Since some types of psychiatric disorders and medications may affect car driving performance and driving fitness, patients are required to obtain medical certificates regarding their fitness to drive when they obtain or renew their driver’s licenses. It is, however, difficult for physicians to evaluate patients’ driving fitness. We propose a method for evaluating the driving fitness of patients with psychiatric disorders based on driving data collected with a driving simulator (DS). We developed a machine learning (ML) model that discriminates whether a subject’s driving is normal or performance-compromised, which was trained with normal and performance-compromised driving data collected from 25 healthy participants with and without consumption of alcohol. In this study, driving under alcohol administration is defined as blood alcohol concentration (BAC) 0.5 mg/ml. We tested the trained model using healthy participants’ driving data that were not used for training, and a sensitivity of 100%, a specificity of 97%, and an AUC of 0.98 were achieved. In addition, we applied the trained model to driving data collected from 43 patients with schizophrenia as a polit study. The performance of the model was verified from the viewpoint of the patient’s clinical information. The model correctly discriminated the driving data of patients with factors affecting driving, such as severe psychiatric symptoms, high medication dosage, and lack of driving experience. The proposed method contributes to assisting patients with schizophrenia disorders, as well as their attending physicians in making decisions about patients’ driving fitness.
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号