Background Occupational accidents remain a major occupational health problem, particularly in high-risk sectors such as agriculture, including the palm oil industry. However, evidence from this sector remains limited, and existing studies often lack the integration of occupational health perspectives and predictive modelling approaches. This study aimed to explore occupational risk factors associated with occupational accidents among palm oil workers using a mixed-methods approach and to examine the early-stage development of a predictive model. Methods An exploratory mixed-methods design was conducted, integrating quantitative data from 191 workers and qualitative data from 12 informants. Quantitative analysis included descriptive and bivariate analyses to identify candidate risk factors, followed by exploratory multivariate modelling. Qualitative data were analysed using thematic analysis to provide contextual insights and inform variable selection and interpretation. Results Occupational accidents were reported by 41.9% of workers, with most injuries classified as mild. A quantitative analysis identified work experience, prior accident history, and personal protective equipment (PPE) supervision as key factors associated with accident occurrence. Qualitative findings highlighted additional influences, including workload-related fatigue, barriers to PPE use, normalisation of hazardous conditions, and selective reporting of injuries. The integration of findings indicated that additional contextual and behavioural factors contributed to occupational risk beyond statistically significant variables. Conclusions Occupational accidents among palm oil workers are influenced by complex interactions among individual, occupational, and organizational factors. This study provides an exploratory framework for integrating occupational health concepts to inform the conceptual development of a predictive model and highlights the importance of combining quantitative and qualitative approaches in occupational risk assessment.
Waren et al. (Fri,) studied this question.