Early uptake and adoption of the NHS App (quantitative study 1)

Introduction

This chapter describes the first study of the quantitative workstream which used observational data from the NHS App dashboard to examine the uptake and adoption of the NHS App in England. The majority of this text have been reproduced and/or adapted with permission from the journal article describing this work. Reproduced with permission from Kc et al.40 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.

This study aimed to evaluate the effectiveness of the NHS App by exploring the trends of uptake by using data metrics from the NHS App dashboard, and routinely available General Practice data to examine early patterns of app adoption across different population groups and in relation to recent time-trends.

Methods

Data

Data from the NHS App were analysed using data metrics from the NHS App dashboard.41 The NHS App dashboard was jointly developed by NHS England and Improvement team and NHS Digital under England’s National Healthcare system. Metrics currently available through the dashboard include aggregate counts of the number of registrations, downloads, appointment bookings and cancellations, GP health records viewed, prescription requests, visits to NHS 111 online, organ donation registrations and withdrawals, visits to the health A–Z page, and to the national data opt-out page. The research team received monthly aggregated data from the NHS App team at NHS Digital and did not have access to any personal or personally identifiable information.

Covariates data mapping sociodemographic profile of patients at the GP practice level were obtained from verifiable public health data sources. Socio-demographic data on the age and gender of all GP registered population were obtained from the NHS digital website.42 Data on ethnic composition and deprivation were obtained from Fingertips public health profile,43 and data on long-term health conditions were extracted from the GP Patient Survey (GPPS) database.44

Analysis

Descriptive statistics

The study period for this analysis was from January 2019 to May 2021. Descriptive statistics were used to summarise monthly NHS App metrics at a GP practice level nationally. Absolute monthly counts were presented from January 2019 to May 2021 of total app downloads, registrations, login sessions, appointments booked, GP health records viewed, and prescriptions ordered.

Time series analysis

An ecological interrupted time series (ITS) analysis was used to analyse the impact of the first UK national lockdown due to COVID-19 that occurred on 26 March 2020. The time series analysis was used to evaluate changes in uptake and the longitudinal impact of the pandemic on different functionalities of the app.45 An interruption was added to the time series on 1 April 2020, to estimate changes in uptake before and after announcement of the first lockdown. The ITS explored changes in level and trend for national app logins, appointment bookings, GP health records viewed, and prescriptions ordered between January 2019 and May 2021.

The ITS model was then used to assess if there was a change in app usage immediately after the first lockdown and if a change in trend occurred over the whole study period. Absolute and relative changes of the post intervention trend were also estimated, had the first lockdown not taken place. The absolute change in trend was calculated by taking the difference between the predicted pre-lockdown trend of the outcome and the post-lockdown trend at the end of the study period. The relative change was calculated as the absolute change as a relative proportion. To analyse app usage and uptake before the NHS COVID Pass was introduced, a separate ITS analysis was conducted excluding the month of May 2021. Autocorrelation was assessed by analysing the residuals of the ITS models.45

Cross-sectional analysis

A cross-sectional analysis of the differences in NHS App registration rates at February 2021 across categories of age, sex, ethnicity, Index of Multiple Deprivation (IMD), size of the GP practice and long-term physical or mental health conditions, disabilities or illness was conducted. A negative binomial regression model using incidence rate ratios (IRRs) was applied to explore variations in NHS App registrations across these groups.

For analysis, the sociodemographic covariates were ranked into four quartiles using the following identifiers: age (per cent 15- to 34-year-olds), gender (per cent males), ethnicity (per cent White), chronic care needs (per cent with long-term healthcare needs) and practice size (total number of GP registered patients over 15 years). Quartile 1 included the lowest ranked practice and quartile 4 the highest. The only exception to this was the IMD rank which followed the Office for National Statistics (ONS) method of grouping into quintiles.46 Quintile 1 included the most deprived practices and quintile 5 included the least deprived practices. Cumulative total NHS App registration rate per 1000 GP registered population at February 2021 was considered as the outcome variable.

The Strengthening the Reporting of Observational Studies in Epidemiology guidelines were used for the reporting and analysis of this study.47

Results

NHS App registrations and downloads

There were 8,524,882 NHS App downloads and 4,449,869 registrations from January 2019 to May 2021 (Figure 2). At the start of the study period, there were 31,633 downloads and 1279 registrations. At the onset of the first UK lockdown in March 2020, there was a spike of 532,275 app downloads. The highest number of downloads occurred after the announcement and launch of the COVID Pass in May 2021, with a total of 2,668,535 downloads. During this month, 2,099,234 users registered for the app. At the end of May 2021, there were 51,956,423 GP registered patients in England. Of these registered patients, 8.56% of the population aged 13 + were registered for the NHS App.

