Cancer diagnostic services were badly impacted by the COVID-19 pandemic. Fewer patients reported cancer symptoms to their family doctor, referral numbers fell by more than half, and camera testing services all but stopped. The impact on the numbers of cancers diagnosed is not fully understood. Reports suggest that 3,500 fewer people than expected were diagnosed and treated for bowel cancer in England between April and October 2020. The Cancer Registry is the gold standard resource for the type and number of cancer diagnoses as well as stage at diagnosis. However, data for 2020 will not be available for several years.
We will use Clinical Practice Research Datalink (Aurum) data to explore the immediate impact of COVID-19 on the number of cancers diagnosed, and the recovery back to expected levels. We will report the results by cancer type, age group, sex and region. We will repeat this for type-2 insulin-dependent diabetes mellitus, which is diagnosed without secondary care input. Having this comparison group will help us to start to differentiate the COVID-19’s impacts on recorded cancer incidence associated with primary and secondary care factors. In this way our results will help us to:
1. Understand the size and duration of diagnostic delays.
2. Inform cancer diagnostic service planning about potential increases in demand by cancer site, or patient age or sex.
3. Lay the groundwork for a subsequent validation study of cancer diagnoses in the CPRD around the time of the pandemic once Cancer Registry data for that period become available.
Cancer diagnostic services were severely impacted by COVID-19. Immediately after the March 2020 lockdown, patient consultations in primary care for possible cancer symptoms reduced by one-quarter, urgent referral rates were 40% of the expected level, and endoscopy services were running at 5% of the normal level. Cancer diagnostic services had started to recover by July, but delays in diagnosis are likely. For example, between April and October 2020, 3,500 fewer people than expected were diagnosed and treated for colorectal cancer in England.
Interrupted time series models will investigate the change in recorded cancer incidence in the CPRD across the pre-pandemic (01/03/2017–28/02/2020) and post-pandemic (29/02/2020–28/02/2022) eras. We will analyse recorded incidence of insulin-dependent type-2 diabetes mellitus, which is diagnosed wholly in primary care, to help differentiate COVID-19’s impacts on primary and secondary care factors. The outcome will be the count of incident diagnostic codes by age group, sex and general practice per 28-day aggregation period. A pre-/post-pandemic interruption variable will capture changes in incidence associated with the pandemic. Other variables will capture the pre-interruption trend in incidence and the deviation from it, as well as seasonal effects. The models will include sex, age group, region, plus a random effect for general practice and an offset term for the denominator population at risk.
Understanding any deviations will serve three purposes:
1. To understand potential delays in diagnosis of cancer, in terms of number of delayed diagnoses and the duration of that delay.
2. To inform cancer diagnostic service planning about potential increases in demand by cancer site, or patient demographic based on a previously lower than expected cancer incidence.
3. To lay the groundwork for a subsequent validation study of cancer diagnoses in the CPRD around the time of the pandemic once Cancer Registry data for that period become available.
Cancer diagnosis; type-2 diabetes mellitus diagnosis
Sarah Price - Chief Investigator - University of Exeter
Sarah Price - Corresponding Applicant - University of Exeter
Daniel Jones - Collaborator - University of Leeds
Luke Mounce - Collaborator - University of Exeter
Sarah Bailey - Collaborator - University of Exeter