Introduction
Parts of this text have been reproduced and/or adapted with permission from Kc et al.75 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/.
GP services account for the largest volume of activity with the NHS76 and with approximately 31 million appointments (excluding COVID-19 vaccinations) delivered through general practices in November 2023 alone77 it forms a fundamental component of the UK Primary Healthcare system. However, persistent issues, in part due to changes in healthcare delivery models due to COVID, have impacted patient access to GP services and continuity of primary care.78 There are reports of severe delays to timely access to care with approximately 7.6% of all GP appointments taking place after more than 2 weeks since booking77 and a significant drop in public satisfaction with GP services, which fell to 35% in 2022, marking the lowest levels recorded since 1983.79 These challenges are exacerbated by an ongoing workforce crisis due to chronic reductions in the number GPs available and concurrent increases in demands for servicing an ageing population with complex healthcare needs.80
Strategies to recover access to primary care services include a number of national initiatives, such as the Modern General Practice model in England, which focuses on ensuring equitable and easy access to care across multiple channels using optimised telephone systems, well-designed practice websites, as well as using digital tools such as the NHS App (also referred to as ‘the app’).81 These approaches align with a global commitment to prioritise primary care access, as outlined in the WHO Declaration of Astana, which emphasises the significance of primary health care in achieving universal health coverage and supports the use of technologies to improve access to care and to involve individuals in their healthcare decisions.6 They also reflect on patient and public views regarding priorities for improving the NHS, with ease of access to GP appointments highlighted as one of the top two concerns.79
Online services, including patient portals, have become almost universally implemented by GP practices in England over the past decade, serving as alternatives to traditional contact methods like phone or face-to-face interactions.82 The NHS App is one such service that currently forms a part of a multi-channel approach offering diverse healthcare access options, including in-person, phone and web services alongside online tools. Through the app, several options for primary care interactions have been shifted online using functions that allow booking GP appointments, communicating with care providers, ordering repeat prescriptions etc. Additionally, it grants patients access to health information online through features like NHS 111 symptom checks and GP record access etc. There are also plans for improving secondary care access using features for managing hospital appointments and related documents,64 with targets to achieve a 75% adoption rate among the population in England by early 2024.83
However, uptake alone does not always translate to engagement and not everyone can use digital technologies equally.40 Patient experience of care and familiarity with existing GP services may be important considerations that affect how patients engage with novel technological solutions, as evidence highlights the influence of patient experience of access to care and service interactions on the uptake of GP provides and resultant quality outcomes.84 There are also known overlaps between digital and interpersonal aspects of health care, where positive experiences with both enhance patient self-management and overall user engagement,85 which may foster a culture of trust and motivation to navigate the digital healthcare landscape. Digital tools facilitating patient-provider communication are more effective when positive clinician–patient relationship is present and they aid in overcoming access issues and mitigate past negative experiences, especially for marginalised groups.86 They may also offer alternatives to traditional care approaches, particularly for health information seeking in cases where patients express dissatisfaction with their GP care providers.87
Patient experience of using existing digital channels may also be equally important in understanding public amenability to newer technologies. Digital competence that aids familiarity and satisfaction with broader information communication tools may increase exposure to digital pathways, enhancing patients’ willingness to engage with new services.88 Additionally, where patients remain dissatisfied with their existing GP practice digital ecosystem that greatly varies across practices, tools such as the nationally rolled-out NHS App may offer more standardised alternatives. As research on the uptake of the NHS App uptake is emerging, these factors that relate to the use of the app need to be explored to understand aspects pertaining to access in a national portal context
This study aimed to explore whether patients’ experience of access to their general practices and their overall practice experience influences engagement with the NHS App. This would highlight whether there is a symbiotic relation between GP patient experience and experience of using the NHS App and if there is a need for concerted efforts to enhance the digital patient journey to achieve NHS England’s ‘digital by default’ objective.
Methods
Study design
An ecological study with data at the GP practice level in England exploring differences in the usage of NHS App features by patient experience of care and care access was conducted.
