Patterns of online consultation use in Great Britain, 2019–2023: an observational analysis

Abstract

Objective

To examine online consultation (OC) uptake and usage patterns over time in Great Britain by patient characteristics, practice location and condition/symptom-specific OC template type, for a major provider of OC services (eConsult).

Methods and analysis

An observational study was conducted using records of OCs completed between January 2019 and December 2023. Main outcome measures were OC volume and rates by country, region, deprivation level, template type and user demographics (age group and sex).

Results

Over 43 million OCs were submitted on the eConsult platform within the time frame. The number of OCs per month greatly increased in early 2020 after the introduction of COVID-19 pandemic restrictions on in-person contacts. Sign-ups to the platform were highest between February and May of 2020. At their respective peaks, 43.7% of English practices, 11.6% of Scottish practices and 50.5% of Welsh practices were active in eConsult. Rates of OCs declined throughout 2022 and most of 2023 and were highest in more affluent areas. For England, Scotland and Wales, respectively, OC rates peaked in October 2021 at 17.6, 10.5 and 20.4 OCs per 1000 general practitioner-registered patients. Nearly two-thirds (64.4%) of OCs were completed by female patients. The proportion of OCs for children aged ≤10 has risen over time. 89% of OCs were submitted to the patient’s practice, whereas 7.7% were redirected out towards urgent or emergency care services. Over time, there have been decreases in the number of OCs submitted at the weekend and in usage of generic ‘general advice’ templates.

Conclusions

These findings reveal the substantial and evolving utilisation of eConsult over time across Great Britain. These findings may reflect evolving product functionality and engagement with OC services by patients and clinical staff, or local differences in service coordination.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Online consultations (OCs) between patients and clinicians have emerged as a core component of many primary care services in recent years. Although there have been several small-scale or pilot studies looking at the impact of OCs, there is limited evidence on the large-scale utilisation of OC platforms.

WHAT THIS STUDY ADDS

• This work provides insight into OC rates over 5 years at a national level, by a major commercial provider of OC services in Great Britain. There was substantial variation in coverage and OC rates over time between England, Scotland and Wales, by region of England, and by deprivation level. Declining trends in OC volumes, reduced reliance on general advice templates and growing usage by young children reflect shifting postpandemic patient and practice engagement with OCs.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings highlight the substantial volume of OC usage in British primary care, much of which is largely uncaptured by routine primary care data. These findings underline the need for sustained efforts to optimise these tools within primary care systems as well as standardised recording practices to facilitate evaluation of such services.

Introduction

Most patient interactions in the UK National Health Service (NHS) occur in primary care, where general practitioners (GPs) play a critical role in managing and coordinating patient care and managing access to specialist services.1 However, safely and efficiently responding to patients’ presentations promptly is an increasing challenge due to growing demand relating to an ageing population, with rising multimorbidity and case complexity, and workforce retention and recruitment difficulties. This has led to concerns about access, appointment availability and the sustainability of the traditional consultation model.2 3

Online consultations (OCs) have emerged as a response to these challenges.4 OCs allow asynchronous doctor–patient interactions where people can seek medical advice or submit administrative requests (eg, prescriptions, sickness certificates or letters), through structured online forms.5–8 These forms collect detailed information, enabling prioritisation of problems. While initial uptake of OCs by primary care practices was limited, improvements in digital platforms and increasing digital health literacy have driven their adoption in Great Britain.7–10 By 2021, approximately 95% of practices in England had an OC system,11 driven by the ‘total triage’ model in primary care introduced by NHS England in 2020.12 However, with a range of available providers, limited data on OC utilisation by OECD countries, and little published on OC use and their users, describing the current use of OCs in primary care remains challenging.3 13 Provider-specific usage data can offer valuable insights into OC use, with potential to inform current and future policy.

