A study of unplanned 30-day hospital readmissions in the United States : early prediction and potentially modifiable risk factor identification
Unplanned hospital readmissions greatly impair patients' quality of life and have imposed a significant economic burden on American society. The pressure to reduce costs and improve healthcare quality has triggered the development of readmission reduction interventions. However, existing solutions focus on complementing inpatient care with enhanced care transition and post-discharge interventions, which are initiated near or after discharge when clinicians' impact on inpatient care is ending. Preventive intervention during hospitalization is an under-explored area, which holds the potential for reducing readmission risk. Nevertheless, it is challenging for clinicians to predict readmission risk at the early stage of inpatient care because little data is available. Existing readmission predictive models tend to incorporate variables whose values are only available near or after discharge. As a result, these models cannot be used for the early prediction of readmission. Another challenge is that there is no universal solution to reduce readmissions during hospitalization. Patients can be readmitted for any reason, and their heterogeneous social and clinical factors can further complicate the planning of interventions. The objective of this project was to improve the timeliness of readmission preventive intervention through a data-driven approach. A systematic review of the literature was performed to collect reported risk factors for unplanned 30-day hospital readmission. Using various predictive modeling and exploratory analysis methods, we have developed an early prediction model of readmission and have identified potentially modifiable readmission risk factors, which may be used to guide the development of readmission preventive interventions during hospitalization for different patients.