better outcomes.

Summary findings from analysis of clinic-specific outcomes

In the case of the two sites that do not use DCCs, over time, there was improvement in clinical outcomes in the HIV site and there was deterioration of outcomes in the mental health 2 (CAMHS) site. Among the six sites that use DCCs, there were improvements over time in clinical outcomes in four sites (diabetes mellitus 1, diabetes mellitus 2, dermatology, cystic fibrosis 2 and liver conditions) and deterioration of outcomes in the sexual health site. Thus, comparing sites that use DCCs and those that do not, over time, there are no clear differences in profiles for clinical outcomes.

The possible impact of DNA rates on clinic-specific outcomes was assessed using scatterplots and Pearson’s r correlation coefficients in two sites not using DCC [HIV and mental health 2 (CAMHS)] and two sites that use DCCs (diabetes mellitus 1 and diabetes mellitus 2). In all these sites, low DNA rates were associated with better clinic-specific outcomes. Thus, it seems to be the case that, if young people do not miss appointments, clinical outcomes improve. However, this association was weak in the diabetes mellitus 1 site. From qualitative data, it is known that patients in the diabetes mellitus 1 site send their glucose level results by e-mail. Therefore, it is possible they are getting clinical advice digitally when they miss appointments.

General findings

Based on all outcomes (DNA rates, A&E attendance rates, hospital admission rates and clinic-specific outcomes), there were no clear differences between sites using DCCs and sites that do not use DCCs. This finding may be attributable to the lack of good comparators.

Appointment status data

It could be expected that, in sites where DCCs are used, DNA rates would not increase and, in sites where they are not used, DNA rates would not decrease. This pattern was observed in all but two sites. In one site that does not use DCCs (the HIV site), DNA rates decreased continuously. In one site that uses DCCs (IBD 2), DNA rates increased over time. These unexpected trends may be a result of other changes in the sites. Overall, DCCs were linked to lower DNA rates.

Analysed data suggested that DNA rates are sensitive to changes of practice in sites. For example, in the liver conditions site, when a new digital communication format was introduced, DNA rates decreased. This indicates that the impact of digital communication can be assessed using DNA rates.

Appointments data were the most easy to obtain from sites and were well recorded, possibly because trusts record this information for the national audit. With the exception of sites that are drop-in services, if a site provided data, appointment status data were included. Appointment status data were analysed by computing annual ‘Did not attend and appointment rates’ (DNA rates). DNA rates were calculated using the formula DNA/(attended + cancelled + DNA). Having ‘cancelled’ in the denominator assumes that attended and cancelled appointments are equivalent in terms of young people receiving some form of advice. An alternative would be to have cancelled appointments in the numerator, but this assumes no advice is provided when an appointment is cancelled. Thus, appointments status data could be improved if additional information was available on whether advice was given digitally or via other means when an appointment was cancelled.

In summary, DNA rate is a good measure of the impact of DCCs. However, more information on actions taken following cancellation of appointments and possible confounders, such as other changes that may affect appointments data, need to be accounted for when assessing the impact of DCCs on DNA rates.

Accident and emergency attendance data

No differences between sites using DCCs and those not were demonstrated. However, this may be attributed to the fact that the comparison was made by contrasting profiles over time for clinics with different conditions.

It would be expected that low DNA rates would be associated with low A&E attendance rates. However, in all study sites where relationship between DNA and A&E attendance was assessed, this was not the case. For example, in the HIV site, there was a strong relationship between DNA and A&E attendance rates, with a low DNA rate associated with more A&E attendances. From qualitative data, this result could be explained by the fact that in this site, following a clinic visit, young people were booked an A&E visit to get specialised treatment. In other sites there was no substantial relationship between DNA and A&E attendance rates. Thus, no link was established between DNA and A&E attendance rates. Probably a link would be identified if additional information, such as whether or not an attendance was preceded by a DNA, were available.

Hospital admission data

There were no differences in hospital admission rates between sites using DCCs and those that did not. However, this may be attributed to the fact that the comparison was made by contrasting profiles over time for clinics with different conditions.

It would be expected that low DNA rates would be associated with low hospital admission rates in the three sites where this was assessed. This was the case in two sites (diabetes mellitus 1 and HIV) but not in the IBD 2 site, where a negative relationship was seen. A possible explanation for this surprising result in the IBD 2 site is that when young people are poorly, they do not miss clinic appointments and may also have hospital admissions. Thus, expected links were established between DNA and hospital admission rates in two sites and not in one. Like A&E attendance, expected links between DNA and hospital admission rates may be seen if additional information, such as whether or not an admission was preceded by a DNA, were available.

Clinic-specific outcomes data

Profiles of clinic-specific outcomes did not demonstrate clear differences between sites using DCCs and sites that did not. Improvement in outcomes was found in five of the six sites that use DCCs and in one of two sites that did not use DCCs. This difference may be because of the small number of sites studied, secular trends in patients’ health or differences in use of DCCs.

It is worth noting that, in one site (liver conditions), there were improvements in young people’s outcomes when a new format of digital communication was introduced. This suggests a benefit of using DCCs.

It would be expected that low DNA rates would improve clinical outcomes. This was the case in seven sites where associations between DNA rates and clinical outcomes were assessed. Thus, aiming to lower DNA rates in these sites would lead to better health outcomes for young people.

Objectives

Using published peer-reviewed research literature, this systematic review aimed to identify and examine generic PROMs that have been used to assess the impact of DCC.

Methods

Results

A total of 3802 abstracts were identified from the electronic search. The title, abstract and full-text screening process reduced the potential number of systematic reviews to 56. A summary of the literature identified at each stage of the search process is provided in a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) study flow diagram (Figure 14). Thirty-seven systematic reviews out of the possible 56 were excluded because the RCTs described in them did not meet our eligibility criteria. A total of 19 systematic reviews covering 70 RCTs were assessed for inclusion of validated PROMs.

FIGURE 14.

The PRISMA study flow diagram.

Patient-reported outcome measure selection

Thirty-five RCTs revealed 74 validated PROMs (Figure 15). The 74 validated PROMs were reviewed for relevance in accordance with our inclusion criteria of (1) no specific condition, symptoms or treatments stated in the PROM title, the PROM completion instructions or the items within the PROM; and (2) the PROM had been used in a population aged > 16 years. This resulted in 28 generic candidate PROMs (see Figure 15) for assessing the impact of DCCs across different LTC clinical populations.

FIGURE 15.

Flow diagram of candidate generic PROMs.

We present the identified PROMs which assess,first, patient-level factors (Table 10) and, second, care-level factors (Table 11).

TABLE 10 - Patient-focused validated PROMs

EQ-5D, EuroQol-5 Dimensions; HADS, Hospital Anxiety and Depression Scale; PAM, Patient Activation Measure; SR, systematic review.

