Development and validation of risk prediction model for premenstrual syndrome in nurses: results from the nurses-based the TARGET cohort study

Abstract

ObjectivePremenstrual syndrome (PMS) stands as a significant concern within the realm gynecological disorders, profoundly impacting women of childbearing age in China. However, the elusive nature of its risk factors necessitates investigation. This study, therefore, is dedicated to unraveling the intricacies of PMS by focusing on nurses, a cohort with unique occupational stressors, to develop and validate a predictive model for assessing the risk of PMS.MethodsThis investigation employed a multi-center cross-sectional analysis drawing upon data from the TARGET Nurses’ health cohort. Utilizing online survey versions of the Premenstrual Syndrome Scale (PMSS), a comprehensive dataset encompassing physiological, social, psychological, occupational, and behavioral variables was collected from 18,645 participants. A stepwise multivariate logistic regression analysis was conducted to identify independent risk factors for PMS. Furthermore, a refined variable selection process was executed, combining the Least Absolute Shrinkage and Selection Operator (LASSO) method with 10-fold cross-validation. The visualization of the risk prediction model was achieved through a nomogram, and its performance was evaluated using the C index, receiver operating characteristic (ROC) curves, and the calibration curves.ResultsAmong the diverse variables explored, this study identified several noteworthy predictors of PMS in nurses, including tea or coffee consumption, sleep quality, menstrual cycle regularity, intermenstrual bleeding episodes, dysmenorrhea severity, experiences of workplace bullying, trait coping style, anxiety, depression and perceived stress levels. The prediction model exhibited robust discriminatory power, with an area under the curve of 0.765 for the training set and 0.769 for the test set. Furthermore, the calibration curve underscored the model’s high degree of alignment with observed outcomes.ConclusionThe developed model showcases exceptional accuracy in identifying nurses at risk of PMS. This early alert system holds potential to significantly enhance nurses’ well-being and underscore the importance of professional support.