Article Text

Original research
How is primary care access changing? A retrospective, repeated cross-sectional study of patient-initiated demand at general practices in England using a modern access model, 2019–2022
  1. Paul Chappell1,2,
  2. Alison Dias1,2,
  3. Minal Bakhai1,3,
  4. Jean Ledger1,
  5. Geraldine M Clarke2
  1. 1NHS England, London, UK
  2. 2Improvement Analytics Unit, The Health Foundation, London, UK
  3. 3NHS Brent, London, UK
  1. Correspondence to Paul Chappell; paul.chappell8{at}


Objectives To explore trends in patient-initiated requests for general practice services and the association between patient characteristics including demographics, preferences for care and clinical needs and modes of patient contact (online vs telephone), and care delivery (face-to-face vs remote) at practices using a modern access model.

Design Retrospective repeated cross-sectional study spanning March 2019 to February 2022.

Setting General practices in England using the askmyGP online consultation system to implement a modern general practice access model using digital and non-digital (multimodal) access pathways and digitally supported triage to manage patient-initiated requests.

Participants 10 435 465 patient-initiated requests from 1 488 865 patients at 154 practices.

Results Most requests were initiated online (72.1% in 2021/2022) rather than by telephone. Online users were likely to be female, younger than 45 years, asking about existing medical problems, had used the system before and frequent attenders (familiar patients). During the pandemic, request rates for face-to-face consultations fell while those for telephone consultations and online messages increased, with telephone consultations being most popular (53.8% in 2021/2022). Video was seldom requested. More than 60% of requests were consistently delivered in the mode requested. Face-to-face consultations were more likely to be used for the youngest and oldest patients, new medical problems, non-frequent attenders (unfamiliar patients) and those who requested a face-to-face consultation. Over the course of the study, request rates for patients aged over 44 years increased, for example, by 15.4% (p<0.01) for patients aged over 74 years. Rates for younger patients decreased by 32.6% (p<0.001) in 2020/2021, compared with 2019/2020, before recovering to prepandemic levels in 2021/2022.

Conclusions Demand patterns shed light on the characteristics of patients making requests for general practice services and the composition of the care backlog with implications for policy and practice. A modern general practice access model can be used effectively to manage patient-initiated demand.

  • primary care
  • telemedicine
  • world wide web technology
  • health informatics
  • primary health care

Data availability statement

No data are available.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • This study uses routinely collected real-world data that capture approximately 10.4 million patient-initiated requests (by online, telephone or in-person) at 154 general practices using a modern model of general practice access over a 3 year period.

  • The general practices in the study have similar characteristics to the population of practices in England in each of the 3 years under study.

  • We examine the links between a range of individual level factors (including patient age, gender, type of request and frequent attender status) and modes of care contact and delivery.

  • No information was available at the patient level on ethnicity or socioeconomic status, clinical outcomes or safety of care delivered.

  • The analysis is based on only one online consultation supplier, so the results are not necessarily generalisable to other suppliers or other models of implementation.


In January 2019, the NHS Long Term Plan committed to offering every patient the right to access primary care digitally by 2023/2024.1 ‘Access’ in this context refers to both how a patient contacts their practice (eg, booking or cancelling an appointment, ordering a repeat prescription, viewing their health record) and to how care is delivered (eg, by face-to-face or remote consultation by telephone, video or online message). To support the Long Term Plan objectives, the 5 year general practitioners (GPs) contract reform framework set out a requirement for all general practices to offer online and video consultation systems by April 2021.2

The use of digital health tools and services increased exponentially at the start of the COVID-19 pandemic in March 2020 as government guidance was issued recommending radical changes to the provision of NHS primary care services. Practices were urged to adopt a total triage access model supported by an online consultation system. Using the system, patients contacting the practice, regardless of access route, could share relevant information before an appointment which could be used to ensure that they are provided with the right type of appointment, with the right healthcare professional, first time, and to better manage demand via a single workflow. This model is supported by digital tools that enable general practices to provide remote consultations where clinically appropriate.3 The ability to access care remotely became of paramount importance during COVID-19 to reduce the risk of exposure to infection to patients and staff. Online consultation systems allow patients to contact their practice via an online form4—such as to book an appointment, discuss symptoms, a medication issue or check test results and upload photographs where appropriate. This is often referred to as making an online ‘submission’ or ‘request’.

