key: cord-343982-ymaql0hx
authors: Carr, M. J.; Wright, A. K.; Leelarathna, L.; Thabit, H.; Milne, N.; Kanumilli, N.; Ashcroft, D. M.; Rutter, M. K.
title: Impact of COVID-19 on the diagnoses, HbA1c monitoring and mortality in people with type 2 diabetes: a UK-wide cohort study involving 13 million people in primary care
date: 2020-10-27
journal: nan
DOI: 10.1101/2020.10.25.20200675
sha: 
doc_id: 343982
cord_uid: ymaql0hx

Background: The COVID-19 pandemic has already disproportionately impacted people with diabetes. Timely diagnosis and appropriate monitoring in people with type 2 diabetes (T2D) are necessary to reduce the risk of long-term complications. Methods: We constructed a cohort of 23M patients using electronic health records from 1709 UK general practices registered with the Clinical Practice Research Datalink (CPRD), including 13M patients followed between March and July 2020. We compared trends in diagnoses, monitoring and mortality in T2D, before and after the first COVID-19 peak, using regression models and 10-year historical data. We extrapolated the number of missed or delayed diagnoses using UK Office for National Statistics data. Findings: In England, rates of new T2D diagnoses were reduced by 70% (95% CI 68%-71%) in April 2020, with similar reductions in Northern Ireland, Scotland and Wales. Between March and July, we estimated that there were more than 45,000 missed or delayed T2D diagnoses across the UK. In April, rates of HbA1c testing in T2D were greatly reduced in England (reduction: 77% (95% CI 76%-78%)) with more marked reductions in Northern Ireland, Scotland and Wales (reduction: 84% (83-84%)). Reduced rates of diagnosing and HbA1c monitoring were particularly evident in older people, in males, and in those from deprived areas. Mortality rates in T2D in England were more than 2-fold higher (110%) in April compared to prior trends, but were only 66% higher in Northern Ireland, Scotland and Wales. Interpretation: As engagement with the NHS increases, healthcare services will need to manage the backlog and the expected increase in T2D severity due to delayed diagnoses and reduced monitoring. Older people, men, and those from deprived backgrounds with T2D may be groups to target for early intervention. Funding: National Institute for Health Research (NIHR) Greater Manchester Patient Safety Translational Research Centre

The first wave of the COVID-19 pandemic has had major health and economic effects across the world. So far in the UK, there have been more than 50K COVID-related deaths 1 with disproportionate impacts in people with diabetes; 2 nearly a third of all COVID-related deaths having occurred in people with diabetes. 3 The impact on the NHS, and in particular on diabetes services, has been enormous, with the suspension of much routine care. As we enter the second wave in the UK, there is an urgent need to minimise the harm done through suspension of routine services and to prioritise care and resources to areas of greatest need.

The diagnosis of type 2 diabetes occurs almost exclusively in primary care. 4 Timely diagnosis is critically important as delays will increase the risk of long-term complications.

There is limited data on the indirect consequences of the COVID-19 pandemic on the incidence and monitoring of diabetes in primary care. Likewise, there is limited information on COVID-19 impacts on mortality rates in people with diabetes during and after the first wave of COVID-19. Therefore we used a large primary care longitudinal dataset, broadly representative of the UK population, aiming to compare: i) the UK-wide incidence of type 2 diabetes; ii) the frequency of HbA1c testing; and iii) mortality rates in people with type 2 diabetes, before and after the nationwide COVID-19 lockdown in March 2020. We compared observed and predicted rates using data covering ten years prior to the pandemic.

Since older people and more socially disadvantaged groups have been disproportionally affected by COVID-19 infections, and since the same groups may be more adversely impacted by the unintended consequences of government interventions, we aimed to study variation in outcomes by gender, age group, deprivation level and region.

We examined primary care electronic health records using the Clinical Practice Research Datalink (CPRD) Aurum and GOLD databases. 5, 6 The study population consisted of 19,763,481 patients from 1,368 general practices in England, with a further 36 practices in Northern Ireland (339,153 patients), 195 practices in Scotland (1,804,938 patients), and 110 in Wales (1,277,009 patients).