Data sources and linkage
NHS App usage data were obtained from the NHS App dashboard with access provided by the NHS Digital team as anonymised daily logs for a 27-month period between March 2020 and June 2022. Baseline sociodemographic data on the profile of the GP-registered population in terms of their age, sex, ethnicity, long-term healthcare needs and deprivation were obtained in relation to an earlier publication in this series from public health data sources. The full list of variables considered, and their data sources are detailed in Appendix 2, Table 7.
Data on patient experience of care and care access were extracted from the GPPS database. GPPS is a survey of approximately 2.4 million adult GP-registered patients in the UK and is currently conducted annually. The data used for this study were from a survey conducted in January 2021, which was completed online or via post by 850,000 patients nationally, with an overall response rate of 35.3%.44 GPPS provides population-weighted positive responses at the GP practice level for a number of individual questions,2 which was used to categorise practices into quintiles. For each covariate, quintile 1 included the lowest-ranked practices and quintile 5 the highest. Data related to the study aim included the following indicators and related GPPS questions:
- Ease of phone access: ‘Generally, how easy is it to get through to someone at your GP practice on the phone?’ (GPPS Question Q1)
- Ease of web access/use: ‘How easy is it to use your GP practice’s website to look for information or access services?’ (GPPS Question Q4)
- Overall GP practice experience: ‘Overall, how would you describe your experience of your GP practice?’ (GPPS Question Q30)
The relevance of the GPPS indicators to the research aims, and their appropriateness for use was finalised with inputs from the PPIE group members, the NHS England team and the wider research team.
Statistical analysis
General Practitioner Patient Survey data were linked to the NHS App data using GP practice code as an identifier for each GP practice in England. Data for practices with missing practice size or area codes were removed, as were practices with fewer than 200 registered patients (n = 146) to avoid linkage errors and to improve data reliability. For analysis, complete app data were available for a total of 6386 GP practices across England for the full period between 23 March 2020 and 27 June 2022. Stata 17.0 software was used for the statistical analysis.
The main outcome variables were weekly rates of each NHS App function used (e.g. registrations, logins, prescriptions ordered, medical record views and appointments booked) per 1000 registered GP practice population. Patient responses to the select GPPS variables were used as covariates capturing data related to patient experience of access to their GP practices by phone, website, as well as their overall practice experience.
IRR obtained from negative binomial regression models were used to explore differences in the use of the various NHS App features for each GPPS variable in comparison to the reference group (Q1), with others held constant. The regression models also controlled for all baseline differences in patient sociodemographic profiles.
Charts of use of the various functions over time, showing the different quartiles of practice by overall patient experience, were also plotted to visually demonstrate the temporal trend in use of the different functions for context.
Results
Results for each of the NHS App features in relation to the GPPS variables and their corresponding forest plots using IRRs and confidence intervals are presented below. A full table including percentage changes from adjusted models with complete IRR values is presented in Appendix 4, Table 9.
Registrations
Figures 10 and 11 show weekly rates of NHS App registration by the covariates of interest. Results exploring differences in NHS App registration rates by patient experience of care and care access showed that practices with higher proportion of patients reporting better phone access had lower registration overall compared to those in the lowest quintile (i.e. reference group, Q1 = practices with the lowest proportion of patients reporting ease of phone access). For example, registration rates were 2.9% lower in Q4 than Q1 (p < 0.001) and 1.1% lower in Q5 than Q1, although this was not statistically significant (p > 0.05).
FIGURE 10 Weekly NHS App registration rates per 1000 GP registered population. (Figure not reproduced here; see the original article.)
FIGURE 11 Time series graph showing mean NHS App registration rate over time per 1000 GP registered patients according to overall GP practice experience. Where, Q1 = GP practices the lowest percentage of patients reporting positive practice experience, Q4 = practices with the highest percentage of patients reporting positive practice experience. Time period = Weekly data from 7 January 2019 to 26 June 2022. (Figure not reproduced here; see the original article.)