Outcome measure description	Scale name	Reference
		SR	RCT
Awareness
Health literacy	Health literacy	84–86	87,88
Knowledge of management strategies	Pulmonary rehabilitation knowledge test	42,84,85	89
Cognitive impact
Cognitive appraisal	Appraisal of Diabetes Scale	90	91
Empowerment
Decision-making confidence	Sanders and Courtney Scale	92	93
Patient empowerment	PAM	85,86,94	95,96
	Patient Empowerment Scale	94	95
Patient empowerment (self-care)	Heart Failure Self-Care Behaviour Scale	84	97
Self-confidence for communicating with providers	Perceived Efficacy in Patient–Physician Interactions	98	42
Self-efficacy	Cancer Behavior Inventory	84	99
	General Self-Efficacy Scale	100	101
	Heart Failure Self-Efficacy Scale 30	102	103
	Pain Self-Efficacy Questionnaire	57,100	104
	Self-efficacy instrument	84	105
	Diabetes Empowerment Scale	42,84–86,94	87,88,106,107
Self-worth	Harter’s Self-Perception Profile for Adolescents	85,108	109
Health-related quality of life
Asthma	Modified Marks Asthma Quality of Life Questionnaire	48	110
	Paediatric Asthma Caregiver’s Quality of Life Questionnaire	48	111,112
	Asthma Quality of Life Questionnaire	42,48	113–120
	Paediatric Asthma Quality of Life Questionnaire	42,48	113,121–123
Cancer	Functional Assessment Of Cancer Therapy – Breast	124	125
Diabetes mellitus	Diabetes Distress Scale	42	126
	Problem Areas in Diabetes-2	42,84,85,90,94	91,105,107,127,128
	Diabetes Quality of Life for Youth	44,90,101,129	90,130
Health related	Fifteen-dimensional Health-Related Quality of Life (15D)	84	99
	Healthy Days Measure (Centers for Disease Control and Prevention Health-Related Quality of Life)	42	98
	Short Form questionnaire-36 items	42,84,85,100,131	88,132–135
	Short Form questionnaire-12 items	94	136
	Short Form questionnaire-6 Dimensions	42	137
	Personal Wellbeing Index	100	101
	EQ-5D	42	137
Heart failure/cardiac	Minnesota Living with Heart Failure Questionnaire	84,102	97,103
	Quality of Life Index-Cardiac Version	102	103
Children	Pediatric Quality of Life Inventory 4.0	57	138
	EQ-5D Child Version: Quality of Life	42	137
Mental health
Anxiety	State–Trait Anxiety Inventory	57,100	104,139
Depression	Beck Depression Inventory	102,140,141	142
	Center for Epidemiological Studies Depression Scale	57,84–86,90,92,100,102,131	93,104,107,135,139,143–145
	Revised Children's Anxiety and Depression Scale	57	146
	Short-care	84	105
	Patient Health Questionnaire	42,85,86,94,131	96,106,126–128,135,147
Depression, anxiety and stress	Depression Anxiety Stress Scales	100	101
	HADS	102	103
Distress	Memorial Symptom Assessment Scale	84	99
	Health distress	57,92,100,129	148
	Medical Outcomes Study Health Distress Scale	85,86	96,149
Mental health	Medical Outcomes Study Mental Health Inventory	85,108	109
	Offer Self Image Questionnaire	44,102,129	130
Physical impact
Activity participation	Australian version of the Activity Card Sort	100	101
Fatigue	Fatigue Impact Scale	100	101
Role function	Illness Intrusiveness Scale	57,92,100,129	148
Social impact
Family dynamics	Family Assessment Device	57,102,129	130
Responsibility for care	Parent–Child Responsibility Scale	130	57,102,129
Social acceptance	MacArthur Scale of Subjective Social Status	85,108	109
Social isolation	Instrumental–Expressive Social Support Scale	92	93
Social support
	Diabetes Support Scale	84,85,90	107,143
	Medical Outcomes Study Social Support Scale	42,84,85	89,145
	Duke Social Support Index	100	101

TABLE 11 - Care-focused validated PROMs

ICT, information and communication technology; SR, systematic review.

Outcome measure description	Scale name	Reference
		SR	RCT
Adherence to medical regimens
	General Adherence Scale from the Medical Outcomes Study	150	84,94,100,102,151
	Hill-Bone Compliance to High Blood Pressure Therapy Scale	87,88	84–86
	Morisky Medication Adherence Scale	150	84,94,100,102,151
Care satisfaction
Assessment of care	Patient Assessment of Chronic Illness Care	95	94
Doctor–patient communication	Art of Medicine Questionnaire	150	84,94,100,102,151
	National Committee on Quality Assurance/American Diabetes Association provider recognition programme	126	42
Patients’ perceptions of provider autonomy support	Modified Health Care Climate Questionnaire	126	42
Treatment acceptability and satisfaction	Treatment Evaluation Inventory-Short Form	146	57
	Diabetes Treatment Satisfaction Questionnaire	91,128	90,94
	Consumer Assessment of Healthcare Providers and Systems	95,128,136	94
Decision-making confidence	Sanders and Courtney Scale	93	92
Health-care utilisation
Ratings of outpatient visits	Visit satisfaction	152	57
Risk of readmission	Prediction of Rehospitalisation	147	131
ICT satisfaction
	Client Satisfaction Questionnaire	144	84,86,90,102
	Home Care Client Satisfaction Instrument adapted for ICT	154	132
	Telemedicine Patient Satisfaction Survey	91	90
	Telemedicine Perception Questionnaire	153	131
Internet use	Internet use	152	57

Summary of findings

Nineteen systematic reviews consisting of 70 RCTs were included in the review. Seventy-four validated PROMs were used in the included RCTs; most RCTs used more than one PROM at a time and some PROMs were used in multiple RCTs. Most validated PROMs were patient-focused, assessing quality of life, self-efficacy, depression or anxiety, and social impact. Some validated PROMs were also used to assess care-focused outcomes, such as adherence to medical regimens, doctor–patient communication and treatment acceptability or satisfaction. Finally, some PROMs were specifically developed for younger populations; however, the majority were aimed at adults. Thus, we are unsure whether they are suitable for use by younger people.

Methodological overview

The research approach taken combines (1) further scoping literature reviews, (2) a series of both personal and anonymous online public and patient engagement activities and (3) cognitive interviews with study participants. The reviewing (including the systematic review presented above) and the PPI activities are intertwined, and to aid understanding the methods are illustrated in Figure 16. For each activity, the method is presented first and the findings immediately below.

FIGURE 16.

Flow diagram depicting methods corresponding with research stages.

Public and patient appraisal of systematic review findings

Scoping review for systematic reviews to identify patient-reported outcomes used in intervention studies to improve patient–clinician consultations

Following discussion at the October 2014 PMG meeting we undertook a scoping review of systematic reviews to identify PROs and PROMs used to evaluate interventions aimed at ‘bettering hospital appointments’. The rationale for this was that our early fieldwork indicated that in-person appointments were improved as a consequence of using DCCs in between the appointments. We determined this ‘better’ in relation to:

• data availability for the consultation

• timing of consultation

• people involved: the right people for what the consultation needs to achieve

• the place of the consultation.

The above criteria were used to identify relevant systematic reviews from which we used the reference lists to look at the included studies for the candidate PROs. In the light of evidence gathered from our clinical case sampling interviews and our first case study fieldwork, we focused our attention on looking for PROs that might indicate whether or not patients were having better hospital appointments as a consequence of the additional use of DCCs. Here we focus on the quality of the communication rather than the mode of communication. We scoped the literature to find systematic reviews with particular focus on interventions that aim to increase patient participation in consultations. 179,180 Patient activation PROMs had a good fit with the consultation dynamics, the management of LTCs and health policy directions around supporting self-management. The patient behaviours that these types of interventions target were:

• increasing the number of questions patients ask in consultations

• patients raising their own concerns

• patients asking for clarification

• patients checking their understanding.

Increases in these patient behaviours have been measured180 by:

• immediate impacts (e.g. patient satisfaction and recall of information)

• intermediate outcomes (e.g. agreeing with, and adherence to, recommendations)

• long-term outcomes (could be changes in health status or lifestyle).

Harrington et al. 179 found that when interventions were used to increase patient activation and participation in consultations (none of the interventions was DCCs), the following range of immediate, intermediate and long-term outcomes were observed:

• improvement in patient recall (immediate outcome)

• increases in question asking (immediate outcome)

• no change in consultation length (immediate outcome)

• no change in patient treatment satisfaction (immediate outcome)

• more participatory physician behaviour (immediate outcome)

• reduced anxiety (intermediate outcome)

• increased patient perception of control (intermediate outcome)

• improved appointment attendance (long-term outcome)

• increased medication adherence (long-term outcome)

• increased perception of disease control (long-term outcome).

These outcomes were used to inform online PPI survey B.

Public and patient appraisal of patient-reported outcomes used to assess the impact of the patient–clinician consultation intervention (patient and public involvement survey B)

The social media methods developed in PPI survey A were repeated in relation to the findings from the scoping review of PROs. Online survey B was developed in consultation with LYNC study PPI members and was launched via SurveyMonkey and promoted by the LYNC study Facebook page and Twitter between December 2014 and February 2015. Our PRO topic question to potential PPI respondents in Table 13 asked them to prioritise outcomes identified in the Harrington et al. review. 179

Rapid scoping review to identify candidate patient-reported outcome measures to assess the patient-reported outcomes and final patient and public involvement assessment

Building on the four PPI survey B priorities we went back to the Harrington et al. review179 to identify the included studies that had assessed priorities a–d below.