A total of 22,717,623 patients were included for estimating the expected rates in the pre-COVID-19 period (January 2010-February 2020). In line with guidance from the CPRD's central administration, Aurum and GOLD databases were analysed separately. The CPRD contains anonymised consultation records and includes patient demographic information, symptoms, diagnoses, medication prescriptions, and date of death. We also examined practice-level Index of Multiple Deprivation (IMD) quintiles, 7 a measure representing an area's relative level of deprivation, ranked within each UK nation.

. CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020. 10.25.20200675 doi: medRxiv preprint To enable comparisons of rates before the COVID-19 outbreak, during its peak and after the peak, we included patient records from January 2010 that established long-term trends and patterns of seasonality.

We focussed primarily on reporting observed versus expected rates from 1/3/20 to 10/7/2020. First, we estimated incidence rates of type 2 diabetes diagnoses, new prescriptions for metformin (the most commonly prescribed medication in new-onset type 2 diabetes) and insulin, and rates of HbA1c testing and mortality in people with type 2 diabetes. Incident type 2 diabetes was identified from Read/SNOMED/EMIS codes used in CPRD GOLD and Aurum (see https://clinicalcodes.rss.mhs.man.ac.uk). The CPRD Aurum and GOLD databases were analysed separately, with data from Aurum restricted to English practices and GOLD providing information on practices in Northern Ireland, Scotland and Wales. The use of two discrete data sources also enabled independent replication of our findings. All code lists and medication lists were verified by two senior clinical academics (a diabetologist: MKR, and a senior academic pharmacist: DMA).

For each patient, we defined a 'period of eligibility' for study inclusion which commenced on the latest of: the study start date (1st January 2010); the patient's most recent registration with their practice; the date on which data from the practice was deemed 'up-to-standard' by the CPRD. A patient's period of eligibility ended on the earliest of: registration termination; the end of data collection from their practice; death. For incident diagnoses and prescriptions, we also applied a 'look-back' period during which a patient was required to have been registered for at least a year prior to the event. Flow diagrams illustrating the delineation of the study cohorts using CPRD Aurum and GOLD are presented in supplementary figures 7 and 8 respectively. The denominator for the incidence rates was the aggregate person-months at risk for the whole eligible study population. Mortality and testing rates in people with type 2 diabetes were calculated using the person-months at risk from all those with type 2 diabetes as the denominator.

Incidence, mortality and testing rates were stratified by gender, age group (<18, 18-29, 30-44, 45-64, 65-79 and ≥80 years), practice-level deprivation (IMD quintiles) and region (in England) or nation (in the rest of the UK).

The data were structured in a time-series format with event counts and 'person-months at risk' aggregated (by year and month) with stratification by gender, age group, deprivation quintile and region (or nation in GOLD). Mean-dispersion negative binomial regression models were used to estimate expected monthly event counts from March 2020 onward based on antecedent trends since 2010. The natural logarithm of the denominator (person-months at risk) was used as an offset in each regression model. To account for possible seasonality and long-term linear trends, calendar month was fitted as a categorical variable and time as a continuous variable with the number of months since the start of the study serving as the unit of measurement. For each month studied, observed and expected event counts were converted to rates using the observed person-month denominator. The monthly expected rates, and their 95% confidence intervals, is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint were plotted against the observed rates. As they share a common denominator, differences between expected and observed monthly rates are expressed as a percentage 'rate reduction (or increase)'.

Extrapolated estimates of the number of missed (or delayed) diagnoses of type 2 diabetes were derived using the discrepancy between observed and expected frequencies from March 2020 onward, and approximations of the proportional representation of the populations of England and the rest of the UK (in CPRD Aurum and GOLD respectively) using data from the Office for National Statistics. 8 All data processing and statistical analyses were conducted in Stata version 16 (StataCorp LP, College Station, TX). We followed RECORD (REporting of studies Conducted using Observational Routinelycollected health Data) guidance (see online supplement). 9

The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Our focus was on the impact of the COVID-19 pandemic between March and July 2020. Using the inclusion criteria described in the Study Design, a mixed cohort was utilised consisting of patients whose period of eligibility began before 1st March 2020 and those who became eligible for inclusion between 1st March 2020 and 10th July 2020. The study cohort was comprised of 13,352,550 patients (median (IQR) age: 42 (25, 59) years, 50% female) of whom 707,103 had type 2 diabetes. Of those with type 2 diabetes, the median (IQR) age was 67 (57, 77) years, 44% were female and 25% lived in an area that was in the most deprived quintile compared to the rest of the UK.

