Article Text
Abstract
Objectives To describe the impact of the COVID-19 pandemic on hypertension diagnosis and management in UK primary care.
Design Population-based cohort study.
Setting Over 2000 general practices across the UK contributing to the Clinical Practice Research Datalink.
Participants A cohort of 23 076 390 patients over 18 years of age and registered with their general practice for at least 1 year between 2011 and 2022, who did not have a previous diagnosis of hypertension. From these patients, a subcohort of 712 461 patients diagnosed with hypertension between 2011 and 2022 was selected.
Primary and secondary outcome measures Coprimary outcomes included rates of hypertension diagnosis and rates of antihypertensive treatment initiation, treatment change and blood pressure measurement in patients newly diagnosed with hypertension.
Results In April 2020, the first month of lockdown, incident hypertension diagnosis rates fell by 65% (95% CI 64% to 67%) compared with historical trends and remained depressed until November 2021, leading to 51 000 fewer diagnoses than expected by March 2022. However, by March 2022, there were 2.6% fewer diagnoses than expected in Scotland, compared with 20%–30% fewer in other UK Nations. Rates of treatment initiation and change fell by 47% (95% CI 43% to 51%) and 36% (95% CI 33% to 38%), respectively, in April 2020. However, initiation rates rebounded above expectations and remained elevated until March 2022. Blood pressure measurements fell by 69% (95% CI 65% to 72%) in April 2020, recovering in February 2021.
Conclusions Hypertension diagnosis and management in UK primary care were significantly disrupted during the COVID-19 pandemic. Future studies should investigate the potential clinical implications for the cardiovascular health of the UK population.
- Hypertension
- COVID-19
- Primary Care
Data availability statement
Data are available upon reasonable request. This study is based in part on data from the Clinical Practice Research Datalink obtained under license from the UK Medicines and Healthcare products Regulatory Agency. The data are provided by patients and collected by the UK National Health Service as part of their care and support. The interpretation and conclusions contained in this study are those of the author/s alone. Because electronic health records are classified as “sensitive data” by the UK Data Protection Act, information governance restrictions (to protect patient confidentiality) prevent data sharing via public deposition. Data are available with approval through the individual constituent entities controlling access to the data. Specifically, the primary care data can be requested via application to the Clinical Practice Research Datalink (https://www.cprd.com).
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: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
We provide the most comprehensive account of the disruption to hypertension management in UK primary care during the COVID-19 pandemic.
We used a large electronic medical records database that is representative of the general UK population, allowing our results to be generalised to the UK as a whole.
A long follow-up period allowed us to describe the recovery process after pandemic restrictions were lifted.
We were unable to identify prescriptions received in secondary or tertiary care or to determine whether the prescribed medication was actually dispensed to the patient, as our data only includes prescriptions written by general practitioners.
We were unable to assess variation by ethnicity due to the large amount of missing data in the database.
Introduction
The COVID-19 pandemic has had a profound impact on the provision of healthcare in the UK and around the world.1 2 Policies designed to limit person-to-person contact led hospitals to cancel or delay elective care, altered patient’s health-seeking behaviour and drastically reduced the frequency of in-person visits to general practitioners (GPs).1 3–5 In the primary care setting, this led to a reduction in the number of diagnoses of chronic conditions and compromised the ability of physicians to provide quality care for these conditions. For example, in the UK, diagnoses of circulatory conditions in primary care fell by 43% and diagnoses of diabetes fell by 70% compared with historical trends, while care processes deemed essential for diabetes by the UK’s National Institute for Health and Care Excellence (NICE) fell by 76%–88%.6–8
Hypertension is a leading risk factor for cardiovascular disease and related mortality worldwide and is mostly managed in primary care settings.9 Guidelines for the diagnosis, treatment and monitoring of hypertension in primary care have been published by numerous international organisations.10–13 In the UK, NICE guidelines provide blood pressure thresholds for diagnosing hypertension, procedures for measuring blood pressure and recommendations for antihypertensive treatment.10 Upon a diagnosis of hypertension, the treatment guidelines recommend lifestyle changes, followed by a stepped treatment regimen in which patients are prescribed a single antihypertensive medication, with additional medications prescribed until blood pressure control is achieved.10 To incentivise physicians to be proactive in managing hypertension, NICE therefore includes regular blood pressure monitoring as an indicator of quality hypertension care in its Quality and Outcomes Framework, its pay-for-performance scheme.14–16 However, during the early phase of the COVID-19 pandemic (1 March 2020 to 10 December 2020), rates of blood pressure measurements among patients with diabetes were 47% lower than would be expected based on historical trends, and rates of new prescriptions for antihypertensive medications fell by 22%.8 A similar depression in rates of treatment and monitoring among patients with hypertension would lead to suboptimal blood pressure control and have significant clinical implications for the future cardiovascular health of the UK population.