With an ongoing workforce crisis in primary care and rising levels of demand, the need for a sustainable and more equitable model of general practice is critical. An estimated 40% of appointments in general practice could be managed by healthcare professionals other than GPs, or outside of a general practice setting,5 further reinforcing the importance of care navigation and triage beyond the pandemic response. Implemented appropriately, a model of general practice access using both digital and non-digital (multimodal) access pathways alongside total triage may allow for greater choice and flexibility,6 faster care navigation, clinical assessment and response and for care to be tailored to clinical needs and patient preferences.3 7 In 2023, the Primary Care Access Recovery Plan set out a commitment to systematically support implementing this approach under the banner ‘Modern General Practice Access’.8

From a patient perspective, modern general practice access may enable greater flexibility better matched to individual circumstances and support patient-centred care.9 For example, due to difficulties in spending time on the telephone, travelling to the practice, taking time away from work or caring responsibilities or sitting in a waiting room.4 9 In addition, patients who find in-person interactions difficult may be more comfortable accessing primary care online. From a practice perspective, triage and gathering of relevant information asynchronously may enable more equitable, consistent and effective demand management to support signposting and continuity of care for those that need it as well as provide staff with more control over how they manage their time. However, there are concerns that, depending on how digitally supported models are implemented and used, there is a risk of creating—or exacerbating—inequalities in healthcare.10 11 For example, some patients without digital skills or access to technology, or those with language barriers, specific health needs or disabilities, may experience worse access and struggle with remote consultations and digital tools.9 12 In addition, easier access may potentially result in higher patient-initiated low acuity demand and consequential increases in overall workload.13

In this study, we examine patient-initiated activity at practices using the askmyGP Version 2 online consultation system (hereafter known as askmyGP) to implement the modern general practice access model.14 This is one of 31 online consultation systems approved by the NHS Digital First Online Consultation and Video Consultation framework and available to practices and commissioners in England.15 Each online consultation system varies in design, functionality and implementation.16 AskmyGP is an example of a free-text hybrid care-ready online consultation system. It allows all patient-initiated requests to be recorded in the system, regardless of the route of access, by allowing staff to input details of offline patient-initiated requests. AskmyGP also records the outcome of each request. This provides a unique opportunity to study patient journeys in practices who adopt this system, including comparison across patient subgroups.

At practices using askmyGP, patients can visit the practice website and complete a short questionnaire-based form online in which they report their symptoms using free text as well as indicate preferences for continuity of care and preferred modality of response—either remote (by telephone, online message or video), or face-to-face (in the practice or at home). For patients who telephone or walk into the practice, staff complete the form on behalf of patients. Practices manage all requests via a process of triage and navigation to the right person or service and arrange for care to be delivered in the appropriate time frame and consultation mode. By recording the outcome of the request, askmyGP provides information about the end-to-end access journey—from contact to closure of the request—providing data about preferred patient access routes and consultation type that is unobtainable from other data sources.

In this paper, using patient-initiated request data from askmyGP, we provide estimates of changes in the volume of patient-initiated demand before and during the COVID-19 pandemic and explore demographic, clinical and other differences, including patient preference, in the patient end-to-end access journey through general practice. This allows us to present a unique picture of patient-initiated demand at practices already using the modern general practice access model. Our analysis spans the year before the COVID-19 pandemic began in March 2020 and the first 2 years after.


Data and data sources

This was a retrospective, observational repeated cross-sectional study using data from patient-initiated requests made at general practices using the askmyGP online consultation system between 1 March 2019 and 28 February 2022. Practices were included if they used the system to record all incoming patient-initiated demand, whether initiated online or by telephone or walk-in. Data were examined weekly and across three 12 month rolling periods spanning the period immediately before the COVID pandemic (2019/2020: March 2019–February 2020) and the first (2020/2021: March 2020–February 2021) and second (2021/2022: March 2021–February 2022, respectively) years of the pandemic.

