Aims and objectives
The overall aims of WP2 were to investigate the range of DF currently offered in a sample of primary care practices and to examine how this related to patient awareness and experiences of DF, and patient awareness and use of online services. We also wanted to understand how patient characteristics influenced their awareness and use of both DF and use of online services. Work package 2 comprised four interconnected elements:
- a survey of general practices, with the primary objective of mapping the range of DF services currently offered in a sample of English primary care. Secondary objectives included understanding the impact of the COVID pandemic on digital service provision and the drivers influencing DF efforts
- a survey of patients with the objectives of exploring patients’ views and experiences of different modes of DF identified through the practice survey, and identifying what patient factors predict awareness and uptake of DF
- a patient survey boost sample of patients living with mental health conditions with the primary objective of exploring the experiences of DF in this population and potentially identifying any particular needs for these patients
- analysis of national GPPS data with the objectives of examining, on a large scale, the factors associated with patients’ awareness and use of online services, and how that awareness and use varied according to the modes of DF identified in the practice survey.
Methods
Practice survey instrument
A questionnaire was designed to be completed in approximately 10 minutes by either practice managers or GPs and was produced in paper (postal) and online (using SmartSurvey Online Software and Questionnaire Tool) formats.
The practice survey questionnaire (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024) was developed by a subgroup of the research team, with input and feedback from the PAG, the wider research team and additional GPs. The questions were developed using an iterative process, informed by the findings of the Di-Facto literature review (see Results). The questionnaire included questions addressing the range of online services offered to patients (currently and prior to the COVID pandemic), the activities used to promote online services or to support patients in using them, the staff involved in the provision of any support to patients and whether specific groups were targeted with that support. The questionnaire also sought the respondent’s views on responsibility for support, the influence of various factors on service provision and the rationale behind providing online services. There was also one question which gave the opportunity for a free-text response, inviting comment on anything else that the respondent saw as relevant.
Practice survey sample and distribution
The online version of the practice survey was piloted in 24 practices (November–December 2020) in Cambridgeshire and Peterborough Clinical Commissioning Group (CCG), and East and North Hertfordshire CCG. No changes to the questionnaire were required, but the recruitment process was refined after feedback (e.g. online survey formatting, improving the visibility of the online link). Subsequently, using the latest available NHS Digital data on General Practices,158 a sampling frame was defined comprising 610 practices from 8 CCGs (Devon, Birmingham and Solihull, South Warwickshire, Coventry and Rugby, Cambridgeshire and Peterborough, East and North Hertfordshire, Enfield, and Haringey CCGs with the latter 2 CCGs subsequently forming part of the North Central London CCG) from 4 Clinical Research Networks (CRNs) covering 4 geographical locations (South West England, West Midlands and East England and North London). Five hundred practices were randomly selected to form the sample, in anticipation of receiving one or more responses from 60% of practices. Pilot practices included in the final sample were not asked to complete the questionnaire again and the pilot data were retained. The research team searched www.nhs.uk to find practice e-mail addresses or websites for all 500 practices, and subsequently searched practice websites for e-mail addresses or contact forms. A sequential mixed mode was used, starting in mid-January 2021, whereby practices were contacted by e-mail or practice website contact form (where available) initially to alert them that a questionnaire would be soon sent, followed by the questionnaire itself, and up to three e-mail reminders (over 4 months). Paper versions of the questionnaire were also mailed to the practice manager, and subsequently to up to four named GPs identified from NHS Digital data on GP practitioners.158 When a response was received by a practice no further mailings (electronic or postal) were sent to that practice. Where multiple responses were received, the first complete response per practice was used for analysis. Finally, CRNs were requested to follow-up practices who had not responded. Full details of the recruitment process are provided in (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024). The practice survey was closed to responses in May 2021. Practices that responded and provided contact details were entered into a prize draw for 1 of 10 £250 vouchers.
Patient survey questionnaire
Patient survey questionnaire development commenced with a scoping meeting with the PAG, followed by a series of iterative workshops with the patient survey team. The questionnaire (see additional files www.fundingawards.nihr.ac.uk/award/NIHR128268; accessed 3 April 2024) drew on the literature review findings and adopted (revised) wording from existing questionnaires on eHealth literacy,159 the GPPS and the Get Digital basic skills assessment. Where possible, the patient survey content mirrored the practice survey questions (tailoring wording for patients with PAG input) to allow for direct comparison. It included sections exploring patients’ digital confidence, their awareness and uptake of online services and experience of the support provided by the practice to use online services. The questionnaire was piloted by volunteers. Feedback was that it was straightforward to complete; however, some words and layouts were subsequently changed for clarity.
For the Mental Health Boost sample, the questionnaire was identical to the main part of the patient survey. We consulted with a wider PAG, who had experience of mental health conditions, to identify acceptable language for the patient invitation letter and information sheet, to explain to patients how they would be identified, and to explore how to safeguard privacy when opening the invitations. The researchers involved with the mental health survey were experienced in and acutely aware of mental health issues and dealt very cautiously and sympathetically to the few patients where any concerns were raised.
Patient survey sample and distribution
We aimed to distribute the survey to 12,000 patients across 60 practices, in anticipation of a response rate of 35% (4200 responses). All practices returning at least one practice survey were invited to take part; however, additional practices, outside the original sample, serving deprived populations were later invited to participate to reach our target and to ensure that patients from deprived communities were included. The additional practices were selected with the support of the CRN. Practices were given instructions to run a patient record search and to produce a random sample of adult patients aged 16 and over. To ensure that patients from practices with more deprived populations had fair representation, practices with higher levels of deprivation were asked to identify more patients (320) than those with medium (240 patients) or low deprivation (170 patients). Practices were invited to review their list to exclude those experiencing severe mental illness, recently bereaved, under 16 years old, or incapable of giving informed consent to participate (all as defined by their GP). After exclusions, practices were asked to select the first 285, 220 and 150 reviewed patients, respectively, and to transfer the patient details onto an Excel file (with usernames and passwords added). The Excel file was then securely uploaded by the practice to the mailing company, CFH Docmail. The practice was instructed to initiate three patient mail outs. Each patient should have received an invitation letter, information sheet, patient survey and reply envelope for the first two mail outs and a postcard for the third. Patients were invited to provide contact details solely for the purpose of entering them into a prize draw for 1 of 10 £25 gift cards. Patients returning a survey either by post or online were deemed to have provided consent to participate in the study. Where paper copies were received data were entered by the study team using the same online survey platform. Double data entry was performed for four surveys per practice by a different member of the study team.
For the Mental Health Boost sample, we aimed to distribute 3000 questionnaires across 15 practices (200 each) to patients aged 16 years and over identified as living with anxiety or depression, in anticipation of a response rate of 20% (600 responses). A search was developed to identify a random selection of 220 patients who either had a code for anxiety or depression in their medical records (in the past 2 years) and/or received a recent anxiolytic or antidepressant drug prescription (in the previous 12 months) for anxiety or depression in their medical records. Patients taking Amitriptyline at < 50 mg/day were excluded since such doses may be commonly given for chronic pain management or for other non-mental health conditions. Practice staff were requested to screen the list to ensure patients fulfilled the eligibility criteria and excluded patients for the same reasons as for the main patient survey, but also for those who were known to be actively suicidal, or who were receiving diazepam for back pain or muscle spasm (with no history of anxiety).
General Practice Patient Survey
The GPPS is a national postal questionnaire that is sent to 2.3 million patients annually on behalf of NHS England. Further details of the survey and its development can be found elsewhere.13,160,161 Data from the 2019, 2020, 2021 and 2022 GPPS162 were analysed both on their own and in relation to the practice survey.160,163–166 We restricted the data to those who reported attempting to make a general practice appointment in the last 12 months to ensure that responders had the opportunity to use online services. We made use of two questions about the awareness and use of online services:
- As far as you know, which of the following online services does your GP practice offer? Booking appointments online, ordering repeat prescriptions online, accessing my medical records online, none of these, don’t know.