In April 2020, the rate of new diagnoses of type 2 diabetes in English primary care was reduced by 70% (95% CI 68% to 71%) compared to the expected rates based on 10-year historical trends (figure 1a; supplementary table 1). Prior to March 2020, rates of type 2 diabetes diagnoses in English practices were higher in older individuals, in men, and in people from deprived areas. These groups experienced the greatest reductions in rates for new type 2 diabetes diagnosis at the time of the first COVID-19 peak (supplementary figure 1 ). The reduced rates of type 2 diabetes diagnosis in April were mirrored by reduced rates of new metformin prescriptions in English practices (reduction: 53% (95% CI 51% to 55%; In April, rates of HbA1c testing in England were greatly reduced in people with type 2 diabetes (reduction: 77% (95% CI: 76% to 78%)); is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. The reduced rates of diagnosis, new insulin/metformin prescribing and HbA1c testing increased gradually between May and July 2020 though levels remained well below expected rates based on 10-year historical data ( figure 1a-d) . Overall in English practices, between 1/3/20 and 10/7/20, the rate of diagnosis of type 2 diabetes was reduced by 46% (95% CI: 44% to 49%), metformin prescribing was reduced by 33% (30% to 35%), insulin prescribing fell by 12% (7% to 16%), and HbA1c testing in people with type 2 diabetes was reduced by 48% (46% to 49%); supplementary table 1.

. CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint

In April 2020, mortality rates in people with type 2 diabetes in England were more than 2-fold higher compared to prior trends (mortality rate increase: 110% (95% CI: 102% to 118%); Figure 2a; supplementary table 1). Peaks in mortality were seen particularly in individuals aged over 65 years (supplementary figure 5a). Mortality rates returned to expected levels in people with type 2 diabetes and sub-groups between May and June 2020 (Figure 2a) . Overall, between 1/3/20 and 10/7/20, the rate of mortality in people with type 2 diabetes in English practices was increased by 30% (95% CI: 25% to 35%); supplementary table 1. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint metformin prescribing reduced: 26% vs. 33%; supplementary tables 1 and 2). Over the same 4½ month period, the overall reduction in HbA1c testing in type 2 diabetes was greater in CPRD GOLD practices based in Northern Ireland, Scotland and Wales (GOLD vs. Aurum: 56% vs. 48%) but the mortality rate increase was lower than in England (GOLD vs. Aurum: 16% vs. 30%; figure 2; supplementary tables 1 and 2).

Using primary care data from 13 million people in the UK, and 10-year historical data, we have shown that within the first 4 months of the nationwide 'lockdown' in March 2020, the indirect consequences of the COVID-19 pandemic led to: i) a 69-70% reduction in new diagnoses of type 2 diabetes, with older individuals, males, and people from deprived areas experiencing the greatest reduction in diagnosis rates;

ii) a 77-84% reduction in HbA1c testing; iii) a reduction in metformin and insulin prescribing, particularly in older people with type 2 diabetes, supporting the reduced rates of diagnosis and monitoring; and iv) a short-term 110% increase in mortality rate in people with type 2 diabetes in England and a 66% increased mortality rate across the rest of the UK.

There is limited prior data on the impact of the COVID-19 pandemic on the diagnosis of type 2 diabetes. A study using primary care data from Salford, UK showed 135 fewer diagnoses of type 2 diabetes than expected between March and May 2020, which amounted to a 49% reduction in activity. 10 Here we extend these observations by assessing primary care data across the UK and by providing supplementary data on HbA1c testing and mortality. We show that the reduced rate of diagnosis applies to all areas of the UK and not just to deprived areas of the UK such as Salford. To the best of our knowledge, no study has reported the impact of the COVID-19 pandemic on HbA1c monitoring in diabetes, and no study has described national variation in mortality rates in people with type 2 diabetes following the first peak of the pandemic.