In terms of web access, practices with more patients reporting better access and use of practice websites had higher registration rates compared to practices in the lowest quintile (Q1 = practices with the lowest proportion of patients reporting ease of web access and use). For example, registration rates were 2.6% higher in Q4 than Q1 (p < 0.001) and 2.1% higher in Q5 than Q1 (p < 0.001).
Registration rates according to patient experience varied when comparing practices with higher reported experience to those with the lowest (i.e. reference group Q1 = practices with the lowest reported practice experience overall). For example, registration rates were 1.4% higher in Q3 compared to Q1 (p < 0.05) and 3.5% lower in Q5 compared to Q1 (p < 0.001).
Logins
Figures 12 and 13 show weekly NHS App login rates by covariates of interest. Results exploring differences in NHS App login rates by patient experience of care and care access showed a varied trend. In terms of phone access, practices in the higher quintiles had overall lower login rates compared to the reference group (i.e. Q1 = practices with the lowest proportion of patients reporting ease of access via phone). For example, login rates 2.3% lower in Q3 than Q1 (p < 0.001). However, in practices with the highest proportion of patients reporting ease of phone access (Q5), login rates were 3.5% higher compared to Q1 (p < 0.001).
FIGURE 12 Weekly NHS App login rates per 1000 GP registered population. (Figure not reproduced here; see the original article.)
FIGURE 13 Time series graph showing mean NHS App login rate over time per 1000 GP registered patients according to overall GP practice experience. Where, Q1 = GP practices the lowest percentage of patients reporting positive practice experience, Q4 = practices with the highest percentage of patients reporting positive practice experience. Time period = Weekly data from 7 January 2019 to 26 June 2022. (Figure not reproduced here; see the original article.)
In terms of web access, practices with higher proportion of patients reporting better access and use of practice website had higher login rates overall compared to practices in the lowest quintile (Q1 = practices with the lowest proportion of patients reporting ease of web access and use). For example, login rates were 5.4% higher in Q4 than Q1 (p < 0.001) and 5.1% higher in Q5 than Q1 (p < 0.001).
Login rates according to patient experience varied when comparing practices with lowest reported patient experience (i.e. reference group Q1) with the practices in the higher quintiles. For example, login rates were 1.3% higher in Q4 compared to Q1 (p < 0.05) and 5.2% lower in Q5 compared to Q1 (p < 0.001).
Appointments booked
Figures 14 and 15 show weekly rates of appointments booked by covariates of interest. Results exploring differences in the rates of GP appointments booked using the NHS App by patient experience of care and care access showed that practices with the highest proportion of patients reporting better phone access had higher appointment booking rates overall compared to those in the lowest quintile (i.e. reference group, Q1 = practices with the lowest proportion of patients reporting ease of phone access). For example, rates of appointments booked were 25% higher in Q4 than Q1 (p < 0.001) and 57.8% higher in Q5 than Q1 (p < 0.001).
FIGURE 14 Weekly rates of appointments booked through the NHS App per 1000 GP registered population. (Figure not reproduced here; see the original article.)
FIGURE 15 Time series graph showing mean NHS App appointment booking rate over time per 1000 GP registered patients according to overall GP practice experience. Where, Q1 = GP practices the lowest percentage of patients reporting positive practice experience, Q4 = practices with the highest percentage of patients reporting positive practice experience. Time period = Weekly data from 7 January 2019 to 26 June 2022. (Figure not reproduced here; see the original article.)
In terms of web access, practices with the highest proportion of patients reporting better access and use of practice website had lower appointment booking rates overall compared to practices in the lowest quintile (Q1 = practices with the lowest proportion of patients reporting ease of web access and use). For example, rates of appointments booked were 17.6% lower in Q4 than Q1 (p < 0.001) and 18.9% lower in Q5 than Q1 (p < 0.001).
Appointment booking rates according to patient experience were also lower in practices with more patients reporting better patient experience when compared to the reference group (Q1 = practices with the lowest proportion of patients reporting better patient experience). For example, appointment booking rates were 29.3% lower in Q4 compared to Q1 (p < 0.001) and 27.5% lower in Q5 compared to Q1 (p < 0.001).
