Objectives: This study established a highly accurate financial profitability forecasting
model suitable for hospitals to predict future profits in order to achieve the goal of
sustainable operations.
Methods: This study used the key financial indicators on financial statements of non-
public regional hospitals from 2013-2017 in Taiwan to establish a financial profitability
forecasting model. This research started with constructing a regression model based on
financial theory. After that, we compared the regression model’s forecasting accuracy to
it by the Back-propagation neural network model which used the same variables with the
regression model, then selected the better model which has less mean absolute percentage
error (MAPE). Finally, the validity of the forecasting model was verifi ed by the hospitals’
financial statements of year 2018. Furthermore, the sensitivity analysis was adopted to
analyze the degree of influence of the financial explanatory variables on profitability ratio.
Results: The results showed that the accuracy of the prediction results by the Back-
propagation neural network model was “reasonable” which was better than the regression
model’s. Furthermore, sensitivity analysis showed that net medical operating revenue,
personnel expenses, drug expenses and medical materials expenses were important
financial indicators that affected the gross profit margin of medical services.
Conclusions: The Back-propagation neural network obtained a better predicting
accuracy through an appropriate combination of variables, proper learning and training
than regression analysis. Therefore, the Back-propagation neural network might be a
good analysis tool to obtain better forecasting results when hospital managers want to do
financial profitability forecasting.