The largest UK OC platform provider for NHS primary care is eConsult Health (‘eConsult’), with over 1700 current GP users.14 eConsult is a clinician-curated algorithm-based OC system with over 100 problem-specific or symptom-specific structured online forms (‘templates’). Completion of a template results in a summary document (‘OC’) being submitted to the GP practice for review, triage and response. In-built checks for ‘red flags’ redirect patients to urgent or emergency services if concerning symptoms are entered during a consultation, in which case the OC process ceases. Despite local pilot studies and user surveys,5–8 no comprehensive analysis of eConsult utilisation across Britain’s primary care landscape has been conducted. This study examines patterns in eConsult usage and uptake by practices over time to provide insight into broader trends in OCs in primary care. Specifically, this study’s objectives were to (1) present numbers of OCs over time, by template type, patient user characteristics and outcome destination of submitted OCs, (2) analyse coverage of eConsult among practices in different regions and (3) compare OC rates between regions.

Methods

Study design

An observational study of OC service utilisation was conducted for the period between 1 January 2019 and 31 December 2023. Records of templates (OCs) completed by patients during this period were provided by eConsult. Data consisted of records of OCs along with information on the specific template name, date of template OC completion, patient self-entered age and sex, destination outcome of the template, and information about the associated healthcare provider organisation. This study adheres to Strengthening the Reporting of Observational Studies in Epidemiology guidelines for reporting cross-sectional studies.15

Data cleaning

Templates

Template names were cleaned to correct for spelling inconsistencies. After cleaning, 131 template types remained and were grouped into 27 template groups, categorised according to broad clinical problem area. A complete list of eConsult templates available to patients during the study period and associated template groups is included in online supplemental table S1.

Outcomes of OCs

Completed OCs are sent directly to the patient’s general practice. However, if during the process of completion, a patient is identified by their responses as in need of alternative (more urgent) services, they are redirected to the relevant services (eg, NHS 111, 999 or hospital emergency department) and the OC is not sent to the practice.

eConsult providers

Data were cleaned to identify unique health provider organisations in the dataset (see online supplemental file page 2, online supplemental figure S1). Of the providers identified, most were individual practices. However, health providers also included primary care partnerships and group practices, prescribing centres and other healthcare service providers. It was not possible to distinguish individual practices from groups of practices in some cases. Those OCs identified as occurring in a hospital, urgent care or outpatient setting were excluded. Of the remaining providers, 18 (0.5%) could not be associated with a postcode or one in England, Scotland or Wales, and were excluded from all analyses.

Go live dates

The ‘go live’ date describes the date eConsult services became available to patients at a given provider. Any OCs in the data record that occurred before a provider’s ‘go live’ date were excluded as these represent testing and training submissions. Where the ‘go live’ date was missing, this was set as the date of the first recorded OC. Providers were defined as ‘live’ in eConsult after their ‘go live’ date. Providers were defined as ‘active’ during the continuous period between their ‘go live’ date and the date of their last OC.

Patient characteristics

The age (0–100 years) and sex (male or female) of patients were available at the OC level; no patient-level data were collected or available. In the case of patients under the age of 16, the OC is completed by a parent or legal guardian on behalf of the patient. Age was categorised into 10-year age bands from 0 to 70 years, with all ages over 70 years grouped together.

Details of the full data cleaning process and data sources can be found in online supplemental file pages 2–3.

Statistical analysis

For providers located in England, Scotland and Wales, frequencies of provider sign-ups and OCs were examined, with the latter visualised by country, day of the week, patient characteristics (age and sex), template group and OC outcome destination.

Coverage of eConsult and OC rates was examined for each country and each region of England. List size data are presented for the months of January, April, July and October according to Scottish and Welsh list size data availability. Of the 3843 unique provider organisations in the dataset, 3720 (96.8%) could be linked to data on practice list sizes. Practice Index of Multiple Deprivation Quintile (IMD Quintile) was obtained for the postcode of each practice. Coverage of eConsult was calculated as the proportion of practices identified as active in the platform in each month. The total number of OCs per 1000 registered patients was calculated separately for patients registered to live, active or all practices (ie, regardless of practice utilisation of eConsult). All analyses were performed in R V.4.3.0.16

Patient and public involvement

Neither patients nor members of the public were involved in the development or execution of the research study. Results from this study will be disseminated via conference presentations and through our organisation’s patient and public engagement events.