For each request, we determined the general practice organisational data service code; date and time submitted; patient age and gender; contact mode (online or telephone or walk-in); query type (categorised by patient as existing problem, new medical problem, medication query or other question); user type (defined as frequent attender or not, where frequent attenders were those patients occupying the top 10% of age-adjusted and gender-adjusted requesters in their practice in the yearly period); preferred consultation mode (face-to-face, telephone, online message or video); actual consultation mode (face-to-face, telephone, online message or video) and whether continuity of care was preferred.

The data set was pseudonymised (all patient identifiable data had been removed) but linked at the practice level to routinely available data about general practices across England on patient registration17 and workforce18 from NHS Digital; area deprivation using the English Index of Multiple Deprivation;19 area rurality/urbanisation from the Office of National Statistics20 and GP patient satisfaction survey results21 from NHS England.

Statistical analysis

To examine the generalisability of our results and any potential selection biases, we compared the practices included in our study with those in the rest of England. To examine demand in relative terms, we calculated the proportion of requests split by patient demographics, query type, user type, preference for continuity of care, contact mode, preferred consultation mode and actual consultation mode.

For each practice, we approximated the person-years of observation for each age and gender strata over each 12 month period as the fraction of the average list size over the 12 month period. Request rates were calculated by dividing the total number of requests by the total person-years contributed across all practices for each period or strata. Variation across practices was summarised using practice-size-weighted SD. Partially overlapping samples t-tests22 were employed to test if rates changed significantly between successive 12 month periods.

To examine the steps in the patient journey from contact mode, preferred consultation mode and actual consultation mode, we produced crosstabulations between these variables for each 12 month period and graphed the results as alluvial plots.

Logistic regression models were used to examine the association between patient characteristics and two binary outcomes: contact mode and actual consultation mode in each 12 month period. Contact mode was dichotomised into either online or telephone/walk-in (henceforth referred to as ‘telephone’) and consultation mode was dichotomised as either ‘face-to-face’ or ‘telephone’/‘online message’/‘video’ (henceforth referred to as ‘remote’). Models were adjusted for practice characteristics including practice size, rurality, index of multiple deprivation and number of full-time equivalent GPs. Month of the year dummies were included to control for seasonal effects. Since practice characteristics are likely to have a strong influence on these outcomes, models also included practice-level random effects.23 List-wise deletion was used to handle missing data on gender, request type, preferred consultation mode and actual consultation mode in each regression as required (see online supplemental methods).

All analysis was conducted in R V.

Public and patient involvement



Study sample

We analysed 10 435 465 patient-initiated requests at 154 practices that met our inclusion criteria between 1 March 2019 and 30 February 2022. Of the 154 practices, 33 were using askmyGP to capture all patient-initiated demand in 2019/2020, 140 in 2020/2021 and 132 in 2021/2022 (online supplemental table 1). Practices were spread across 28 of 135 clinical commissioning groups in England and were broadly comparable to practices in the rest of England with the following exceptions: a greater proportion of registered patients with white ethnicity (~90% vs ~83% nationally in each of the 3 years); slightly larger median list sizes and slightly more patients living in rural areas. The proportion of practices in the 40% of most and least deprived areas was similar to national levels. A slightly lower proportion of patients at askmyGP practices reported good or very good overall experience of their practice in the 2020 GP Patient Survey compared with nationally (80.6% vs 82.4%) but the converse was true in 2021 (85.6% vs 83.4%).

Request rates

Request rates remained relatively steady over the study period with an average of 3.8 (95% CI 3.5 to 4.2), 3.7 (3.6 to 3.8) and 4.0 (3.8 to 4.2) requests per person-year (ppy) in 2019/2020, 2020/2021 and 2021/2022, respectively (table 1). Request rates varied seasonally dipping from 4 ppy in March 2020 to an unseasonal low of 2.1 ppy in April 2020 at the start of the pandemic and then climbing to a peak of 4.4 ppy at the end of the third national lockdown in March 2021 (online supplemental figure 1A).

