Qualitative study

Introduction

The qualitative study examined how the NHS App has been used and experienced over time by patients, carers, members of the public, healthcare staff, delivery teams, industry and other stakeholders, and whether and how the app has met varying needs in accessing and organising health care. Parts of this chapter have been reproduced with permission from Reidy et al.25 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

Our research questions were:

1. How did multiple interacting influences shape NHS App use, implementation and roll-out in general practice (and beyond)?
2. How did patients, carers and healthcare staff experience the integration of the NHS App in the health service? What were the perceived implications for access, efficiency and safety in general practice?
3. What transferable learning can we draw from the example of the NHS App to inform the implementation, national roll-out and routine use of health technologies?

Ethical approval for this study was granted by the National Research Ethics Service and Health Research Authority (Reference 21/WS/0031).

Methods

This was a longitudinal qualitative study which employed a range of methods, including interviews, focus groups, observation and document analysis. Technology implementation, roll-out and transferability were explored as nonlinear processes across different settings. We compared and contrasted findings between general practice sites. Data collection were carried out over 2 years between June 2021 and April 2023. This longitudinal approach facilitated investigation of the NHS App as an evolving technology in a policy and public health environment which was changing rapidly. It also allowed exploration of sociotechnical processes over time, such as administrative processes, professional roles, norms around primary care access and patient expectations.

The sites were five GP practices across England. We used a staged approach to site identification and recruitment to achieve maximum variation in characteristics such as geographical location, ethnicity mix, patient list size, deprivation levels and digital maturity status. Digital maturity was assessed using the Greenhalgh et al.26 digital maturity five-point scale, which considers the organisation’s readiness to plan and deliver a digital service (including measures to address digital inequalities), its capability (based on what level of digital measures are already present and running), and the organisational infrastructure (the underpinning material, regulatory, and human resource frameworks) that are in place to support further development of digital services. The characteristics of the five sites are shown in Table 1. Sites were identified via NIHR Local Clinical Research Networks and professional networks. The functionality offered through the NHS App to patients differed by practice (and over time). These are summarised in Box 1 for clarification on the functionality offered to patients in each site.

Participant recruitment

Patients and carers were primarily identified through GP practice screening and selection or through invitation via practice patient groups in the five study sites (2–6 patients per site). We used theoretical and maximum variation sampling: as data collection progressed, patients and carers were selected based on views that were either missing, or views that we wanted to explore in more depth, for example, patients who had attempted to gain proxy access, or patients who had supported a member of their family to access the app. We also recruited some patient and public participants directly through voluntary and community organisations (such as groups representing different health conditions) or social media [Facebook groups (Facebook, Inc., Menlo Park, CA, USA) and Twitter (Twitter, Inc., San Francisco, CA, USA)], for example, those using the #NHS App or using condition-specific groups.

Interviews and focus groups involved people living with a range of health conditions (Table 2), including those living with, or carers of people living with, conditions such as HIV, long COVID (and those with comorbidities relating to long COVID), type 1 and type 2 diabetes, Parkinson’s, colitis, epilepsy, arthritis, skin conditions, cancer, asthma, amputations, high blood pressure, high cholesterol, fibromyalgia, Crohn’s disease, immune compromised, kidney failure, allergies, severe intolerances, sight loss, mental health conditions and those having investigations done for yet undetermined conditions (e.g. spinal stroke, eye conditions).

We used purposive sampling to recruit NHS staff (selecting participants from different professional roles to ensure a varied sample). NHS staff participants were identified and invited to participate through the case study sites as staff working within the site.

Policy, commissioning, industry and other stakeholders were recruited using purposive and snowball sampling, considering relevant expertise in their professional role and involvement in the introduction and roll-out of the NHS App. We contacted participants across a range of relevant organisations, drawing on authors’ professional networks and publicly available information.

Data collection

We recruited 88 participants to 62 semistructured and think-aloud interviews, and 4 focus groups (Box 2). These 88 participants included patients, carers and members of the public (who used the app to different degrees or not at all), clinical and non-clinical NHS staff, and other stakeholders from the digital health industry, policy and commissioning, other public bodies, and civil rights organisations. Some of the participants discussed their experiences with the NHS App in a dual capacity, in terms of professional involvement but also personal use.

• Interviews:
  • 62 interviews with 66 participants comprising
  • semistructured interviews (some involving more than one participant) with
• 23 patients, carers and members of the public
• 23 healthcare staff (practice managers, GPs, nurses, healthcare assistants, admin and finance)
• 16 stakeholders (from commissioning, NHS agencies, policy, data privacy organisations, public bodies, industry)
  • Think-aloud interviews with four patients
• Four condition-specific focus groups with 22 patients/carers living with diabetes, long COVID, Parkinson’s, HIV
• Observations in five GP surgeries (60 hours)
• Fieldnotes (notes, photos, videos, screenshots)
• Document analysis of approximately 100 documents (policy blogs, GP documents/forms, government reports, social media discussions)

All interviews and focus groups were conducted by Claire Reidy (CR) (PhD, experienced qualitative researcher), and CP (PhD

