Comparison of risk-assessment tools for cardio-cerebrovascular diseases (CVD) in male shipyard workers: a cross-sectional study

Ha, Jea Chul; Son, Jun Seok; Kim, Young Ouk; Chae, Chang Ho; Kim, Chan Woo; Park, Hyoung Ouk; Lee, Jun Ho; Shin, Young Hoo; Park, Hyun Woo

Abstract

BACKGROUND
Periodic revision of assessment tools is essential to ensure risk assessment reliability and validity. Despite the recent revision of the Korea Occupational Safety and Health Agency (KOSHA) 2018, there is no evidence showing that the revision is superior to other cardio-cerebrovascular diseases (CVDs) risk-assessment tools for workplace health management. We conducted a comparative analysis using the Framingham risk score (FRS) as a gold standard to identify the most relevant CVDs risk-assessment tool for workplace health management.
METHODS
We included 4,460 shipyard workers who had undergone a workers' health examination during January-December 2016. Risk levels for CVDs were calculated based on the FRS, KOSHA 2013, KOSHA 2017, KOSHA 2018 (2 methods), National Health Screening Program health risk appraisal (NHS HRA) 2017, and NHS HRA 2018. Study participants were categorized into low-risk, moderate-risk, or high-risk groups. Sensitivity, specificity, correlation, and agreement of each risk-assessment tool were calculated compared with the FRS as a gold standard. For statistical analyses, Spearman's rank correlation coefficient and the linearly weighted kappa coefficient were calculated.
RESULTS
Sensitivity of the risk assessments was highest in the KOSHA 2018 (health risk appraisal [HRA]). The FRS showed correlation coefficients of 0.354 with the KOSHA 2013, 0.396 with the KOSHA 2017, 0.386 with the KOSHA 2018, 0.505 with the KOSHA 2018 (HRA), 0.288 with the NHS HRA 2017, and 0.622 with the NHS HRA 2018. Kappa values, calculated to examine the agreement in relation to the KOSHA 2013, KOSHA 2017, KOSHA 2018, KOSHA 2018 (HRA), NHS HRA 2017, and NHS HRA 2018 with the FRS, were 0.268, 0.322, 0.352, 0.136, 0.221, and 0.559, respectively.
CONCLUSIONS
The NHS HRA 2018 risk calculation method is a useful risk-assessment tool for CVDs, but only when appropriate classification criteria are applied. In order to enhance the risk-group identification capability of the KOSHA guideline, we propose to apply the classification criteria set in this study based on the risk group definition of the 2018 Korean Society of Hypertension guidelines for the management of hypertension instead of the current classification criteria of the KOSHA 2018.