Table 1

Request rates per person-year

Annual request rates in women consistently averaged more than two-thirds higher than men (online supplemental figure 1C). Age-specific rates had a consistently J-shaped distribution (online supplemental figure 1B): high in infants aged 0–4 years, dipping to the lowest in children aged 5–14 years and rising to the highest in patients aged 75 years or older. Changes in rates over the study period varied for the different age groups. Most notably, rates for the youngest children (aged less than 15 years) decreased by over 28% (p<0.001) in 2020/2021 compared with 2019/2020 before almost recovering to prepandemic levels in 2021/2022. Conversely, rates steadily increased for people aged over 74 years by 11.8% (p<0.05) and 15.4% (p<0.01), respectively, in 2020/2021 and 2021/2022, compared with 2019/2020.

Request characteristics

The greatest proportion of requests were about existing medical problems (49.4%, 44.9% and 42.5% in 2019/2020, 2020/2021 and 2021/2022, respectively) (table 2, figure 1A). Requests about new medical problems accounted for 40.2% in 2019/2020, decreased to 32.2% in 2020/2021 but recovered to 36.2% in 2021/2022. Requests about ‘other questions’, which may include administrative and medication queries, accounted for approximately 10%–11% of all requests. Medication queries made up around 11% of queries after being introduced in April 2020.

Figure 1

Percentage of all patient-initiated requests by (A) request type and (B) user type. Note: n=10 435 465. Shaded grey areas indicate timings of national COVID-19 lockdown.

Table 2

Descriptive statistics showing request characteristics by 12 month period

Although frequent attenders represent just 10% of the patient population, they consistently accounted for almost a third of all requests (32.8%, 35.9% and 34.7%) and a surge in requests at the start of the pandemic (figure 1B). Less than one quarter of requests were from patients with a preference for continuity of care (20.9%, 20.8% and 18.5%).

The patient journey

More requests were initiated online than by telephone—increasing from 64.7% of all requests in 2019/2020 to 74.9% in 2020/2021 and 72.1% in 2021/2022 (figure 2, online supplemental figure 2A and table 2).

Figure 2

Flow of patient-initiated requests by contact mode, preferred consultation mode and actual delivery mode. Total number of requests=9 243 859 (see online supplemental table 2).

The proportion of patient requests stating a preference for face-to-face consultation dropped dramatically from 30.9% before the pandemic to just 7.4% in 2020/2021 but increased slowly to 12.5% at the end of our study period in February 2022 (online supplemental figure 2B). Telephone was the most preferred consultation modality (42.5%, 54.6% and 53.8%). Requests for online messages increased from 26.5% in 2019/2020 to 37.5% and 35.5% in 2020/2021 and 2021/2022, respectively. Video consultation was seldom requested (<0.5% in each 12 month period).

Before the pandemic, 42.1% of all requests were delivered face-to-face. During 2020/2021, this dropped to 7.8% but increased steadily during 2021/2022. In the last week of our study period in February 2022, 15.4% of requests were delivered face-to-face (online supplemental figure 2C). Across the 132 practices studied in 2021/2022, the proportion of requests delivered face-to-face ranged from 3.0% to 32.6% and was only greater than 20% in 13 of these practices.

We also examined the flow of requests from contact mode to actual consultation mode and how that varied by patients’ preferred consultation mode (figure 2 and online supplemental tables 2 and 3). Mode of contact did not dictate mode of delivery—more than half of the requests delivered face-to-face were initiated online (58.4%, 66.3% and 62.9%). More than 60% of all requests were consistently delivered in line with patient preference (60%, 68.2% and 65.4%). For example, in 2021/2022, 66.3%, 40.9% and 71.4% of requests were delivered as requested for a telephone call, a face-to-face consultation and an online message, respectively.

