Residential setting as an effect modifier of intellectual and physical disability associations with COVID-19 infection, hospitalisation and mortality in Scotland.

BACKGROUND

People with intellectual and physical disabilities experienced disproportionately adverse COVID-19 outcomes, and communal residential settings were major sites of transmission. However, whether residential setting modifies the associations of these specific disability groups with recorded infection, hospitalisation and mortality has not been examined at population scale. We assessed interaction on both multiplicative and additive scales.

METHODS

We conducted a population-wide retrospective cohort study of 3,719,651 adults aged 16 years or older who were alive and resident in Scotland on 1 March 2020. Disability status was derived from the 2011 Scottish Census and classified as intellectual disability, physical and/or sensory disability without intellectual disability, or no recorded intellectual or physical disability. The latter category was the prespecified comparison group and could include people with other long-term conditions. Residential setting was classified as private dwellings, small communal establishments (<50 residents) or large communal establishments (50 or more residents). Outcomes were first laboratory-confirmed SARS-CoV-2 infection, first COVID-19 hospitalisation and COVID-19-related mortality from 1 March 2020 to 30 April 2022. Cox proportional hazards models estimated joint category hazard ratios and departures from multiplicative and additive no-interaction models.

RESULTS

For recorded infection, interaction was positive on both scales. Multiplicative interaction was significant (χ²(4) = 541.42, p < 0.001), with joint hazard ratios for intellectual disability increasing from 2.54 (95% CI 2.46-2.62) in private dwellings to 9.42 (95% CI 8.08-11.00) in large communal establishments; corresponding estimates for physical disability increased from 1.55 (95% CI 1.53-1.57) to 5.67 (95% CI 5.37-5.98). RERI was 6.09 (95% CI 4.63-7.54) for intellectual disability and 3.31 (95% CI 3.00-3.62) for physical disability in large communal establishments. For hospitalisation, there was no clear evidence of interaction (χ²(4) = 9.32, p = 0.050; all RERI confidence intervals included zero). For mortality, interaction was negative on both scales (χ²(4) = 106.58, p < 0.001), with predominantly negative RERI estimates in communal establishments, most notably for intellectual disability in small communal establishments (RERI -2.07, 95% CI -3.21 to -0.93).

CONCLUSIONS

Residential setting modified associations between intellectual and physical disability and COVID-19 outcomes, with positive interaction for infection, no clear interaction for hospitalisation, and negative interaction for mortality. These findings support pandemic preparedness that considers disability and residential setting together, including tailored infection prevention in communal settings.