Kudos to Michael Cary, associate professor, and his entire team for the submission of his NIH R01 proposal entitled: “Development of an electronic clinical prediction tool to guide successful community discharge following inpatient rehabilitation for older adults." This proposal request funding for a five-year period with a start date of July 1, 2021.

Nearly 1 in 8 of the 339,000 Medicare beneficiaries who receive post-acute care (PAC) in Inpatient Rehabilitation Facilities (IRFs) is rehospitalized within 30 days; with 1-year rehospitalization rates as high as 60%. Current prediction models lack sufficient accuracy to identify patients at high risk for re-hospitalization. Many of these models incorporate demographics, comorbidities, and prior-year hospital use; few consider key drivers of rehospitalization among IRF patients such as social and functional status variables. Furthermore, emerging population-based research suggests factors not previously considered in IRF rehospitalization prediction models are important predictors of 1-year outcomes among community-dwelling older adults including co-occurring functional limitations (mobility, self-care, cognition) and geriatric syndromes (depression, incontinence, low cognition).

The purpose of this study is to build a real-time, electronic clinical tool accessible via the web that predicts successful community discharge from IRFs for older adults. Using Inpatient Rehabilitation Facility-Patient Assessment Instrument (IRF-PAI) and Medicare claims data from more than 200,000 older adults discharged from 1,200 IRFs, we propose a 5-year project with the following aims and sub-aims: Aim 1. Quantify the impact of functional limitations (mobility, self-care, cognition) and geriatric syndromes (incontinence, pressure ulcer, depression, and falls) on successful community discharge (no re-hospitalization or death within 31-days) following inpatient rehabilitation for older adults using IRF-PAI and Medicare claims. Aim 1a. Explore the impact of these factors on 1-year outcomes (days spent in the community, nursing home institutionalization, emergency department visits, re-hospitalization, and death). Multilevel linear and survival models both adjusting for facility-level variation will be conducted. Aim 2. Develop and validate a prediction model to estimate the real-time probability of successful community discharge in a large sample of older adults treated in IRFs. Predictor variables will be derived from routinely collected sociodemographic and clinical IRF-PAI data.

We will perform an 80:20 split of the sample to derive the tool (80% test dataset) and validate the tool (20% validation set). Aim 3. Build an electronic clinical tool of the validated prediction model in Aim 2. Design Thinking methodology will be used to emphasize, ideate, prototype and iteratively revise the tool. The co-creation of this tool will involve a panel with stakeholders (a nurse, therapist, physician, social worker, and administrator) from IRFs. Based entirely from clinical information collected from patients during their IRF stay, our study will be the first to build a real-time, electronic clinical decision support tool that will identify high risk patients and facilitate referral to community-based care following IRF discharge.