FIGURE 2 Monthly cumulative and new NHS App registrations and downloads. Reproduced with permission from Kc et al.40 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.)

Subgroup differences in app registration

The negative binomial regression model measured a difference in the NHS App registration rate, measured for each of the covariates with others held constant. The IRR comparing the practices in the highest versus the lowest quantile for each variable found 25% lower registrations in the most deprived practices [IRR 0.75; p < 0.001, 95% confidence interval (CI) 0.74 to 0.75], and 44% higher registrations in the largest-sized practices (1.44; p < 0.001, 95% CI 1.42 to 1.45).

App registration was 36% higher in practices with the highest percentage of White patients (1.36; p < 0.001, 95% CI 1.34 to 1.38) and 23% higher in practices with the highest percentage of 15- to 34-year-olds (1.23; p < 0.001, 95% CI 1.21 to 1.25). Practices with the highest percentage of males had 13% lower registrations (0.87; p < 0.001, 95% CI 0.87 to 0.88), which was also 2% lower in practices with the highest percentage of people with long-term care needs (0.98; p < 0.001, 95% CI 0.97 to 0.99), when compared to the reference groups (Table 3).

NHS App metrics

Login sessions

National NHS App login sessions per month ranged from 5157 to 16,730,430 from January 2019 to May 2021. There was a steady increase of login sessions from January 2019 to April 2021, with the highest number of logins occurring after introduction of the COVID Pass in May 2021 (16,730,430 login sessions) (Figure 3). The ITS showed that before implementation of the first lockdown, the number of NHS App login sessions was increasing over time at a rate of 52,093 logins a month and decreased immediately after implementation of the first lockdown. However, only the change in trend over time was found to be positive and significant, indicating an average slope change of 602,124 (p = 0.004) logins a month. Twelve months after the first lockdown, the average number of NHS App logins was 5,663,224 more than would have been expected if lockdown did not occur. This represented a 441% increase (Table 4).

FIGURE 3 Login sessions and appointments booked time series and ITS (January 2019–May 2021). Reproduced with permission from Kc et al.40 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.)

Appointments booked

Appointments booked using the NHS App ranged from 298 to 42,664 per month during the study period (see Table 4 and Figure 3). There was a surge of appointment bookings in January 2020 (33,003 appointments booked), September 2020 (32,335 appointments booked), and in May 2021 (42,644 appointments booked). There was a significant drop in appointment bookings after the first lockdown was announced in April 2020 (7674 appointments booked). The ITS showed that before the first lockdown, there was significant evidence of an increase in the average number of appointment bookings 2247 (p < 0.001), followed by an immediate decrease in bookings after the lockdown was announced −21,315 (p < 0.001). Twelve months after the first lockdown, the average number of appointments booked using the NHS App was 21,601 fewer than would have been expected if lockdown did not occur. This represented an overall 38% decrease in appointment bookings.

GP health records viewed

GP health records viewed in the NHS App ranged from 2212 to 9,324,546 per month from January 2019 to May 2021. There was a steady increase in records viewed up until April 2021 and a threefold increase from April 2021 (3,309,586 records viewed) and May 2021 (9,324,546 records viewed) (Figure 4). The ITS showed a non-significant decline in the average number of health records viewed right after lockdown −1,441,297 (p = 0.11). The sustained effect over time was significant (p = 0.001) indicating that the number of records viewed increased on average by 371,656 views per month post lockdown (Table 5). Twelve months after the first lockdown, the average number of GP health records viewed was 3,390,234 more than would have been expected if lockdown did not occur. This represented a 548% increase.

FIGURE 4 GP health records viewed and prescriptions ordered time series and ITS (January 2019–May 2021). Reproduced with permission from Kc et al.40 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.)

Prescriptions ordered

Repeat prescriptions ordered using the NHS App ranged from 655 to 530,382 per month from January 2019 to May 2021. There was 1.3 times increase in prescriptions ordered after announcement of the first lockdown, March 2020 (98,692 prescriptions ordered) and April 2020 (128,028 prescriptions ordered) (see Figure 4). The ITS showed that pre lockdown, there was significant evidence that the number of prescriptions ordered was increasing at an average rate of 5255 (p = 0.001) a month (see Table 5). Post lockdown, there was a positive and significant increase in trend of 19,934 (p < 0.001) prescription orders per month (see Table 5). Twelve months after the first lockdown, the average number of NHS App prescription orders was 286,630 more than would have been expected if lockdown did not occur. This represented a 225% increase.