- Which of the following general practice online services have you used in the past 12 months? Booking appointments online, ordering repeat prescriptions online, accessing my medical records online, had an online consultation or appointment (e.g. completed an online form or had a video call).
The question on awareness of online services was removed from the questionnaire in 2021 and so responses from 2019 to 2020 were used for this question, while data from 2021 to 2022 were used for the analysis concerning use of online services. In addition, we used responses to questions addressing ease of getting through on the phone, helpfulness of practice receptionists, patients having a preferred GP and ease of using the GP practice website, as well as patient reported demographics, health status and the Index of Multiple Deprivation (IMD) based on the patient postcode of residence.
Data analysis
Descriptive statistics
Practices that took part in the practice survey were compared with other practices in England considering deprivation, rurality, list size and the percentages of registered patients over the age of 65 and (separately) with a non-white ethnicity. Deprivation and ethnicity data were obtained from the Public Health England Fingertips website167 based on 2021 data; list size and patients over 65 were obtained from NHS Digital December 2020 data, and rurality was based on practice postcodes as used in the 2019 GPPS. Descriptive statistics were produced for responses for the practice and two patient surveys. Following factor analysis indicating a single construct underlying the responses to Q2–Q6 of the patient survey (see Appendix 2, Tables 30–32 and Figure 17), a composite measure of digital confidence was formed from these questions. For each question, a response of ‘Not very confident/I can’t do this’ scored 0, ‘Quite confident’ scored 1 and ‘Very confident’ scored 2. Scores for the five questions were summed giving a score between 0 and 10. This was further split into three categories of 0–3 (not confident), 4–7 (quite confident) and 8–10 (very confident). For the subsample of responses to the patient surveys where double data entry was performed, the number and percentage of data items where there was disagreement between the primary and double data entry were recorded overall and by question.
Comparisons of the percentage of patients reporting being aware of and using different modes of facilitation in the main patient survey were made between practices who reported using the corresponding mode of facilitation in the practice survey. Similar comparisons were made for patients being aware of, and of using, any mode of facilitation between practices who reported undertaking ad hoc promotion or support and those who did not, and between those who reported employing a practice champion with specific responsibility for supporting or facilitating online services and those who did not. Finally, comparisons were made of awareness and use of any modes of facilitation between patients registered at practices that reported targeting older adults (65 years or older), people with physical health conditions, mental health conditions, limited or no internet access, non-English speakers/English as a second language, ethnic minorities and carers, restricting comparisons to patient respondents reporting that they belonged to that group.
Frequencies and percentages of awareness and use of online services by patient groups were calculated using national GPPS data using weights supplied with the data which accounted for sampling design and non-response bias.
Regression analyses
Main patient survey
Four binary outcomes were created from patient survey responses: (1) awareness of DF; (2) use of DF; (3) being told by someone from the practice about online services and (4) being helped by someone from the practice to use online services. Awareness/use of DF was based on the patient respondent endorsing one or more of the following modes of facilitation in the appropriate question: displays in the practice, leaflets, e-mail/text messages, practice website, social media, workshops/events and making IT equipment available to access online services.
Mixed-effects logistic regression models, with a random effect for practice, were used to examine variability in these outcomes by patient factors. For each of the binary outcomes, univariable associations were calculated along with a multivariable model adjusting for age, gender, deafness/hard of hearing, parental status, ethnicity, physical or mental long-term health condition, working status, first language and whether they had repeat prescriptions (adjusted Model 1). A second model (adjusted Model 2) augmented adjusted Model 1 with the digital confidence variable. A further regression was run using a binary version of the digital confidence scale (confident vs. quite or not confident) as the outcome and the same covariates as adjusted Model 1.
Sensitivity analyses were performed excluding respondents who reported having helped to complete the survey, and for the digital confidence model, using an alternative cut-point of confident or quite confident versus not confident.
Patient survey: mental health boost sample
The regression models performed on the data from the main patient survey were repeated on data from the mental health boost sample survey combined with data from responders to the main patient survey registered at those practices that took part in the mental health boost sample survey. These models were augmented with an additional variable indicating if responders were part of the main sample or the boost sample.
General Practice Patient Survey
Mixed-effects logistic regression models were applied to GPPS data with awareness and use of each of the services as binary outcomes (seven outcomes in total). Models included patient reported age, gender, ethnicity, the presence of long-term condition, not working due to long-term sickness/disability, hearing impairment with use of sign language, as well as deprivation based on national quintiles of IMD. All models included a random intercept to quantify variation between practices.168
Another series of models extended the previous models to include other factors reported in the GPPS and to examine their influence on awareness and use of the online services. These other factors were: ease of getting through on the phone, helpfulness of practice receptionists, patients having a preferred GP and ease of using the GP practice website. The latter was chosen as a proxy for facilitation/enabling efforts of online services. To account for the fact that individuals’ general capacity to use websites in general may influence their response to the question about ease of use of practice websites, we repeated these models using a categorical practice-level version of the question on ease of use of practice website. This was obtained through the quintiles of the random effects of a mixed-effects logistic regression model with the question (Q6) as a binary outcome and adjusted for all known and available confounding factors with awareness and use of online services.
Finally, we explored the potential for DF to impact health inequalities, introducing pairwise interaction terms of ease of use of practice website with each of the demographic and health status variables. These models aimed to indicate whether there was evidence for different levels of association between the proxy for facilitating online services and different patient groups.
National GPPS data from those practices who responded to the practice survey were analysed to ascertain the extent to which the use of DF is associated with patients’ experiences (awareness and use) of online services and the ease of use of the GP website. Comparisons were made between practices using different approaches. Mixed-effects logistic regression models similar to those used to explore variation in awareness and use of online services were augmented with indicator variables derived from the practice survey and restricted responses of patients registered at practices responding to the practice survey.
Complete-case analysis was used throughout. All analyses were performed in Stata v17.0.
Analysis of free-text responses in the patient survey
We examined all the free-text responses for the main patient survey (i.e. not mental health boost sample) to the following questions:
- Q16: What can the practice do to help you access the online services?
- Q17: Is there anything else you would like to add about online services and the support of the practice to help you use them? If so, please tell us about it here.
We conducted a text-based analysis to reflect the topics emerging from the data. One researcher (NK) initially read and re-read all of the responses to familiarise herself with the data. The data were then coded to capture key concepts from the free-text responses. A second researcher (EP) independently read and re-read and double coded the responses. Both researchers met to discuss the initial coding to compare and contrast their findings and to resolve discrepancies. Codes were then organised into higher level themes based on the content and relationship between the codes. We used NVivo 12 (QSR International, Warrington, UK) qualitative analysis software to manage the data.
Findings
Practice survey
Of the 499 practices invited (2 were latterly found to have merged at the time of invitation), 156 (31%) sent back one or more questionnaires (in total 168 questionnaires were returned). This included 19 practices who had participated in the pilot version of the survey. Compared with other practices in England, participating practices were more likely to serve less deprived populations (26% in the least deprived quintile vs. 20% in other practices), to be in rural areas (30% vs. 15% in other practices) and have over 12,000 registered patients (35% vs. 23%). Smaller differences were seen in the age profile and ethnicity of the registered populations (see Appendix 2, Table 33).
Summaries of the responses to the practice survey are shown below (full details Appendix 3, Tables 34–43). Figure 4 and Table 34 show the online services that practices reported offering relative to the current COVID pandemic. All services were currently offered by over 80% of practices with the exception of online access to test results (77%) and online appointment booking (51%). There was a substantial reduction in online appointment booking provision associated with the pandemic (44% of practices offering the service pre-pandemic, but not at the time of the survey, due to temporary suspension). All other services had seen an increase in provision with between 25% (online access to test results) and 86% (video consultations) of practices providing a service that was not available pre-pandemic.