Our data have important clinical implications. In early March 2020, GPs were advised to minimise the number of face-to-face contacts they had with their patients, including NHS health-checks. 11 Our data suggests that this reduction of clinical services has led to major reductions in the diagnosis and monitoring of type 2 diabetes. The concomitant reductions in new prescriptions issued for metformin and insulin further support these findings. Type 2 diabetes develops over many years, so it seems unlikely that people's behaviour during the pandemic has reduced the true incidence of these conditions. Assuming that the true incidence of type 2 diabetes has remained constant from March 2020, our data suggest that, across the UK, the indirect consequences of the pandemic have led to more than 45K missed/delayed diagnoses of type 2 diabetes in the 4½ months between 1/3/20 and 10/7/20. This figure may be an underestimate if sedentary lifestyles and adverse dietary changes during lockdown have increased obesity rates in the general population. 12 These data are a clinical concern because undiagnosed type 2 diabetes will cause potentially serious long-term complications.

The huge reduction in the rate of HbA1c testing is another important concern for people with type 2 diabetes, because they, and their clinicians, often rely solely on HbA1c data to make decisions about . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint treatment. The reduction in new prescriptions for insulin was largely observed in older individuals suggesting this reduction was explained by a failure to intensify therapy in people with poorly controlled long-duration type 2 diabetes. There are already concerns in the UK about clinical inertia in diabetes management, with frequent failures to escalate care when glucose control is poor. 13 These HbA1c data indicate potential further delays in the management of type 2 diabetes that are predicted to cause avoidable diabetes-related long-term complications. A reduced frequency of HbA1c testing in primary care might also contribute to missing people with non-diabetic hyperglycaemia who might benefit from referral to the NHS Diabetes Prevention Programme.

The higher COVID-related death rate in people with diabetes has been well-documented, 2,3,14 and our data support these observations. Here, we add to these data by showing national differences in the impact of COVID-19 on mortality rates in people with type 2 diabetes, with higher rates observed in England compared to the rest of the UK. Further research is required to understand how population characteristics including ethnicity, population density and deprivation might explain these differences.

As engagement with health services increases, and hopefully is maintained during the second COVID-19 peak, our data predict a marked increase in presentations with incident type 2 diabetes. Should this occur, then healthcare services will need to manage this backlog, and the expected increase in the severity of diabetes brought about by delayed diagnoses. Older individuals, males and people from deprived backgrounds appear to be most adversely affected by reductions in rates of diagnosis and monitoring of type 2 diabetes. As outpatient diabetes services start to open up, these individuals may be a group to target for early intervention, and in particular, for HbA1c testing and treatment intensification when appropriate. If a second full lockdown occurs, then effective public communications should ensure that patients remain engaged with diabetes services including HbA1c screening 15 and monitoring, and the use of remote consultations. 16, 17 Our study had several strengths: this is the first UK-wide study reporting the indirect impact of the COVID-19 pandemic on the diagnosis of type 2 diabetes, related prescribing and HbA1c testing in primary care.

Our findings in English practices were replicated using data from other parts of the UK. By combining assessments of diabetes coding and prescribing, our data supports the conclusion that reduced rates of diagnoses are genuinely explained by missed diagnoses. Our study has some limitations: First, ethnicity coding is not adequately captured in primary care and therefore we had limited ability to explore ethnicityrelated variation in care and outcomes. Future studies will incorporate linked secondary care data that has more complete capture of ethnicity data. Second, it is possible that some diabetes diagnoses may have been made in a hospital setting following an acute presentation and that the related primary care coding had not been updated at the time of our data extraction. While hospital presentation of incident diabetes may have occurred in some instances, it would not explain the reductions in new prescribing for metformin and this potential explanation does not fit with our local experience. In general, people have avoided hospital attendance during the pandemic. For example, one study documented a 23% reduction in emergency admissions in the UK. 18 Finally, although our results and conclusions are relevant to the UK population, generalisability to other healthcare systems may be limited. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020. 10.25.20200675 doi: medRxiv preprint In conclusion, we highlight marked reductions in the diagnosis and monitoring of type 2 diabetes as indirect consequences of the COVID-19 pandemic. Over the coming weeks, healthcare services will need to manage this predicted backlog, and the expected increase in the severity of diabetes due to delayed diagnoses. Older people, men and those from deprived backgrounds with type 2 diabetes may be specific groups to target for early HbA1c testing and intervention. Should a second full national lockdown occur, then effective public communications should ensure that patients remain engaged with diabetes services including HbA1c screening and monitoring and the use of remote consultations. The lead authors (MJC and AKW: the manuscript's guarantors) affirm that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained. The corresponding author had full access to all of the data and the final responsibility to submit for publication. 