Prescriptions ordered
Figures 16 and 17 show weekly rates of prescriptions ordered by covariates of interest. Results exploring differences in the rates of prescriptions ordered using the NHS App by patient experience of care and care access showed a varied pattern. In terms of phone access, practices with better phone access had higher rates of prescription orders overall compared to those in the lowest quintile (i.e. reference group, Q1 = practices with the lowest proportion of patients reporting ease of phone access), although this slightly varied for Q2. For example, the rate of prescriptions ordered was 27.0% higher in Q5 than Q1 (p < 0.001) but was 0.4% lower in Q2 than Q1 although the difference was non-significant (p > 0.05).
FIGURE 16 Weekly rates of prescriptions ordered through the NHS App per 1000 GP registered population. (Figure not reproduced here; see the original article.)
FIGURE 17 Time series graph showing mean rates of NHS App prescription orders per 1000 GP registered patients according to overall GP practice experience. Where, Q1 = GP practices the lowest percentage of patients reporting positive practice experience, Q4 = practices with the highest percentage of patients reporting positive practice experience. Time period = Weekly data from 7 January 2019 to 26 June 2022. (Figure not reproduced here; see the original article.)
In terms of web access, practices with higher proportion of patients reporting better access and use of practice website had higher rates of prescription orders overall compared to practices in the lowest quintile (Q1 = practices with the lowest proportion of patients reporting ease of web access and use). For example, rates of prescriptions ordered were 7.8% higher in Q4 than Q1 (p < 0.001) and 7.7% higher in Q5 than Q1 (p < 0.001).
With regards to patient experience, rates of prescription orders were lower in practices with better patient experience when compared to those with lowest reported patient experience (i.e. reference group Q1). For example, prescription order rates were 8.3% lower in Q4 than Q1 (p < 0.001) and 16.1% lower in Q5 than Q1 (p < 0.001).
Medical record views
Figures 18 and 19 show weekly rates of medical record views by covariates of interest. Results exploring differences in the rates of medical record views using the NHS App by patient experience of care and care access showed that practices with higher proportion of patients reporting better phone access had higher rates of medical record views compared to those in the lowest quintile (i.e. reference group, Q1 = practices with the lowest proportion of patients reporting ease of phone access). For example, rates of medical record views were 5.4% higher in Q4 than Q1 (p < 0.001) and 13.8% higher in Q5 than Q1 (p < 0.001).
FIGURE 18 Weekly rates of medical record views through the NHS App per 1000 GP registered population. (Figure not reproduced here; see the original article.)
FIGURE 19 Time series graph showing mean rates of NHS App medical record views per 1000 GP registered patients according to overall GP practice experience. Where, Q1 = GP practices the lowest percentage of patients reporting positive practice experience, Q4 = practices with the highest percentage of patients reporting positive practice experience. Time period = Weekly data from 7 January 2019 to 26 June 2022. (Figure not reproduced here; see the original article.)
In terms of web access, practices with higher proportion of patients reporting better access and use of practice website also had higher rates of medical record views compared to practices in the lowest quintile (Q1 = practices with the lowest proportion of patients reporting ease of web access and use). For example, rates of medical record views were 3.4% higher in Q4 than Q1 (p < 0.001) and 0.5% higher in Q5 than Q1, although this was not significant (p > 0.05).
With regards to patient experience, rates of medical record views varied when comparing practices in the highest quintiles to practices with the lowest reported patient experience (i.e. reference group Q1). For example, rates of medical record views were 3.2% higher in Q4 compared to Q1 (p < 0.001) and 3.3% lower in Q5 compared to Q1 (p < 0.001).
Discussion
Our analysis of GP patient experience of care and care access on public engagement with NHS App features highlighted a varied pattern of uptake and revealed a nuanced relationship between access modalities, patient experiences, and the utilisation patterns of NHS App features. It might have been anticipated that better patient experience was associated with higher NHS App use, or possibly lower use. We found a more complex and variable relationship.