While several studies have investigated the impact of the COVID-19 pandemic on acute cardiovascular outcomes, no study has assessed the impact on primary care for hypertension.17 18 We therefore used a large, representative UK primary care database to assess the impact of the pandemic on primary care for hypertension—including new diagnoses of hypertension, initiation or change of antihypertensive medications and rates of blood pressure measurement.
Methods
Data source
We used the UK Clinical Practice Research Datalink (CPRD) GOLD and Aurum databases.19 The CPRD is a large anonymised primary care medical records database of over 60 million patients enrolled in more than 2000 general practices and is broadly representative of the UK population in terms of geographical distribution, age and sex.19–21 The CPRD contains demographic information, medical diagnoses, procedures, prescriptions and referrals to specialists and hospitals. Medical diagnoses and procedures are recorded using the Read and SNOMED-CT clinical coding classification systems.20 21 Prescriptions written by a general practitioner are automatically recorded in the computerised medical file using a drug dictionary based on the British National Formulary. Numerous studies have shown the validity and high quality of the recorded data.20 21
Study population
To examine trends in hypertension diagnosis over time, we first formed a cohort of all adults aged 18 years and older registered at general practices contributing to the CPRD between 1 January 2011 and 31 March 2022. Cohort entry was defined as the patient’s eighteenth birthday, 1 year after the patient’s registration at a contributing practice, 1 year after the date their practice received data quality standards approval for research, or 1 January 2011, whichever occurred later. We excluded patients who had a hypertension diagnosis, defined as the presence of a diagnostic code for hypertension in the patient’s record at any time before cohort entry. As described above, all patients had a minimum 1-year look-back period to identify past hypertension diagnoses but some patients may have had longer look-back periods depending on their registration date with a general practice contributing to the CPRD. Moreover, the key past medical events and chronic diseases of a patient are typically recorded by their GP during their first medical visit. Thus, even if a patient was diagnosed with hypertension before their first visit to a CPRD contributing practice, their hypertension diagnosis would likely be recorded at their first visit. Patients were followed until the date they transferred out of a contributing practice, death, or the study end date (31 March 2022), whichever occurred first.
Within this cohort, we identified a subcohort of all patients newly diagnosed with hypertension during the study period to examine trends in antihypertensive treatment initiation and blood pressure measurements, with cohort entry defined as the date of diagnosis. Patients were required to have at least 1 year of medical history prior to diagnosis and no prescription of antihypertensive medication in the year before cohort entry.
Outcomes
The primary outcome was incident hypertension diagnosis, defined as the first diagnostic code for hypertension in a patient’s record during follow-up.22 23 Diagnostic codes used to identify hypertension are presented in online supplemental table S1. Among the subcohort of patients newly diagnosed with hypertension, the coprimary outcomes were (1) initiation and change (switch or add-on) of antihypertensive medication (calcium channel blockers, angiotensin II receptor blockers, ACE inhibitors, beta-blockers, thiazide/thiazide-like diuretics, alpha-blockers, potassium-sparing and loop diuretics, central acting agents, alpha-2 receptor agonists, peripheral adrenergic inhibitors and vasodilators) and (2) any form (clinic, ambulatory and home) of blood pressure measurement, identified using relevant Read codes and SNOMED terms. British National Formulary codes for antihypertensive medications are presented in online supplemental table S2. Treatment initiation was defined as the date of a patient’s first prescription for an antihypertensive medication after their hypertension diagnosis. Among patients previously prescribed an antihypertensive medication, treatment change was defined as the first prescription for a new class of antihypertensive medication during follow-up.