, experienced qualitative researcher). Neither CR nor CP had any prior relationship with study participants. Semistructured interviews explored individual experiences, while focus groups (conducted with patients and carers only) allowed participants to collectively debate their views and perceptions. We followed a flexible topic guide adapted for different participant groups to gain an understanding of experiences with and perceptions of the app, how people used it, as well as background design and development processes, and the role the app played in organising care in general practice. Think-aloud interviews with four patients enabled direct in-depth observation of interaction with the app (e.g. viewing or using features on the app), facilitating articulation of how tasks were accomplished and challenges encountered. We captured these interactions in field-notes, photos (of participants’ use of the NHS App), screenshots, and videos, with participant permission. Interviews and focus groups took place on Microsoft Teams, over telephone or face-to-face at different stages of the pandemic and lasted between 16–72 minutes (average = 41 minutes) and 59–87 minutes (average = 73 minutes), respectively. All interviews and focus groups were digitally recorded and transcribed verbatim for analysis (with written or digitally recorded consent from participants).

Claire Reidy conducted focused observations (face-to-face, totalling ≈ 60 hours) within each GP surgery, to better understand the back-end operational and technical processes required to integrate the app in the service (e.g. managing digital appointment booking or prescription ordering). Observations facilitated assessment of digital maturity within each practice site, the context of the practice, collection of NHS App-related and digital health promotion documents in GP surgeries and supported participant recruitment.

We collected a range of documents (e.g. policy blogs, Twitter discussions, government reports and plans, newspaper articles, documents in GP surgeries – e.g. forms for patient digital health access, and GP survey examples) which helped inform recruitment but also contributed to data interpretation and context setting.

Data analysis

Theoretical framework

Data collection and analysis was informed by Greenhalgh et al.’s NASSS (Figure 1) an evidence-based framework developed to guide thinking on the implementation, roll-out and routine use of technology in health care.24,26 By drawing on NASSS, we have taken a complexity-informed approach27 to embrace rich theorising, generative learning, as well as pragmatic adaptation to changing contexts. In this sense, we acknowledge the dynamic, interconnected, and adaptive nature of healthcare systems, and the implementation and roll-out of a complex app in a complex system. For example, we used NASSS to explore multiple interacting influences to NHS App implementation and use, including differences in engagement with the NHS App depending on clinical needs and service configuration, the complexity of the technology itself and modifications to it over time, the value generated for different stakeholders, the role of NHS staff in its introduction, the underlying organisational structures in place and the complexity of wider NHS service provision.

FIGURE 1 NASSS framework. Reproduced with permission from Greenhalgh et al.24 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure above includes minor additions and formatting changes to the original text. (Figure not reproduced here; see the original article.)

Analysis

Analysis took place in parallel to data collection, enabling us to refocus our methods over time, to account for a range of experiences and emergent themes until we were able to build a sufficient understanding.28 It enabled us to consider changes as the app was rolled out, and who we needed to speak to and what data we would need to capture to build on our evaluation of the NHS App roll-out, for example, the development of the COVID-19 Pass. Braun and Clarke’s29 six-step framework guided our thematic analysis, which began with data familiarisation, generation and iterative refinement of initial ideas and codes from transcripts and fieldnotes. This was followed by inductive and deductive development of broader themes within and across NASSS domains, including comparing and contrasting similarities, differences, and connections in the data (especially across sites). Combining an inductive and deductive approach allowed us to pay attention to emergent as well as anticipated themes and to identify cross-cutting areas of complexity. We used NVivo 12 (QSR International, Warrington, UK) for data management.

Our patient public involvement group was involved in the process of data analysis and sense-making. We held six patient public involvement and engagement (PPIE) workshops at project initiation and set-up, through recruitment, and as data emerged. The workshops enabled a space for sharing, discussing and challenging insights gained, and considering and prioritising approaches and directions for continuing data collection. Further, weekly team meetings with the PPIE lead facilitated continued engagement throughout the study.

Results

The analysis draws on the NASSS framework to identify the various ways in which complexity manifested as part of implementation, use and roll-out of the NHS App, including how multiple interacting influences shaped its trajectory. We categorised these against the seven domains of the framework, as described below. Illustrative quotes are provided under each NASSS domain.

Domain 1: clinical condition

Participants with health conditions described different care pathways for the management of their health and shared different ways of engaging with the NHS App. For example, those with conditions primarily managed in general practice found more useful functionality through the app (following up on investigations and test results from within the GP Practice, and letters between GP practice and secondary care). Many of those primarily receiving care in hospital or community settings (e.g. patients living with Parkinson’s, HIV or type 1 diabetes) found the app less useful in their self-management as they were unable to access the majority of their test results or health records (unless their practice included a personal health record function add-on in the app through a commercial provider). This led some patients to express frustration with the app which they saw as reflecting system fragmentation and a lack of joined-up care;

I’ve got now a multidisciplinary team behind me … [so] using those different authorities in the NHS, my medical records are not too complete … if I look at my [GP] medical record, I would be very surprised if it even says I’ve got Parkinson’s because I don’t interact with my GP surgery at all with Parkinson’s … I think [the NHS App] is great to have, [but] it’s gotta be able to interface with all the other potential medical people I use.

