Predictive modeling for dengue patient’s length of stay (LoS) using big data analytics (BDA)

Abstract

Big data analytics (BDA) in healthcare has become increasingly popular as it offers numerous benefits healthcare stakeholders including physicians, management and insurers. By using dengue epidemic as a case, we identified patient’s length of stay (LoS) as a parameter for the efficiency of care and potentially optimize hospital costs. This paper reports findings from two healthcare facilities based in Malaysia, which recorded 9,261 dengue patients in the year 2014. The main purpose of this study is to provide descriptive analysis and propose big data analytics modeling technique to determine and predict LoS of dengue patients. Demographic data such as age, gender, admission and discharge date have been identified as factors that contribute to the prediction of LoS. The suggested predictive modeling technique may improve resource planning through the use of simple decision support system. Recommendations of this study may also assist the expectation of healthcare facilities on their patient’s LoS.