Abstract: Objective To evaluate the significance of serum cortisol in identification of suboptimal health state(SHS).Methods Totally 702 physical examinees aged 18-60 years and without serious diseases were recruited at Xuanwu Hospital of Capital Medical University and surveyed with a self-designed SHS questionnaire (SHSQ) and serum cortisol and catecholamine detection between March to July 2012.Latent class method was used to evaluate the significance of the variables in identifying SHS.Probabilities of latent categories and conditional probability of the indicators were estimated by using expectation-maximization (EM) algorithm and the sensitivity,specificity,and critical value of the indicator were calculated.Results The established latent potential classification model showed a good fitness,with the consistent Akaike's information criterion (CAIC) of 67.02 and Akaike information criterion (AIC) of 28.14.The sensitivity and specificity of serum cortisol were 0.7781(95% confidence interval95%CI:0.747 4-0.808 8)and 0.964 4(95%CI:0.950 7-0.978 1),respectively.The sensitivity and specificity of the SHSQ were 0.736 3(95%CI:0.703 7-0.768 9)and 0.701 7(95%CI:0.667 9-0.735 5),respectively.Conclusion Serum cortisol is a good indicator for identifying SHS.The SHSQ is more suitable for the screening of SHS.