FIGURE 4 Online services offered by practices responding to the practice survey pre coronavirus and at the time of the survey. (Figure not reproduced here; see the original article.)
There was a clear division in the survey between the endorsement of modes of DF that might be described as passive (displays, leaflets, text messages, e-mails, social media and material on practice websites) and those that might be described as active (ad hoc support, employing a practice champion, holding workshops/events and the provision of tablets/computers). Passive modes of facilitation were reportedly used for either promotion of online services and/or supporting patients to use them in a majority of practices (see Figure 5, also Appendix 3, Table 35). In contrast, most active modes of facilitation (employing a practice champion, holding workshops/events and the provision of tablets/computers) were reported by only a minority of practices. An exception to this observation was the provision of ad hoc support which was reported in 91% of practices. Most practices reported using DF across a wide range of online services with 97% reporting using it to promote or support online repeat prescription ordering (see Appendix 3, Table 36).
FIGURE 5 Digital facilitation activities used to promote online services and support patients to use them. (Figure not reproduced here; see the original article.)
The vast majority of practices reported DF being carried out by administrative (86%) and reception (86%) staff, with doctors (62%), nurses (53%) and other healthcare professionals (42%) also frequently reported as being involved. IT staff (28%) volunteers (20%), external contractors (4%) and those with specific responsibilities around DF were reported less often (see Appendix 3, Table 37). Around 22% of practices reported targeting their facilitation activities at every patient group considered, with most practices reporting targeting multiple groups. However, the only specific group targeted by most practices were older adults (56% of practices) (Table 5).
| Patient group | N (%) | Patient group | N (%) |
|---|---|---|---|
| Older adults | 87 (55.8) | Non-English speakers or those for whom English is a second language | 49 (31.4) |
| People with physical health conditions | 68 (43.6) | People from ethnic minority communities | 45 (28.9) |
| People with mental health conditions | 64 (41.0) | Lower-income populations | 38 (24.4) |
| People with limited or no internet access | 45 (28.9) | People with low literacy levels | 37 (23.7) |
| People with limited computer skills | 50 (32.1) | Socially isolated individuals | 53 (34.0) |
| People living in rural communities | 35 (22.4) | Patients with caring responsibilities or patient carers | 67 (43.0) |
| Other | 22 (14.1) |
Most practices reported either reducing or removing in person access to services (including the provision of paper prescriptions) following the national lockdown starting 23 March 2020 (see Appendix 3, Tables 38 and 39). However, although most practices had not changed phone access, 10% of practices reduced access to booking appointments by phone, and 28% reported removing (13%) or reducing (15%) access to ordering repeat prescriptions on the phone.
FIGURE 6 Level of agreement with statements about DF by practices responding to the practice survey. (continued) (Figure not reproduced here; see the original article.)
Most practices agreed in the survey (selecting ‘either agree’ or ‘strongly disagree’ when asked if they agreed with a statement) (see Figure 6, also Appendix 3, Tables 40–42) that:
- COVID had been a crucial driver in the uptake of online services by patients
- COVID had led to an increase in the support they gave to patients to use online services
- it is the practices’ and NHS’s responsibility to inform patients about online services and provide support
- support provided to patients was in response to patient demand rather than from the CCG, practice staff or monetary incentives
- supporting patients to use online services benefited the practice and patients
- they lacked the adequate capacity to provide this support
- they had increased the uptake of online services by supporting patients to use them and that they hoped to further increase uptake through support
- some patients were unlikely to use online services regardless of the support provided
- some patient groups required more support than others
- increased patient uptake of online services led to operational efficiencies for the practice
- online access to primary care services was complementary to traditional forms of access.
Analysis of free-text comments made by (staff) responders throughout the practice survey identified two main themes which illustrated issues reported:
- COVID had had an impact on digital services: many services made a shift to being mainly online. However, COVID also curtailed a few attempts to provide support (e.g. provision of workshops prior to COVID).
- There was a relatively negative attitude to implementing digital services in terms of limited practice capacity, finances and motivation:
GP practices do not have the capacity or resources to facilitate online access.
- There was a recognition that guidance, support and funding from CCGs or at the national level was required in order to provide more help to patients:
[F]unding would be useful to provide tablets/pc for waiting rooms where patients could access online/self- help/information/advice.
It makes sense for support to be delivered at a national level and outside of GP practice, as is the case with the NHS App.
Patient survey
Sixty-two practices participated in the patient survey, sending invitations to 12,822 patients, of whom 3051 (23.8%) responded (see Appendix 5, Table 49). Analysis of double data entry detected a difference of 0.4% in data points entered which was deemed acceptable. Response rates were lower in practices serving deprived populations; however, due to the approach to sampling we adopted, responses were reasonably representative in terms of deprivation (Figure 7).
FIGURE 7 Main patient survey: invitations and response by deprivation quintile. (Figure not reproduced here; see the original article.)
Of responders to the main patient survey, slightly more were female than male, 45% were 65 or older, the vast majority were white (93%), 9% reported that English was not their first language, 44% were working either full-time or part-time, and 43% were retired (Table 6).
| Demographic | Total (%)(n = 3051) | Demographic | Total (%)(n = 3051) | ||
|---|---|---|---|---|---|
| Gender | Male | 1312 (43.4) | Working status | Full-time paid | 906 (31.2) |
| Female | 1710 (56.6) | Part-time paid | 363 (12.5) | ||
| Age (years) | 16–24 | 140 (4.6) | Full-time education | 75 (2.6) | |
| 25–34 | 195 (6.4) | Unemployed | 59 (2.0) | ||
| 35–44 | 262 (8.7) | Permanently sick | 72 (2.5) | ||
| 45–54 | 436 (14.4) | Fully retired | 1244 (42.8) | ||
| 55–64 | 619 (20.5) | Looking after family | 104 (3.6) | ||
| 65–74 | 727 (24.0) | Doing something else | 75 (2.6) | ||
| 75–84 | 515 (17.0) | Furloughed | 6 (0.2) | ||
| ≥ 85 | 131 (4.3) | ||||
| Deaf/hearing impediment | Yes | 358 (12.0) | English as first language | Yes | 2705 (91.2) |
| No | 2631 (88.0) | No | 261 (8.8) | ||
| Blind/partially sighted | Yes | 54 (1.8) | Carer | Yes | 700 (23.8) |
| No | 2909 (98.2) | No | 2237 (76.2) | ||
| Ethnicity | White | 2741 (92.7) | Parent | Yes | 420 (14.2) |
| Mixed | 31 (1.1) | No | 2536 (85.8) | ||
| Asian/Asian British | 107 (3.6) | ||||
| Black/African/Caribbean/black British | 57 (2.0) | Repeat prescriptions | Yes | 1957 (65.7) | |
| No | 1021 (34.3) | ||||
| Help to complete the survey | Yes | 183 (6.2) | |||
| Other | 22 (0.7) | No | 2780 (93.8) | ||
| Mental health condition/disabilities | Yes | 472 (16.4) | Physical health condition/disabilities | Yes | 1046 (36.1) |
| No | 2281 (79.3) | No | 1737 (60.0) | ||
| Don’t know | 72 (2.5) | Don’t know | 74 (2.6) | ||
| Prefer not to say | 51 (1.8) | Prefer not to say | 43 (1.5) |
Nine per cent of responders reported not having access to the internet at home, and nearly all of these completed a paper version of the survey (see Appendix 5, Table 50). The percentage of responders not confident or unable to complete digital tasks ranged from 17% for using search engines to 31% for installing apps (see Appendix 5, Table 50). The lack of confidence was strongly associated with response mode, with at most 4% of online responders lacking confidence in these tasks.
