Impact of the Dry Law on road traffic mortality in Brazilian states: an interrupted time series analysis

The comparative analysis of SARIMA, Facebook Prophet, and LSTM for road traffic injury prediction in Northeast China

Objective

This cross-sectional research aims to develop reliable predictive short-term prediction models to predict the number of RTIs in Northeast China through comparative studies.

Methodology

Seasonal auto-regressive integrated moving average (SARIMA), Long Short-Term Memory (LSTM), and Facebook Prophet (Prophet) models were used for time series prediction of the number of RTIs inpatients. The three models were trained using data from 2015 to 2019, and their prediction accuracy was compared using data from 2020 as a test set. The parameters of the SARIMA model were determined using the autocorrelation function (ACF) and the partial autocorrelation function (PACF). The LSTM uses linear as the activation function, the mean square error (MSE) as the loss function and the Adam optimizer to construct the model, while the Prophet model is built on the Python platform. The root mean squared error (RMSE), mean absolute error (MAE) and Mean Absolute Percentage Error (MAPE) are used to measure the predictive performance of the model.

Findings

In this research, the LSTM model had the highest prediction accuracy, followed by the Prophet model, and the SARIMA model had the lowest prediction accuracy. The trend in medical expenditure of RTIs inpatients overlapped highly with the number of RTIs inpatients.

Conclusion

By adjusting the activation function and optimizer, the LSTM predicts the number of RTIs inpatients more accurately and robustly than other models. Compared with other models, LSTM models still show excellent prediction performance in the face of data with seasonal and drastic changes. The LSTM can provide a better basis for planning and management in healthcare administration.

Implication

The results of this research show that it is feasible to accurately forecast the demand for healthcare resources with seasonal distribution using a suitable forecasting model. The prediction of specific medical service volumes will be an important basis for medical management to allocate medical and health resources.

Association of the new zero-tolerance drinking and driving law with hospitalization rate due to road traffic injuries in Brazil

We investigated the association of the new zero-tolerance drinking and driving law (Law 12,760, Dec. 2012) with hospital admissions due to road traffic injuries in Brazil by using interrupted time series from 2008 to 2019. We used linear regression designed to adjust for autocorrelation and Cumby–Huizinga test for residual autocorrelation. Newey–West standard errors was used to handle heteroscedasticity. We used ICD-10 codes for land transport accidents (V01–V89). The hospitalization rate was calculated per 100,000 inhabitants. The sources were the Hospital Information System and the Brazilian Institute for Geography and Statistics. Pre- and postintervention consist of 59 and 85 months, respectively. For Brazil, the hospitalization rate was associated with a reduction of 0.34 (p = 0.097; 95% CI − 0.74 to 0.06) in the first month of the intervention (Dec. 2012), followed by a significant change in the hospitalization trend. Compared to the period prior to the intervention, the monthly trend was associated with a reduction of 0.05 (p < 0.01; 95% CI − 0.06 to − 0.04) in the post period. These results stand in agreement with subgroup analyses for the Brazilian regions, although North and Northeast regions did not immediately reduce hospitalization rates (level change). Our results suggested that 440,599 hospitalizations for land transport accidents would be averted by the new zero-tolerance drinking and driving law from Dec. 2012 to Dec. 2019 in Brazil. Even using a quasi-experimental approach, our findings must be interpreted with caution due to observational design and registration flaws surrounding our data.