# Methodology for predicting hospital admissions and evaluating recovery rates for coronavirus disease in Japan

## TL;DR

This study developed a reliable method to predict hospital admissions and evaluate recovery rates for coronavirus in Japan using a combined SIR model.

## Contribution

A novel method combining SIR model with case-admission relationships to predict hospitalizations and assess recovery rates.

## Key findings

- The model achieved 99% accuracy in predicting hospital admissions.
- It enabled evaluation of recovery rates and effectiveness of measures to reduce treatment days.
- The model estimates hospitalization peaks and provides statistically valid recovery rates for countermeasure analysis.

## Abstract

In this study, we aimed to propose a method to predict the number of patients needing hospitalization using a combination of available technologies. We developed a method to predict the number of hospital admissions by combining a simple susceptible-infected-recovered (SIR) model with the relationship between the number of new positive cases and the number of hospital admissions, increasing the reliability of each prediction. The accuracy of the concordance between the actual number of patients and the predicted number of hospitalized patients was 99%. Owing to the high accuracy, we were also able to establish a method to evaluate recovery rates. This facilitated determination of the effectiveness of measures implemented throughout Japan to reduce the number of treatment days. The model developed in this study facilitates immediate estimation of the maximum number and timing of hospitalizations based on the peak of new positive cases. Moreover, it provides a statistically true value of the recovery rate required by the mathematical model for investigating countermeasures.