Results

Between 1 January 2019 and 31 December 2023, there were 43 657 891 OCs in England, Scotland and Wales across 3843 provider organisations (online supplemental table S2). Monthly OCs rose sharply in early 2020, from 3.8 to 12.6 per 1000 registered patients between January and April (figure 1). The number of OCs per year fell over time from a total of 14 240 548 in 2021 to 9 421 649 in 2023. Monthly OCs peaked at 35.1 per 1000 registered patients in October 2021 but subsequently declined to a low of 18.7 per 1000 registered patients in April 2023. Numbers of OCs per month for each country are shown in online supplemental figure S2.

Figure 1.

Total online consultations (OCs) per 1000 registered patients in Great Britain and frequency of OCs by day of the week, using data for the whole GP registered population between January 2020 and December 2023. GP, general practitioner.

Across days of the week, OCs declined in frequency from Monday to Saturday, with 24.8% (n=10 813 714) of OCs overall occurring on Mondays. The proportion of OCs occurring on Saturday or Sunday declined over time (figure 1).

Patient characteristics

The mean age associated with OC submission was 42.0 (SD 21.0) years and 28 122 457 (64.4%) were for female patients. OCs were most common for patients aged 21 to 40, who contributed 40.9% (n=17 859 321) of OCs. One tenth (n=4 377 436) of OCs occurred for patients aged 16 years or younger, while 8.9% (n=3 900 271) occurred for patients aged 65 and older.

The mean age was similar for males (44.0±21.6) and females (43.5±20.6) (p<0.001) overall; however, a greater proportion of OCs for males were for the age groups 0–10 and 60+ years (figure 2). There has been an increase over time in the proportion of OCs for young children aged 10 years or less, particularly among male patients (figure 2).

Figure 2.

Percentage of online consultations by patient user age group in years, over time for the period of January 2019 to December 2023.

Template usage

Usage of condition-specific template types varied over time (figure 3). The most common template groups were ‘administrative help’ (n=9 757 841; 22.3%) and ‘general advice’ (n=8 476 894; 19.4%), both of which declined in frequency over time (figure 3). The number of distinct templates used per month increased over time from 79 in January 2020 to a high of 93 in June 2023 (online supplemental figure S3). Frequency of respiratory template submissions exhibited peaks in the winter periods of each year.

Figure 3.

Frequency of template types as a proportion of total online consultations (OCs) per month for the years 2019–2023, for the top nine template groups, with all remaining templates grouped into ‘all other’. A full list of template names can be found in online supplemental table S1.

eConsult outcomes

Of the 43 657 891 million total OCs, 89.0% (n=38 860 124) were submitted to the patient’s practice, and 9.6% (n=4 181 029) were diverted out of eConsult to other services. A total of 3 377 822 (7.7%) were directed to urgent or emergency care, of which 90.2% (n=3 047 133) were directed to urgent care, 5.7% (n=191 131) to emergency care, and 1.9% (n=62 514) were urgent mental health redirects. Diversions to urgent care varied in frequency over time with peaks between May and October of 2020, and in October 2021, and December 2022 (online supplemental figure S4).

Sign-ups and coverage

Between January 2015 and December 2023, 3843 new providers signed up to eConsult. Of these, 1635 (42.5%) signed up between March and May 2020, with 21.0% (n=824) signing up in April 2020 alone (online supplemental figure S5). The numbers of live and active practices per month are shown in online supplemental figure S6.

The number of providers active in eConsult in Great Britain peaked at 3204 in April 2021 (online supplemental figure S6), covering 40.2% of total primary care practices (figure 4). At their respective peaks, 43.7%, 11.6% and 50.5% of practices in England, Scotland and Wales, respectively, were active in eConsult (figure 4). As of October 2023, 27.3%, 8.6% and 26.8% of English, Scottish and Welsh practices, respectively, were active in eConsult. For England and Scotland, coverage of eConsult was lower in more deprived areas (online supplemental figure S7).

Figure 4.

Percentage of primary care practices in Great Britain active in eConsult between 2019 and 2023, by country and region. Practice data were not available for Wales prior to 2020.

There was regional variation in eConsult coverage in England (figure 4). Coverage of eConsult was highest in the Northeast of England, where between 91.1% and 99.1% of practices in the region were active in eConsult for any month between April 2020 and October 2023. Coverage was overall lowest in the East Midlands, ranging between 4.3% and 12.9% between April 2020 and October 2023. Throughout 2023, there were declines in the number of active providers across all English regions, excepting the East of England, which continued to increase over this period.

