Factors associated with COVID-19-related death using OpenSAFELY

Abstract

Coronavirus disease 2019 (COVID-19) has rapidly affected mortality worldwide¹. There is unprecedented urgency to understand who is most at risk of severe outcomes, and this requires new approaches for the timely analysis of large datasets. Working on behalf of NHS England, we created OpenSAFELY-a secure health analytics platform that covers 40% of all patients in England and holds patient data within the existing data centre of a major vendor of primary care electronic health records. Here we used OpenSAFELY to examine factors associated with COVID-19-related death. Primary care records of 17,278,392 adults were pseudonymously linked to 10,926 COVID-19-related deaths. COVID-19-related death was associated with: being male (hazard ratio (HR) 1.59 (95% confidence interval 1.53-1.65)); greater age and deprivation (both with a strong gradient); diabetes; severe asthma; and various other medical conditions. Compared with people of white ethnicity, Black and South Asian people were at higher risk, even after adjustment for other factors (HR 1.48 (1.29-1.69) and 1.45 (1.32-1.58), respectively). We have quantified a range of clinical factors associated with COVID-19-related death in one of the largest cohort studies on this topic so far. More patient records are rapidly being added to OpenSAFELY, we will update and extend our results regularly.

Figure 1. Estimated log hazard ratio by age in years

Footnote: From the primary fully adjusted model containing a 4-knot cubic spline for age, and adjusted for all covariates listed in Table 2 except for ethnicity.

Figure 2. Illustration of data flows in the OpenSAFELY platform

Figure 1. Flow diagram of cohort with numbers excluded at different stages and identification of cases for the main endpoints.

Figure 2. Kaplan-Meier plots for COVID-19 related death over time by age and sex

Figure 3. Estimated Hazard Ratios (shown on a log scale) for each potential risk factor from a multivariable Cox model

Footnote: Error bars represent limits of the 95% confidence interval for the hazard ratio. Obese class I: 30-34.9kg/m², class II: 35-39.9kg/m², class III: >=40kg/m². OCS = oral corticosteroid. All HRs are adjusted for all other factors listed other than ethnicity. Ethnicity estimates are from a separate model among those with complete ethnicity data, and are fully adjusted for other covariates. Total n = 17,278,392 for non-ethnicity models, and 12,718,279 for ethnicity model.