Abstract
Polycystic ovary syndrome (PCOS) is the most common endocrine disorder affecting women of reproductive age. It increases the risk of type 2 diabetes, cardiovascular disease, endometrial cancer, infertility, gestational diabetes, preeclampsia, and preterm birth. Accurately identifying predictors of these health risks is crucial. Electronic health records (EHRs) offer an affordable approach, however, the validity and reliability of EHRs for PCOS diagnosis are unclear. A scoping review of the literature on the prevalence and reliability of the diagnosis of PCOS using EHRs was performed. An analysis of the feasibility of obtaining diagnostic variables from a PCOS patient database was also carried out. Eight studies met the criteria. The prevalence of PCOS ranged from 0.27% to 5.8%. Reliability varied, with one study reporting a sensitivity of 50% and a specificity of 29%. Another study found a 74.4% agreement between international classification of disease (ICD) codes and clinical criteria. The database analysis found only 13.7%, 8%, and 7.5% of women had all the necessary variables for an objective diagnosis of PCOS using the Rotterdam, National Institutes of Health (NIH), and Androgen Excess and PCOS Society (AEPCOS) criteria, respectively. Using EHRs results in an underestimation of PCOS prevalence compared to other diagnostic criteria, and many women identified may not meet the complete diagnostic criteria. These findings have implications for future research studies on PCOS prevalence and related health risks.