The United States healthcare industry has witnessed a number of hospitals declare bankruptcy. This has a meaningful impact on local communities with vast implications on access, cost, and quality of care available. In our research, we seek to determine what contemporary structural and operational factors influence a bankruptcy outcome, and craft predictive models to guide healthcare leaders on how to best avoid bankruptcy in the future. In this exploratory study we performed, a single-year cross-sectional analysis of short-term acute care hospitals in the United States and subsequently developed three predictive models: logistic regression, a linear support vector machine (SVM) model with hinge function, and a perceptron neural network. Data sources include Definitive Healthcare and Becker's Hospital Review 2019 report with 3121 observations of 32 variables with 27 observed bankruptcies. The three models consistently indicate that 18 variables have a significant impact on predicting hospital bankruptcy. Currently, there is limited literature concerning financial forecasting models and knowledge detailing the factors associated with hospital bankruptcy. By having tailored knowledge of predictive factors to establish a sound financial structure, healthcare institutions at large can be empowered to take proactive steps to avoid financial distress at the organizational level and ensure long-term financial viability.
Keywords: bankruptcy; financial distress; forecast; hospital; predictive models.