The COVID mortality gap in severe mental illness widened after vaccines arrived

Design

retrospective whole-country linked electronic health record cohort study with Cox regression, stratified by three pandemic stages defined by vaccination roll-out

Population

13,463,945 people in England with COVID-19 between January 2020 and June 2023; 160,190 (1.19%) with severe mental illness

Primary outcome

all-cause and COVID-19-specific mortality following SARS-CoV-2 infection, by severe mental illness status and pandemic stage

Effect

all-cause aHR 1.56 (95% CI 1.53-1.59) overall; by stage 1.27 (1.22-1.32), 1.56 (1.51-1.61), 1.55 (1.50-1.59); full vaccination 79.4% vs 87.2% by study end

This whole-country study linked English primary care records covering over 99% of the population, identifying 13,463,945 people who had COVID-19 between January 2020 and June 2023, of whom 160,190 (1.19%) had severe mental illness — schizophrenia, schizoaffective disorder, bipolar disorder, or another affective or non-affective disorder with psychosis.

All-cause mortality after infection was higher in severe mental illness overall (adjusted HR 1.56, 95% CI 1.53 to 1.59). What the staged analysis shows is that this did not improve. The gap was 1.27 (95% CI 1.22 to 1.32) in the pre-vaccination stage, then widened to 1.56 (1.51 to 1.61) during roll-out and stayed at 1.55 (1.50 to 1.59) afterwards. Excess COVID-19-specific mortality persisted and was only partially attenuated by vaccination (aHR 1.25, 95% CI 1.17 to 1.34).

The vaccination figures explain part of it and are the most actionable finding. People with severe mental illness were initially vaccinated at higher rates than the general population — 73.2% against 67.4% during roll-out, reflecting their priority-group status — and ended lower, 79.4% against 87.2%. In other words, a system that prioritised them successfully at first failed to sustain it, and the advantage inverted as the programme moved from active outreach to routine access. That is a service design finding as much as a clinical one: when vaccination shifts from being brought to people to being available to people, this population loses ground. It is observational and cannot prove the coverage gap caused the mortality gap, but the two moved together.

• Check and record vaccination status at psychiatric review rather than assuming primary care has it in hand
• Recognise that active outreach worked and passive availability did not for this population
• Do not read the widening gap as vaccines helping others more; unvaccinated status concentrated in severe mental illness over time
• Physical health monitoring in severe mental illness should include immunisation, not only metabolic parameters
• Whole-country coverage removes selection bias but coded severe mental illness in primary care records misses undiagnosed and disengaged patients

The statistics, in plain English

The confidence intervals here are extremely narrow because the cohort is essentially the whole country, so differences between pandemic stages (1.27 versus 1.56) are real rather than noise. Adjusted hazard ratios account for the sociodemographic and clinical differences the investigators measured, but people with severe mental illness differ in smoking, deprivation and comorbidity in ways records capture imperfectly, so some residual confounding remains. A partially attenuated COVID-19 mortality ratio of 1.25 after accounting for vaccination means vaccination explains part of the gap and something else explains the rest.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.