Using Facility Administrative Data to Inform Prevention and Improve the Clinical Management of COVID‑19 Infection, Illness, Hospitalization for Nursing Homes and Retirement Homes
Older adults are most vulnerable to developing complications from COVID‑19 leading to hospitalization or death. Among older adults, the COVID pandemic has had a profound impact on long-term care/nursing home residents who account for 60% of all COVID related deaths in Canada, and internationally. By virtue of living in congregate settings nursing home residents are more likely to contract COVID, and the underlying medical conditions that cause them to need nursing care make them particularly susceptible to COVID complications. This context requires knowledge of both i) individual-level chronic conditions, cognitive and physical impairments, and age-related disabilities, and ii) home-level factors including the physical environment, and staffing levels and mix. Moreover, individual- and home-level characteristics interact and must be studied jointly to understand, and reduce, COVID infections and complications. To date, these forces affecting resident vulnerability have not been addressed jointly. There is a need to disentangle individual-level features such as functional and physical impairments, and home-level features such as the percentage of part-time workers in a unit. However, not all features can be disentangled. Some operate jointly, not independently, and characterizing the interactions between individual and home characteristics is a key goal. We will address this by building a data repository combining both home-level and resident-level data. This will allow us to build prediction tools, including machine learning algorithms addressing interactions, to differentiate patients who may or may not do well with COVID, or its variants, in real-time to inform rapid decision-making. The unique contribution of our study is to employ electronic medical record (PointClickCare) data, including clinical frailty measures that are not available from other public health and provincial data centres, together with human resource and other operational administrative data.