Supplemental material
Statistical analysis
To assess the impact of the pandemic on hypertension management, we compared observed monthly rates of hypertension diagnosis, antihypertensive treatment initiation and change, and blood pressure measurements between 1 March 2020 (given the first lockdown in the UK began on 23 March 2020) and 31 March 2022, to those expected based on 10-year historical trends. We calculated the observed monthly crude rates of patients newly diagnosed with hypertension as the number of patients with a first recorded hypertension diagnosis divided by the total person-time of follow-up (up to the first hypertension diagnosis) in each month. Among the subcohort of patients newly diagnosed with hypertension, we also estimated the rate of patients newly prescribed an antihypertensive medication in the 3 months following hypertension diagnosis. Thus, follow-up was limited to a maximum of 3 months after cohort entry (hypertension diagnosis) for this analysis. We calculated monthly rates of antihypertensive treatment initiation as the number of patients with a first prescription of an antihypertensive medication divided by the total person-time of follow-up for each month. Next, to calculate monthly rates of antihypertensive medication change, we considered the subcohort of patients newly diagnosed with hypertension who initiated an antihypertensive drug, with cohort entry defined as the first prescription. Monthly rates corresponded to the number of patients prescribed a new antihypertensive drug class divided by the total person-time accumulated in this subcohort each month. Finally, we calculated monthly rates of blood pressure measurement among patients newly diagnosed with hypertension as the number of blood pressure measurements divided by the total person-time accumulated by patients with hypertension each month. Patients could contribute multiple measurements in each month, but multiple measurements occurring on the same day were counted as a single event.
Expected rates were estimated using a negative binomial regression model fit to monthly event counts between 1 January 2011 and 31 December 2019, using the natural logarithm of person-time at risk as an offset. The offset, therefore, differed for each outcome. All models included terms for calendar month and the number of months since January 2011 to control for seasonal and secular trends. We also conducted stratified analyses by age (≤50, 51–70 and >70 years), sex, and UK Nation (England, Northern Ireland, Scotland and Wales), and fit models separately to the stratum-specific event counts. Fitted models were then used to predict expected monthly rates between 1 March 2020 and 31 March 2022.
All analyses were performed by using SAS V.9.4 (SAS Institute).
Patient and public involvement
The conduct of our study was motivated by the information available in the scientific literature. As such, no patients were included during the design and conception of the study. Similarly, no patients were involved in setting the research question or the outcome measures, nor were they involved in the design and implementation of the study.
Results
Between 1 January 2011 and 31 March 2022, we identified 23 076 390 adults in the CPRD, of which 712 461 were diagnosed with hypertension during follow-up (online supplemental figure S1). Prior to the pandemic, between January 2011 and December 2019, hypertension incidence rates were stable with minor seasonal variation, averaging 74 per 100 000 person-months (SD: 9.3). As shown in figure 1A, in April 2020, the first full month of lockdown, the incidence rate fell to 28 per 100 000 person-months, a 65% (95% CI 64% to 67%) reduction compared with historical trends. Rates did not return to expected levels until November 2021, 20 months after the initial lockdown, with approximately 51 000 fewer hypertension diagnoses than expected between January 2020 and March 2022 (figure 1B). Little variation was observed by age group in either diagnosis rates (figure 2A) or cumulative incidence (online supplemental figure S2A) relative to their expected values. Diagnosis rates in both men and women were approximately 65% lower than expected in April and May 2020, but from July 2020 onward, women were consistently diagnosed at a rate closer to their expected levels than men (figure 2B), leading to a smaller deficit in cumulative diagnoses among women than men (online supplemental figure S2B). The immediate impact of the lockdown varied between UK Nations: Northern Ireland had the greatest reduction in hypertension diagnoses in April 2020 (74%; 95% CI 72% to 76%) and Scotland the smallest (54%; 95% CI 52% to 57%), while England and Wales were similarly impacted (66%; 95% CI 65% to 67% and 68%; 95% CI 66% to 70%) (figure 2C). After September 2020, hypertension was diagnosed at consistently higher rates in Scotland than the other nations (figure 2C). As a consequence, by March 2022, the observed cumulative incidence was only 2.6% (95% CI 1.5% to 3.7%) below expectation in Scotland while ranging between 20% and 30% lower than expected in other nations (online supplemental figure S2C).
Observed and expected monthly incidence rates of new hypertension diagnoses per 100 000 person-months (A); difference between observed and expected cumulative incidence of hypertension diagnosis (B).
Difference between observed and expected hypertension incidence rates expressed as a percentage of expected rates, stratified by age group (A), sex (B) and UK Nation (C).
Among patients with hypertension, rates of antihypertensive treatment initiation fell by 47% (95% CI 43% to 51%) in April 2020 (figure 3A). However, by July 2020, treatment was initiated at a rate 78% higher than expected, and rates remained above historical trends thereafter (figure 3A). By March 2022, the sustained elevation in treatment initiation resulted in approximately 26 000 more initiations than expected, which were similarly distributed among age groups (figure 3B) and sexes (figure 3C) but showed substantial variation by nation (figure 3D). Indeed, cumulative incidences in England and Scotland were 37% (95% CI 35% to 39%) and 34% (95% CI 31% to 37%) higher than expected, respectively, while in Wales and Northern Ireland, the numbers were 9% (95% CI 9% to 11%) and 13% (95% CI 9% to 17%), respectively (figure 3D).