Rates of OCs by region

While OC rates for patients registered to live practices declined across 2022, no declining trend was observed among those registered to active practices (figure 5). In 2023, rates increased from April to October in Scotland, Wales and all regions of England except the West Midlands and Yorkshire and the Humber.

Figure 5.

Total online consultations (OCs) per 1000 registered patients in Great Britain, per month and region of England. ‘All practices’ includes the whole GP registered population, while only those patients registered to practices using eConsult services are included in the ‘live’ and ‘active’ plots. GP, general practitioner.

OC rates among the whole GP-registered population peaked in October 2021 for England (17.6 OCs per 1000 patients), Scotland (10.5 OCs per 1000 patients) and Wales (20.4 OCs per 1000 patients). As of October 2023, there were 11.2, 8.6 and 11.3 OCs per 1000 registered patients in England, Scotland and Wales, respectively. Among those registered to live or active practices, OC rates were consistently higher in Scotland than in Wales or England (figure 5).

Similar rates of decline in OC rates for non-active practices were exhibited by most English regions (figure 5). OC rates in the Northeast of England increased throughout the study period across all practice denominators. There were differences in regional OC rates between the whole GP registered population and only those registered to live or active practices (figure 5), likely reflecting regional differences in eConsult coverage (figure 4). In 2023, while OC rates increased from April to October in most regions of Great Britain, rates in the West Midlands greatly declined as practices in these areas halted use of the service (figure 5).

Rates of OCs by deprivation level

Among the whole GP-registered population, patients registered to the most affluent areas (IMD quintile 5) often had higher OC rates relative to more deprived areas (figure 6). Overall for England, OC rates were lowest among patients registered to the most deprived areas (IMD quintile 1), but OC rates were similar for IMD quintiles 2–5. OC rates by deprivation level varied greatly between regions of England (online supplemental figure S8). For Scotland, OC rates were 2–3 times higher for the most affluent (IMD quintile 5) relative to the least affluent areas (IMD quintile 1), but their OC rates decreased with greater affluence between quintiles 2 and 4. For Wales, although OC rates were highest in the most affluent areas, OC rates were lower in less deprived areas between quintiles 1 and 4. The difference in OC rates between quintile 5 and all other quintiles in Wales increased over time between July 2021 and December 2023 (figure 6). Estimates of OC rates for live and active practices are presented in online supplemental figure S9.

Figure 6.

Total online consultations (OCs) per 1000 registered patients in Great Britain, per month and deprivation quintile, using data for the whole GP registered population. Index of Multiple Deprivation rank is calculated separately for England, Scotland and Wales. GP, general practitioner.

Discussion

Principal findings

This study summarises primary care OC activity from a major provider of OC services in Britain. Between January 2019 and December 2023, over 43 million OCs were submitted on the eConsult platform across Wales, Scotland and England. There was substantially greater use of eConsult following the introduction of COVID-19 pandemic restrictions on in-person contacts in early 2020. The greatest volume of sign-ups to the eConsult platform occurred between February and May of 2020, to reach a peak coverage of 43.7% of English, 11.6% of Scottish and 50.5% of Welsh practices. Following the loosening of pandemic restrictions, OC rates and the number of active providers in eConsult declined throughout 2022 and most of 2023. Almost 90% of OCs were submitted to the patient’s primary care provider while 7.7% were diverted out of eConsult towards urgent or emergency care services. A fifth of OCs submitted were administrative requests. Changes in manner and frequency of eConsult utilisation were noted over time by factors including location and patient age.