1. Question asking: you feel confident in asking more questions during your appointments (see Table 3 in Harrington et al. 179).

2. Doctor encourages patient engagement: your clinician feels more engaged with you (see Table 3 in Harrington et al. 179).

3. Recall: you remember more of the discussion when you leave the appointment (see Table 4 in Harrington et al. 179).

4. Disease control: you feel more in control of your condition (see table 4 in Harrington et al. 179).

In response to the LYNC study PPI feedback regarding the datedness of the survey A topics and associated PROMs, our eligibility criteria applied an earliest publication date limit of 1995 for the included studies, given that the PROM will have been published prior to the publication of the study which used it. Three studies were potentially useful181–183 but on assessment none of them used a patient outcome measure to assess priorities a–d. A further search of the literature was subsequently undertaken to identify systematic reviews assessing clinician–patient communication interventions and patient activation interventions. The purpose of this activity was to identify PROMs used to assess PROs of interest (priorities a–d) associated with interventions to increase patient engagement with consultations. Using key words, Google Scholar (Google Inc., Mountain View, CA, USA) identified three further reviews (Table 14).

Rao et al. 184 revealed one candidate PROM and Griffin et al. 192 revealed two candidate PROMs. The Krantz Health Opinion Survey,190 used by Joos et al. ,189 reported two subscales – desire for information (seven items scored 1–7) and behavioural involvement (nine items scored 1–9) – that were a close match to the PPI survey B priorities. Beck et al. 191 reviewed 22 studies about physician verbal and non-verbal communications. They looked at patient outcomes associated with the presence and absence of these verbal and non-verbal primary care communications. Analysis of the full text of the studies in Beck et al. 191 did not reveal PROMs addressing PRO priorities of our PPI work. Griffin et al. 192 identified seven studies to potentially reveal a candidate PROM. Thom et al. 196 used two potentially relevant scales: the Physicians’ Humanistic Behaviours Questionnaire (PHBQ) and their own Trust in Physicians Scale. 199,200

A scoping search of Google Scholar using patient + activation + outcome identified the Patient Activation Scale and the Patient Activation Short-Form Scale. 178,201 A search for patient activation systematic reviews identified the Griffin et al. review. 192 We consequently added the two PAMs and, therefore, had five candidate scales (Table 15) that assess aspects of our priorities: (a) patient question asking; (b) clinician engagement with patient; (c) post-consultation recall; and (d) patients feeling more in control of their own condition. Although these PROMs had not been linked in experimental studies to DCCs but instead to clinician–patient communications, they were nonetheless all themes that were emerging from our LYNC study fieldwork. Consequently, we appraised them as having topic validity to underpin further consultation with the LYNC study PPI members.

The PMG PPI members were consulted to assess the face validity of the PROMs for evaluating the effects of DCCs. They were asked for their fit-for-purpose votes as follows: (1) fit for purpose, (2) may be fit for purpose or (3) not fit for purpose. The full PAM201 and the PHBQ199 emerged as having the greatest face validity for assessing the impact of DCC-enhanced care (see Table 15).

Cognitive interviews with Long-term conditions Young people Networked Communication study case site clinicians and young people

Published psychometric evaluations of the patient-reported outcome measures

Principal findings

This represents a large body of activity consisting of literature reviewing, PPI work and empirical research aimed at addressing our research questions. Research question 1 sought to identify existing measures that had been used to assess digitally enhanced care outcomes. We identified 28 such PROMs which our PPI activity consultation outcomes identified to be lacking in relevance and credibility. We answer question 1 by stating that the published evidence relating to previous evaluations of digitally enhanced services did not reveal a suitable outcome measure.

Research question 2 sought to look prospectively at future evaluations of digitally enhanced care using PPI-informed scoping reviews based on the emerging strengths of such care provision from the LYNC study fieldwork. This process resulted in the identification of two PROMs with potential clinical and research utility. Both measures were satisfactorily appraised by clinicians and patients as having sufficient degrees of face and content validity to proceed with further experimental use in a digitally enhanced service. Psychometrically, the PAM has stronger and more detailed validity in demonstrating clear face, content, construct and criterion validity and psychometric reliability. The PHBQ has less documented detail about the face and content validity. However, the cognitive interviews suggest that it has sufficient face and content validity to proceed to reliability testing in a DCC evaluation. Both scales have demonstrated themselves to be credible clinical evaluation measures. Subsequent research studies need to expose them to real clinical evaluations of digitally enhanced services to examine how they perform psychometrically.

Strengths and limitations: patient and public involvement activity and literature reviews

Particular strengths lie in the PPI activities, which have involved > 150 people with experience of living with a LTC and using specialist health services. Approximately 30% of these individuals have direct experience of receiving a digitally enhanced service, with the remainder reporting no experience of this. This offers a good spread of considerations of the PROM topics from across the NHS service user community. Systematic reviews very often do not incorporate PPI activity and this was the case in our initial systematic review. Interestingly, it was only when we began to incorporate PPI perspectives into our literature reviewing that we began to make some progress in answering our research questions. Our first evidence review was systematic and based on our previously published reviews44,45,212 and we are confident that we captured the available evidence. Our scoping reviews, intertwined with our PPI activities, were much less so and were more PPI question led. It is likely that we missed published PROMs that might be of future relevance for evaluating digitally enhanced services.

Strengths and limitations: cognitive interviews

The cognitive interviews provided in-depth analyses of two PROMs with both clinicians and patients. We recruited a broad range of health professional disciplines to participate and four of the six health professionals evaluated both measures and in this respect showed commitment to the endeavour. These health professionals were experienced users of digitally enhanced care and so were evaluating the measures’ real clinical relevance for use in a service evaluation or audit context. In relation to the young people, they were all receivers of digitally enhanced care and all patients evaluated some aspects of both measures.

There are a number of limitations, most important of which is that most young people and some clinicians did not understand the exercise of doing cognitive interviews. The interviewer had to work with the time and level of understanding that the participant presented. In this respect, it could be argued that some interviews were not sufficiently accessing cognitions on the topic, particularly those of young people; when they did, it was following considerable prompting from the researcher in an attempt to facilitate the young person’s understanding of the interview purpose. Only two young people answered each item on both measures as a result of an alternating strategy implemented to maximise response rates and the young people were recruited from just two clinical areas which might affect their views and experiences. With regard to health professional interviews, one participant was unable to provide cognitions for their PROM use in a digitally enhanced service. In addition, they sometimes answered the questions in the third person, on behalf of the patient and, consequently, did not truly represent the level of assessment that patients would make about the two instruments.

In the cognitive interviews, the PAM and the PHBQ have demonstrated both strengths and weaknesses. The weaknesses lie in the scale in general, and some items in particular, and not in their usefulness in assessing digitally enhanced care. The clinicians interviewed were looking at the PAM as a clinical tool and they had not been asked to do otherwise; this may have contributed to some negative appraisals regarding item or topic repetition within the scale. Asking the same question in different ways is often a strength in a research tool to enable the research team to tease out response congruity. In a clinical setting, where time for both clinicians and patients is at a premium, repetition may appear wasteful and irritating. The PAM is one of the few PROMs that serves twin purposes of performing well as both a clinical and research tool (URL: www.england.nhs.uk/2016/07/person-centred-care/; accessed 5 March 2017). The shorter 13-item PAM, which was slightly less well appraised by our LYNC study PPI team, has considerably more citations and psychometric validity studies, and its brevity as both a clinical and research tool may be the reason for the greater level of interest in it than in the full 22-item scale.

