Development of a Confined-Space Suitability Index (CSSI) using Least Absolute Shrinkage and Selection Operator (LASSO) regression: a pilot study for structured fitness-for-duty screening

Abstract

Background Confined space work exposes workers to complex risks, including oxygen deficiency, hazardous gases, and physical strain, requiring structured criteria for fitness-for-duty screening. This study aimed to develop a Confined-Space Suitability Index (CSSI) based on routinely available health examination and functional test data. Methods A total of 111 workers were analyzed. Risk factors were classified into anthropometric and metabolic factors, functional and physiological factors, and lifestyle factors. Low cardiorespiratory fitness was defined using age- and sex-specific VO₂max reference thresholds from Korea Occupational Safety and Health Agency (KOSHA) H-43–2021 and Canadian Public Health Association reference values. Risk-factor weights were derived using LASSO regression, and internal validation was performed using leave-one-out cross-validation and bootstrap stability analysis. Results According to the revised CSSI criteria, 86 workers were classified as suitable, 19 as caution, and 6 as unsuitable. The independent specialist assessment classified 7 workers as unsuitable. In leave-one-out cross-validation, the CSSI showed an AUROC of 0.940 and an AUPRC of 0.567, with sensitivity of 0.833, specificity of 0.933, and negative predictive value of 0.990. In bootstrap stability analysis, dyslipidemia, low cardiorespiratory fitness, and hypertension were consistently selected. Conclusions The CSSI may serve as a structured screening and referral-support index for identifying workers who require additional specialist evaluation before confined space work. However, it should be interpreted as an auxiliary index rather than a replacement for specialist fitness-for-duty judgment.