All clinical codes used in the study are published on Clinicalcodes.org. Electronic health records are, by definition, considered "sensitive" data in the UK by the Data Protection Act and cannot be shared via public deposition because of information governance restriction in place to protect patient confidentiality. Access to data are available only once approval has been obtained through the individual constituent entities . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint Supplementary figure 2: Comparison of monthly HbA1c testing rates in people with type 2 diabetes in primary care by age, gender, deprivation level and by region before and after the first COVID-19 peak in England (CPRD Aurum) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint Supplementary figure 3: Comparison of monthly incidence rates for metformin prescribing in primary care by age, gender, deprivation level and by region before and after the first COVID-19 peak in England (CPRD Aurum) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint Supplementary figure 4: Comparison of monthly incidence rates for insulin prescribing in primary care by age, gender, deprivation level and by region before and after the first COVID-19 peak in England (CPRD Aurum) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint Supplementary figure 5: Comparison of monthly mortality rates in people with type 2 diabetes in primary care by age, gender, deprivation level and by region before and after the first COVID-19 peak in England (CPRD Aurum) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint Supplementary figure 6: Comparison of observed and expected monthly incidence rates for type 2 diabetes in primary care, HbA1c monitoring and new prescriptions for metformin and insulin before and after the first COVID-19 peak in Northern Ireland, Scotland and Wales (CPRD GOLD) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020. 10 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20200675 doi: medRxiv preprint

UK's record on pandemic deaths

Risk factors for COVID-19-related mortality in people with type 1 and type 2 diabetes in England: a population-based cohort study

Associations of type 1 and type 2 diabetes with COVID-19-related mortality in England: a whole-population study

Van den Bruel A. Opportunities for earlier diagnosis of type 1 diabetes in children: A case-control study using routinely collected primary care records

Data Resource Profile: Clinical Practice Research Datalink (CPRD)

Diagnosis of physical and mental health conditions in primary care during the COVID-19 pandemic: a retrospective cohort study

Letter to general practices

The silent pandemic: How lockdown is affecting future health

Legacy benefits of blood glucose, blood pressure and lipid control in individuals with diabetes and cardiovascular disease: Time to overcome multifactorial therapeutic inertia?

Factors associated with COVID-19-related death using OpenSAFELY

H(ome)bA1c testing and telemedicine: High satisfaction of people with diabetes for diabetes management during COVID-19 lockdown

Endocrinology in the time of COVID-19: Remodelling diabetes services and emerging innovation

Video consultations for COVID-19

What is happening to non-COVID deaths?

Reasons: (a) registration ended prior to study period due to death or migration from practice (n = 12,440,573); (b) registration end date defined on or prior to registration start date (n = 21,315); (c) aged under 5 on study period end

) and/or compare expected and observed rates

Acceptability defined by CPRD as meeting certain quality standards

Reasons: (a) registration ended prior to study period due to death or migration from practice (n = 2,159,261); (b) registration end date defined on or prior to registration start date (n = 146,723); (c) aged under 5 on study period end

) and/or compare expected and observed rates

Supplementary table 1: Comparison of observed and expected monthly incidence rates for type 2 diabetes in primary care, HbA1c monitoring in type 2 diabetes, new prescriptions for metformin and insulin, and deaths in people with type 2 diabetes between 1/3/20 and 10/7/20 and at the COVID-19 peak in April 2020, in England (CPRD Aurum)