Scotland cohort to test how disability and household setting shaped COVID-19 risk
Scotland is setting up a population-wide retrospective cohort study to examine how intellectual disability and physical disability intersect with the domestic spaces where people live during the COVID-19 pandemic. The protocol focuses on a question with direct public health relevance: whether residential setting and household size altered the excess risks already seen in disabled populations for infection, hospitalisation and mortality.
The study is anchored in the long-recognized reality that disability can change exposure patterns, care needs and the ability to isolate. A person living in a crowded household, or in a communal establishment, faces a different transmission environment from someone in a smaller private dwelling. For people who depend on close personal assistance, those differences can matter at every stage of the infection chain, from first exposure to severe disease and death.
Population-wide linkage across census, address and health records
Investigators will build the cohort from people recorded in the 2011 Scottish Census who can be linked to the Scottish population spine. That spine includes everyone registered with a general practice who received a Community Health Index number, creating a health-linked denominator that is unusually broad for a national study of this kind. The eligible cohort will include people who were alive, resident in Scotland and at least 16 years old on 1 March 2020.
Disability classification will come from Census records and will separate people with intellectual disability, people with physical disability without intellectual disability, and a comparison group without reported disability. Housing characteristics will then be constructed from administrative address data. General practice address information will be linked with Ordnance Survey AddressBase Premium to identify Unique Property Reference Numbers, while Basic Land and Property Unit codes will classify residences as private dwellings or communal establishments. Household size will be derived from resident counts at each property.
This design matters because it moves beyond broad social descriptors and into the material settings where viral exposure occurs. A communal establishment can amplify contact intensity, while larger households can make isolation difficult even when someone is already symptomatic or exposed. In people with disability, those pressures may be compounded by care dependence, frequent contacts with paid or unpaid caregivers and reduced flexibility in changing routines.
Endpoints cover infection, hospitalisation and death through April 2022
Following linkage to the Public Health Scotland COVID-19 Research Database, the study will ascertain COVID-19 infection, hospitalisation and mortality from 1 March 2020 to 30 April 2022. That window spans the early waves of the pandemic and extends into later phases when vaccination and changing public health measures altered population risk. The protocol does not promise a single static estimate; instead, it is built to assess how disability-related inequality evolved across a prolonged period of transmission.
Investigators will use Cox proportional hazards models to estimate associations between disability status, living arrangements and the three COVID-19 outcomes, adjusting for relevant covariates. The statistical plan also includes interaction testing to determine whether residential setting and household size modify the association between disability and COVID-19 outcomes on both multiplicative and additive scales. That approach is especially important here. If disability interacts with crowded or communal living conditions, the combined risk may exceed what would be predicted by either factor alone.
One sentence captures the clinical stakes plainly: household structure can convert vulnerability into measurable harm.
The analysis is designed to separate where inequality originates from social and environmental exposure and where it reflects downstream illness severity. Infection risk may rise because of contact patterns, while hospitalisation and mortality may also reflect pre-existing health burden, reduced access to preventive care or delayed recognition of deterioration. By tracing all three endpoints, the cohort can identify whether the household environment modifies risk mainly at the point of acquisition or also at the point of clinical progression.
Why the household lens matters for disability epidemiology
Disability research in pandemics often describes elevated risk, but less often asks where that risk is produced. Household composition is a plausible mechanism because it shapes proximity, caregiving routines and the practical ability to separate an infected person from others in the home. In a private dwelling, fewer residents may lower contact frequency and ease isolation. In a communal establishment, repeated close interactions can create persistent opportunities for transmission, particularly when residents share space, staff or support networks.
For disabled people, those pathways are not abstract. Reliance on assistance with daily activities can make strict distancing impossible, and physical barriers in the home can limit rapid self-isolation. The protocol is built to capture those realities at scale, using administrative data rather than small samples or convenience cohorts. That gives the study the chance to detect gradients of risk that may be invisible in routine clinical reporting.
The emphasis on both intellectual and physical disability is also important. These groups can face different combinations of care dependence, communication barriers, mobility limitations and comorbidity burden. Separating them may reveal distinct patterns of susceptibility and service need, rather than collapsing all disability into a single category. The planned analyses should therefore offer a more precise view of how disadvantage accumulated during the pandemic.
Ethics approvals and planned dissemination
Ethical approval came from the College of Social Sciences Research Ethics Committee at the University of Glasgow, reference 400200099. Data linkage was approved by the Scottish Public Benefit and Privacy Panel for Health and Social Care and the Scottish Government & National Records of Scotland Data Access Panel, reference 2021-0119. Those permissions are central to the study because the project depends on secure linkage between census, address and health data across the national system.
Findings will be shared through peer-reviewed open-access publications and conference presentations. The protocol itself does not report outcome data, and no effect estimates are available yet. What it does establish is a rigorous framework for testing whether the domestic environment amplified or buffered COVID-19 inequalities among disabled people in Scotland.
The unanswered empirical question is not whether people with disability faced excess pandemic risk. It is how much of that risk was shaped by where and with whom they lived, and whether those settings changed the course of infection, admission and death across two years of the pandemic.