Figure 8 shows a clear bimodal distribution in the digital confidence score constructed from responses to questions 2–6. Almost 40% of responders scored 10, implying they feel very confident in undertaking all of the tasks considered, while around 12% of responders scored zero, implying that they are not very confident/cannot complete any of the tasks considered.
FIGURE 8 Distribution of digital confidence score for responders to the patient survey. (Figure not reproduced here; see the original article.)
Table 7 provides a summary of responses to the core patient survey questions (i.e. not those concerning digital confidence or patient characteristics). Over a third of responders (37%) had not tried to use their general practice’s website, though most of those that had done so found it easy to use. Awareness of online services was generally low (≤ 37% of respondents were aware) for all online services apart from booking appointments and ordering repeat prescriptions online (55% and 64% respectively) and use of online services was universally low with at most 33% of patients reporting having ordered repeat prescriptions online. Awareness of DF efforts was low, with 17% or less being aware of any modes apart from texts or e-mails for which 39% of responders were aware of their use. The use of those facilitation efforts was even lower, with < 10% of patients making use of any mode apart from e-mail or text messages (28%).
| Questions | Total (%) (n = 3051) | Questions | Total (%) (n = 3051) |
|---|---|---|---|
| Q7. Ease of practice website use | Q10. Awareness of facilitation | ||
| Very easy | 541 (18.0) | Displays | 522 (17.1) |
| Fairly easy | 920 (30.7) | Leaflets | 178 (6.0) |
| Not very easy | 286 (9.5) | Text/e-mails | 1205 (39.4) |
| Not at all easy | 145 (4.8) | Web content | 229 (7.5) |
| Haven’t tried | 1106 (37.0) | Social media | 118 (4.0) |
| Q8. Awareness of online services | Workshop/events | 17 (0.6) | |
| Appointment booking | 1675 (55.0) | Tablets/computers | 16 (0.5) |
| Repeat prescriptions | 1944 (63.7) | Q11. Use of facilitation | |
| Access records | 945 (31.0) | Displays | 291 (9.5) |
| Access test results | 663 (21.7) | Leaflets | 93 (3.0) |
| E-mail enquiries | 1126 (37.0) | Text/e-mails | 860 (28.2) |
| eConsults | 945 (31.0) | Web content | 152 (5.0) |
| Video consults | 458 (15.0) | Social media | 79 (2.6) |
| Q9. Use of online services | Workshop/events | 15 (0.5) | |
| Appointment booking | 469 (15.4) | Tablets/computers | 15 (0.5) |
| Repeat prescriptions | 1003 (33.0) | Q12. Told about online services | |
| Access records | 393 (13.0) | Yes | 883 (29.8) |
| Access test results | 260 (8.5) | No | 2801 (70.2) |
| 457 (15.0) | Q13. Helped to use online services | ||
| eConsults | 501 (16.4) | Yes | 392 (13.4) |
| Video consults | 118 (3.8) | No | 2543 (86.6) |
| Q14. Usefulness of help | Q14. Usefulness of help | ||
| 14_1. Booking appointments online | 196 (6.4) | 14_7. Online video consultations | 81 (2.7) |
| Not helpful | 40 (20.4) | Not helpful | 28 (34.6) |
| Quite helpful | 77 (39.3) | Quite helpful | 19 (23.5) |
| Very helpful | 79 (40.3) | Very helpful | 34 (42.0) |
| 14_2. Ordering repeat prescriptions online | 211 (7.0) | 14_8. Other | 33 (1.1) |
| Not helpful | 22 (10.4) | Not helpful | 10 (30.3) |
| Quite helpful | 66 (31.3) | Quite helpful | 8 (24.2) |
| Very helpful | 123 (58.3) | Very helpful | 15 (45.5) |
| 14_3. Accessing medical records online | 121 (4.0) | Q15. Reasons why not using online services | |
| Not helpful | 24 (19.8) | No internet access | 247 (8.1) |
| Quite helpful | 42 (34.7) | Security concern | 228 (7.5) |
| Very helpful | 55 (45.5) | Confidentiality | 145 (4.8) |
| 14_4. Accessing test results online | 111 (3.6) | Not knowing how to register | 440 (14.4) |
| Not helpful | 23 (20.7) | Registration too difficult | 175 (5.7) |
| Quite helpful | 34 (30.6) | Not knowing how to get support | 259 (8.5) |
| Very helpful | 54 (48.6) | Practice too busy to help | 262 (8.6) |
| 14_5. E-mail enquiries to the practice | 154 (5.0) | Prefer to speak to person | 1432 (47.0) |
| Not helpful | 27 (17.5) | Other (with comments): | 811 (27.1) |
| Quite helpful | 53 (34.4) | Unable to get what I need online | 73 (2.4) |
| Very helpful | 74 (48.1) | Unable to access (e.g. no phone or computer, non-English speaking, errors, physical, learning or mental disabilities)Not aware of other online services than repeat prescriptions/booking appointments or of online services per se | 60 (2.0) |
| 14_6. Help from GP using online form | 171 (5.6) | ||
| Not helpful | 28 (16.4) | 55 (1.8) | |
| Quite helpful | 59 (34.5) | ||
| Very helpful | 84 (49.1) | ||
Only 30% of patients reported that they had been told about online services by someone at their general practice and only 13% reported being helped to use them. However, of those that did have help, they generally rated it as helpful (90% said ‘quite’ or ‘very helpful’ for repeat prescriptions, and 65% said ‘quite’ or ‘very helpful’ for video consultations). Various reasons were selected for not using online services, but most common was preferring to speak to a person (47%) with 20% either not knowing how to register (14%) or finding registration too difficult (6%).
The adjusted (Model 1) logistic regression analyses are shown in Table 8 with unadjusted results shown in Appendix 4 (see Table 44). Age was positively associated with a lower likelihood of all four outcomes (p < 0.05 for all), particularly for being helped to use online services (adjusted OR for ≥ 85 vs. 55- to 65-year-olds 0.08, 95% CI 0.02 to 0.36; p = 0.006). Patients in receipt of repeat prescriptions were more likely to have experienced all four facilitation outcomes (p < 0.05), for example, the adjusted odds ratio (OR) for being told about online services was 1.70 when compared with those not in receipt of repeat prescriptions (95% CI 1.36 to 2.13; p < 0.001). Patients of non-white ethnicity were more likely than those patients describing themselves as white to be aware of DF, have made use of DF and been helped to use online services, but no difference was seen between white and non-white patients for being told about online services (adjusted OR = 0.91, 95% CI 0.61 to 1.36; p = 0.651). Similarly, people for whom English was not their first language were more likely to use DF and be told about online services, and with very weak evidence (p < 0.1) of similar differences for the other two outcomes. Gender, being deaf or having a hearing impairment, and working status were not associated with any of the four outcomes (p > 0.1). Including the digital confidence measure in the models (see Appendix 4, Table 45) showed that lower digital confidence was associated with lower awareness and use of facilitation, and a lower chance of being told about, or being helped to use, online services. Comparing Model 1 with 2 we found that digital confidence explained some of the difference attributable to older age for awareness and use of facilitation and being told about online services. However, the reductions in odds ratios were fairly small (e.g. OR for 85 + awareness of facilitation reduced from 0.32 to 0.59). In contrast, digital confidence hid some difference attributable to older age in being given help indicated by ORs becoming larger when adjustment for digital confidence was made. Digital confidence also masked some differences in the four outcomes between those who spoke English as a first language and those who did not. Sensitivity analyses excluding those who reported having helped to complete the questionnaire did not lead to materially different findings (see Appendix 4, Tables 46 and 47).