Strengths and limitations

This is the first study to present countrywide longitudinal data for a major provider of OC services in Great Britain, covering a critical period in the changing utilisation of remote consultations. Our data highlight the substantial underestimation of OC service utilisation and, therefore, a large volume of uncaptured workload for GPs. NHS Digital statistics from English practice appointment systems reported approximately 125 000 online or video consultations consistently for each month between 2020 and 2022 (online supplemental figure S10). In contrast, our data show substantially higher OC volumes and large monthly variation in OC numbers (figure 1, online supplemental figure S2, online supplemental table S2). In their analysis of electronic health records from a large selection of English GP practices, Fonseca et al report a rate of just over 10 OC coding events per 1000 registered patients in November 2020.11 Here, for eConsult alone, this study reports a higher rate for the same period (figure 5). These discrepancies may be partly attributable to the presented data covering all activity types, while GP recording practices may distinguish between types of OC activities, resulting in certain OC activities being recorded in clinical systems, others being recorded in administrative or scheduling systems and others not being documented at all.11 Beyond Great Britain, there are limited data among OECD countries on utilisation of telemedicine services, particularly OCs.13 This may be due to the novelty and varied nature of OC services, and the perception that they are an adjunct to the traditional consultation process.6 Our findings, therefore, provide unique and comprehensive insight into engagement with OCs on a national level, reinforcing the need for improved recording of OC activity in primary care records.

Our study is subject to some limitations. First, eConsult does not collect patient-level data and, therefore, it was not possible to identify OCs submitted by the same patient. Utilisation rates may, therefore, be skewed by high-frequency users. It was also not possible to examine patient-level preferences and the degree of choice available to them in how they accessed primary care. Nor was it possible to examine factors such as sensory and intellectual impairments, language or digital health literacy which may limit the accessibility of OC tools to certain patients.17 Second, it was not possible to distinguish individual practices from groups or partnerships of practices in some instances. This may have resulted in an underestimation of the number of provider organisations active in eConsult and potentially inflated consultation rates where multiple providers shared the same identifying information. To help mitigate this, OC rate distributions were inspected to identify where individual provider organisations had outlying values (online supplemental file page 2). Third, this analysis lacked information on how the receiving practice handled the OC, such as the time to or type of response, or whether recommendations to seek emergency care were appropriate or adhered to by patients. This information would be valuable in assessing the effectiveness and impacts of OCs. Lastly, this analysis used data from one OC provider; consequently, these findings do not represent the totality of OCs in Great Britain and may not be generalisable to other OC providers. This study’s findings are, however, consistent with previous data from other OC service providers,18 and more general national data on OC rates,11 that show there is higher OC usage on Mondays and by female and young adult patients.

Interpretation of findings and comparison with existing evidence

Offering online access routes to primary care has been an objective of the NHS Long Term Plan since 2019, with the aim of OCs improving efficiency of care delivery.3 4 Despite this, initial uptake by practices was low until the advent of COVID-19 restrictions.11 19 Being an early major provider of OC services, eConsult experienced a large and rapid rise in use following the sudden introduction of in-person contact restrictions in early 2020 and the subsequent ‘total triage’ model in primary care.11 12 During this period, patients had to first undergo a remote triage, which included telephone calls and OC tools, to obtain a primary care appointment.12 Over time, several competing eConsult service providers have emerged,3 20 introducing a greater range of options for practices, which has resulted in the loss of some practices from eConsult and fewer new sign-ups. This, together with the move away from the total remote triage model, contributed to declines in eConsult OCs in 2021 and 2022. Nevertheless, there was a rise in OC rates in late 2023 (figure 5). This may reflect policy drivers introduced under NHS England’s ‘Delivery plan for recovering access to primary care’, published in May 2023, which promotes continued use of online tools for routine aspects of primary care delivery.21 However, this does not include any contractual requirements in terms of duration of online access or proportions of OCs, allowing for continuing significant variation between practices.

There was variation in coverage of eConsult services and OC rates between England, Scotland and Wales, and between English regions (figures 4 and 5). For example, nearly all practices in the Northeast of England were participating in eConsult by April 2020; however, OC rates were relatively low for patients registered to live and active practices in the Northeast compared with other regions (figure 5). Low OC rates despite high availability of the OC tool imply low utilisation by the end users, possibly reflecting top-down selection of a product which does not align well with the preferences, workflow, accessibility needs or digital skills of all patients or clinical staff. In contrast, coverage in the Northwest was approximately 32% between 2021 and 2022, but live and active practices in the Northwest often had the highest OC rates of any region for this period. These differences likely reflect regional variation in service organisation between primary care networks, and therefore, varied implementation policy and funding structures for OCs.21 Additionally, the differences in OC rates by deprivation level are suggestive of variation in unmeasured factors, such as education, urbanicity and internet coverage, which vary geographically and would influence OC utilisation.10 11 22 A more comprehensive analysis of geographical and other factors linked to eConsult rates is an area for future work.