While the NHS App was designed to be used by anyone registered for NHS care (aged 13 or over), some patients and staff posited that it would be less relevant for those not managing long-term conditions (e.g. it was suggested young people were more likely to belong to this category) as they may not need to use its features, or the health service, as regularly.

Yeah I use like a health app on my phone like for my steps … I think my dad told me about it [the NHS App] … I’ll do it later … I just don’t feel like I need it at the minute … I don’t wanna do it … [it’s] for everyone and also for patients that might want to track anything on there … it can be for me, I just, I don’t know … I just don’t feel like I would use it very often that’s why …

Domain 2: the technology

The app and its features, over time and in different settings

Compared to alternative patient portals already in use by GP practices, the NHS App required a more elaborate registration and authentication process (e.g. filling in an online or, usually, a physical form), plus requirement to provide identity documents (see Appendix 1, Figure 20, take a photo or video, facial recognition, and for the practice to enable a digital ‘linkage key’) which in many cases made the app initially appear complex and difficult to navigate for patient users.

Apart from core functionality provided directly through the app, the rest of its features relied on background systems and infrastructures (e.g. GP IT systems, access to the Spine, Electronic Prescription Service data, Summary Care Record) with the NHS App only functioning as gateway. This increased complexity as not all features were available by default and consistently to all users. For example, GP practices made decisions independently on whether they would enable appointment booking (including what types of appointments and how these would be displayed) through the app which resulted in significant variability in whether and how this feature became available to patients;

… nobody understands how complicated booking practice nurse appointments are … and so we would never let them [patients] do that [online booking] … Because they would get it wrong … it depends on the nurses skill set if she’s the right nurse and how long it needs depends on certain things … there’s well over a 100 different types of [Practice Nurse] appointments … if [patients] speak to our staff they can signpost you …

Appointment booking was in fact disabled by most surgeries given triage protocols during the pandemic and demand pressures following consecutive lockdown periods (see Appendix 1, Figure 21). There were some accounts of patients being able to use appointment booking on the app, especially before the pandemic, although this was not the case for most people we interviewed (who reported they would benefit from such a feature):

pre-COVID you had to put 75% of appointments available online for booking … Now the trouble with that is that goes against all sign posting practice … we spend our life trying to triage … to get things to the right clinician in the right time frame … [booking] straight into appointments online goes completely against all of that … While booking an appointment that they don’t need, you know, for a hay fever medication which they should just go and get from the chemist. There’s no way of filtering that.

Own record access was controlled by local protocols and procedures which in some cases differed (at least in terms of their interpretation) between, and sometimes even within, GP practices [especially in relation to limiting access due to General Data Protection Regulation (GDPR) concerns];

there’s certain things in your records where it says … you don’t have access to this or you don’t have permission so I’ve actually gone into the surgery three or four times and said I want access to this part and quite often they say oh we don’t know how to do that … and for some reason my husband he’s got access to more sections on his than I have.

Even in cases where patients were provided with access to their own records, sometimes these were in a coded format, rather than designed for sharing with patients. Therefore, some patients found them to be lacking in meaning and relevance;

not that I can understand this [coded entry on the app].

Over time the app became integrated with external commercial platforms, such as a personal health record platform, a digital triage and online consultation service (where patients submitted symptoms or requests to their GP practice), and other communication solutions. Again, these options were only available in GP practices or secondary care services using such external providers through local contracts and arrangements (e.g. independently or through their CCG). This increased variability in the features on the app, with additional options available for some patients and practices, but not others. It also complicated our interviews with patients as it was often unclear to us (and to them) what features were available to each of them and at what point.

Knowledge required to access and use the app

Our data indicated a certain level of knowledge was needed to be able to request access to the app in the first place and make good use of available features. Yet, awareness of resources and guidance available to support use of the app varied among staff and patients. Some suggested that the app should facilitate access in a more patient-facing way that does not assume advanced health literacy to interpret, for example, normal ranges for blood test results (something that was addressed in a later version of the app);

… having an app like this does assume a level of health literacy that I don’t think the general population had, I mean again even somebody who has some medical awareness there’s some bits of it that I kind of look at and I go ‘what does that mean?’ … [so] what is a blood test what is a normal range just having some access to some of those sort of information.

In interviews and focus groups, there was confusion over what the NHS App was in the first place. Some patients thought it was the same as the COVID track and trace app used for contact tracing at the time. There was general lack of awareness of what other features (other than NHS COVID Pass) were available on the app and so patients and staff often articulated that more publicity and promotion was required.

Complexity emerged not just because practices had different processes and protocols in terms of enabling or disabling specific features on the app. Staff also needed to know how to interpret and activate these protocols (e.g. enabling proxy access or access to own records) so that features of the app would become usable for patients.

Domain 3: the value proposition

Should this be the only app that practices use or one of many?