In concluding this chapter, it is important to remind ourselves of the increasing levels of interest across health economies in PROs. The US Food and Drug Administration213 has produced guidance on their use by the pharmaceutical industry in recognition that the evolution of science and health-care products and practices should not come at any price. Although this advice is specifically intended for the pharmaceutical industry, Speight and Barendse214 remind us that the principles underpinning the US Food and Drug Administration’s guidance applies across other contexts. Health care is awash with technological solutions and in the area of DCCs and the context of our LYNC study focus, NHSmail 2 is currently undergoing roll-out across the NHS with the specific objective of promoting secure DCCs. With such fundamental shifts in the way that health care will be delivered and received on the horizon, it is imperative that we ensure that we have clinically and scientifically fit-for-purpose PROMs available that have credibility with patients, clinicians, researchers and technological innovators.

Background

The use of digital communication technologies has become ubiquitous in our societies over the last two decades. These technologies encompass a wide range of tools such as social networking sites [(e.g. Facebook, Myspace (Viant Technology, Beverly Hills, CA, USA)], mobile phones and smartphones, VoIP system (e.g. Skype), forums, e-mail, or short message services (SMSs).

Digital communication technologies have also been extensively implemented into health care because of their capacity to modernise and facilitate communication between health-care professionals and patients. As young people are prolific users of new technologies, a large number of studies have assessed the feasibility of DCC between young people with LTCs and health professionals as a way to meet the special needs of this population. Indeed, the management of young people with LTCs is challenging, as they tend to disengage from health services and experience poorer health outcomes when compared with the adult population with the same conditions. 217 Saywer et al. 218 have emphasised that this population of patients have distinct health needs. This is of particular importance during the transition from paediatric to adult health services. 219 Although promising and increasingly used, the effectiveness of digital communication technologies on young people’s outcomes remain unclear. 44 Some studies are currently ongoing to review the evidence of the effectiveness of mobile phone and tablet applications in this subset of patients. 220

For young people with health conditions, parental involvement and parent–health professional relationships are an important part of the treatment journey. At present, it is unclear what impact digital communication has on these relationships. Some authors have emphasised that parents play an important part in promoting the development of young people’s self-management skills in LTC management, but they may be less confident than young people in using technology. 221,222

There is an emerging body of literature on the use of digital communication for families and carers (e.g. childminders, home care workers) within the scope of management of children and young people with short-term conditions or LTCs (e.g. web-based consultations for parents of children with atopic dermatitis223 and the use of SMSs as a tool to reduce the social exclusion experienced by caregivers of people with disabilities224).

In order to bring more clarity to this area, this review aims to evaluate evidence on the impact of the use of DCCs between families/carers and health professionals.

Aim and objectives

The aim of this review was to evaluate evidence on the impact of the use of DCCs for the families and carers of young persons with LTCs.

Objectives were to:

• describe existing DCC use in families and carers of young persons with LTCs

• evaluate the feasibility of using DCCs with the families and carers of young persons with LTCs

• evaluate the impact of using DCCs with the families and carers of young persons with LTCs.

Methods

Results

Search

The search identified 1156 potentially relevant records. We removed 956 records that did not meet our inclusion criteria at title/abstract stage, leaving 200 articles to be examined at full text. Among these, 168 were excluded leading to a total of 32 publications meeting our inclusion criteria. The flow diagram describing the process of identifying relevant literature can be found in Figure 17, including the reasons for exclusion of papers at full-text level.

FIGURE 17.

Study PRISMA flow diagram.

Strengths and weaknesses

All the included studies presented results suggesting the feasibility of using DCCs, although several authors recognised these results to be preliminary and that they necessitated further confirmatory studies. Many studies within the review examined interventions composed of more than one component, making it difficult to determine whether the results were attributable to the use of DCCs or other aspects of the intervention. As we conducted a rapid review and not a systematic review, we were not able to undertake a grey literature search aimed to identify unpublished studies that could have failed to demonstrate the feasibility of such communication means. Consequently, we cannot exclude the absence of publication bias in this area. As the majority of included studies were conducted in the USA and overall came from only nine different countries, there is limited generalisability of the findings beyond these settings.

Comparison with other reviews

We are aware of a recently published systematic review that assessed telehealth tools and interventions to support family caregivers of paediatric, adult and older patients with chronic diseases. 257 This review, which included 64 articles published over the period 1997–2014, concluded that telehealth can positively affect care in family caregivers. Our work differs from the previous study as we were interested only in the impact of using DCCs for families or carers of young people with conditions and we excluded adults and older patients.

Background

Diabetes mellitus is a common metabolic disorder. It affects > 3.2 million people in the UK, of whom approximately 35,000 are young people aged < 19 years. 258 NICE has recently lowered target HbA_1C levels to 48 mmol/mol for children and young people in order to minimise risks of long-term complications. 259 However, in 2013–14, only 18.4% of children and young people with diabetes mellitus achieved the previous target for glycaemic control of < 58 mmol/mol, and 23.9% of children and young people have very poor glycaemic control (HbA_1c levels of > 80 mmol/mol), predisposing them to dangerously high risks of health complications, such as diabetic ketoacidosis (DKA). 260 Research has indicated that clinical outcomes are worse in young diabetes mellitus patients than in older patients. 260,261 This has led to increased focus for the NHS as poor glycaemic control has been linked to increased risks of both microvascular and macrovascular complications. 260 Disorganised transition of patients from child to adult services also has a negative impact on clinical outcomes. The NHS identified the phenomenon and subsequently published a guide to facilitate a smoother transition from paediatric to adult services. 262

The use of communication technologies, such as mobile phones/smartphones, e-mail, SMSs/text messaging, VoIP (e.g. Skype) and social networks (e.g. Facebook and Twitter), has exploded in recent years. Now 90% of people in the UK own a mobile phone. 263 Among young adults aged 16–24 years, smartphone internet use has doubled over the past 5 years to 87%, and this age group is the highest subscriber (91%) to social networks. 264 Concurrently, digital communication systems are being increasingly developed specifically for health-care purposes. A Cochrane review of published evidence demonstrated the feasibility, and acceptability to patients, of establishing technologies in health care. 265 Evidence suggests that DCCs could help overcome some barriers to health-care access for marginalised groups. 266 These barriers include practical difficulties, such as travelling to a health-care centre and waiting for an appointment, and interpersonal difficulties, such as stigmatising and negative reactions to patients and lack of continuity of care. A review of the use of digital communication interventions between parents/carers of children with a LTC and health-care providers found that the technologies addressed issues of continuity of care, time constraints and geographical barriers. 55 The interventions commonly focused on medication management, education and physiological monitoring.

Several reviews have focused on the use of digital communication technologies in the management of diabetes mellitus in both children and adults. Telehealth is known to improve glycaemic control in adults,267 children and adolescents. 268 This can be facilitated through remote monitoring and treatment advice269 and diabetes mellitus self-management education. 270 Researchers have recently started to explore the use of more novel technologies in the care of diabetes mellitus patients. Text message (SMS)-based interventions have been used in the care of children and adolescents with diabetes mellitus to monitor blood glucose levels remotely and to send reminders and advice about medication and lifestyle choices. A review of such studies found that the impact of such interventions on glycaemic control and self-management was unclear. 271 Furthermore, although patients reported that using text messages was feasible and enjoyable, some did experience technical difficulties, resulting in data loss and affecting patient motivation. Applications on smartphones have been explored with adults with diabetes mellitus. Applications providing advice and assistance on how to manage diabetes mellitus help to improve glycaemic control. 272 There is, however, a lack of evidence for the impact of smartphones and applications on the glycaemic control of young adults and children with diabetes mellitus. 273 Studies exploring the use of social networking services have not yet studied well-known sites such as Facebook or Twitter, focusing instead on sites purposely designed for research. 274 Such interventions have provided secure messaging between patient and clinician (via e-mail or online chat rooms); self-management education; and a way of sharing medical records and blood glucose readings. A review of these studies reported that such interventions have a positive impact on glycaemic control and are a novel way of improving diabetes mellitus care. 274 Overall, digital communication technologies have the potential to positively affect the health of people with diabetes mellitus; however, the technology used must be appropriate to the patient group. 270

To date, only one review has focused on interventions that included two-way communications between clinicians and young adults with diabetes mellitus. 44 This review found that, although contact between clinician and patient increased, there was uncertainty about whether or not this led to improved clinical outcomes. As part of the wider LYNC study, this review will update this previous work.

