"A cohort study of mortality in the Clinical Practice Research Datalink (CPRD) primary care database - CPRD Aurum"

Date of Approval: 
2020-12-15 00:00:00
Lay Summary: 
Deaths from any cause are routinely monitored as a measure of population health. Until recently mortality has been declining rapidly in the UK. However, over the past decade there is evidence that this trend is flattening out and life expectancy may no longer be increasing. The reasons for this are poorly understood. Now, we also have the pandemic of Covid19 which is acknowledged to have increased mortality in the UK in 2020. The Clinical Practice Research Datalink (CPRD) Aurum is a new large population-based data resource that enables us to gain new insights into population mortality in the UK. The study initially aims to compare mortality rates in CPRD Aurum with those from national registration systems. We will then evaluate changes over time in mortality in different age groups and in men and women. We will specifically evaluate how the Covid19 pandemic has impacted on mortality. Finally, we will evaluate how much mortality varies between general practices and whether this overall variation can be explained by variables in the dataset. These analyses will help to show whether CPRD Aurum is a suitable data source for mortality analysis. The analyses will provide substantive evidence concerning changes in mortality over time, with reference to the pandemic, and variation among general practices.
Technical Summary: 
This study aims to explore all-cause mortality in CPRD Aurum. The specific objectives are: i) to compare mortality rates between CPRD Aurum and published mortality rates from national registration data in England and Wales and Scotland; ii) to evaluate trends in mortality taking into account age, calendar year and year of birth; iii) to conduct an interrupted time series analysis to estimate the excess mortality associated with Covid19; and iv) to evaluate the extent of variations in mortality among general practices taking into account region, ethnicity, deprivation, comorbidity, influenza, Covid19 and preventive medical interventions. A cohort study will be conducted in CPRD Aurum using data for person time and deaths from any cause that are aggregated across general practices, age-groups, gender and calendar year. We will also analyse aggregated data for proportion of population coded as of black or minority ethnic origins, having morbidities including malignant neoplasms, cardiovascular diseases, respiratory diseases, metabolic diseases (including diabetes) and dementia, proportion of practice population treated with antihypertensive drugs or statins, and annual incidences of influenza and Covid-19. Mortality rates will be compared with national registration data. Trends over time and variation among general practices will be evaluated from hierarchical Poisson regression models. An interrupted times series analysis will be used to evaluate the impact of the pandemic. Variation among practices will be evaluated from the distribution of random effects. The analyses will evaluate how CPRD Aurum data can be used to evaluate mortality trends. The analyses will also contribute to understanding recent changes in mortality including excess mortality form Covid-19.
Health Outcomes to be Measured: 
All-cause mortality based on the CPRD death date in CPRD Aurum.
Application Number: 
20_000200
Collaborators: 

Martin Gulliford - Chief Investigator - King's College London
Martin Gulliford - Corresponding Applicant - King's College London
Emma Rezel-Potts - Collaborator - King's College London

Linkages: 
Practice Level Index of Multiple Deprivation