Estimation of patient safety culture in private and public hospitals using machine learning methods

BackgroundPatient safety is a critical component of health care systems. Large groups of patients, as a result of medical errors, are at risk of harm. OBJECTIVE: This study evaluated the patient safety culture (PSC) between different work groups in both public and private hospitals, using machine learning approaches.MethodsThe HSOPSC questionnaire was used for evaluating safety culture, and the artificial neural network (ANN), random forest (RF) and linear regression (LR) algorithms were used for data modeling. Orange Data Mining version 3 and SPSS software were used for analysis.ResultsThe overall PSC score in public and private hospitals was 41.99 and 40.96, respectively. According to the results, the examined hospitals have a weak PSC. The safety culture level was correlated with education level, work experience, gender, income, and organizational position of the workers. The ANN showed that the issues mostly effecting PSC, in order of priority, include the feedback and communication about errors, organizational learning and continuous improvement, and management support for patient safety. Also, based on the findings LR model showed better performance for PSC prediction than RF model.ConclusionsThe healthcare experts and policymakers can improve PSC in hospitals through training and allocation of resources. Considering the importance of PSC in preventing accidents and reducing injuries, the results of the present study and the presented models can be used to predict PSC in hospitals.