Aim

The specific aim of this review is to present evidence for the effectiveness of digital communication technologies on clinical outcomes in young people with diabetes mellitus.

Methods

Results

Search

Initial searching yielded 775 unique records, of which 539 were removed based on their title and abstract. The full text of 236 was read and 223 were excluded: 120 did not meet the inclusion criteria, 68 were not full-text papers, 32 were not empirical papers and three were unobtainable in the time frame of the rapid review. To check reviewer agreement, 10% of papers were screened by two reviewers. Agreement was 94%. A total of 13 papers was included in the review (Figure 18).

FIGURE 18.

Study flow diagram.

Discussion

This review has explored the impact of DCCs on clinical outcomes in young people with diabetes mellitus. There was no clear indication as to how digital communication technologies affected glycaemic control; some studies reported decreased HbA_1c levels in the intervention groups,275,278,283,285 whereas others reported no significant impact of DCCs on HbA_1c levels. 250,276,279,281,282,284,286 Similarly, findings relating to psychological outcomes were mixed. There was a trend for reporting other benefits of the interventions, such as increased glucose testing277 and decreased admissions for DKA,279 but these were isolated findings not supported by other studies. These findings update a 2011 review exploring the use of communication technologies with young people with diabetes mellitus44 and support the authors’ original findings. Overall, there is no clear picture as to how DCCs affect clinical outcomes in young people with diabetes mellitus.

Potential benefits of DCC interventions for health-care professionals and health service delivery were described; however, these were not supported by data. 275 Advantages to the intervention described focused on the transmission of health records rather than two-way communication between the health professional and patient.

Conclusion

This review analyses recent evidence for the impact of digital communication technologies on clinical outcomes in young people with diabetes mellitus. From the studies analysed, there is no clear benefit to the use of digital communication technology. However, the quality of the research is rated as moderate, highlighting the need for further research in this area.

Background

Worldwide, mental illness is a significant source of disease burden. 289 Its impacts are far-reaching and it negatively affects physical health, educational attainment, social capital and emotional well-being. 287,288,290 In young people, this is particularly significant. It is estimated that the economic cost amounts to £105B a year in the UK alone. 291

Most mental health problems begin in childhood or young adulthood,292 and their prevalence increases with age, affecting 7.7% of young people aged 5–10 years, 11.5% of young people aged 11–16 years 293 and around 23% of young people aged 18–20 years. 294 The prevalence of conditions such as depression increased among young people over the 20 years 1986–2006. 295 Early intervention and good care have been shown to be cost-effective and improve both short- and long-term outcomes. 296 Early intervention in psychosis, for example, has a reasonable evidence base297 and has been widely implemented in the UK.

There are a number of challenges associated with providing health care for young people, including barriers to access, behavioural issues such as risk-taking behaviour and problems associated with transitioning from paediatric to adult services. 298–300 This is particularly relevant in mental health, where there are significant differences between CAMHS and Adult Mental Health Services, and there is concern that young people may get lost as a result of the gaps in service provision. 301–303

The previous review, which examined diabetes mellitus, outlined the dramatic increase in utilisation of the communication technologies in recent years by young adults and the increasing development of digital communication systems for health-care purposes.

To date, systematic reviews focusing on two-way communication between clinician and young adults with mental health conditions are scarce. A previous review found that networked communication technologies can increase the opportunity for communication between patient and health-care professionals, but that limited improvements in quality of life and continuity of care were reported. 45 As part of the wider LYNC study, this review will update this previous work.

Aim

The aim of this review is to present evidence for the effectiveness of digital communication technologies on clinical and process outcomes.

Methods

Results

Search

The search resulted in 3503 papers. A total of 3281 were excluded based on title or abstract. Two hundred and twenty-two papers were read in full and a total of six papers met our inclusion criteria and form part of the review as presented in Figure 19.

FIGURE 19.

Study PRISMA flow diagram. CCBT, computerised cognitive–behavioural therapy.

Discussion

Based on this systematic review, it seems likely that networked communication technologies for the treatment of mental illness are generally acceptable to young people. It seems likely that interventions provided using these technologies may be effective, and the studies published between 2009 and 2016 are generally positive in their findings. Nevertheless, there is limited evidence available as far as young people are concerned and the effectiveness of most interventions for the majority of mental health conditions in this group has not yet been adequately investigated. Although we found a large number of studies resulting from our search, many of the articles focused on computerised CBT, perhaps indicating the current relative interest in this technology. Computerised CBT is an interactive programme of CBT delivered via computer that may or may not include therapist feedback. As it was not possible to extract the contribution of the therapist in relation to the computer program from these studies, these articles were excluded from this review.

Conclusion

Networked communication technologies show promise in the treatment of mental health conditions, but further research is needed into a wider range of conditions and methods of treatment.

Background

Adolescence is not normally characterised by ill health. The prevalence of chronic disease in this population is low, with the leading cause of death in middle- and high-income countries being road traffic accidents. 314 Nevertheless, one in five young people aged < 16 years must regularly take medications. 315 Many conditions for which medication is taken do not resolve quickly (e.g. some mental health conditions) or are permanent (e.g. diabetes mellitus); therefore, numerous young people will continue to take medication beyond the age of 16 years. The number of young people with LTCs has increased in recent years, mainly because of improvements in survival for previously fatal childhood conditions, such as cystic fibrosis314 and childhood cancers,316 and improved diagnosis of conditions, such as asthma. 317 At the same time, as children grow up, transition from paediatric to adult health services is a vulnerable time and can result in disengagement. 318

In this review we aim to evaluate new approaches to delivery of health care mediated by new communication technologies and the value of such technologies in aiding effective communication with health professionals and self-management of health care. The normal concerns about image and identity in the context of social development among the peer group in this age group319 and the transition to adult health services is an additional challenge. 318 It is hypothesised that the use of digital communication with health-care staff may help with the disengagement from health care often experienced by young patients at this time.

Aim

The specific aim of this review is to present evidence for the effectiveness of digital communication technologies in young people with long-term physical conditions.

Methods

Inclusion and exclusion criteria

Study type/language of publication

All types of study designs published in all languages, from 2009 onwards, were included.

Participants

Young people/adolescents (aged between 12 and 25 years). Studies where the mean/median age of the population was > 12 and < 25 years were included. Clinicians included all health-care professionals.

Condition/disease

Asthma, HIV infection, cancer (all types), cerebral palsy, Duchenne muscular dystrophy, physical disability, cystic fibrosis, juvenile rheumatoid arthritis, IBD, traumatic brain injury, chronic kidney disease, rheumatological condition, haemoglobinopathy, liver disease, dermatological condition, gynaecological condition, spina bifida, epilepsy and myalgic encephalopathy.

Intervention

All identified digital communication technologies involving two-way communication between patient and clinician were included. Examples include e-mails, SMSs, social networking websites, mobile telephony, videoconferencing and teleconferencing. When the intervention involved several components, such as health-care information transfer and access to websites for information, studies were included only if the two-way digital communication between patient and clinician was evaluated separately.

Comparator

We compared the effectiveness of communication technology with traditional in-person intervention and, where relevant, telephone intervention.

Outcomes

All outcomes were included. We reported the outcomes according to the following groupings:

• patient outcomes (e.g. treatment outcomes, health behaviours, health status and well-being, communication)

• health professional outcomes (e.g. knowledge and understanding, consultation processes)

• health service delivery outcomes (e.g. adverse events, health economic outcomes).

Screening

One reviewer screened the identified papers at title and abstract stage, then two independent reviewers screened the full texts. Any discrepancies that arose between the reviewers were resolved by a third reviewer. The study flow and reasons for exclusion of full-text papers were documented in a study flow diagram (Figure 20).

FIGURE 20.

Study PRISMA flow diagram.

Results

Discussion

In the limited number of studies identified, the use of DCCs was an acceptable method of communication with a health-care professional. In the one study in which clinical outcomes were assessed using clinical tools, teleconsultation provided clinical equivalence, but fewer patients perceived a benefit than in the control group. This highlights a possible gap between what is a clinically relevant benefit and what patients feel is a benefit. We did not identify any studies that reported health-care professional outcomes. The included studies had small sample sizes, which limits the reliability and generalisability of the findings.

