An Exploratory Mixed-Methods Study of Risk Factors and Predictive Modelling of Occupational Accidents in Palm Oil Workers

Study design and setting

This study employed an exploratory mixed-methods design integrating quantitative and qualitative approaches to investigate occupational risk factors and inform the early-stage development of a predictive model. The study was conducted between August 2025 and April 2026 at a palm oil plantation company in Indonesia. The mixed-methods design was selected to enable a comprehensive understanding of occupational accidents by combining statistical associations with contextual insights derived from workers’ experiences. The quantitative component was used to identify potential risk factors, whereas the qualitative component was used to explore underlying mechanisms and inform interpretation and variable selection.

Study population and sampling

The study population consisted of palm oil workers involved in plantation and processing activities. A total of 200 workers were recruited, of whom 191 were included in the final analysis after excluding incomplete data. Participants were eligible if they were aged 18–55 years, employed for at least one year, and provided informed consent. Workers who were unable to complete the assessment were excluded.

For the qualitative component, purposive sampling was applied to ensure the representation of different roles within the occupational system, including field workers, processing workers, supervisors, and occupational health personnel. A total of 12 informants were included, and recruitment continued until thematic saturation was achieved.

Data sources and variables

Data were obtained from multiple sources to capture a comprehensive perspective on occupational health. Quantitative data were collected using structured questionnaires, workplace observations and company records.

The variables were grouped into several domains, including individual factors (age, sex, educational level, and health-related conditions); occupational factors (work unit, years of service, working hours, and physical workload; environmental exposures, including heat, noise, and ergonomic strain); and organizational factors (safety training, supervision, and use of personal protective equipment).

Workload was assessed using the NASA Task Load Index (NASA-TLX), while occupational accidents were defined based on established criteria and categorised according to type and severity. Qualitative data were collected through in-depth interviews using a semi-structured guide to explore perceptions of occupational risk, safety practices, and contributing factors to occupational accidents.

Data collection

Quantitative data were collected by trained researchers using interviewer-assisted questionnaires where necessary to accommodate varying literacy levels among workers. Observational assessments were performed during routine working activities to evaluate PPE use and workplace conditions. Qualitative interviews were conducted in a private setting and audio-recorded with participants’ permission after obtaining written informed consent. All interviews were transcribed verbatim. Field notes were also taken to capture contextual observations. Secondary data, including accident reports and occupational health records, were obtained from company databases to complement primary data sources.

Exploratory analysis strategy

Quantitative analysis was conducted using Stata in an exploratory framework to identify candidate risk factors for model development. Descriptive statistics were used to summarise participant characteristics and study variables. Bivariate analysis was performed to screen for factors potentially associated with occupational accidents. Variables with a p-value <0.25 were considered for inclusion to minimise the risk of excluding potentially relevant predictors at this early stage. Importantly, variable selection was not based solely on statistical significance. Instead, variables were retained based on a combination of statistical criteria and conceptual relevance, informed by occupational health principles and supported by qualitative findings. This approach reflects a deliberate shift from purely data-driven selection toward a more conceptually informed modelling strategy. All candidate variables meeting these criteria were entered simultaneously into the multivariate model.

Qualitative analysis

Qualitative data were analysed using thematic analysis. The analysis process involved transcription, coding, categorisation, and theme development. Coding was conducted to identify patterns related to occupational risk, safety practices, and organizational factors. Emerging themes were grouped into broader categories reflecting key dimensions of occupational accident risk.

The qualitative findings were used not only to describe contextual factors but also to support the interpretation of quantitative results and to inform the selection and prioritisation of variables for modelling.

Integration of mixed-methods findings

The integration of quantitative and qualitative findings occurred during the interpretation and model development stages. Quantitative results were used to identify statistically associated factors, whereas qualitative findings provided insights into the mechanisms and contextual drivers underlying these associations. For example, factors such as PPE use and supervision were interpreted in light of qualitative evidence on usability barriers, compliance behaviour, and organizational practices. This integrative approach enabled a more comprehensive understanding of occupational risk and supported the development of a conceptually grounded preliminary model.

