Prevalence of Musculoskeletal Disorders in Heavy Vehicle Drivers and Office Workers: A Comparative Analysis Using a Machine Learning Approach

Abstract

PURPOSE: Job profiles such as heavy vehicle drivers and transportation office workers that involve prolonged static and inappropriate postures and forceful exertions often impact an individual's health, leading to various disorders, most commonly musculoskeletal disorders (MSDs). In the present study, various individual risk factors, such as age, weight, height, BMI, sleep patterns, work experience, smoking status, and alcohol intake, were undertaken to see their influence on MSDs. METHODS: The modified version of the Nordic Questionnaire was administered in the present cross-sectional study to collect data from 48 heavy vehicle drivers and 40 transportation office workers. RESULTS: The analysis revealed low back pain (LBP), knee pain (KP), and neck pain (NP) to be the dominant pains suffered by the participants from both occupational groups. LBP, KP, and NP were suffered by 56%, 43.75%, and 39% heavy vehicle drivers and 47.5%, 40%, and 27.5% transport office workers, respectively. From the insignificant value of Chi-square, it can be inferred that the participants from both occupations experience similar levels of LBP, KP, and NP. The Bayesian model applied to the total sample showed that NP influenced KP, which further influenced the LBP of the workers. Age was predicted as LBP's most significant risk factor by the logistic regression model when applied to the total sample, while NP was found to decrease with an increase in per unit sleep. CONCLUSIONS: The overall results concluded that heavy vehicle drivers and office workers, irrespective of their different job profiles, endured pain similarly.

Keywords: Bayesian network modelling; knee pain; lower back pain; machine learning; neck pain; transportation industry.

Figure 1

MSD prevalence among the drivers and office workers by body part. (Bar graph showing the MSD among drivers and office workers by body part). Note: Data from Mohammad Raza et al., 2024 [16].

Figure 2

Bayesian network modelling (Bayesian network causal association model between LBP, NP, and KP).

Figure 3

ROC curve on the test set for LBP. (Curve between the true positive rate and the false positive rate).

Figure 4

ROC curve on the test set for KP. (Curve between the true positive rate and the false positive rate).

Figure 5

ROC curve on the test set for NP. (Curve between the true positive rate and the false positive rate).