We excluded studies in which the effect of the DCCs was not separately assessed, for example where DCCs were used to supplement a web-based intervention. It is possible that DCCs play an important role in the effectiveness of these interventions, but it is not possible to discern this in a multifaceted intervention.

Conclusion

Digital clinical communication appears to be an acceptable method of communication for young people with LTCs. Conclusions about clinical outcomes cannot be drawn given the limited evidence base to date.

Background

Internationally, there is interest in using digital communication to improve access to health care, particularly in countries aiming to achieve universal health coverage. 322 Universal health coverage means that all people can use the promotive, preventative, curative, rehabilitative and palliative health services they need, of sufficient quality to be effective, while also ensuring that the use of these services does not expose the user to financial hardship. 323 Some countries, such as Brazil, are opting to use open-source software for digital health systems324 and Latin American countries have recognised the growing importance of DCCs by offering training in teleheath. 325 Although South American countries have universal health-care systems, they do not administer them universally and so are not classified as offering this coverage, being instead regarded as moving towards doing so. 326 In conducting this review we wanted to identify studies based in countries that were not already represented in our other reviews and might offer evidence over and above that already identified.

Aim

To identify the use of DCCs for young people in countries with or moving towards universal health coverage.

Methods

Inclusion and exclusion criteria

Study type/language of publication

All types of study designs published in all languages, from 2009 onwards, were included.

Setting

A study set in a country with or moving towards universal health coverage. Countries were identified according to the World Health Organization’s Global Health Observatory data. 329

Participants

Young people/adolescents (aged between 12 and 25 years). Mean/median age of the sample population of < 25 years or > 12 years were also included. Clinicians included all health-care professionals.

Condition/disease

We excluded mental health conditions, diabetes mellitus and long-term physical health conditions [asthma, HIV infection, cancer (all types), cerebral palsy, Duchenne muscular dystrophy, physical disability, cystic fibrosis, juvenile rheumatoid arthritis, IBD, traumatic brain injury, chronic kidney disease, rheumatological conditions, haemoglobinopathy, liver disease, dermatological conditions, gynaecological conditions, spina bifida, epilepsy and myalgic encephalopathy], as these were included in the other reviews.

Articles relating to health promotion or disease prevention were excluded. We excluded studies that focused on neurodevelopmental disorders, such as attention deficit–hyperactivity disorder and autism, and non-chronic conditions such as acute stress reaction.

Intervention

All identified digital communication technologies involving two-way communication between patient and clinician were included. Examples include e-mails, SMSs, social networking websites, mobile telephony, videoconferencing and teleconferencing. We excluded computerised CBT. Interventions allowing only one-way communication were excluded. Multifaceted interventions were also excluded.

Comparator

We were interested in comparisons with traditional in-person interventions and, when relevant, telephone interventions.

Outcomes

All outcomes were to be included. We intended to report outcomes according to the following groupings:

• patient outcomes (e.g. treatment outcomes, health behaviours, health status and well-being, communication)

• health professional outcomes (e.g. knowledge and understanding, consultation processes)

• health service delivery outcomes (e.g. adverse events, health economic outcomes).

Screening

One reviewer screened the identified papers at title and abstract stage and the same reviewer read the full text. A second reviewer analysed 10% of the articles from each of the exclusion and inclusion lists. The study flow and reasons for exclusion of full-text papers were documented in a study flow diagram (Figure 21).

FIGURE 21.

Study PRISMA flow diagram. ISMEAR, Index Medicus for the South East Asia Region; LILACS, Latin American and Caribbean Health Sciences Literature.

Results

After reviewing abstracts and titles against inclusion criteria, as described in Methods, 297 titles remained. After screening 297 full-text articles, none met the inclusion criteria. Six were already identified in the other reviews.

Discussion

The review failed to identify any studies meeting the inclusion criteria, even when searching for articles published in languages other than English. This is due in part to having already identified relevant published literature in the other reviews. Rapid reviews do not search the grey literature; it may be that information and evidence about DCCs used in countries such as Brazil are published outside of the peer-reviewed literature (e.g. in policy documents and news sources). During the screening of the abstracts and titles, it was noticeable that a large number of studies did not meet the criterion of the mean age being ≤ 25 years. This may have led us to exclude articles that could provide us with knowledge useful for implementation of DCCs, based only on age. This is a disadvantage of the rapid review method, which utilises a focused question and, in this case, linked closely with the overall research question for the LYNC study. The setting of interest is likely to yield rich information on the use of DCCs, but at present this is not available in the published literature.

Aim

This review aims to give an up-to-date overview of the evidence concerning the use of videoconferencing between the clinician and patient from the patient’s own home (or mobile device), by summarising the existing reviews of the literature.

Search strategy

A set of searches was developed for this rapid scoping review aiming to capture reviews on the use of this specific technology for clinical communication, irrespective of the age of the study participants (see Appendix 5). First, the EndNote database of results from the review of systematic reviews undertaken to identify generic outcome measures populated in 2014 was searched using the terms ‘skype’, ‘videoconferenc*’, ‘Video-conferenc*’ and ‘Google AND (talk or hangouts)’ in any field and sorted by publication date to select only those records published from 2009 onwards (see Appendix 5). Second, an update search for systematic reviews of this specific technology was conducted in MEDLINE in January 2016 (see Appendix 5). Third, a search of MEDLINE combining thesaurus and free-text terms for the concepts of ‘internet videoconferencing technology’ and ‘clinical communication’ was conducted.

Inclusion and exclusion criteria

Screening

One reviewer screened the identified papers at the title and abstract stage and two reviewers read and screened the full texts. Any issues were resolved by a third reviewer. The study flow is documented in Figure 22.

FIGURE 22.

Study PRISMA flow diagram.

Data extraction strategy

One reviewer (CB) extracted relevant data from papers using a standardised data collection form.

Search

A total of 254 relevant studies were identified in the MEDLINE search, which included several systematic reviews. After screening, in total, 123 abstracts were selected for full-text review and 24 review articles were included in the review.

Characteristics of studies

A total of 24 reviews were included. The articles covered a wide range of LTCs, including heart failure, depression, schizophrenia, spinal cord injury and chronic pain. The majority of the articles were systematic reviews. Only seven reviews looked at videoconferencing exclusively, with the remainder looking at a range of telehealth interventions (including videoconferencing) (Table 20).

Quality assessment

Table 32 in Appendix 6 provides a summary of the AMSTAR results. Only three reviews were methodologically strong,342,353,362 with the remaining 21 obtaining a score of ≤ 6, thus being rated as being of poor quality.

Mars et al. 351 and Shore358 provide limited details of included studies’ characteristics, such as clinical outcomes and participants’ demographics. Although McGeary et al. 352 searched for grey literature, they excluded it later without giving any reasons. This is significant as the authors note that unpublished material could have provided additional insights. 352 Although Peeters et al. 354 supply funding details, they do not specifically report any conflict of interest. For Conway et al. ’s342 subanalysis, full search terms, the inclusion of unpublished literature and list of excluded studies, and in-depth risk and publication assessments were appraised in the original systematic review and meta-analysis publication. 343

Patient outcomes

Almost all the reviews found evidence of increased patient satisfaction, and some found improvement in at least some health outcomes. A review related to mental health indicates that a good therapeutic alliance between clinician and patient is possible via video. 360 In a review of the use of telemedicine in diabetes mellitus, 23 of the 27 RCTs reported improved metabolic control. Twelve of these 23 studies produced significant results, whereas only two studies observed negative health outcomes. 359 The review of telepsychiatry analysed results from a total of 1054 patients with various mental illnesses. The 10 RCTs reviewed mostly included patients from psychiatric services. The conclusion of the review was that telepsychiatry is safe to use and may be a useful alternative when a face-to-face encounter is not possible, but there is not yet sufficient evidence regarding its effectiveness in routine management of mental illness. 347 In our own review of the use of digital communication between clinicians and young people requiring mental health care, 5 of the 12 studies concerned videoconferencing, but significant improvement in health outcome was seen only with e-mail contact and web-based technologies. 45 Although the review of videoconferencing in oncology included 19 studies, the authors found no conclusive evidence of a difference between video consultation and face-to-face consultation. 350 A review of telemedicine for asthma concluded that there was no improvement in asthma quality of life, but there was a reduction in hospital admissions. 353 There is also some evidence of increased satisfaction among individuals from minority ethnic communities. To illustrate, the review of telecounselling for depression pooled results from 498 adults of African-American, Spanish and Asian origin. 345 Three independent studies found significant short-term improvements with internet-mediated consultations. Limited data also pointed towards longer-term health benefits. 345 Some reviews found differences in outcome depending on the communication medium. For example, a review of telemedicine for heart failure concluded that videophone did not improve outcome but structured telephone follow-up and telemonitoring did improve health outcome, including all-cause mortality. 342 Furthermore, a review of information communication technologies for depression found that videoconferencing is as beneficial as in-person care. 348 The authors also concluded that, as self-help internet programs have shown to improve symptoms and quality of life, they could be utilised in situations when conventional health care is unavailable.