Preliminary model development

A preliminary predictive model was developed using a modified Poisson regression model with robust variance to explore the combined effects of the selected variables on the occurrence of occupational accidents. Model development followed an exploratory approach, focusing on identifying patterns of association rather than establishing a final predictive tool. The purpose of this stage was to examine how the selected variables jointly contribute to occupational accident risk and to identify potential predictors for future model validation and refinement. Model performance was assessed at a preliminary level based on its discriminatory ability.

Ethical considerations

This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. Ethical approval was obtained from the Health Research Ethics Committee of Universitas Padjadjaran (No. 724/UN6. KEP/EC/2025) on August 19, 2025. Written informed consent was obtained from all participants prior to the study. For participants with limited literacy, the consent form was explained verbally by the researcher, and consent was documented by a signature or thumbprint. Participation was voluntary, and confidentiality and anonymity were strictly maintained throughout the study period.

Participant characteristics

Finally, 191 palm oil workers were included in the final analysis. The demographic and occupational characteristics are summarised in Table 1. Most participants were aged ≥30 years (90.6%) and predominantly men (88.0%). Regarding nutritional status, obesity was observed in 36.1% of workers, followed by normal BMI (34.6%) and overweight (25.1%). Most had secondary-level education, with 37.2% completing senior high school. Occupationally, most workers had ≥5 years of work experience (84.3%) and worked ≥8 h per day (94.8%). Workers were primarily distributed across processing (35.1%), administration and management (27.7%), and harvesting units (25.7%). Regarding occupational health practices, 57.6% of workers reported complete use of personal protective equipment (PPE), while 42.4% reported incomplete use. Additionally, 22.5% of participants had a history of previous occupational accidents.

Occupational accident profile

Overall, 41.9% of workers reported experiencing at least one occupational accident within the past year ( Table 2). Most injuries were minor, with “other wounds” accounting for 52.5% of cases, followed by multiple injuries (16.3%) and superficial injuries (13.8%). The upper extremities were the most frequently affected body region (47.5%), followed by the lower extremities (22.5%). In terms of severity, 62.5% of injuries were classified as minor, 22.5% resulted in lost workdays, and 15.0% required medical treatment. No fatal injuries were reported. These findings suggest that occupational accidents in this setting are relatively frequent but often underrecognized because of their predominantly minor nature.

Exploratory risk factor analysis

Bivariate analysis identified several candidate factors associated with occupational accidents ( Table 3). Younger workers (<30 years) demonstrated a higher risk of accidents than older workers (PR 1.865; 95% CI 1.324–2.626; p < 0.001). Similarly, workers with <5 years of experience had a markedly higher risk (PR 2.439; 95% CI 1.867–3.188; p < 0.001), suggesting that adaptation to the work environment may play an important role. Occupational factors also showed variability in risk. Work unit was significantly associated with accident occurrence (p = 0.001), indicating differences in exposure across job types. Compared to administrative workers, harvesting, maintenance, and processing units demonstrated higher risks, while no significant difference was observed for security personnel. Heavy lifting activities were also associated with an increased risk (≥23 kg: PR 1.671; 95% CI 1.157–2.412; p = 0.006), reflecting the contribution of physical workload. A strong association was observed between prior accident history and current accident occurrence (PR 2.178; 95% CI 1.619–2.926; p < 0.001), suggesting a pattern of repeated vulnerability among certain workers. From an organizational perspective, both PPE use and PPE supervision were significantly associated with occupational accidents. Workers with incomplete PPE use had a higher risk (PR 2.037; 95% CI 1.445–2.871; p < 0.001), while poor supervision was associated with an even greater increase in risk (PR 2.290; 95% CI 1.622–3.234; p < 0.001). In contrast, several variables, including education level, body mass index, working hours, chronic disease status, and environmental factors such as temperature and noise, were not significantly associated with accident occurrence. Overall, these findings indicate that occupational accident risk is distributed across multiple domains, including individual, occupational, and organizational factors.