Association between contact mode and request characteristics

Results from multiple logistic regression (table 3, online supplemental table 4) show very similar results across each of the 3 years. Patients initiating their request online rather than by telephone were more likely to be female (male vs female ORs=0.88, 0.90 and 0.85 in 2019/2020, 2020/2021 and 2021/2022, respectively, p<0.001 for each 3 years); younger rather than older—for example patients aged over 74 years were almost 10 times less likely to use the online route than those aged 25–44 years (ORs=0.12, 0.11, 0.11, p<0.001); asking about an existing medical issue; a frequent attender rather than a non-frequent attender (ORs=1.69, 1.58, 1.52, p<0.001); overwhelmingly prefer an online message over a face-to-face consultation (ORs=97.8, 10.2, 13.2, p<0.001) and; more likely to contact online next time (ORs=1.70, 1.2 and 1.19, p<0.001). There were some differences in prepandemic and during-pandemic results in terms of the associations between contact mode and preference for continuity of care. In 2019/2020, patients who preferred a telephone consultation were more likely to contact via telephone (prefer telephone OR=1.15, p<0.001) and there was no difference in contact mode between patients requesting continuity of care and those not requesting (OR=1.01, p>0.05). However, during the pandemic, patients who preferred a telephone consultation, or who asked for continuity of care, were less likely to contact the practice by telephone—in 2020/2021 and 2021/2022, respectively, preference for telephone ORs=0.22 and 0.22 (p<0.001) and continuity of care ORs=0.79, 0.88 (p<0.001). Patients rarely indicated a preference for an online message when contacting the practice by telephone (0.8%, 3.5% 2.8%) (online supplemental table 3).

Table 3

Association between request characteristics and contact mode (online message vs telephone initial request) and actual consultation mode (face-to-face vs remote consultation)

Association between actual consultation mode and request characteristics

Results from multiple logistic regression (table 3 and online supplemental table 5) show that associations with actual consultation mode were very consistent across each of the 3 years. Remote consultation was most likely to be used for patients aged 25–44 years—for example, patients aged less than 5 years were almost twice as likely (ORs=1.92, 1.63, 2.17 in 2019/2020, 2020/2021 and 2021/2022, respectively, p<0.001 for each of the 3 years) and patients 75 years and over were up to 25% more likely (ORs=1.06, 1.13, 1.25, p<0.001) to be seen face-to-face compared with those aged 25–44 years (table 3). Compared with queries about existing medical problems, face-to-face consultation was more likely to be used for new medical problems (ORs=2.08, 1.67, 1.66, p<0.001) but less likely to be used for medication queries (ORs=0.20, 0.18 for 2020/2021 and 2021/2022, respectively, p<0.001) or ‘other questions (ORs=0.53, 0.64 and 0.63, p<0.001). Frequent attenders were less likely to have a face-to-face consultation compared with non-frequent attenders (ORs=0.83, 0.86, 0.89, p<0.001). Before COVID, patients with a preference for continuity of care were less likely to be seen face-to-face than those that did not, but after 2020/2021, the trend reversed (ORs=0.88, 1.19, 1.22, p<0.001). Consultations tended to be delivered in line with patient preference—those requesting a face-to-face consultation were significantly more likely to have one compared with those who requested an online or telephone consultation (online ORs=0.15, 0.12, 0.13, p<0.001; telephone ORs=0.22, 0.18, 0.20, p<0.001).


This study examined over 10 million patient-initiated requests for general practices services at 154 practices between March 2019 and February 2022 using a modern general practice access model.

Much research in this area to date has been from the health professional perspective and this study is the first to look at patient-initiated demand using online consultation system data and to include how patient preferences, as well as characteristics and clinical needs, relate to how patients contact their practice and how their care is delivered.25

Demand for care

The patterns in patient-initiated request rates across age, gender, frequent attender status and existing medical problem status are consistent with those in GP consultation rates in other research which quantify demand for GP appointments in aggregate terms.26 27 However, rates here are likely to be lower because the askmyGP online consultation system is focused on patient-initiated demand and is not employed to capture all types of consultation activity—for example, practice-initiated and proactive work such as long-term condition reviews, discussing abnormal test results or arranging tests and prescriptions following advice from other settings. Hence, no conclusions can be drawn about overall rates of general practice activity based on this study.