Health professional outcomes

Some of the reviews found that clinician satisfaction depended on the particular type of communication medium. A review of forensic telepsychiatry in mental illness351 reported that health-care staff were less satisfied with using videoconferencing than the prisoners they were treating. Martin et al., however, found that clinicians prefer e-mail and web-based approaches. 45 Despite the equivocal evidence for telemedicine, Shore358 argued that health-care professionals can feasibly implement this modality. Nevertheless, both reviews recommend that staff reflect on their communication styles in order to ensure they interact naturally with patients as they would during any face-to-face consultation. Shore358 has also maintained that psychiatrists should learn how to deal with the various clinical, regulatory and administrative issues surrounding videoconferencing which, if overlooked, could potentially compromise patient confidentiality and safety. In a review of technology-based interventions for anorexia and bulimia nervosa, only one study out of the 40 identified (outcomes noted in 45 publications), observed differences between the preferences of patients and health-care professionals. 356 The authors of the review concluded that videoconferencing provides little benefit for clinical staff in terms of cost or time savings, but may assist patients who live in remote places and find it difficult to access specialist psychological services.

Health service delivery outcomes

In the review of the use of telemedicine for older patients, 36 of the 50 studies (that included a medical end point) discussed various health service outcomes (i.e. cost, hospitalisation, health-care utilisation). 131 The conclusion of the review was that video consultation may be an effective method for decreasing health-care expenditure. The authors of a review focusing on telepsychiatry reported limited data supporting the cost-effectiveness of video technologies. 349 There is some evidence pointing towards the feasibility and acceptability of telemedicine in different sociocultural contexts. A review of telecounselling in minority ethnic communities concluded that it is a highly adaptable and appealing health-care medium that could be implemented in different spatial locations. 345 The evidence indicates that the impact of video consultation is a highly contested issue. Although the subanalysis342 of a larger systematic review and meta-analysis into heart disease343,344 found that structured telephone follow-up and telemonitoring reduced heart failure-related hospitalisation admittances, the authors found no conclusive evidence that this occurred with videophone. Furthermore, a review of the financial benefits of videoconferencing noted that eight of the nine included studies were methodologically weak and reported no advantages compared with usual care. 354 The remaining study, with the highest methodological quality, observed that video communication was more costly than conventional health care. Taking into consideration the limited data surrounding the financial implications of telemedicine, some reviews were unable to draw any meaningful conclusions. 45,140,362

Strengths and limitations of the review

To our knowledge this is the first review of reviews conducted in this topic area. Conducting a review of existing reviews presents challenges in synthesising large amounts of information. We have aimed in this rapid review to summarise the existing literature, taking into account the quality of the reviews. However, we are conscious that summarising reviews can lead to loss of information and that our results are a high-level summary.

Several of the reviews looked at multiple telehealth interventions, which included videoconferencing. These reviews did not identify much evidence relating to videoconferencing. A review on diabetes mellitus included 35 RCTs, with videoconferencing being just one of a number of telehealth interventions included. 362

A review of telehealth for patients with heart disease found no studies that had examined the use of videoconferencing in this population. This limits the findings as we cannot be sure that they would hold for video alone.

Owing to the time restraints of a rapid review, we adapted point 11 on the AMSTAR checklist, ‘Was the conflict of interest reported?’. Ordinarily every included study within a review would be checked for a conflict of interest statement. This was not possible in our time frame and so we assessed conflict of interest in relation to the review itself.

Choosing the mode of communication

Each of the different digital modalities is particularly well suited to different purposes. Across the services we studied, patients and staff used telephones for urgent contact. E-mail was considered useful for giving complex or lengthy information and links to external sources, whereas text messaging was good for appointment changes or for quick queries. Upsetting news and complex issues would be discussed in person in a clinic appointment. PHRs were useful for sharing data prior to clinic appointments. There was little mention in our data of specific training in clinical communication using these communication modalities. 367,368 Our study participants considered the benefits, and limitations, of the use of the different modalities. For example, clinicians mentioned the loss of tone of voice, which may be important when assessing mental health, and young people mentioned that they preferred to raise embarrassing or sensitive issues with their clinician by e-mail or text message rather than in person. Young people liked to receive complex instructions by e-mail or text message so that they could refer back to them. Clinicians struggled with how to record their digital communication with young people in their clinical records. IG officers suggested that they would like to avoid dictating to patients the mode of communication to be used with the NHS.

Understanding and mitigating risks to patients from the use of timely digital clinical communication

Concern from clinicians about risk to patients varied across clinics depending on the health condition, the potential for harm to the patient in terms of how quickly harm might occur and the severity of the harm, and whether or not urgent care would normally be provided elsewhere (by a local hospital for young people attending a tertiary service). For example, those working in mental health were concerned about patients sending an e-mail out of hours when feeling suicidal and not getting a response. The young people who agreed to be interviewed did not echo this concern. The freedom to be able to write an e-mail when they want to was valued, but an immediate response out of hours was not expected. The young people interviewed described using different communication modalities according to the urgency of the health problem. Very few incidents could be recalled by patients or clinicians where the use of digital communication had adversely affected patient care.

However, the use of IT in health care does have its hazards that need to be addressed. 369 Unanticipated novel forms of hazards can occur. 370 For example, it has been shown that electronic prescription systems may improve the quality of prescribing, but also introduce new sources of error. 371,372 Hazards from the use of digital communication in health care include the manipulation of identity, the misinterpretation of digital information, breaching confidentiality and overreliance on digital communication as a means of communication at the detriment of traditional clinic visits. 372,373 The IG officers interviewed in our study recognised the importance of addressing these vulnerabilities. 374 Clinicians and young people described a number of mitigation strategies that they use to eliminate or reduce risks. Perhaps most importantly, timely DCCs were used between young people and clinicians who knew each other. There was a pre-existing relationship and a level of trust which was often enhanced through the use of this communication. This reshaping of relationships was identified in a realist review of telehealth interventions for self-management support for LTCs. 375 Participants also used common-sense strategies, such as seeking advice from other services (e.g. emergency department), when they required urgent help but communication with their clinic had failed (e.g. no response or delayed response). They limited or restricted the use of digital communication to reduce the risk of breaching confidentiality, for example by not including sensitive information in e-mails and text messages, or by using digital communication only for simple issues, such as rearranging an appointment. Not all the clinical teams we studied used communication systems with encryption. This is now becoming more widely available in the NHS providing a technological solution for this risk.

Most health-care professionals in our study strongly supported the need to inform patients of the risks of digital communication, although not all sought explicit consent from patients. Specific guidance for clinicians from their trusts on obtaining consent for the use of this form of communication was lacking in some places. This was a source of concern among clinicians. Young people tended to be less concerned about confidentiality than clinicians, but this may reflect their lack of awareness of the potential risks. Some young people were concerned about confidentiality of messages on their own mobile phones, whereas others were not.