By designating the NHS App as the digital ‘front door’ to the NHS, policy-makers aimed to promote access to services and support self-management, enabling patients to ‘be more equal partners’ in managing their health care. Members of the NHS App development teams saw it as a way to address inefficiencies, generate better (and less) use of appointments, and reduce burden on services, such as through patients ordering prescriptions, monitoring test results or checking on referral progress. It was publicly pronounced as a ‘thin’ platform on which other services and third-party suppliers would connect to, perhaps due to push-back from industry about the NHS App’s potential to ‘stifle the market’:

Yes but if you’ve seen the App you will see that, so we got so much push back from third party suppliers about the NHS trying to stifle the market and Matthew [Gould]’s blog was very much the NHS App would remain very thin but nobody ever really kind of said what that meant … I think he meant it would be a platform that other services would connect to etc. … people would choose whatever, whatever kind of user interfacing they preferred.

However, the strategy or concept of a ‘thin’ app was not clearly characterised or explained to practices. Over time, and with integration of popular features such as the COVID Pass, value was articulated in different ways to meet different needs and interests, given broader policy intentions and targets for widespread use.

Demand-side value: ‘empowerment’ and ‘control’ for patients, reduced workload for practices?

For most patient participants, value is often derived from relatively simple features, such as ordering prescriptions on the app (compared to calling or visiting the surgery). Security features, such as fingerprint scanning and face id to access the app, were considered appealing to many users we spoke with and was especially important for those with health diagnoses they did not want shared outside the health service:

I’m really worried about my HIV status coming out but I would rather have my results on my phone, I don’t pass my phone to anybody so nobody can go onto it, you can’t go onto the App without your fingerprint so they can’t just see that App and think oh I’ve have a look and see what that is because they can’t get on it.

Having features such as security codes for accessing the app sent through text message provided reassurance (see Appendix 1, Figure 20).

When patients could access their own (or where they had proxy access rights to others’) health records, they reported an increased sense of control, for example, knowing what information was held about them, what consultations had occurred when, and what record was kept regarding those consultations. Some patients also saw such access to records as providing the ability to hold the service to account:

Well I suppose the purpose of the App should be for me to be able to know what information is kept about me, you know, with regards to whatever interaction I’ve had with the NHS … [if] the medical institution has any information about me, you know, I want to know what that information is.

However, given the availability of different features varied between settings, what patients expected as added value from the app often did not materialise. This included quick and easy access to booking appointments through the app, viewing own health records or providing a history of encounters across the health service, rather than just in general practice.

With promised features not always realised, some patients were reluctant to shift to using the NHS App as they compared its value to other apps or portals with similar features they had already been using (e.g. pharmacy apps). They suggested added value would be generated if the NHS App could provide more interactive features, such as pop-up notifications when prescriptions were ready to collect, or when new test results, letters or referrals were issued. However, when pop-up notifications were enabled in newer versions of the app, technical issues meant some notifications appeared without any content.

From the perspective of general practice staff, the app did not always appear to have a clear and sustained value proposition in terms of supporting their practice, workload or engagement with patients, apart from the prescription ordering feature. They did not always perceive the app as having substantial clinical need or utility (at the time), for example, not tracking referrals, or messaging patients, or other ways for the app benefiting patients directly, although general practice staff recognised it held significant potential in the future (and in some cases awareness about current app features was limited).

With the announcement of the COVID Pass being added to the NHS App in April 2021, the value proposition became quickly associated with this feature and resulted in rapid increase of downloads as well as increased awareness by practice staff. In addition, many patients that we spoke to had only heard about the NHS App in terms of the COVID Pass, and some patient participants described being unsure what value the NHS App would hold for them when they were no longer required to use the COVID Pass.

Domain 4: the adopter system

Tensions between top-down implementation and staff capacity in general practice

At first, introduction of the NHS App and its basic functions did not seem to require significant effort by GP staff. Yet, over time it became apparent that patients saw their surgery as the first point of contact, for example, how to register on the app, use it and request access to different features (such as proxy access, personal health records), also correcting personal details for travel purposes using the COVID Pass. This meant surgery staff needed to become increasingly familiar with the app (and its various iterations) from a user and IT infrastructure perspective so they could answer such queries, demonstrate to patients how to use the app or identify sources of external guidance for them, even on matters unrelated to the practice such as the COVID Pass. Staff perceived a lack of technical support being available, and some proposed the development of a toolkit to resolve issues, such as app error codes, or a central team or helpline to support troubleshooting. Without a clear value proposition in terms of addressing workflow and demand pressures, and with little training or support to take on new roles and responsibilities related to the NHS App, some staff, especially practice managers, expressed frustration that the onus had been placed on them in a top-down way without meaningful consultation:

[M]y staff haven’t had any help, haven’t had any training in how to support people [with the NHS App], most of them only know about it because they set it up for themselves … Or because they’re interested in NHS stuff and they do it out of the goodness of their hearts so they can help patients … And giving themselves a tutorial.

Staff also wanted to ensure the NHS App did not signal a complete move away from face-to-face engagement and digital approaches were used in the context of already established, trusted relationships with patients. This was particularly relevant for specific patient groups (e.g. those with language barriers or from particular cultural backgrounds) and those who needed more support to navigate the system through face-to-face contact rather than through complex online tools:

I think a lot of us are quite pro integrating technology into our systems it’s just integrating it in a way that doesn’t advantage the people that are already good at getting access to healthcare … So [digital triage tool]s are great, I would say a lot of the time they are used by people that already receive good healthcare … [unless] English isn’t your first language and you don’t have family support around you.