Our finding that a greater proportion of demand comes from older cohorts is consistent with a recent analysis of consultation rates over the same period,28 and suggests that these groups are driving workload pressures in general practice. This is despite the fact that these patients were the least likely to use the online system, suggesting that inclusive implementation of digital and non-digital access pathways as part of a modern general practice model does not disadvantage non-digital users. The increased demand is also reflective of a growing complexity of care,29 a disproportionate impact of COVID and an increased workload burden, perhaps due to transfer of work from other healthcare settings to general practice.30

Patterns in the distribution of request characteristics over time shed light on the key users of primary care and the composition of the care backlog; 40% of requests were about existing medical problems which may include episodic conditions but are most likely to be about long-term conditions known to account for most consultations in primary care.31 32 Steady increases in the proportion of requests from non-frequent attenders, about new medical problems and a recovery in the rate of requests from younger patients during 2021/2022 suggest that some patients may have been reluctant to seek care earlier in the pandemic, or were affected by delays in outpatient or elective care33–35 further increasing pressure on general practice.

Contacting the practice

Having used the online route, patients were more likely to do so next time but flexibly and dynamically. They were more likely to choose telephone /in-person access for new problems (issues they are likely to have perceived as more urgent). Requests initiated online tended to be from patients asking about existing medical problems, who were female, 45 years or younger, without a preference for continuity of care, frequent attenders and from those with a preference for an online message response. Remarkably, there were no differences in these trends before and during the pandemic.

Before the pandemic, increases in the proportion of requests made online did not increase in line with demand suggesting a genuine shift in patient behaviour from telephone/in-person to online access. During the pandemic, there was no increase in the ratio of requests made online versus by telephone suggesting that the ability to contact the practice online was not stimulating the additional demand. This further confirms the shift in patient behaviour and, in contrast to anecdotal reports,36 suggests no evidence of supply induced demand fuelled by easier access.

Care delivery

Only a minority of requests at these practices indicated a preference for a face-to-face consultation. Variation in care delivery mode across different patient subgroups suggests that clinicians were tailoring care according to patient characteristics, clinical risk and patient preference. Importantly, mode of contact did not determine how care was delivered. Face-to-face consultations were more likely to be used by clinicians for the youngest and oldest patients, those asking about new medical problems, non-frequent attenders (likely to be less familiar to the practice), a preference for continuity of care (during the pandemic) and a preference for face-to-face consultation. These findings are consistent with other evidence suggesting that face-to-face consultations are more likely to be used for the most vulnerable or complex patients.37 38 Patients with existing problems and frequent attenders were more likely to have a remote consultation. These patients may be less likely to require a physical examination and more likely to be known to their general practice and have a pre-existing doctor–patient relationship.

Video was rarely requested or used at these practices. Although some patients may have positive views of video consulting,39 issues related to patient access (equipment, skills, reliable internet connection and comfort with technology) and loss of physical presence (to have a physical examination) might outweigh the convenience of their increased accessibility for patients.40 From the clinician perspective, video consulting in primary care may have few advantages over telephone and face-to-face consulting41 and technical issues may make the process unsatisfactory.

Strengths and limitations

Practices studied here are broadly representative of practices in England, and while the analysis is based on only one online consultation system and model of implementation and therefore not necessarily generalisable to practices using other online consultation systems and models of implementation, it captures the key changes to general practice services accelerated by the pandemic: multimodal access, navigation and triage, use of a multiprofessional team and flexible consultation options with all the practices using a mostly consistent model of implementation.

In multiple regression analyses, we were able to adjust for a range of factors that captured differences between the general practices included in the study, such as adopter status, deprivation and practice size. Estimated effects were highly significant and similar across the three different time periods studied. This effectively provides both a sensitivity check and indication that findings are robust.

This data set allows for the examination of a combination of rich contextual factors capturing information about the patient at both the point of contact and delivery not obtainable from other data sources. However, no information was available at the patient level on ethnicity or socioeconomic status, clinical outcomes or safety of care delivered. Nor could we infer any details about who is not using primary care at these practices, and how that might relate to the access model.