Observed and expected rates of antihypertensive treatment initiation per 100 000 person-months (A); difference in observed and expected cumulative incidence of treatment initiation as a percentage of the expected value, stratified by age group (B), sex (C) and UK Nation (D).
Among patients with hypertension prescribed antihypertensive drugs, rates of treatment change fell by 36% (95% CI 33% to 38%) in April 2020 but quickly recovered, reaching 36% (95% CI 30% to 42%) higher than expected in July 2020, and settling back to expected levels thereafter (figure 4). Rates of treatment switching were similar regardless of age, sex and UK nation (online supplemental figure S3).
Observed and expected rates of antihypertensive treatment change per 100 000 person-months.
The rate of blood pressure measurements among patients with hypertension fell by 69% (95% CI 65% to 72%) in April 2020 and remained reduced until February 2021 (figure 5). Overall, from January 2020 to March 2022, the cumulative incidence of blood pressure measurements was 13% lower than expected. There was little variation in rates of blood pressure measurements between age groups, sexes or UK nations (online supplemental figure S4).
Observed and expected monthly rates of blood pressure measurements per 100 000 person-months among patients with hypertension.
Discussion
In this large population-based cohort study, we documented the disruption and adaptation of primary care for hypertension in the UK during the COVID-19 pandemic. Compared with historical trends, the number of new diagnoses of hypertension fell by 65% in April 2020, the first full month of lockdown. Diagnosis rates remained depressed until November 2021, suggesting a sizeable backlog of undiagnosed cases across the UK. While no age group was significantly more affected than another, there were substantially fewer missed or delayed diagnoses among women than men by March 2022 (relative to sex-specific expectations). National variation was also evident, as the backlog was only 2.6% of expected diagnoses in Scotland by March 2022 but 20%–30% in England, Wales and Northern Ireland. Beyond the reduction in diagnosis rates, there was a notable change in how physicians managed hypertension after the lockdown. Following a 47% reduction in the rate of new antihypertensive treatment initiation in April 2020, rates remained elevated from July 2020 to March 2022. This apparently lasting change in prescribing practice was clearest in England and Scotland and less so in Wales and Northern Ireland. The effect of the lockdown on rates of treatment change was transient, with rates returning to expected levels by August 2020. Finally, blood pressure monitoring was also affected and did not recover until February 2021.
Several studies have shown the adverse impact of the COVID-19 pandemic on primary care in the UK. Weekly primary care contacts for acute physical and mental health conditions fell dramatically in the first few months of the pandemic, with acute cardiovascular conditions such as stroke, heart failure and myocardial infarction being especially affected.24 Similar results for chronic mental and physical conditions, including circulatory system diseases and type 2 diabetes, were found in a deprived urban population in Salford, UK early in the pandemic.6 An analysis of UK-wide primary care records between March and December 2020 further showed that rates of all essential health checks for diabetes fell well below historical levels, underscoring the effect of the pandemic on primary care.8
Using a detailed primary care medical records database, we found a similar substantial impact of the pandemic on the diagnosis, treatment and monitoring of hypertension in UK primary care. One study reported a reduction in new diagnoses of circulatory system diseases, a composite outcome that included hypertension, in the first 2 months of the pandemic.6 Another study described trends in the dispensing of antihypertensive medications as a proxy for the management of hypertension in England, Scotland and Wales between March 2020 and July 2021.25 It reported only a small decline in overall dispensing of antihypertensive medications during the pandemic, similar to findings in other countries.25 26 However, the absolute number of incident dispensing in the UK declined significantly following the first lockdown in March 2020, and monthly dispensing counts remained below prepandemic levels as late as July 2021, the end of their study period.25 Within a well-defined cohort, we separately described trends in diagnostics and prescriptions, thereby providing a more comprehensive description of hypertension management during the pandemic. Indeed, absolute numbers of incident dispensing will naturally decline if diagnosis rates decline, which will confound any inference about changes in prescribing practices. The excess 26 000 initiations observed in our study must be understood, therefore, not as an increase in absolute number, but rather an increase relative to expectated numbers given the amount of person-time accumulated in our cohort of individuals with hypertension. Indeed, by estimating treatment initiation rates among patients with established hypertension, we were able to account for changes in the timing of treatment initiation that cannot be inferred from changes in absolute dispensing numbers. Our results on diagnosis rates and treatment initiation rates, therefore, also complement those found using aggregate data to investigate small-area variation in the impact of the pandemic in England.27
