Care.Coach Avatars for Improvement of Outcomes in Hospitalized Elders, Including Mitigation of Falls and Delirium: a Multi-Site Clinical Study (AvatarHELP)

Care.Coach Avatars for Improvement of Outcomes in Hospitalized Elders, Including Mitigation of Falls and Delirium: a Multi-Site Clinical Study

Care.Coach Avatars for Improvement of Outcomes in Hospitalized Elders, Including Mitigation of Falls and Delirium, Through Psychosocial Support and Protocol-Driven Interventions Based on the Hospital Elder Life Program: a Multi-Site Clinical Study

Through NINR project 1R44NR017842-01 which preceded the present study, the investigators enhanced the care.coach avatar platform to incorporate a robust suite of evidence-based protocols based on the Hospital Elder Life Program (HELP), and to leverage an integration with hospital-based electronic medical record (EMR) systems. In the present study, the investigators seek to validate the efficacy of the new avatar platform, as measured by reduction in falls, delirium, and patient sitter utilization. Also, the investigators seek to gather patient and outcomes data at a scale sufficient to begin developing machine learning algorithms for intelligent, automatic assignment of protocols to maximize patient engagement and clinical efficacy, and for intelligent, automatic screening of delirium to assist care teams in positive identification of delirium. Therefore, the present study comprises a two-year randomized between-groups comparison across multiple hospitals to compare outcomes with the new generation of care.coach avatars as the intervention versus usual care only as the control. Each study group will be geographically distributed across participating research sites: initially MediSys Health Network's Jamaica Hospital Medical Center in New York, with additional hospitals to join the study over the course of two years.

care.coach human-in-the-loop avatar system with software-directed protocols based on the Hospital Elder Life Program (HELP).

For each patient without prevalent delirium (each patient who screens negative per the Confusion Assessment Method [CAM] upon study enrollment, which is assumed to be within 24 hours of admission to the hospital unit), Incident Delirium is indicated by any positive CAM screen during the patient's study enrollment period, with CAM screens being performed at least daily (the final CAM screen is assumed to be within 24 hours of discharge from the hospital unit). Incident Delirium is thus a binary variable (true/false) for each patient, and the rate of Incident Delirium will be compared across study groups, with a lower rate being desirable.

For each patient with prevalent delirium (each patient who screens positive per the Confusion Assessment Method [CAM] upon study enrollment, which is assumed to be within 24 hours of admission to the hospital unit), delirium will be considered resolved if the patient's final CAM screen (which is assumed to be within 24 hours prior to discharge from the hospital unit) indicates no delirium. Delirium Resolution is thus a binary variable (true/false) for each patient, and the rate of successful Delirium Resolution will be compared across study groups, with a higher rate being desirable.

For each patient, the Short Portable Mental Status Questionnaire (SPMSQ) is administered to measure cognitive function (0-10 errors, more errors indicating greater cognitive impairment) upon enrollment and upon dis-enrollment from the study. Change in Cognitive Function is the difference from the enrollment SPMSQ score to the dis-enrollment SPMSQ score. Change in Cognitive Function will be compared across study groups, with a larger numerical reduction being desirable.

For each patient screened delirium positive on the Confusion Assessment Method (CAM), the Memorial Delirium Assessment Scale (MDAS) is administered to measure delirium severity upon enrollment and upon dis-enrollment from the study. Change in Delirium Severity is the difference from the enrollment MDAS score to the dis-enrollment MDAS score. Change in Delirium Severity will be compared across study groups, with a larger numerical reduction being desirable.

Inclusion Criteria: fall/delirium risk (may be based purely on age at the election of each site, e.g. 65+) Exclusion Criteria: aggression or combativeness with intent to harm self or others severe hearing impairment and simultaneous severe vision impairment, despite assistive devices no comprehension of either Spanish or English

care.coach human-in-the-loop avatars for psychosocial support, health coaching, and care coordination