Grant: Project Grant - PA: Patient-Oriented Research: Early-Career Investigator

Delirium is a state of acute confusion that affects up to 1 in 3 hospitalized adults. It is a leading cause of preventable harm in hospitals, increasing morbidity, mortality, and cost. Delirium can be very distressing to patients, their caregivers, and their healthcare providers. Up to 40% of delirium cases can be prevented through a bundle of interventions that includes improving sleep, mobility, nutrition, and cognitive stimulation. Implementing these interventions relies on nursing and clinical team members, who are often already overburdened. Ideally, resources to support delirium prevention would be targeted to hospital wards where they are needed the most. However, delirium is difficult to measure because clinical documentation is poor and routinely-collected data captures only 25% of cases. As a result, hospitals cannot reliably measure delirium rates, prioritize its prevention, or evaluate the success of delirium prevention efforts. Thus, there is a well described failure to prevent delirium in hospitals. This project aims to overcome the barriers to delirium prevention in 8 hospitals. First, we will use artificial intelligence methods to estimate rates of delirium based on data stored in hospital electronic medical records. We will use these estimates to produce a Delirium Scorecard for hospital managers and clinicians, which reports the rates of delirium and how well prevention strategies are being implemented on each hospital ward. Next, we will adapt established delirium prevention interventions to meet the specific needs of each hospital. Finally, we will evaluate whether an automated Delirium Scorecard paired with locally-adapted delirium prevention efforts can reduce delirium in a clinical trial. If successful, this strategy could be implemented in hospitals across Canada and internationally to prevent delirium and improve health outcomes, patient experiences, and provider experiences, and reduce the costs of care.