A data-informed multidimensional composite score for stress assessment

Occupations such as the military, emergency responders, and athletes demand acute physical, cognitive and psychosocial effort. Individuals in these roles face task-specific stressors, including extreme heat, sustained mental focus and performance under social pressure, resulting in variable stress responses. Comparing individual stress responses is challenging due to the complex physiological, cognitive, and emotional interactions. This study presents a novel, data-focused multidimensional approach to assessing stress by quantitatively integrating cognitive, psychosocial, and physiological indicators. Participants completed one of four stress trials: exercise in hot conditions, musculoskeletal exertion, psychosocial stress, and sleep deprivation. Stress was assessed using self-reported measures of mood, fatigue and stress levels, as well as with performance on reaction time (Psychomotor Vigilance Task) and inhibitory control tasks (Go/No-Go), and postural balance (sway measures). A three-phase analysis was employed for stress assessment. Variables were clustered into three domains, self-report, cognitive, and physical, based on similar patterns of relative change across participants. A score was derived from each domain using Principal Component Analysis, and these scores were combined into a single composite stress score using the same approach. The analysis revealed that participants in the sleep deprivation trial exhibited the highest stress score. A random forest model predicted participants with high stress responses using baseline lifestyle, cognitive, and psychosocial characteristics, with balanced accuracy exceeding 90% for cognitive performance. Resilience, risk-taking, BMI, and sleep trial participation emerged as strong predictors. This study demonstrates that integrating multi-domain performance changes with individual baseline characteristics provides an effective approach to measuring and predicting stress responses across different stressors taking a data-focused approach.