| Awareness of any facilitation effortsa | Use of any facilitation effortsb | Being told about online services | Being helped to use online services | |||||
|---|---|---|---|---|---|---|---|---|
| Adjusted OR (95% CI) | p-value | Adjusted OR (95% CI) | p-value | Adjusted OR (95% CI) | p-value | Adjusted OR (95% CI) | p-value | |
| Q18. Gender | ||||||||
| Male | Reference | |||||||
| Female | 0.92 (0.78 to 1.09) | 0.342 | 0.90 (0.76 to 1.07) | 0.226 | 1.14 (0.95 to 1.36) | 0.161 | 0.86 (0.68 to 1.10) | 0.237 |
| Q19. Age | ||||||||
| 16–24 | 0.84 (0.50 to 1.42) | < 0.001 | 0.64 (0.37 to 1.09) | < 0.001 | 1.58 (0.93 to 2.70) | 0.043 | 1.00 (0.47 to 2.17) | 0.006 |
| 25–34 | 1.08 (0.73 to 1.58) | 1.07 (0.74 to 1.56) | 1.14 (0.77 to 1.69) | 1.11 (0.65 to 1.91) | ||||
| 35–44 | 0.98 (0.67 to 1.43) | 1.01 (0.70 to 1.46) | 1.08 (0.73 to 1.60) | 1.42 (0.85 to 2.37) | ||||
| 45–54 | 0.94 (0.70 to 1.27) | 0.94 (0.70 to 1.26) | 1.01 (0.74 to 1.38) | 1.31 (0.87 to 1.98) | ||||
| 55–64 | Reference | |||||||
| 65–74 | 0.71 (0.53 to 0.95) | 0.88 (0.66 to 1.17) | 0.77 (0.56 to 1.05) | 0.81 (0.52 to 1.24) | ||||
| 75–84 | 0.43 (0.30 to 0.60) | 0.62 (0.44 to 0.87) | 0.66 (0.45 to 0.95) | 0.50 (0.30 to 0.84) | ||||
| ≥ 85 | 0.32 (0.19 to 0.54) | 0.27 (0.15 to 0.49) | 0.38 (0.20 to 0.71) | 0.08 (0.02 to 0.36) | ||||
| Q20. Deaf/hearing impairment | ||||||||
| No | Reference | |||||||
| Yes | 0.93 (0.70 to 1.22) | 0.595 | 1.16 (0.87 to 1.53) | 0.314 | 0.82 (0.60 to 1.13) | 0.229 | 1.32 (0.88 to 2.00) | 0.184 |
| Q22. Parent | ||||||||
| No | Reference | |||||||
| Yes | 1.28 (0.96 to 1.72) | 0.095 | 1.03 (0.78 to 1.37) | 0.831 | 1.28 (0.95 to 1.73) | 0.105 | 1.03 (0.70 to 1.52) | 0.880 |
| Q23. Ethnicity | ||||||||
| White | Reference | |||||||
| Other | 1.48 (1.00 to 2.20) | 0.050 | 1.48 (1.03 to 2.15) | 0.036 | 0.91 (0.61 to 1.36) | 0.651 | 1.80 (1.14 to 2.86) | 0.012 |
| Q24–25. Long-term physical or mental health condition | ||||||||
| No | Reference | |||||||
| Yes | 1.25 (1.03 to 1.50) | 0.021 | 1.17 (0.97 to 1.41) | 0.092 | 1.07 (0.88 to 1.31) | 0.478 | 1.29 (0.98 to 1.70) | 0.069 |
| Q26. Working status | ||||||||
| Work | Reference | |||||||
| Education | 0.73 (0.38 to 1.41) | 0.118 | 1.05 (0.54 to 2.04) | 0.186 | 1.09 (0.56 to 2.12) | 0.978 | 1.23 (0.49 to 3.10) | 0.623 |
| Other | 0.72 (0.54 to 0.95) | 0.78 (0.59 to 1.04) | 0.95 (0.70 to 1.28) | 1.23 (0.83 to 1.82) | ||||
| Retired | 0.96 (0.72 to 1.26) | 0.77 (0.59 to 1.02) | 1.00 (0.74 to 1.35) | 1.24 (0.82 to 1.89) | ||||
| Q27. First language | ||||||||
| English | Reference | |||||||
| Other | 1.38 (0.96 to 1.96) | 0.078 | 1.79 (1.28 to 2.52) | 0.001 | 1.61 (1.12 to 2.32) | 0.010 | 1.52 (0.97 to 2.37) | 0.068 |
| Q29. Repeat prescription | ||||||||
| No | Reference | |||||||
| Yes | 1.28 (1.04 to 1.56) | 0.019 | 1.46 (1.19 to 1.79) | < 0.001 | 1.70 (1.36 to 2.13) | < 0.001 | 1.45 (1.06 to 1.97) | 0.018 |
- Awareness of any facilitation efforts includes respondents who ticked any of the first seven options of Q10.
- Use of any facilitation efforts includes respondents who ticked any of the first seven options of Q11.
Considering digital confidence as an outcome (see Appendix 4, Table 48), we found a strong age gradient, with over 85-year-olds much less likely to be very confident than 55- to 64-year-olds (adjusted OR = 0.04, 95% CI 0.01 to 0.11; p < 0.001). We also found that parents were much more likely to be digitally confident than non-parents. People with long-term health problems, whose first language is not English, who are retired, or not in work/education were less likely to be digitally confident.
Patient survey free text
A total of 2246 responses were provided to the free-text questions 16 and 17 (Table 9).
| Q16 free-text responses | Q17 free-text responses | Total | |
|---|---|---|---|
| Paper survey | 905 | 695 | 1600 |
| Online survey | 396 | 250 | 646 |
| Total | 1301 | 945 | 2246 |
Respondents highlighted the following six areas of importance relating to DF in primary care:
- Personal barriers to accessing digital services
- Respondents described barriers including age, poor eyesight, learning difficulties, lack of internet literacy and poor access to computers or smartphones as limiting their ability to access digital services.
- Practice barriers to accessing digital services
- Several respondents felt that their practice was not supportive of online services or discouraged patients from using them. Other respondents felt that they needed help from practice staff to help with getting access to and using the digital services but commented that this help was not routinely offered.
- How concerns about data security impacted their views on digital services
- Some participants raised concerns about data confidentiality, data security and fears about online ‘scams’ and how this impacted their trust in the digital services offered by the practice.
- How technology design impacted on access
- Several respondents described that they found practice websites and mobile applications confusing and difficult to navigate, and this limited the use of the services.
- Barriers in getting the right online credentials to access digital services
- Several respondents found the process of getting access to online services difficult, especially in terms of verifying their identity with the practice and getting the correct login codes.
- Potential facilitators to adopting digital technology
Respondents wanted to have more visibility about what was on offer from their practice, and also described unmet training needs about how to access and use online services.
The themes are described in more detail in Appendix 5.
Combined analyses of the practice survey and main patient survey
We found only limited evidence of differences between the reported awareness, and use of, different modes of DF (as reported in the patient survey) by patients registered at practices who reported using the corresponding mode for either promotion of, or supporting patients to use, online services (as reported in the practice survey – Table 10). There was strong evidence (p = 0.004) that patients registered at practices who reported using in-practice displays to support DF were more likely to be aware of such displays (18% vs. 11% of patients). Similarly, awareness and use of social media for DF were higher in those practices that reported using it (6% vs. 1% for awareness and 4% vs. 1% for use) and awareness and use of DF workshops/events were higher in those practices that reported using them (1.5% vs. 0.5% for awareness and 1.5% vs. 0.4% for use). No evidence of differences in patient awareness and use was seen between practices reporting using leaflets, text messages/e-mails, putting content on practice websites and providing tablets or computers, and those that did not. There was no evidence of differences in the patient awareness and use of any mode of DF between those practices that reported using ad hoc support for DF and those that did not (see Appendix 6, Table 51). Employing a practice champion was associated with higher patient use of facilitation but no greater awareness of that form of facilitation. Furthermore, there was no evidence of differences in the patient awareness and use of any mode of DF within any particular patient groups where practices reported targeting that group (see Appendix 6, Table 52).
