Evaluation of the UK's COVID-19 public health policy “Shielding: Results of a linked data matched cohort study

Objective
To assess outcomes associated with shielding, introduced during the COVID-19 pandemic across the UK to protect those at highest risk of harm.

Study design
Linked data and questionnaires in matched cohorts from the population of Wales, UK.

Methods
We compared individual-level linked routine and self-reported outcomes between people identified for shielding (n = 123,293) and comparators (n = 120,997) matched by age, sex, and previous health service utilisation. We sent questionnaires to 1500 randomly sampled people in each cohort.

Results
At one year 6·1 % of shielded people had contracted SARS-CoV-2 compared to 6·2 % in the matched cohort (Adjusted Odds Ratio [AOR] 0·970; 95 % confidence interval [CI] 0·937 to 1·004). Suspected healthcare associated infections were more likely in shielded people (1·1 % vs 0·6 %; AOR 1·678; 95 % CI 1·529 to 1·842). All-cause and COVID-19 related deaths were higher in the shielded cohort (7·0 % vs 3·5 %; AOR 2·280; 95 % CI 2·190 to 2·374; and 1·1 % vs 0·8 %; AOR 1·430; 95 % CI 1·308 to 1·563, respectively).
About 1/3 completed questionnaires (n = 1015), with linkage possible in 752 cases (shielded: n = 411; matched: n = 341). Shielded respondents reported lower physical and mental health (SF12 PCS difference: −3·752; 95 % CI -4·823 to −2·682; SF12 MCS difference: −1·217; 95 % CI -2·580 to 0·145). They were more likely to have strictly avoided contact; stayed at home; felt scared to go outside; and were less likely to have gone out for shopping, leisure or travel.

Conclusion
We found no evidence of a protective effect of shielding on SARS-CoV-2 infections or COVID-19 related mortality, an increased rate of hospital acquired infections and increased self-isolation. Shielding during a future pandemic should only be considered alongside effective measures to reduce healthcare associated infections.