Using Updated Health Care Data to Build Predictive Model on Hospital Readmission Rate

Hospital readmission being so high has always been a serious problem in the healthcare field. A lot of prior actions was taken either by different hospitals, or directly from the government in order to reduce the readmission rate. While in a lot of intervention programs, there has been a misunderstanding that most high cost patients usually came from high cost in the previous year. In our research project this summer, we used the health care data from 110,000 patients with over 3 million claims in order to analyze the trend for patients cost transitions over the years. Some prior work has been done on similar topic by professor Ian Duncan, but this time we used the latest updated data and specialized in the indicator of diseases. We focused on comparing the different cost level transitions for patients in several common chronic disease groups. In that case, our analysis will be really useful in building a predictive model in the future for reducing readmission rate purpose in general.

Faculty Advisor: Ian Duncan