| Mode of facilitation | % patients aware in practices using mode of facilitation | % patients aware in practices not using mode of facilitation | p-value | % patients using in practices using mode of facilitation | % patients using in practices not using mode of facilitation | p-value |
|---|---|---|---|---|---|---|
| Displays (N = 2935) | 17.55% (466/2655) | 10.71% (30/280) | 0.004 | 9.64% (256/2655) | 7.50% (21/280) | 0.244 |
| Leaflets (N = 2776) | 5.39% (92/1707) | 6.27% (67/1069) | 0.333 | 2.75% (47/1707) | 3.18% (34/1069) | 0.515 |
| Text/e-mails (N = 2828) | 39.14% (1068/2729) | 43.43% (43/99) | 0.390 | 28.22% (770/2729) | 25.25% (25/99) | 0.519 |
| Web content (N = 3032) | 7.43% (197/2650) | 8.38% (32/382) | 0.514 | 5.09% (135/2650) | 4.45% (17/382) | 0.590 |
| Social media (N = 2975) | 6.06% (105/1733) | 0.97% (12/1242) | < 0.001 | 3.64% (63/1733) | 1.21% (15/1242) | < 0.001 |
| Workshops/events (N = 2874) | 1.50% (4/267) | 0.46% (12/2607) | 0.030 | 1.50% (4/267) | 0.38% (10/2607) | 0.013 |
| Tablets/computers (N = 2921) | 0.65% (2/306) | 0.54% (14/2615) | 0.791 | 0.98% (3/306) | 0.42% (11/2615) | 0.180 |
Mental health boost sample
Of the 26 practices (who had already participated in the main patient survey) invited to participate in the mental health boost sample element of the patient survey, 18 agreed to take part (3 serving deprived populations). In total, 3456 patients were sent questionnaires (maximum of 200 per practice) of which 756 (21.85%) responded (see Appendix 7, Table 53). Analysis of double data entry detected a difference of 0.4% in data points entered.
Table 11 shows the demographics of responders to the patient survey mental health boost sample along with the responders to the main survey registered at the same general practices. Sixty-one per cent of those responding to the mental health boost sample survey reported that they had a long-term mental health condition. A higher percentage of responders to the mental health boost sample survey were female, aged 25–64, reported a long-term physical health condition and were much more likely to be in receipt of repeat prescriptions (94% vs. 66% in the main survey).
| Characteristics | Main patient survey (n = 944) total (%) | Mental health survey (n = 756) total (%) | |
|---|---|---|---|
| Gender | Male | 388 (41.4) | 231 (30.8) |
| Female | 550 (58.6) | 519 (69.2) | |
| Age | 16–24 | 45 (4.8) | 30 (4.0) |
| 25–34 | 49 (5.2) | 49 (6.5) | |
| 35–44 | 78 (8.3) | 83 (11.1) | |
| 45–54 | 136 (14.5) | 134 (18.0) | |
| 55–64 | 205 (22.0) | 174 (23.2) | |
| 65–74 | 228 (24.3) | 154 (20.5) | |
| 75–84 | 159 (17.0) | 98 (13.0) | |
| ≥85 | 38 (4.1) | 29 (4.0) | |
| Deaf/hearing impairment | Yes | 99 (10.7) | 100 (13.5) |
| No | 829 (89.3) | 642 (86.5) | |
| Blind/partially sighted | Yes | 21 (2.3) | 23 (3.1) |
| No | 899 (97.7) | 717(96.9) | |
| Parent | Yes | 120 (13.0) | 104 (14.1) |
| No | 800 (87.0) | 636 (85.9) | |
| Ethnicity | White | 892 (97.4) | 732 (98.3) |
| Other | 24 (2.6) | 13 (1.7) | |
| Mental health condition | Yes | 148 (17.5) | 402 (60.5) |
| No | 698 (82.5) | 262 (39.4) | |
| Physical health condition | Yes | 346 (39.8) | 371 (54.0) |
| No | 523 (60.2) | 317 (46.1) | |
| Working status | Work | 396 (44.0) | 312 (43.6) |
| Education | 22 (2.4) | 13 (1.8) | |
| Retired | 399 (44.3) | 249 (34.8) | |
| Other | 83 (9.2) | 141 (19.7) | |
| English as first language | Yes | 857 (93.7) | 721 (97.0) |
| No | 55 (6.3) | 22 (3.0) | |
| Carer | Yes | 205 (22.8) | 204 (27.7) |
| No | 695 (77.2) | 532 (72.3) | |
| Repeat prescriptions | Yes | 603 (65.6) | 703 (94.1) |
| No | 316 (34.4) | 44 (5.9) | |
| Help to complete survey | Yes | 47 (5.1) | 68 (9.1) |
| No | 871 (94.9) | 678 (90.9) | |
A summary of responses given by respondents to the patient survey mental health boost sample are shown in Table 12. Responders to the mental health boost sample survey reported higher levels of digital confidence, and they were considerably more likely to have used practice websites than responders to the main survey. Reported awareness and use of online services were generally higher for those responding to the mental health boost sample survey compared with those responding to the main survey (with the exception of awareness of booking appointments online) as was awareness and use of DF efforts (with the notable exception of social media). Responders to the mental health boost sample survey were also more likely to have been told about online services and helped to use them. Finally, a higher percentage of responders to the mental health boost sample survey endorsed all reasons for not using online services, with the exception of not having internet access.
| Characteristics | Main patient survey (n = 944) total (%) | Mental health survey (n = 756) total (%) |
|---|---|---|
| Digital confidence | ||
| Very confident | 488 (53.0) | 416 (56.0) |
| Quite confident | 219 (23.7) | 163 (22.0) |
| Not confident | 215 (23.3) | 165 (22.1) |
| Ease of practice website use | ||
| Very easy | 156 (16.8) | 167 (22.1) |
| Fairly easy | 291 (31.3) | 251 (33.3) |
| Not very easy | 89 (9.6) | 91 (12.1) |
| Not at all easy | 34 (3.7) | 52 (6.9) |
| Haven’t tried | 359 (38.6) | 193 (25.6) |
| Awareness of online services | ||
| Appointment booking | 518 (55.0) | 398 (52.6) |
| Repeat prescriptions | 621 (65.8) | 560 (74.1) |
| Access records | 280 (29.7) | 286 (37.8) |
| Access test results | 205 (21.7) | 215 (28.4) |
| E-mail enquiries | 340 (36.0) | 342 (45.2) |
| eConsults | 272 (28.8) | 302 (40.0) |
| Video consults | 124 (13.1) | 116 (15.3) |
| Use of online services | ||
| Appointment booking | 125 (13.2) | 131 (17.3) |
| Repeat prescriptions | 301 (32.0) | 380 (50.3) |
| Access records | 109 (11.5) | 137 (18.1) |
| Access test results | 71 (7.5) | 109 (14.4) |
| E-mail enquiries | 152 (16.1) | 152 (20.1) |
| eConsults | 167 (17.7) | 187 (24.7) |
| Video consults | 29 (3.1) | 32 (4.2) |
| Awareness of facilitation | ||
| Displays in the practice | 163 (17.3) | 193 (25.5) |
| Leaflets | 53 (5.6) | 64 (8.5) |
| Text/e-mails | 336 (35.6) | 324 (43.0) |
| Web content | 65 (7.0) | 74 (9.8) |
| Social media | 28 (3.0) | 21 (2.8) |
| Workshop/events | 5 (0.5) | 13 (1.7) |
| Tablets/computers | 5 (0.5) | 8 (1.1) |
| Use of facilitation | ||
| Displays in the practice | 89 (9.4) | 85 (11.2) |
| Leaflets | 28 (3.0) | 24 (3.2) |
| Text/e-mails | 238 (25.2) | 224 (29.6) |
| Web content | 41 (4.3) | 46 (6.1) |
| Social media | 23 (2.4) | 12 (1.5) |
| Workshop/events | 5 (0.5) | 4 (0.5) |
| Tablets/computers | 6 (0.6) | 3 (0.4) |
| Told about online services | ||
| Yes | 286 (31.2) | 285 (38.6) |
| No | 631 (68.8) | 453 (61.4) |
| Helped to use online services | ||
| Yes | 123 (13.6) | 112 (15.1) |
| No | 784 (86.4) | 626 (84.8) |
| Usefulness of online help | ||
| Booking appointments | 66 (7.0) | 60 (8.1) |
| Not helpful | 12 (18.2) | 14 (23.3) |
| Quite helpful | 25 (38.8) | 24 (40.0) |
| Very helpful | 29 (44.0) | 22 (36.7) |
| Ordering repeat prescriptions | 70 (7.4) | 66 (9.0) |
| Not helpful | 5 (7.1) | 5 (7.6) |
| Quite helpful | 18 (25.7) | 16 (24.2) |
| Very helpful | 47 (67.1) | 45 (68.2) |
| Accessing medical records | 43 (4.6) | 45 (4.0) |
| Not helpful | 10 (23.3) | 12 (26.7) |