Some functions (e.g. own record access) engendered more complexities than others, including fears that patients may be affected negatively by viewing their records in a system-facing rather than patient-facing format, leading to misunderstandings, additional appointments, as well as safeguarding concerns.

Variability in patient and carer experiences depending on features available

Patients, especially those living with long-term conditions, used the app to navigate the service and support self-care. In interviews they discussed ordering prescriptions, checking symptoms, viewing test results, and (where available) booking appointments. However, many patients described app usage as occasional rather than regular, in contrast to banking, shopping or neighbourhood apps (where people can connect with, communicate and exchange information or buy and sell items based on their geographical area). Users with access to additional features (e.g. being able to view appointments and secondary care letters) appeared to have an enhanced experience of the app, as they perceived their care as more joined up (see Appendix 1, Figures 23 and 24).

Those with proxy access on the NHS App (e.g. for family members) described being able to view health records more easily and conveniently ‘on the go’, ordering prescriptions simply, which supported managing other responsibilities (including employment and caring for others). Yet the process of acquiring access was depicted as long and complex requiring persistence (taking more than 6 months or resulting in failure), as not all GP practices were familiar with protocols and technical processes (e.g. which boxes to select in IT systems) or carers were not always able to procure necessary identification documents. Proxy rights were also an issue when access requirements changed (e.g. COVID Pass age restriction).

Disparities around equitable use and access

While the NHS App was deemed by some to be more relevant to those living with and managing health conditions, some older adults who were managing (potentially multiple) health needs, or those with some specific cultural needs, described not being able to easily use the NHS App, or the NHS App not being something they wanted to use. Personal social support networks (such as volunteers in the surgery, or family members) were deemed especially valuable for those needing specific support to hear about, access and use the NHS App (such as those with language barriers), and a distinct barrier for those who did not have such assistance:

I would say it is a barrier, so when you’ve got patients that can’t speak that can’t read English it’s no good for them really … We still offer the services to them, [we] just ask them to have family or friends to help them with ordering their prescriptions online.

Some staff also expressed fear around digital access more generally, or around add-on app features (such as a digital triage tool) that could potentially mask coercion and control efforts being exercised over patients accessing the health service, and prevent practices from being able to distinguish such cases.

Domain 5: the organisation

In the context of increasing demand pressures within practices during and following the pandemic, organisational capacity from general practices to innovate varied between case sites and was not always at the forefront of their efforts (see Table 1 for case study site characteristics, and digital maturity rating). In one of the large practices already championing active engagement in digital innovations before the pandemic (site 1), there appeared to be more organic and unified processes for integration of new ways of working including dedicated practice staff with experience introducing new technologies.

… because of our involvement in that [working with industry] and other tools when COVID hit we were in quite a good position to be able to manage that … our patients were used to us communicating with them in that way [SMS messaging] we also I think because of my University background … where everything was very digital … I’d already been working over the previous year to try and get … people set up with things like VPN’s with remote access … And so actually and using more digital tools.

In other practices, there was apprehension from practice managers that introduction of digital innovations would conflict with a more ‘traditional’ organisational culture prioritising patients telephoning or attending face-to-face (site 2 and site 5).

… I know some practices have tried to introduce, not a barrier as such, but they have tried to really direct people sort of to, to remove the telephone option to some degree … I think because of our traditional approach, that would be something where I definitely will get resistance from the GPs … I don’t think they would want that as, as part of their practice.

There were also fears expressed from some staff around a lack of capacity to take in, organise and act on potential extra engagement or data as a result of digital innovations and augmented access to test results and health records. Organisational capacity and willingness to innovate, including leadership focus on innovation, led to varying levels of readiness for the NHS App and impacted on NHS App use and promotion within practices. Yet, when the app (or its add-on services) offered clear value by resolving or improving workflow, such as patients sending in photos in a secure and confidential way, it enabled organisational willingness to promote use of the app even in the context of more ‘traditional’ organisational cultures (site 2):

[S]o particularly sort of dermatology type things, we’ll encourage them [patients] to send photos and that’s, that’s where we’ve tried to use the [NHS] App, because that’s a secure way and the patient has had to prove who they are to gain access to the app etc. and so that’s the one we favour.

Nature of adoption and/or funding decisions

The introduction of the NHS App appeared to generate tension, confusion or conflict in practices already using products (apps or patient portals) with similar or additional (e.g. messaging or video consultations early on) functionality, with little support over how to navigate the digital landscape or reach a decision to switch to the NHS App. Integration of the COVID Pass became a deciding factor as it increased awareness of the NHS App for patients and staff, yet some NHS App functions were still not promoted by practices with a history of using commercial apps or portals.

While the NHS App, in its basic form, does not require funding to use it, some add-on features, (which could potentially add unique or more valuable functions for the practice) do require funding. Yet, there was often a lack of dedicated income stream for practices, and so more innovative add-on features were seen as ‘out of reach’, even for practices who were more digitally focused. As a result, there were tensions between the value of the app set against funding, as well as infrastructure, organisational capacity, and staff roles.

