Assessment of a Digital Symptom Checker Tool's Accuracy in Suggesting Reproductive Health Conditions: Clinical...

Area of Science:

• Digital health
• Reproductive medicine
• Clinical informatics

Background:

• Reproductive health conditions like endometriosis, uterine fibroids, and PCOS affect many women, with diagnostic delays up to 12 years.
• These delays lead to complications and increased healthcare costs.
• Symptom checker apps offer potential to shorten diagnosis times by empowering users with information.

Purpose of the Study:

• Evaluate agreement between clinicians and 3 Flo Health symptom checkers for endometriosis, uterine fibroids, and PCOS.
• Demonstrate a method for testing and validating digital health symptom checker tools using vignettes.

Main Methods:

• General practitioners created and reviewed clinical vignettes for testing symptom checkers.
• Vignettes included mixed outcomes (condition-positive and -negative).
• Symptom checker outputs were compared to clinician classifications to calculate accuracy metrics.

Main Results:

• Overall agreement between symptom checkers and clinicians was high: 83% for endometriosis, 83% for uterine fibroids, and 88% for PCOS.
• Sensitivity ranged from 81.8% to 100%, and specificity from 75% to 84.6%.
• Positive and negative predictive values were also strong, indicating reliable performance.

Conclusions:

• Single-condition symptom checkers demonstrate high agreement with general practitioner diagnoses for key reproductive health conditions.
• These digital health tools show promise for improving care pathways and reducing diagnostic delays.
• Innovative apps can potentially mitigate health complications and costs associated with delayed diagnosis.