Anthropometric and fitness predictors of operational preparedness among Malaysian firefighters: a clustering and multivariate logistic regression approach

Introduction: The evolving scope of responsibilities within the Malaysian Fire and Rescue Department necessitates a high level of physical and anthropometric readiness among firefighter recruits. Modern firefighting now encompasses complex land, sea, and air rescue operations alongside conventional fire suppression, requiring rigorous preparedness. Objective: This study aimed to evaluate the fitness and anthropometric profiles of Bomba recruits, identify key variables distinguishing excellent from average fitness performers, and develop a predictive model to classify future high-performing firefighters. Methodology: A total of 746 recruits underwent a final assessment of anthropometric and fitness parameters. K-means clustering was utilised to categorise recruits into Excellent Fitness Readiness (EFR) and Average Fitness Readiness (AFR) groups. Mann-Whitney U tests were then employed to determine significant differences in the measured variables between the groups. Subsequently, a logistic regression model was developed to predict the likelihood of recruits achieving EFR status. Results: Results indicated that five out of nine variables, the 2.4 km run, shuttle run, inclined pull-ups, standing broad jump, and sit-ups, significantly differentiated the two groups (p < 0.05). The regression model demonstrated strong predictive power (AUC = 0.94, sensitivity = 0.94, specificity = 0.77, accuracy = 89%). Notably, improved performance in pull-ups and standing broad jumps increased the likelihood of being in the EFR group by 66.8% and 4.3%, respectively. Conversely, slower shuttle and 2.4 km run times markedly reduced the odds by 85.8% and 33%. Discussion: These findings emphasise the operational relevance of targeted fitness parameters and provide a data-driven framework for optimising firefighter recruitment and training. Conclusions: These findings emphasise the operational relevance of targeted fitness parameters and provide a data-driven framework for optimising firefighter recruitment and training.