Predicting initial aftercare appointment adherence and rehospitalization for individuals with serious mental illness discharged from an acute inpatient stay

Aftercare nonadherence and rehospitalization of individuals with serious mental illnesses has personal, economic, and clinical costs. Seventy-four participants were recruited from a hospital-based psychiatric unit to investigate factors associated with initial aftercare nonadherence, and rehospitalization in a 3-month post-discharge follow-up period. In addition to demographic, clinical, and system risk factors, this research used the Health Belief Model (HBM) and the Transtheoretical Model of Change (TTM) as theoretical frameworks to predict health-care decision making. Risk variables were abstracted from participants' charts. Prior to discharge, each participant completed questionnaires that were selected from the literature to approximate the constructs of the HBM, the TTM, and internalized and externalized motivation. Aftercare service contacts and rehospitalization data were obtained from the local behavioral health entity. Two separate logistic regression analyses were conducted to establish which model best accounted for the two outcomes. Approximately 58% of participants did not have an aftercare service contact in the 3-month follow-up period. Of the risk factors entered on the first step of a sequential logistic regression analysis, case management services significantly increased the odds of aftercare contact. Neither the variables testing the HBM nor the motivational constructs significantly added to model improvement. Rehospitalization data indicated that approximately 27% of participants were rehospitalized at least once in the 3-month follow-up period. Logistic regression analyses showed that the variables testing the HBM significantly improved a risk factor model. Motivational variables did not add to the model. Participants with more favorable attitudes toward psychiatric medications (measured with the DAI-10) had significantly reduced odds of rehospitalization, holding all other variables constant. Implications, study limitations, and future directions are discussed.