Extent of change needed to organisational routines

Introduction of the NHS App required co-ordination between staff members to enable patient access to the app, respond to queries, and integrate processes within IT systems. Yet, at organisational level, allocation of staff roles and relevant training was inconsistent between practices; some provided staff with training or information on the practice intranet while others relied on staff own initiative to train themselves out of interest in technology or ‘out of the goodness of their own hearts so they can help patients’ (S27). Some practices emphasised the unexpected responsibility on the organisation to support change suggesting that ‘as usual GPs are the sponges that soak up all the, everything that drips out of all the cracks elsewhere’ (S27) and advocated for further commissioning support.

Domain 6: the wider context

The NHS App follows a long history of more or less successful initiatives to digitise care in the NHS. This includes attempts to at a policy level provide all patients with access to their own health records which has been one of the most contested features at policy level.

I think this the, the patient, patient access programme has been a very challenging piece of work and, and it has yeah it’s become, become sort of fairly fraught … a lot of tension of kind of ways to make it happen but also to consider, consider all of the eventualities to consider safety.

Controversy around this feature continued as the popularity of the NHS App was growing, with policy-makers and professional bodies engaging in lengthy and difficult negotiations on how appropriate access would be enabled and how safeguarding issues would be avoided:

They [practices] think it’s massive safeguarding issues … they should be enabling patients to access their full records but the majority have not done that. This was in the GP contract way before the pandemic … policy wise … we’re leaving the burden of you [practices] having to do that by us doing it, they’re saying … [it] will increase burden because people will start seeing things that they didn’t see before and phoning up … and GP’s feeling like the clinical software to redact isn’t strong enough.

National mass roll-out of access to prospective records was stalled and replaced by a less centralised approach with practices required to switch on access locally (including as part of new contractual requirements). Yet professional bodies continued to express concern and practice managers and professional bodies highlighted issues such as the range of considerations required to provide access while also protecting information that was not originally intended for sharing with patients:

We did have concerns about how well the programme had been communicated and practices, had some of them given enough time to prepare … and … checking the records for vulnerable patients who might need to have that … you can’t just drop a kind of a new piece of technology in the laps of extraordinary busy practices and expect them to be able to pick it up and make the best use of it … there’s a real need for the change management support … to see technology adopted but also … make sure that that it provides the outcomes and the benefits that it’s designed to.

While a rapid shift towards digital solutions was observed early on in the pandemic, increased demand pressures on practices impacted their ability to take on new technology-mediated ways of working post COVID lockdowns:

So to be honest our main focus at the moment is to, after COVID … I think it’s probably across the whole of England we’re, GP practices are just absolutely bombarded with telephone calls … Patients trying to get through so at the moment our priority is to get everyone on board with the phone calls … Trying to answer as many calls as they can … Take each call as it is but that’s our priority at the moment …

The idea of ‘cross-selling’ (as mentioned by one of the interviewees) also surfaced as one of the principles applied in the context of roll-out meaning that users registering on the NHS App for one purpose were then expected to find value in other functionality provided, such as organ donation and online prescriptions. In this respect, the COVID Pass was seen as a major draw by some interviewees:

And we were hearing at the start of our work that people wanted it to be in the COVID-19 App because, well we’re already using [it] … people still might say that that’s the best place but for the app it’s had a huge amount of unintended consequences in terms of like organ donation registrations have gone through the roof.

However, others described the integration of the COVID Pass on to the NHS App as ‘manipulative’ and setting a ‘terrible resonance’ (Data privacy organisation Co-ordinator). Their concerns centred primarily around data privacy and security when presenting the COVID Pass (and subsequently the NHS App) in settings outside health care:

Rather than having a COVID passport App they kludged it onto the NHS App, clearly they wanna force more people or encourage more people to, depending on your perspective, to use the NHS App. That’s a long-term win for them … [but] you’re talking about handing a, or showing a, a logged in device that’s having access to at least part of your medical records to a third party.

From the perspective of the patients interviewed, there appeared to be less concern around data privacy, security and advertising compared to commercial products, given the NHS branding.

Wider context influencing roll-out of the app included efforts at inter-organisational networking to support learning. In some instances, clinical commissioning groups became invested in the NHS App which they saw as a ‘nation asset’ they did not need to procure for, and provided dedicated support for roll-out, including GP staff training, discussion forums between practices, and as point of contact for queries. At early stages of implementation, some areas even became designated ‘beacon sites’ which received additional assistance such as extra support and coaching but only over a limited time period. In other areas, commissioners did not perceive they had a mandate for the NHS App and engaged less in supporting roll-out.

Domain 7: embedding and adaptation over time

Given the NHS App primarily provides a gateway to different services and platforms, there is significant scope (and policy will) to add or remove functionality so the app remains relevant over time. This has been demonstrated already with integration of the COVID Pass, and subsequently with the addition of popular commercial services (at least in some areas) such as those facilitating online consultations and access to personal health records. This brings an inherent risk, however, in that the NHS App may end up incorporating too many features, making it difficult to sustain a coherent identity.

