AI-powered Pre-Admission Cost of Hospitalization Estimation
Pre-admission hospitalization cost estimates provide patients with a better understanding of the cost and financial responsibility associated with medical treatment sought. To this end, this study assesses several different approaches to pre-admission cost prediction and reports on the results. Data from four different hospitals for the period of 2015 – 2018 was used, demonstrating extensibility and external validity. Overall, simple linear regression models were a good baseline though more recent, non-linear models provided significant performance improvements, reducing error on existing benchmarks by 50 – 75%. The paper also shares the methodology for developing and deploying the AI-powered Pre-Admission Cost of Hospitalization Estimation (APACHE) system, moving it from proof-of-concept (POC) to production, providing a starting point for other researchers, industry practitioners, healthcare providers and payers. This is an important step towards (i) standardization of healthcare cost estimation, especially for countries and medical facilities that lack the necessary know-how or formal estimation systems and (ii) price transparency to build trust between healthcare providers, payers and patients.
Neal Liu, Founder and CTO of UCARE.AI, has close to 30 years of experience in the hi-tech industry. He built his expertise from late night hacking at MIT Media Lab, architecting large-scale solutions at Microsoft, to working on the latest ad-tech at Google. After more than five years at Google, he left in 2016 to jumpstart UCARE.AI, chasing his lifelong quest of using data and machine learning to improve lives. Neal holds an MBA from Wharton, University of Pennsylvania, and a BS in Electrical Engineering and Computer Science from MIT.