| Quite helpful | 14 (32.6) | 9 (20.0) |
| Very helpful | 19 (44.2) | 24 (53.3) |
| Accessing test results online | 38 (4.0) | 39 (6.1) |
| Not helpful | 12 (31.6) | 10 (25.6) |
| Quite helpful | 9 (23.7) | 12 (30.8) |
| Very helpful | 17 (44.7) | 17 (43.6) |
| E-mail enquiries to the practice | 44 (4.7) | 59 (8.0) |
| Not helpful | 11 (25) | 13 (22.0) |
| Quite helpful | 15 (34.1) | 14 (23.7) |
| Very helpful | 18 (41.0) | 32 (54.2) |
| Help from GP using online form | 56 (6.0) | 58 (7.8) |
| Not helpful | 10 (17.8) | 11 (19.0) |
| Quite helpful | 18 (32.1) | 17 (29.3) |
| Very helpful | 28 (50.0) | 30 (51.7) |
| Online video consultations | 23 (2.4) | 23 (3.1) |
| Not helpful | 11 (47.8) | 8 (34.8) |
| Quite helpful | 4 (17.4) | 2 (8.7) |
| Very helpful | 8 (34.8) | 13 (56.5) |
| Reasons why not using online services | ||
| No internet access | 76 (8.1) | 52 (7.0) |
| Security concern | 60 (6.4) | 61 (8.1) |
| Confidentiality | 44 (4.7) | 50 (6.6) |
| Not knowing how to register | 136 (14.4) | 100 (13.2) |
| Registration too difficult | 45 (4.8) | 56 (7.4) |
| Not knowing how to get support | 61 (6.5) | 77 (10.2) |
| Practice too busy to help | 82 (8.7) | 97 (12.8) |
| Prefer to speak to person | 439 (46.5) | 379 (50.1) |
The higher awareness and use of facilitation efforts by responders to the mental health boost sample survey, along with increased likelihood of being told about online services, were confirmed by univariable logistic regression (see Appendix 7, Table 53), but the difference in terms of being helped to use online services was not statistically significant. In adjusted models the only difference between responders to the mental health boost sample survey and the main survey that remained statistically significant was awareness of facilitation (see Appendix 7, Tables 54 and 55) indicating that observed differences in use of facilitation and being told about digital services may be related to differences in the demographics of the samples.
Findings from analysis of national General Practitioner Patient Survey data
Of the 1,510,149 responders to the 2019 and 2020 national GPPSs and the 1,569,343 responders to the 2021 and 2022 national GPPSs, 1,212,801 (77.3%) and 1,304,574 (86.4%) reported trying to make an appointment at their GP practice in the 12 months prior to responding to the survey, respectively.
Among this subset of responders, awareness of online appointment booking and ordering repeat prescriptions online was moderate (2019–20 percentages weighted for non-response = 49% and 45%, respectively), but awareness of the ability to access medical records online was low (2019–20 weighted percentage = 18%). Use of all three services (based on 2021–2 data) was lower at 23%, 32% and 13%, respectively. Around a quarter (25% in 2021–2) of responders reported use of online consultations.
The results of the adjusted logistic regression models are shown in Figures 9–13 and in Table 56, Appendix 8. There was a strong decrease in the awareness and use of all services (making appointments online, ordering repeat prescriptions online, accessing medical records online, and online consultations) with older age; individuals aged over 85 years were much less likely to be aware of, or to report using, online services than younger people (see Figure 9). For ordering repeat prescriptions online, there was a smaller reduction in awareness and use in younger responders than in 55- to 64-year-olds, but for the other services awareness plateaued, and use increased, in all but the youngest responders. Similarly, there were very strong deprivation gradients in the awareness and use of all online services (see Figure 10). Females were more likely to be aware of all three services considered (appointment booking, repeat prescriptions and access to medical records) and more likely to use online consultations and order repeat prescriptions online, but less likely to actually book appointments online or access medical records online (see Figure 11). All non-white ethnic groups were less aware of all three online services that were considered than white responders and were generally less likely to actually use the services, with the exception of booking appointments online where higher use among ethnic minorities was seen (see Figure 12). Those with long-term conditions were more likely to be aware of and use all online services, particularly ordering repeat prescriptions (see Figure 13). Finally, we found that the practice in which a patient was registered was, by far, the strongest predictor of awareness and use of online services (Table 13). For example, the OR comparing the practice with the highest awareness of online booking with the lowest among the middle 95% of practices (i.e. ignoring the most extreme) was 16.2 (95% CI 15.2 to 17.0).
FIGURE 9 Variation of awareness (red) and use (blue) of online services by age obtained from adjusted logistic regression models (odds ratios and 95% CIs). (Figure not reproduced here; see the original article.)
FIGURE 10 Variation of awareness (red) and use (blue) of online services by deprivation obtained from adjusted logistic regression models (odds ratios and 95% CIs). (Figure not reproduced here; see the original article.)
FIGURE 11 Variation of awareness (red) and use (blue) of online services by gender obtained from adjusted logistic regression models (odds ratios and 95% CIs). (Figure not reproduced here; see the original article.)
FIGURE 12 Variation of awareness (red) and use (blue) online services by ethnicity obtained from adjusted logistic regression models (odds ratios and 95% CIs). (Figure not reproduced here; see the original article.)
FIGURE 13 Variation of awareness (red) and use (blue) of online services by long-term condition status obtained from adjusted logistic regression models (odds ratios and 95% CIs). (Figure not reproduced here; see the original article.)
| Online booking appointment | Online repeat prescriptions | Online medical records | Online consultations | |||||
|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | |
| Awareness | 16.18 | (15.34 to 17.02) | 8.40 | (8.07 to 8.75) | 12.32 | (11.7 to 12.95) | N/A | N/A |
| Use | 11.85 | (11.30 to 12.44) | 5.80 | (5.60 to 6.02) | 8.47 | (8.08 to 8.89) | 7.03 | (6.76 to 7.32) |
Augmenting the regression models with patients’ experiences of other aspects of the practice (see Appendix 8, Table 57) found that:
- Patients who say it is easy to get through on the phone tend to have lower awareness and use of online services than those who say it is not at all easy.
- Patients who report receptionists being helpful tend to have higher awareness but lower use of online services, than those who report receptionists being not at all helpful.
- Awareness and use of all online services is most often reported for patients who have a preferred GP for some appointments. With the exception of online consultation use, this is followed by those who report a preferred GP for all appointments, with those who do not have a preferred GP least likely to report awareness or use of online services.
- Patients who say the practice website is easy to use are much more likely to report awareness and use of online services.