There is significant policy interest in maintaining interest and regular use of the NHS App, increasing functionality for purposes such as ‘day-to-day health management’ and securing engagement from patients and practices alike, beyond the COVID Pass. Improvements and adaptations are being made over time, such as with the symptom checker (e.g. relevance for dermatological questions and different skin colours), increased security features, direct booking of vaccine appointments, and the addition of notifications to the app, although not all additions were perceived as useful by patient users:

Sometimes I get an email or notification saying my app has updated, like blood tests … and I go to check and sometimes the, where it tells me it has updated, are not on there … it doesn’t clarify what it means … I have to guess.

Issues with digital inclusion will have to be considered, especially around embedding different languages and increasing patient-facing terminology as the user base grows.

Discussion

Summary of findings

Using the NASSS framework, this article highlights the multiple layers of complexity manifesting when introducing a constantly shifting technology into a challenging environment such as English general practice, during and after a period of considerable societal turbulence because of the COVID-19 pandemic. Several interacting influences shaped the trajectory of the NHS App and we reflect on these below.

Roll-out of the NHS App reflects and illuminates the complexity and fragmentation of the NHS. For example, when patients questioned why their information on the app appeared so disjointed across primary and secondary care. The app appeared to be less able to support those accessing ongoing health care outside of general practice (e.g. patients we spoke to who were living with Parkinson’s, HIV and diabetes) to manage their condition(s), yet useful for those with conditions primarily managed in primary care (for access to prescription ordering, NHS number, GP-related letters, records of GP consultations, test results). The app was also perceived by some staff and patients as being less useful for those without long-term conditions.

There was a level of health system literacy required to comprehend how to access, or request support to access, the app, and understand what the NHS app was, and what functions it encompassed (in addition to being the COVID Pass). To a certain extent, adoption relied on general practice and its staff, who assumed responsibility in supporting patients to access the app but also in safeguarding patients, by restricting access to health records where needed. The app also required integration and access to several background systems and IT infrastructures, which meant that not all features were available by default by all users (e.g. own health records). Further, practices made decisions about what features would be available, for example, appointment booking, so this function, which was desired by many patients, was not available for all. The app design offered a transactional, mechanistic approach to functionality such as appointment booking, which did not translate well into complex practice processes that often relied on relationships with patients. This was also the case for patient access to own records, which were controlled by practices, and for which procedures for access was variable and not understood by all staff, or where staff were concerned about GDPR (and requirements for redaction).

While policy-makers saw the app as a ‘front door’ to the NHS for patients, the purpose and utility of the app, or how this shifted over time, was perceived as not being satisfactorily communicated or fully co-designed in collaboration with practices. Although, the COVID Pass made the presence and value of the app to practices and the wider public more prominent, and coincided with a drive to increase the user base. This was despite development teams having carried out substantive user research, including on accessibility, that remained focused, however, on use by patients rather than the complex work of other implicated groups such as healthcare staff. There was also heterogeneous support provided to, and by, CCGs, which created inconsistent experiences and approaches to roll-out.

For patients, features such as ordering prescriptions online held a strong value proposition. Security features, for example, fingerprint scanning and face recognition were appreciated, especially by those with health diagnosis they did not want shared. Accessing own health records (including proxy access), engendered an increased sense of control over health for many patients, as well as enabling a sense of ensuring accountability of the health service. However, there was complexity and variability in the value proposition of the NHS App, as different features were added, removed or became more or less relevant over time (e.g. COVID pass). NHS App use has been fundamentally influenced by external events such as the COVID-19 pandemic. However, events of such magnitude are rare, which means part of the learning around how the app was received may not be directly transferable, yet most of the lessons still apply beyond the pandemic context.

Comparison with other literature

Despite interest in patient access to their own records through the NHS App, finding a balance between providing access and safeguarding, without imposing additional workload on practices, has been challenging. Related issues have been highlighted in previous research. For example, Louch et al.,30 discuss how clinical and non-clinical staff in primary care supported patient access but also voiced concerns around the type of historical information that would become available and how this might be interpreted or used by those accessing this information. Along similar lines, Nøst et al. found healthcare professionals feared that increased access could increase workload due to additional need for patient follow-up.31 From a patient perspective, McMillan et al.32 report the need to better promote availability of access, as well as introduce joined up records with community and secondary care. However, complexity is exhibited when providing access to features such as health records, which does not automatically lead to patient ‘empowerment’ (as shown elsewhere33) especially when considerable work is required from patients to access such features.

From an organisational perspective, several studies have identified barriers and facilitators related to the implementation of patient portals.34,35 Our study further exemplifies how multiple interacting influences coshape the trajectory of a complex technology with evolving components reaching various parts of the healthcare service, particularly so in a crisis context. Involvement of health professionals in the design and implementation of patient portals plays an important role in mediating their experiences.36 GPs have previously expressed that they want to feel more supported with digital change (in procurement or implementation) and that innovation focused practice leadership (as found in this study), and support from the PCN or commissioning body around implementation has enhanced implementation.7 Finding an optimal balance between top-down and bottom-up implementation remains challenging but is necessary to enable alignment with local care processes and priorities.37

The motivation at organisational level to put in place the processes, staff roles and capacity required for the change effort changed in the context of increasing demand and work pressures during and following on from the pandemic. The availability, interest and capacity of healthcare staff to support adoption also varied in the context of what was perceived as top-down implementation, and in terms of the cultural context of practices, for example, apprehension from practices with a more ‘traditional’ organisational culture, as opposed to those practices who were more innovation focused and digitally strategic (according to Greenhalgh’s digital maturity scale26).

