Mortality Risk Prediction Tool in Hospitalized Dementia Patients

Patients with dementia are frequently hospitalized. Despite having specialist palliative care needs, they have infrequent and late access to corresponding services. Accurate and personalized predictions of mortality risk among hospitalized patients with dementia could inform clinical care decisions made during admission and upon discharge, including whether or not to engage specialist palliative care services. Existing prognostic tools have limitations. We seek to develop and validate a clinical prediction tool to estimate 1-year mortality in this patient population. The derivation cohort will comprise about 235,000 patients with dementia, who were admitted to a hospital in Ontario from April 1st, 2009 to December 31st, 2017. Predictor variables have been fully prespecified based on a literature review and on subject-matter expertise, and were categorized as follows: sociodemographic factors, comorbidities, previous interventions, cognitive status, functional status, nutritional status, admission information, and previous health care utilization. The outcome variable will be mortality within 1 year of admission, which will be modelled as a binary variable, such that a logistic regression model will be estimated. Predictor and outcome variables will be derived from linked population-based administrative databases. The validation cohort will comprise about 63,000 dementia patients, who were admitted to a hospital in Ontario from January 1st, 2018 to March 31st, 2019. Model performance, measured by predictive ability, discrimination, and calibration, will be assessed in the validation cohort. The final model will be based on the full cohort.