AbstractBackgroundPolycystic ovary syndrome (PCOS) is an endocrine disrupting disorder affecting at least 10 percent of reproductive-aged women. Women with PCOS are at increased risk for diabetes and cardiovascular disease. In North America and Europe, the diagnosis of PCOS may be delayed several years and may require multiple doctors resulting in lost time for risk-reducing interventions. Menstrual tracking applications are one potential tool to alert women of their risk for PCOS while also prompting them to seek evaluation from a medical professional.ObjectiveThe objective of this study was to develop the Irregular Cycles Feature (ICF), an adaptive questionnaire, on the mobile phone application (app) Clue® to generate a probability of a virtual test subject’s risk for PCOS. The secondary objective was to assess the accuracy of the ICF by comparing the probability of risk generated by the app to a probability generated by a physician.MethodsFirst, a literature review was conducted to generate a list of signs and symptoms of PCOS, termed variables. These include, but are not limited to, hirsutism, acne, and alopecia. Probabilities were assigned to each variable and built into a Bayesian network. The network served as the backbone of the ICF, which identified potential subjects through self-reported menstrual cycles and answers to medical history questions. Upon completion of the questionnaire, a Result Screen summarizing the virtual test subject’s probability of having PCOS is displayed. For each eligible virtual test subject, a Doctor’s Report containing information regarding tracked menstrual cycles and self-reported medical history is generated. Both of these documents share information about PCOS and detailed explanations for facilitating a diagnosis by a medical provider. Virtual test subjects were assigned probabilities by a) the ICF and b) a board-certified reproductive endocrinology/infertility physician-scientist, which served as the gold standard. The ICF was set to recommend individuals with a score greater than or equal to 25% to follow-up with their physician. Differences between the network and physician probability scores were assessed using a t-test and a Pearson correlation coefficient. An additional iteration was performed to improve the ICF’s prediction capability.ResultsThe first iteration of the ICF produced only one false positive compared to the physician screening score and had an absolute mean difference of 15.5% (SD= 15.1%) amongst virtual test subjects. Upon modification of the ICF, the second iteration had two false positives as compared to the physician screening score and had an absolute mean difference of 18.8% (SD = 13.6%). The majority of virtual test subjects had an ICF score that over predicted PCOS when compared to the physician. However, there was strong positive significant correlation between the ICF and the physician score (Pearson correlation coefficient= 0.69; p < 0.01). The second iteration performed worse with a Pearson correlation coefficient of 0.54; p > 0.01).ConclusionThe first iteration ICF, as compared to the second, was better able to predict the probability of PCOS and can potentially be used as a screening tool to prompt a high-risk subject to seek evaluation by a medical professional.
Aging women with polycystic ovary syndrome: menstrual cycles, metabolic health, and health-related quality of life