There were considerable influences from the wider context on NHS App roll-out, which shifted from a ‘remote-by-default’ orientation to an emphasis on in-person care, as well as controversies related to specific features (such as centralised patient access to own records). Where inter-organisational working, for example, between practices and commissioning groups occurred, the additional support encouraged roll-out in practices. In this landscape, the NHS App offered a means of extending previous efforts to introduce online access to services in the UK. Improvements and adaptations informed by user research were being made to the app throughout data collection, although there were challenges with digital inclusion as well as maintaining user interest in an offer that would remain relevant over time.

Issues of health equity have also been debated in previous literature which explains some of our findings, that disproportionate emphasis is placed on individual responsibility to access and use patient portals, without sufficient acknowledgment of system-level barriers and support for underserved populations.16 This transcended through patient participant’s experiences of requiring support to log in to and access the NHS App and its various features. Further, those from more affluent, White and middle-aged backgrounds with high health literacy appear to be more likely to use digital tools such as patient portal.16 Such trends of inequalities implicate a real risk of a new digital Inverse Care Law whereby attempts at digital transformation directly or indirectly compromise health equity.38,39 With primary care being the access point for many to the health service, there is a pronounced need to consider the complex factors influencing adoption and roll-out and to ensure that the adoption of digital health tools do not create or exacerbate inequities in health service access or use.

Strengths and limitations

The strengths of this study include the diverse sample of research participants in terms of patients, carers, healthcare staff and stakeholders from a range of environments. Participants varied in terms of age, ethnicity, health conditions, location (in terms of urban/rural and across England), health and digital literacy. Study sites covered a wide range of characteristics and patterns of technology engagement across the country. We ensured that we specifically engaged with individuals and practices that were more likely to be affected by issues such as health inequalities and the ‘digital divide’. The heterogeneous sample meant that we were able to garner a wide range of experiences and perceptions to explore how the app had been deployed and used in depth. We also worked closely with the PPIE lead for the study and held a number of workshops with the study PPIE lead and group to develop the research questions, recruitment literature and recruitment practices and case study sites over time, as well as synthesise and evaluate or validate findings as they emerged. The study utilised several qualitative research methods to collect findings, which enabled exploring the roll-out of the NHS App in different contexts and for different users, and to consider societal events in which the roll-out occurred. This led to rich data which could consider the layers of complexity surrounding the roll-out of the NHS App. Collecting data over time also meant that we were able to capture many different aspects of roll-out.

Limitations of this study include the fact that due to collecting data over time, and during many changes at macro, meso and micro levels, the technology itself was also changing. This meant it was difficult at times for those we interviewed, and the researchers, to ascertain exactly what research participants did have access to, either directly in the app, or in the practice settings (in terms of integration or support available). However, capturing this experience of patients also gave more emphasis to the complexity of this app, the app’s integration with practices, the presentation to patients, and the lack of comprehension, at times, about what the app’s functions and purpose were. Further to this, we also held regular discussions with the research team throughout the study, collection of data from different sources (including the media), and analysis of data as it emerged helped to catalogue and consider the different contexts and experiences of users, staff and stakeholders over time to avoid any narrow interpretation of the findings, and to consider the wider context of the app and what was occurring at a higher level.

It also became evident that there were a number of people who were excluded from using the NHS App and that there were important cultural considerations that we were not able to explore in more depth, such as language barriers (with the app only available in English) some functions being less relevant for non-White skin, and around the promotion of digital access to the health service potentially excluding some minority and vulnerable populations where relationships and trust need to be built up with healthcare staff. Future research should focus on the specific needs of defined ethnic groups to explore culturally appropriate ways to support access to primary care through the NHS App.

Conclusions

In this qualitative study, we aimed to unpack the complexity underpinning implementation, use and roll-out of the NHS App, learning from the experiences of patients (users and non-users), healthcare staff and other relevant stakeholders. We found that patients had diverse (positive and negative) user experiences as the app evolved, with some of its features described as more useful (e.g. prescription ordering or access to records). Yet the app was primarily a gateway to general practice systems and infrastructures, therefore, not all features were available by default and consistently to all users, with information often appearing fragmented or presented in a system-facing way (e.g. coded). Using the app as a ‘front door’ to the health service also assumed a certain level of system literacy which was not always realised. For practices, although the app did not initially appear to have a clear value proposition in terms of supporting their work or engagement with patients, scope for potential future development as core NHS infrastructure made it appealing. Thus the NHS App remains a complex intervention in a complex landscape, it is clear ongoing work is needed to ensure the app can achieve its potential, and will continue to be developed to meet patient, service and policy needs.