The effect of ease of use of practice websites was particularly strong within the models of awareness and use of online provision; the ORs comparing those who rated practice websites as ‘very easy’ to use compared with those who said it was ‘not easy at all’ being 5.29 (5.12, 5.48), 5.73 (5.54, 5.93) and 4.07 (3.92, 4.24) for awareness of online appointment booking, ordering repeat prescriptions and accessing medical records online, respectively. Because these ORs may reflect individuals’ general difficulty with using websites we also repeated models replacing person-level responses with a practice score for ease of website use. In that case the odds ratios were more modest, but still substantial at 1.40 (95% CI 1.32 to 1.47), 1.51 (95% CI 1.45 to 1.58) and 1.45 (95% CI 1.38 to 1.52), respectively.
In the final set of models examining GPPS data in isolation (see Appendix 8, Table 58) we looked at whether reported ease of use of practice websites modified the effect of sociodemographic factors. We found evidence of interactions between age, deprivation, gender, having a long-term condition, rurality and one or more of awareness of online appointment booking, use of online appointment booking, or awareness of repeat prescriptions. In all cases good websites were associated with higher awareness and use and the direction of demographic associations did not change for when practices had good or poor websites, though the magnitude of the associations did change, but only to a modest extent. Age, ethnicity, long-term condition and rurality differences were slightly larger when websites were rated as good compared with when they were rated poorly. In contrast, gender and deprivation differences were slightly smaller when websites were rated as good compared with when they were rated poorly.
The results of the analysis of GPPS data from practices who participated in the WP2 practice survey are shown in Appendix 8, Table 59. After adjustment for patient characteristics, we find that 7 out of 63 associations examined are statistically significant at the p < 0.05 level and as such care must be taken with the interpretation of these findings. However, we do see a consistent effect whereby patients registered at practices who report the use of leaflets for DF are less likely to report being aware of online services (appointment booking, medical record access and repeat prescriptions) than those registered at practices that did not report use of leaflets.
Patient and public involvement and engagement
Box 4 describes how the PAG influenced WP2.
Summary of main findings: work package 2
Practice survey
Most practices reported offering a range of online services. With the notable exception of online appointment booking (which was withdrawn nationally in response to the COVID pandemic), the number of practices offering them had substantially increased during the COVID pandemic (particularly video consultations). While most practices reported using passive modes of DF to promote or support patients’ use of online services (social media), more active modes of DF (e.g. employing a practice champion), were rarely reported. In the vast majority of practices, administrative and reception staff conducted DF efforts, with clinical staff contributing in more than half of practices. Most practices reported targeting their facilitation efforts at older adults. While most practices were generally positive about DF, many of them felt that they lacked the capacity to deliver it.
Patient survey
Among responders to the main survey there was bimodal distribution in digital confidence, with many patients indicating very high confidence, but a substantial number indicating very low confidence – highlighting a group of patients who lack the capability to use digital tools unaided. Furthermore, around a third of patients had not used their general practice’s website. While more than half of the patients were aware of online appointment booking and potential for ordering repeat prescriptions online, awareness of other online services (e.g. accessing test results) was much lower, and use of online services was low apart from online appointment booking. Patients endorsed a range of reasons for not engaging with online services, in particular reporting their preference to speak to someone and also reporting having difficulties registering to use online services. Such difficulties indicate the potential lack of effective DF. Older patients, and patients with lower digital confidence, were less likely to be aware of or to use DF efforts offered by practices, or to have been informed about or helped to use online services. However, it was encouraging to find that those of non-white ethnicity and those for whom English was not their first language were more likely to report experiences consistent with engagement with or being engaged in DF. Being in receipt of repeat prescriptions was also a factor associated with such engagement. Free-text responses indicated a range of barriers when accessing online services. These included personal barriers (e.g. access to resources), concerns about data security, difficulties with technology design, practice-level barriers and difficulties with credentials. Broadly similar findings were found in responders to the mental health boost sample survey, although these patients generally had higher awareness and use of online services, higher awareness and use of DF efforts, while also reporting more barriers to using online services. While there were some associations between the DF efforts reported in practice and patient survey responses, they were not always present, indicating a potential lack of effectiveness.
Responders to the mental health boost sample survey were more likely to be digitally confident, were more likely to have used general practice websites and were more likely to be aware of and use online services than responders to the main survey (though this may in part reflect the fact that responders to the mental health boost sample survey were younger, on average, than responders to the main survey). Responders from the mental health boost sample were also generally more likely to report that they were aware of, and made use of, DF undertaken by their practice. However, these associations (except for awareness of facilitation) did not persist once adjustments were made for the demographic composition of responders indicating that differences between responders to the main survey and the mental health boost sample survey were largely attributable to factors other than their mental health status.
General Practice Patient Survey
Of patients who recently tried to make an appointment, just under half were aware of online appointment booking and ordering repeat prescriptions. Considerably fewer were aware of the potential to access medical records online, which was consistent with the findings of our patient survey. Between 13% and 32% of these patients reported use of these online services, with around a quarter reporting the use of online consultations. Older patients, those from deprived areas and those from non-white ethnic groups were generally less likely to be aware of and use all online services. In contrast, those with long-term conditions were more likely to be aware of or to use all online services. While differences between patient groups were substantial, the practice with which a patient was registered was, by far, the strongest predictor of patient awareness and use of online services. Some of this between-practice variation may be attributed to the ease of use of the practice website, indicating the potential for DF.
Deviations from the original proposal
Our original proposal stated that we would aim for 60% of general practice responding to the practice survey. In reality, only 31% did so. This deficit can largely be attributed to the COVID pandemic which placed demand on practice staff, reducing the capacity to take part. Furthermore, we had planned for CRN staff to make phone calls to non-responding practices to encourage participation. However, at the time this research was undertaken, CRN staff were prioritising COVID-related studies and some CRN staff were redeployed to other duties resulting in relatively few practices actually being contacted by CRNs. In light of the smaller than anticipated number of practices participating in the practice survey, with the agreement of NIHR, we reduced the target number of practices participating in the patient survey from 120 to 60. In tandem the target number of patient surveys mailed from each practice was increased from an average of 100 to 200. To ensure representativeness, the number of patients mailed a survey was higher in practices serving deprived populations than those serving less deprived populations. Despite these changes, we were unable to recruit 60 practices to undertake the patient survey from within those practices who had responded to the practice survey. Consequently, we invited practices from outside of our original practice survey sample in order to achieve our target. When doing so we focused on practices in deprived areas to ensure representativeness overall. To facilitate comparisons between patient and practice responses, these additional practices completed a practice survey questionnaire, but were not included in the main practice survey data analysis (as they were not part of the original random sample).
A difference occurred with the analysis of data from the GPPS. One of the questions we had intended to analyse was removed from the questionnaire in 2021. So instead of using the most recent data, analysis of responses in relation to awareness of online services related to data from 2019 to 2020.
Strengths and limitations
There are a number of strengths to this quantitative work. First, we achieved a broadly similar number of responses from patients across all deprivation quintiles by over-sampling patients in deprived areas. Second, we adopted a multimodal approach, collecting data at the practice and patient levels and using national GPPS data, and combining data from the different sources. Third, we refined the survey procedures by piloting the survey and obtained extensive PPIE input. Finally, this work was undertaken as part of a large-scale research project commenced at the start of the COVID pandemic which presented multiple challenges. However, our agile approach allowed us to deliver the WP despite the challenges faced.
There were also some limitations. We had a low response rate from the patient and practice surveys. The response rate could have been an artefact of our efforts to target patients in more deprived areas where the response rate is typically lower. However, our patient survey response rate (24%) is comparable to other primary care-based surveys169–172 and should be considered credible considering the challenges presented by the COVID pandemic.
Furthermore, we did not have information on the socioeconomic status (e.g. deprivation) of patients or the rurality classification of where they lived. This was largely due to not having access to patient addresses which would allow linkage to data sets containing this information.