Implications for policy and practice

This study indicates that a modern general practice access model which uses digital alongside non-digital pathways supported by an online consultation system and a single workflow to enable consistent navigation and triage can be used effectively to manage patient-initiated demand in general practice, and remain sensitive to patient preferences.8 Our findings suggest that such a model can support choice and flexibility, help prioritise care based on need, support more equitable access and optimise use of the available workforce. We also provide evidence that patients are proactively engaging with online consultations and, after using the online consultation system, may alter their future preferences to use it again. Moreover, the dynamic nature of how patients used the different contact modes at these practices, for example, depending on the type of problem and perceived urgency, suggests the multimodal approach allowing for both digital and non-digital modes of contact, is essential with no evidence to suggest that patients who prefer the traditional access routes are disadvantaged.42

Variation across patients and over time suggest that patient preferences for how care is delivered is multifactorial and shaped by service responsiveness, flexibility and choice (eg, between digital and non-digital access routes and consultation types), the type and complexity of the problem, continuity and communication and the wider social context. In order to better meet people’s needs, we must take a holistic view of how general practice services are delivered, rather than focusing on any single aspect or target, such as the mode of consultation.7

Our data do not tell us about the social and cultural factors in these practices but positive team dynamics, leadership, psychological safety, innovative and collaborative cultures, a quality improvement mindset, motivation and external support are all strong levers for successful change.43–46 It will be important to understand more about the practice context and staff and wider team characteristics in order to support others in successfully adapting access and triage models to meet local population needs6—especially given variety in implementation approaches.

A large proportion of requests to general practice are from frequent attenders and patients asking about existing medical problems, and the largest increases in demand for care compared with prepandemic are from older patients. Although many of these patients may have complex needs and need to be seen regularly, an understanding of who they are and why they are seeking care is important to allow general practice to think about how services can be more effectively provided for these groups. In particular, research indicates that many frequent attenders may have wider psychological and social issues that may be better supported in a non-medical environment and by working with communities.47

More work is required using a larger sample of practices and different online consultation systems to understand variation in processes and models of implementation across practices, different design and functionality of online consultation systems and to look at impacts relating to issues of safety and consequences for health inequalities.

Data availability statement

No data are available.

Ethics statements

Patient consent for publication

Ethics approval

This study used an anonymised extract of routinely collected data about patients. This study is classed as service evaluation rather than research (using the HRA Decision Tool), so NHS Research Ethics Committee and Health Research Authority approval have not been sought for this work.


We would like to acknowledge the contributions of the following individuals from the IAU: Richard Brine, Arne Wolters, Emma Vestesson, Liz Crellin, Paris Pariza and Edvige Bordone. The IAU works in partnership with NHS England and we would like to thank Louise Croney, Ejemen Asuelimen and Dilwyn Shears from NHS England for their support and advice while progressing this work. We are also grateful to Jenny Newbould, RAND Europe and Steve Black, Black Box Data Science for their advice while progressing this work. Alison Dias left the IAU during the analysis and write-up period of the study and is now based at the Ministry of Health, NSW Health, Sydney, Australia.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors GC, PC, AD designed the study; derived the indicators and analysis data set and performed the analysis. MB, GC, PC, AD and JL contributed to the interpretation of the work. GC and PC drafted the paper. MB, GC, PC, AD and JL contributed to the paper; read and approved the final manuscript. GC is the guarantor.

  • Funding This research was conducted by the Improvement Analytics Unit (IAU) and received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. The IAU is a partnership between NHS England and the Health Foundation. The IAU evaluates complex initiatives in healthcare in order to support learning and improvement and is funded jointly by NHS England and the Health Foundation, with funding jointly assigned to jointly agreed priorities.

  • Competing interests AD and GC declare no competing interests. PC worked as a data analyst for NHS England during the course of this project. JL worked as a freelance consultant to NHS England and academic researcher during the course of this project. JL and MB receive financial support from NHS England for attending meetings.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.