Many practices will use a combination of OC tools to suit clinician and patient needs and preferences.3 21 These preferences have likely evolved over time due to greater availability of products, as well as evolving functionalities of existing products and services. Several of the changes in utilisation patterns observed in this dataset reflect such product evolution. For example, the proportion of OCs submitted at weekends declined over time. This can likely be attributed to eConsult introducing (in response to customer demand) the ability to turn off eConsult availability to their patients on weekends.

Similarly, the number of redirections has varied over time. One concern raised by clinicians to eConsult was that patients would ‘game’ the system, downplaying symptoms to avoid redirection to urgent or emergency services resulting from red flag symptoms.19 In response, eConsult implemented changes to the sensitivity of diversion responses and included strongly worded warnings about the risks of altering question responses and/or downplaying symptoms, with likely consequent impact on user behaviour. There was a decline over time in the generic template type (‘general advice’), a ‘catch-all’ template, designed to collect information from individuals for whom there was no available template appropriate to their presenting problem. Improvements in the platform search facility, renaming of multiple templates to reflect a more consistently problem or symptom-based approach, as well as continued development of new templates in response to patient and clinician feedback (online supplemental figure S3), are likely to have resulted in this decline. Usage rates of the administrative pathway, comparatively higher than in other OC studies,23 also declined over time. Many patients on becoming familiar with the eConsult system, will realise that the shortest pathway, with fewest questions, is the one designed for administrative requests. Recognising that this constituted a risk for patients and practices due to reduced specificity of clinical questions, eConsult over time strengthened warnings in this pathway to discourage users from submitting clinical queries via this route, likely contributing to the observed usage decline. The findings presented in this study should, therefore, be interpreted as reflecting aspects of platform design as well as patient and practice utilisation habits.

Implications for policy and practice

The large volume of OCs presented here demonstrates the substantial role of OCs within the primary care system of Great Britain. The large regional variation observed in consultation rates may reflect local variation in organisational policies, such as differences in integrated care board versus practice-level funding for implementation of digital health products, as well as the impact of other local policy decisions, healthcare infrastructure and patient engagement levels.

As primary care continues to grapple with growing demand in the context of significant resource challenges in funding and workforce, it is critical that clinical activity, including the use of tools such as OCs which may contribute to improved efficiency in healthcare provision, is accurately recorded and reported. Comparative analyses across different OC providers could further elucidate best practices and inform policy decisions. Policy-makers must use these insights to ensure that OCs contribute effectively to addressing workforce pressures and improving the efficiency, equity and outcomes of primary care.10 22 24

Conclusions

This study highlights the substantial uptake and evolving utilisation patterns of OCs in primary care across Great Britain from 2019 to 2023. The findings emphasise the important role that digital health tools, such as eConsult, played during the COVID-19 pandemic in maintaining access to primary care while adapting to new healthcare delivery models. Subsequent decline in usage underlines the need for sustained efforts to integrate and optimise these tools within primary care systems.

Significant regional and demographic variations in eConsult uptake and use point to potential inequities and require targeted interventions, such as improving digital health literacy, enhancing the accessibility of OC platforms and tailoring implementation strategies to regional needs. Additionally, the lack of comprehensive documentation for OCs in clinical systems underscores the necessity for standardised recording practices to better understand their impact on healthcare delivery and outcomes. Future research should explore the long-term implications of OCs on patient access, satisfaction with care, patient safety, clinical workload and healthcare equity.

Data availability statement

Data may be obtained from a third party and are not publicly available. Data reported and analysed in this study are derived from platform usage statistics for the eConsult primary care online consultation product. These data were shared with Imperial College London by eConsult Health for the purpose of this analysis; a data protection impact assessment was carried out and a data sharing agreement made between parties. Data may be accessible through eConsult Health subject to separate agreement.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

Ethical approval was granted by the Imperial College Research Ethics Committee (ICREC), which oversees health-related research with human participants. ICREC Reference number: 7109745. Through their role working with eConsult Health, BWJH and AH facilitated data acquisition and provided context regarding the functionality and development of the eConsult product. All data preparation, statistical analysis and presentation were conducted independently by GK.