The form of communication between clinician and patient does not change the ethical duty of clinicians to obtain informed consent and maintain confidentiality, or their legal obligations under the Data Protection Act. 376 However, General Medical Council guidance does not specifically consider the specific context of DCCs and it appears that few NHS trusts have clear policies to guide clinicians. Our data suggest that attention needs to be given not just to the obtaining of initial consent, but for confirmation of that consent over time as the health state and the life situation of the young person changes. What a young person finds acceptable to receive on their smartphone may change over time.

The importance of consent and confidentiality in using non-in-person communication has been highlighted in the literature, specifically in the context of e-mail communication,377 and the American Medical Association issued ethical guidelines for e-mail communication between patients and physicians in 2003. 378 In 2004, Car and Sheikh379 noted that ‘early email use in health care has grown without an adequate supporting infrastructure to address security issues’. They also noted the need for informed consent. 379 IG officers in our study varied in their views on the need for specific local guidance in addition to generic professional guidance about clinical communication and information. Professor Wachter has suggested that his report, published in September 2016,37 may encourage resistance to over-regulation of NHS IT. This would place the responsibility with individual clinicians and clinical teams to interpret generic professional guidance in this new context of DCC.

Setting the boundaries of timely digital clinical communication

Both clinical teams and young people talked about the importance of clear boundaries and achievable expectations when using timely DCC. These boundaries and expectations are likely to vary between clinics depending on the nature of the health condition. Of particular importance to clinicians and young people is clear communication about boundaries and continual reinforcement about the boundaries, particularly in clinical specialties dealing with young people who are at high risk (e.g. mental health). Young people wanted clear boundaries, but they also wanted a commitment from clinical teams that expectations would not only be clear but could be, and would be, met by the clinical team. This is important for effective service delivery and for patients to manage their anxiety. Patients and carers become anxious if communication is not received when expected, as they think it signifies complications or bad news.

Digital communication raises challenges for health-care professionals in relation to setting professional boundaries and understanding the implications for the professional duty of care. These challenges generated ethical discomfort in some of our participants. Boundary issues included their availability to respond to communications from patients, particularly out of work time, and the nature and content of these communications. This particular ethical concern related to the use of DCCs is less discussed in the literature. There is currently professional guidance for clinicians on maintaining professional boundaries with the use of social media,380 but not in relation to direct DCCs.

Managing the workload of clinicians providing timely digital clinical communication

Most clinical teams welcomed the ability to use digital communications, and none wanted them reduced or removed, even though they were using the digital communication in parallel with traditional in-person contact. However, the teams recognised real issues with workloads. Services that best managed this had planned systems in place. IG officers were anxious to ensure that all communication was logged so that their NHS trust would receive payment for all services provided. It is important to remember that most of the clinics we studied were providing care to young people who were known to the clinic, setting a limit on demand that would not be the case for a service providing care at first point of contact. In the clinics we studied, there was potential for time spent on responding to digital communication to be saved in other clinic activity, although our health economic analysis was unable to quantify this.

Path dependence

The cases we examined were empirical rather than experimental. As with most ‘bottom-up’ initiatives, the sample will tend to emphasise the most attractive and obvious situations and approaches. Therefore, the observed cases may not involve the most difficult issues, or may not seek to address them systematically. A well-designed quasi-experimental (RCT) approach can help to complete the picture, establish the robustness of the findings and identify specific characteristics that can help to inform decisions, such as if, and how, to implement DCC in other contexts.

Endogeneity

The data routinely collected for unrelated purposes had insufficient link to the DCCs. Interviews consisted of data that seemed significant to the participants in view of their motives for adopting DCCs, the time and attention needed to capture and record data, and the sensitivities that cause those involved to share and discuss their experiences or to keep them to themselves. These data reflect experiences with DCCs. Therefore, particularly ‘good’ or ‘bad’ (to a much lesser extent) experiences are more likely to be captured, which limits external validity.

Within the limitations of the study design, there were strengths of the study. It captured a wide range of perspectives obtained from both clinicians and young people from across different clinical specialties, from clinics of various sizes and from across the UK. The size of our data set ensured that we had a diversity of participants and reached data saturation for each of our major analysis approaches and in most clinics. The opportunity to develop focused interview schedules for ethics and patient safety for use in selected sites contributed to richness of these data. For the health economic analysis, we obtained a good response rate to the questionnaire from staff for quantifying the burden of DCCs. The qualitative data provided an in-depth understanding of the pathways by which DCC affects health, well-being and costs from a diverse range of sites. Our PPI activity gave us insight into the language related to health and digital communication that young people use which we incorporated into our data collection.

There were also limitations of the study. The young people who participated were recruited through their health-care provider. We were unable to recruit for interview young people who do not engage with their service so the voice of young people who have disengaged is missing from our data. These young people may be less positive about DCCs than those we interviewed, or do not have access to it. For example, they may not own a mobile phone, be unable to buy phone credit, live in a location with poor mobile phone coverage or have disabilities making the use of a mobile phone difficult.

We were surprised that health literacy did not arise as a theme from our data. It is possible that we failed to recruit young people for whom this is an issue. We did not directly ask clinicians about this issue.

In some of the small clinics we were unable to recruit for interview a diverse range of patients. We had no examples of Skype communication, as we were unable to identify any clinics using this during our recruitment phase. This was probably because there was still concern about security when using Skype, and NHSmail 2 which includes Skype as an option, had not yet started.

The main weakness of the study from an ethical analysis perspective was the difficulty in obtaining explicit ethical reflection from participants. The method of analysis allows us to look at implicit ethical concerns and issues, but in a study whose primary focus was ethics a Socratic approach to interviewing would be likely to elicit more nuanced ethical reflection.

Our study methodology did not permit a formal economic evaluation. We did not obtain quantitative data on the magnitude of impacts identified qualitatively. The missing responses to the questionnaire may disproportionately represent clinic staff whose engagement with DCCs is minimal, biasing the analysis of responses. Some staff found it difficult to quantify their time spent on DCCs. Some young people were reluctant to engage with the stated preference elicitation.

Most of the young people engaging in our PPI activities did not have a LTC, which may account for some of the contrasts between their views of DCC and that of the study participants.

Test-relevant patient-reported outcome measures for use in evaluating digital clinical communication

The impact of the provision of timely clinical communication on young people with LTCs could potentially be evaluated using a non-disease-specific outcome measure. Our literature review and PPI consultation suggested two potential measures. As discussed in Chapter 1, there is evidence that young people with LTCs have improved health outcomes if they engage with health care. There is also evidence that young people are less likely to engage with health care where there is poor patient–clinician communication, inflexible access to people and information, lack of person-centred health care, and a lack of continuity and relationship development. 16–19 Many of these aspects of care are assessed by the Physicians’ Humanistic Behaviors Questionnaire. 199 There is growing evidence that patient activation, the other of our identified measures, is associated with improved health outcomes. 386 Given this evidence, these generic outcome measures may provide sufficient evidence of impact or not of timely DCCs.

Evaluate, with an experimental design, the impact of timely digital clinical communication for people living with long-term conditions

Our study provides a framework for clarifying the digital communication for testing. Based on current practice and commonly used technology, the intervention most likely to be acceptable to clinicians and patients is event-driven communication using communication technology, provided issues, such as confidentiality and linkage with clinical record, have been resolved. The design of such an evaluation would need to take account of the different ways in which digital communication has an impact on different LTCs.

Explore and evaluate the benefits and disbenefits of replacing traditional clinical consultations with digital communication

Our research suggests the type of issue that needs attention when setting up DCCs to replace traditional clinical encounters. Further research is needed on how this is undertaken most effectively and the benefits and disbenefits for patients and clinical teams evaluated. The IG officers interviewed in our study thought that Skype-type communication is the modality most likely to replace traditional clinical encounters.

Evaluate the impact of the changing digital landscape on how digital clinical communication can achieve continuity of care and patient-centred care in an ethical, safe and cost-effective way

The changing digital landscape offers the opportunity to rethink traditional ways of providing clinical care. What type of continuity of communication through digital channels would enhance continuity of care, and at what cost?