Preliminary multivariable model

A multivariate analysis using a modified Poisson regression model with robust variance was conducted to explore the combined effects of the selected variables on the occurrence of occupational accidents ( Table 4). Workers with shorter tenure (<5 years) had a higher risk of accidents than those with longer tenure (≥5 years) (adjusted PR 2.240; 95% CI 1.633–3.073). A prior history of occupational accidents was also strongly associated with subsequent accidents (adjusted PR 1.847; 95% CI 1.385–2.464). Poor PPE supervision was associated with a significantly higher risk of accidents than good supervision (adjusted PR 2.329; 95% CI 1.544–3.512). Risk varied across work units, with maintenance workers demonstrating a significantly higher risk than administrative workers (adjusted PR 2.048; 95% CI 1.080–3.885).

Qualitative findings

The characteristics of the qualitative informants are shown in Table 5. The participants included workers in various roles across the occupational system, ranging from field and maintenance workers to supervisory and administrative staff, with diverse ages and years of service.

Thematic overview

The seven major themes that emerged were safety perception, workload, hazardous environment, injury characteristics, reporting systems, training and experience, and management and supervision.

Key themes

Safety perception and practices

Workers recognise PPE as essential for preventing injuries; however, consistent use is often hindered by discomfort and reduced functionality. For example, participants reported that protective equipment interfered with their ability to work effectively.

In addition, safety practices were influenced by individual factors, such as attention and focus, with participants noting that lack of concentration could increase the risk of accidents.

Workload and job demands

A high workload, long working hours, and production targets were consistently identified as key contributors to occupational risk. Workers described fatigue as a central mechanism linking workload to accidents:

The time pressure to meet targets was also reported to encourage rushed and potentially unsafe practices.

Hazardous working environment

Participants described the work environment as inherently hazardous, with risks arising from tools, terrain, and machinery. Weather conditions, particularly rain, were perceived to amplify these risks.

Importantly, many workers normalised these hazards as unavoidable aspects of their jobs:

“Memang sudah risiko kerja” (R5)

(“It’s already part of the job risk”).

Injury characteristics and impact

Workers reported a wide range of injuries, from minor cuts and punctures to severe injuries and fatal incidents. Minor injuries were often self-managed and not formally reported.

In contrast, severe injuries resulted in reduced work capacity, temporary absence, and, in some cases, long-term impairment, highlighting the broader impact of occupational accidents on workers’ productivity and well-being.

Accident reporting system

The reporting system followed a structured hierarchy; however, reporting was selective. Minor injuries were often not formally documented, as illustrated by the participants’ statements:

This selective reporting may contribute to the underestimation of the true burden of occupational accidents.

Training and experience

Training and competency were perceived as important for improving safety awareness; however, not all workers had access to formal training programs. Some participants reported a lack of structured training.

Work experience plays a dominant role in shaping safety practices. While experience helps workers adapt to job demands, it also leads to reliance on informal practices rather than standardised procedures.

Management and supervision

Management contributes to workplace safety through policies, standard operating procedures (SOPs), and supervision mechanisms. Regular briefings and inspections were reported as part of safety enforcement.

However, implementation was not always consistent, particularly under the pressure of production. This suggests a gap between the established safety policies and actual practices in the field.

Summary of qualitative insights

Overall, the qualitative findings highlight that occupational risk is shaped not only by measurable factors but also by behavioural, cultural, and organizational dynamics, including fatigue, risk normalisation, and gaps between safety policy and practice.

Integration of findings

The integration of quantitative and qualitative findings revealed several converging patterns. Work experience, prior accident history, and PPE supervision, identified as key variables in the quantitative analysis, were consistently supported by qualitative insights, highlighting mechanisms such as adaptation, behavioural patterns, and supervision practices. In contrast, factors such as workload and environmental hazards, which were not consistently significant in the quantitative analysis, emerged as dominant themes in the qualitative findings. This suggests that these factors may operate indirectly or interact with other variables; therefore, they may not be fully captured through conventional statistical models.

Implications for model development

The combined findings informed the identification of candidate variables for the preliminary model. Variables, such as work experience, prior accident history, and PPE supervision, emerged as key candidate components owing to both statistical association and conceptual relevance. Simultaneously, qualitative findings highlighted additional dimensions, such as fatigue, production pressure, and safety culture, that may influence risk but require further investigation in future modelling. This exploratory process provides a foundation for subsequent model validation and refinement.