Generally, practices with better-reported phone access had higher usage overall for the different app functions but had lower registration rates and a varied pattern of logins. In particular, appointment booking rates were 57.8% higher in the practices that were reported as easiest to access via phone. In terms of web access, practices with websites that were reported as easiest to access and use had a mixed pattern of NHS App feature usage. Although registrations, logins and prescription orders were generally higher among these practices, appointment booking rates were substantially lower compared to the reference group. Also, rates of medical record views were overall higher among practices with websites that were easier to access and use, with modest but non-significant increase in the highest quintile. Similarly, the use of the different NHS App features also varied by patient experience with lower rates overall for appointment bookings and prescription orders among practices with better reports of patient experience. Although registrations, logins and medical record views were generally higher among these practices, usage rates were significantly lower in the highest quintiles, suggesting that practices with the best patient experience had lower rates of use of these different functions.
We found that better phone access meant people were signing up less to use the app, but where they had signed up, uptake of the different features remained high. Better phone access also corresponded to high rates of app appointment booking which may indicate that these practices were particularly focused on improving patient access or had lower demand than other settings. In contrast, appointment booking rates were significantly lower in practices with better practice websites which may suggest that patients prefer to utilise the practice websites to access this service and may be associated to availability of other commercial provides such as Accurx (Accurx Limited, London, UK) that allow patients to book their GP appointments directly using links provided by their providers or through the practice website. However, use of other transactional services such as prescriptions and medical record views remained generally high indicating that patients may prefer to engage with the NHS App for certain functions but find it easier to use their practice websites for others. It may also be indicative of practice provisions where certain practices had disabled the option to book appointments during certain time periods. A longitudinal analysis of the use of the different features and patient experience over multiple time points may provide a clearer overview.
In particular, patient experience and use of the different app features did not follow a clear gradient and highlighted a mixed picture making it difficult to ascertain the relation between the two. Significantly lower usage rates of all app features among the highest rated practices may suggest that where patients remained satisfied with their existing digital providers, they had less impetus to explore other digital channels. This prompts a critical reflection on the varying motivations and expectations that drive patient interactions with digital technologies and their different features.
Limitations
The GPPS is the largest validated survey of patient-reported measures that exists in England but has a low completion rate (35.3% in 2021). The sensitivity of the GPPS may be insufficient to detect any impacts resulting related to a portal scaled at a national level but perhaps only reported a small subset of patients.
Using practice-level GPPS data for assessing individual-level patient satisfaction with their primary care services in relation to the NHS App use may not be appropriate and the cross-sectional nature of the analysis limits establishing any causal relationships. The primary objective of this study is to assess whether patient experience of care and access provisions has any influence on the NHS App use; however, the practice-level assessment did not support subgroup stratification to explore potential impacts. It is plausible that data on patient experience of access directly in relation to the NHS may exhibit significant differences in the association.
Our study focused on the NHS App, even though there are alternative commercial providers that offer similar services within the NHS framework. While this approach is effective for capturing a national population sample, it carries the potential for bias since it is limited to a specific portal which still exhibits differential trends of adoption among different population subsets, therefore excluding some marginalised groups.
As the study explores UK general practice, which is free at the point of access, it does not consider whether patient experience of technology enabled care fosters uptake of novel solutions offering multiple functions for improved access through reduction of costs to the patients.
General Practitioner Patient Survey data used relates to January 2021, whereas the use of the NHS App peaked around May 2021.40 Using this time point allows measuring patient satisfaction with web and phone routes before alternatives options through the app were more widely used. As the GPPS survey was redeveloped in 2021 to account for pandemic-related changes to service delivery,44 we could not assess patient satisfaction before the app was released and after using multiple time points. Future studies should consider a longitudinal approach by capturing data directly in relation to NHS App use for a more direct measure of association.
Conclusions
In conclusion, this study provides an evaluation of the multifaceted dynamics that shape public engagement with digital health solutions such as the NHS App, and how app usage interacts with patient experience. The findings underscore the complexity of patient interactions with novel digital solutions in relation to their experience of care and care access, highlighting the need for a tailored and nuanced approach to digital health technology interventions. To maximise patient engagement with digital tools, a flexible and individualised approach is likely to be required.