Artificial intelligence based prediction of psychological distress among ride-hailing drivers in a low resource setting

BACKGROUNDPsychological distress is an important mental health concern among gig economy workers, particularly ride-hailing drivers who often work under conditions involving long working hours, income instability, and limited social protection.Evidence on psychological distress among ride-hailing drivers in African settings remains limited, and the use of artificial intelligence for mental health prediction in such contexts is still developing. AIMTo exam psychological distress among ride-hailing drivers and evaluated the performance of artificial intelligence-based models for predicting distress.It also identified occupational and psychosocial factors associated with psychological distress among the drivers. METHODSA cross-sectional predictive design was used with 400 ride-hailing drivers operating in Lagos, Abuja, and Port Harcourt.Data were collected using structured questionnaires covering demographic, occupational, psychosocial, and psychological distress variables.Descriptive statistics, correlation analysis, and multiple regression were used to examine the data.Four machine learning models, logistic regression, decision tree, random forest, and support vector machine, were developed and evaluated using Python.Model performance was assessed using accuracy, precision, recall, F1 score, and area under the receiver operating characteristic curve. RESULTSPsychological distress was prevalent among the participants.Work-related stress, financial pressure, sleep disruption, and occupational conditions were significant positive predictors of psychological distress, whereas social support was negatively 4 / 52 associated with distress.Among the machine learning models, random forest showed the strongest predictive performance in the study sample.Feature-importance analysis identified work-related stress, financial pressure, and sleep disruption among the variables that contributed most to the model's predictions. CONCLUSIONPsychological distress among ride-hailing drivers was associated with a combination of occupational and psychosocial factors, with work-related stress, financial pressure, and sleep disruption emerging as particularly important predictors.The machine learning findings indicate the potential value of combining these factors for risk classification within the study population.However, given the context-specific sample and absence of external validation, further research using larger and independent samples is needed before such models can be applied more broadly for occupational or digital mental health screening.