Our results have important clinical implications. First, the large backlog of missed hypertension diagnoses in England, Wales and Northern Ireland must be addressed to protect the future cardiovascular health of the UK population. Delays in diagnosis mean greater time spent with uncontrolled blood pressure, increasing the risk of myocardial infarction, stroke and mortality associated with hypertension.28–31 Further research to explain Scotland’s disproportionate success in diagnosing hypertension may be of value to policy-makers. Second, there was a persistent elevation in the rate of antihypertensive treatment initiation after July 2020 that was highest in England and Scotland. Antihypertensive treatment was, therefore, prescribed earlier after a diagnosis than before the pandemic. There are several possible explanations for this finding. Patients seen during the height of lockdown may have been sicker and therefore require earlier treatment. However, if this were true, the rates would likely drop as restrictions ease, which was not observed. Alternatively, physicians would be aware of the delays in diagnosis caused by the lockdowns and could be initiating treatment earlier to counteract those delays, or because they anticipate delays in scheduling follow-up appointments. Finally, hypertension was also established as a risk factor for severe COVID-19 early in the pandemic which may have increased the perceived urgency of lowering blood pressure.32 33 Rates of antihypertensive treatment change were only minimally affected, which suggests little to no increase in clinical inertia despite the reduction in blood pressure monitoring rates.34 35 Indeed, blood pressure monitoring did not return to expected levels until approximately February 2021, several months after the NHS began distributing home blood pressure monitors in England (in October 2020).36
Our study has several strengths. First, we used a large, representative primary care medical records database to identify incident hypertension diagnoses, antihypertensive treatment prescriptions and blood pressure measurements. Second, the cohort design made it possible to investigate the impact of the pandemic on prescribing practices among patients with hypertension. Third, the follow-up lasted until March 2022, which allowed a more comprehensive description of the recovery process after pandemic-related restrictions were lifted. However, several limitations must also be noted. To identify hypertension diagnoses, we relied on diagnostic codes entered in a patient’s record. If physicians were less likely to enter such codes during the pandemic, the estimates of diagnosis rates may be underestimated. Similar considerations apply to the blood pressure measurement outcome. Also, the CPRD only includes data on prescriptions written by primary care physicians. We were therefore unable to identify prescriptions received in secondary or tertiary care, or to determine whether the medication was actually dispensed to the patient. However, hypertension is primarily managed in primary care, and we expect that a patient prescribed an antihypertensive medication outside of primary care would renew that prescription with their primary care physician. We therefore do not expect this limitation to significantly affect the results. Finally, we were unable to assess variation by ethnicity due to the large amount of missing data in the database.
Overall, the COVID-19 pandemic had a significant impact on the diagnosis and management of hypertension in primary care in the UK, with potential implications for the cardiovascular health of the UK population. Future research should assess the long-term effect of this disruption on cardiovascular event rates. Investigation of the variation observed between UK nations and between the sexes, in terms of policy and/or behavioural differences, may aid in the development of pandemic response strategies. Understanding the lasting impact of the pandemic on primary care may help to adequately plan for possible future disruptions on the scale of COVID-19 and minimise the indirect, adverse consequences of public health measures taken in response.
Data availability statement
Data are available upon reasonable request. This study is based in part on data from the Clinical Practice Research Datalink obtained under license from the UK Medicines and Healthcare products Regulatory Agency. The data are provided by patients and collected by the UK National Health Service as part of their care and support. The interpretation and conclusions contained in this study are those of the author/s alone. Because electronic health records are classified as “sensitive data” by the UK Data Protection Act, information governance restrictions (to protect patient confidentiality) prevent data sharing via public deposition. Data are available with approval through the individual constituent entities controlling access to the data. Specifically, the primary care data can be requested via application to the Clinical Practice Research Datalink (https://www.cprd.com).
Ethics statements
Patient consent for publication
Ethics approval
The study protocol was approved by the CPRD Research Data Governance (No. 22_001921) and the Research Ethics Board of the Jewish General Hospital, Montreal, Canada.
References
Footnotes
Contributors KJ contributed to the conceptualisation, formal analysis, methodology, visualisation, validation, writing—original draft, and writing—review and editing. SB contributed to the formal analysis, methodology, software, visualisation, validation and writing—review and editing. JB contributed to the validation, writing—review and editing. RWP contributed to the validation and writing—review and editing. CR contributed to the conceptualisation, methodology, supervision, validation, writing—original draft and writing—review and editing. CR is the guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests RWP has done consulting for Biogen, Boehringer Ingelheim, Merck, Nant Pharma, Pfizer and Vanda Pharmaceuticals, unrelated to this work.
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.