Discussion

Summary

We found that there was strong adoption of the NHS App even before the onset of the COVID-19 pandemic. Before the first national lockdown in the UK was announced, there were almost 1.5 million downloads of the app. From April 2021 to May 2021, there was a fourfold increase in app downloads, showing the significant impact the introduction of the COVID-19 Pass had on app uptake. However, a disparity between app downloads and registrations exists. At the end of the study period, there were 8,524,882 downloads and 4,446,286 registrations. There are also differences in the NHS App registration rates across the different sociodemographic groups highlighting unequal trends of adoption, with higher usage in less deprived and less ethnically diverse practices, with a generally younger population.

In terms of app functions, COVID-19 and the introduction of the COVID-19 Pass service had significant impacts on their use. Introduction of the COVID-19 Pass accounted for over a threefold increase in login sessions in May, indicating that users could have been logging into the app to retrieve their vaccination status. This pattern was also observed for the number of users using the app to access their GP health records. Appointment bookings fell substantially after the first lockdown in line with a fall in overall primary care NHS activity,48 but continued on the same gradient from March to May 2021. In person GP appointments fell in March 2020 due to the pandemic, and some GPs switched off the ability for patients to book appointments via the NHS App at this time. Prescriptions ordered via the NHS App also significantly increased after the first lockdown, suggesting that more users were using the app to place prescription orders rather than in person or via the phone. These findings could be related to patients and carers seeking alternative options to in person care (e.g. digital prescription ordering).49 They may have also been influenced by changes in health service delivery due to the risk of COVID-19 transmission and due to the reliance on online prescriptions to support patients who had relocated during the national lockdown.50

Strengths and limitations

Our ecological analysis was only able to analyse the NHS App at the GP practice level. As person-level data are currently unavailable, these preliminary results indicate an inequality in adoption, which may be influenced by patient and provider-related factors. Furthermore, this study also presents limitations in the use of ecological data to understand subgroup differences in the uptake of and engagement with an app rolled out at a national level. Even though data linking allows us to overcome some of these challenges, differences in the format, structure and content of different data sources influence biases due to linkage errors. These differences, although potentially negligible, can still result in overestimation or underestimation of results.51 Availability of person-level data would overcome some of these challenges by allowing us to understand the interactions between sociodemographic factors and app use patterns more directly.

Also, the ITS analysis also does not account for other confounding factors, or control for the rapid changes in healthcare delivery during COVID, which may have contributed to the results seen in this study. This could include changes in seasonality, other existing cointerventions and changes in healthcare delivery model that may have occurred in tandem during the study period.

Comparison with existing literature

The public health impact of digital health interventions is dependent upon real-world uptake, and engagement.52 Although our analysis showed a strong interest in the NHS App, variations in the adoption rates and differences in the use of functionalities may not reflect high levels of engagement. A recent study looking at uptake and engagement of health and well-being apps found that one of most important factors for engagement were apps coupled with health practitioner support.52 At the time of this analysis, the NHS App does not directly incorporate any health practitioner support. This may be an important consideration as research highlights the importance of provider endorsement and support, along with technical training and usage assistance as effective digital facilitation strategies in primary care, particularly to support those from marginalised groups.53

There are also concerns about sustained app use as highlighted by the differences in app downloads and registrations, which corroborates the findings of a recent study which showed that 25% of all mobile apps were only accessed once after download.54 The barriers for sustained app use could be for several reasons, from users not being able to successfully register for the app, or never using the app once it is downloaded, which need to be further explored. Furthermore, the differences in the NHS App registration rates support earlier research findings and they indicate that use of the NHS App, as with other digital health technology differs according to gender, ethnicity, socioeconomic status and healthcare needs.50,55

Implications for research and/or practice

Further research is needed to identify whether the NHS App is an effective digital health tool and the extent to which it has met the goals set out by NHS England. The existing literature on technology adoption and diffusion of innovation tells us that this process is difficult and complex. Adoption is not just based on the technology, but a complex mixture of how the public and staff interact with it, what they see as the benefits, organisational culture and wider influences on the system including the policy and regulatory context.56 Also, the evidence of usage inequalities warrants facilitation efforts to promote equitable use of digital services in primary care settings. Provider support, patient assistance and training are identified as useful strategies,53 but a more comprehensive understanding of the impact of these initiatives within the broader organisational context, including technology usability and provider burden is essential.

Our analysis showed that there has been strong adoption of the app, and that COVID has significantly expedited uptake, but further research is needed to evaluate continuous usage of the app over time across different population groups and if it yields any benefits.

Conclusion

This is the first ecological study that has analysed a nationwide intervention rolled out by NHS England. This analysis has shown that the uptake of the NHS App has significantly increased post lockdown, driven by COVID-related events. Usage trends for the different app functions varied and the patterns of app registration was unequal between different subsets of the population. Further research is needed to measure the extent to which it influences inequities in health and whether it impacts care outcomes.