A brand new Flo Health study, published in the JMIR mHealth and uHealth journal, has identified solid accuracy levels for Flo’s groundbreaking Symptom Checker feature, signaling an exciting step towards helping reduce time to diagnosis for conditions like polycystic ovary syndrome (PCOS), endometriosis and uterine fibroids by educating users. Symptom Checker feature is not intended to be used for diagnosis or treatment, but only for education and providing information.
Reproductive health conditions like endometriosis, uterine fibroids, and polycystic ovary syndrome (PCOS) impact a significant number of women and individuals who menstruate globally, with estimates ranging from5% to 40% of reproductive-age women. Currently, it can take up to 12 years to diagnose these specific conditions, subsequently contributing to health complications and increased healthcare costs.
Using a vignette type of study, Flo Health tested the accuracy of its Symptom Checker feature associated with these conditions, and found it performed with a very solid level of accuracy. The findings suggest that Flo’s Symptom Checker can provide users with information to better understand their symptoms, and help them determine whether these symptoms could match a specific health condition like PCOS, endometriosis and - soon - uterine fibroids. This means that Flo’s Symptom Checker has the potential to help significantly reduce the time to diagnosis for reproductive health conditions by educating women about common symptoms of these reproductive health conditions.
Aidan Wickham, Lead Research Scientist from Flo Health, commented: "The results of our study hold vast potential for Flo's Symptom Checker feature. We hope to empower our users to take control of their health and ensure optimal well-being with their healthcare providers. Vignette-patient case studies are an important first step in algorithm testing for digital symptom checkers before their release and ensuring they perform with high accuracy. Having received such fantastic results, we're excited to keep building upon our Symptom Checker feature."
To test its Symptom Checker, Flo’s Science team recruited multiple panels of general practitioners, independent from Flo. Separate panels of independent GPs then created clinical case vignettes of simulated patients, classified them for each condition, and entered the symptoms of each case into the Symptom Checker. Finally, the output of the Symptom Checker was compared to the designation from the independent GPs for each vignette case and the results indicated very high levels of agreement at 83-88%, which implies a solid level of accuracy.
The published study can be found here: https://mhealth.jmir.org/2023/1/e46718