(AIMS 2020) Rehabilitation XAI to Predict Outcome with Optimal Therapies

Abstract: Value-based payment is becoming general in healthcare. In rehabilitation medicine, medical services are becoming to be paid depending on the outcome obtained from hospitalization period and dependency score called as FIM (Functional Independent Measurement). The optimal therapies to maximize the outcome differs by each patient's age, sex, disease, handicap, FIM and therapies. Non-experienced hospitals have a difficulty in improving the outcome. Therefore, there are needs to maximize the outcome by optimizing therapies. We developed a rehabilitation XAI system to predict outcome with optimal therapies. Our system piles up medical records into vectors and predicts the outcome with optimal therapies using machine learning based on vector distance that can explain the basis of prediction in the same way as doctors suggesting optimal therapies to patients based on similar past cases. The interface not only displays optimal therapies but also predicts outcome by each patient. We used data from multiple hospitals and evaluated the adaptability of our system. In case of using the data from one hospital, the pattern achieving high outcome, which was most important because it was used to suggest optimal therapies, occupied the proportion of 31.1% in the actual record while the precision and recall were 64.5% and 73.4%. In case of using the data from another hospital, they were 64.4% and 66.1% against the actual proportion of 35.7%. In case of using the data from both hospitals, they were 63.6% and 71.0% against the actual proportion of 33.3%. Our system achieved similar performance and adaptability between two hospitals. Correlation coefficient between actual and predicted outcome were 0.681 using 204 patients' record. We compared the accuracy to predict outcome between our XAI and humans. Average outcomes of top 70% patients predicted at hospitalization by our XAI and humans were 43.0 and 42.4. Our XAI could predict outcome at higher accuracy than humans.

Authors: Takashi Isobe (Hitachi High Tech America, Inc., USA)

Email: takashi.isobe.sw@hitachi-hightech.com