Latent class analysis of symptoms for sexually transmitted infections among Iranian women: Results from a population-based survey

Abstract

A preliminary symptom-based screening test would lower the financial burden of sexually transmitted infections (STIs) caused by clinical testing. To develop such a screening method, we should first identify the most specific STI symptoms. We aim to distinguish the specific STI symptom(s) that are most likely to be found in the truly infected individuals. We used data from a population-based survey that was conducted in Iran, in 2014. Using Latent Class Analysis (LCA) in R software, we classified 3049 Iranian women, 18-60 years old, with reference to seven self-reported STI-associated symptoms. Using LCA, we categorized nearly 1 of women as "probably STI-infected". Above 70 of participants reported the "seven symptoms" that are associated with STIs, except for genital ulcer. These symptoms could be used to distinguish healthy participants from infected ones. The "probably healthy" class incorporated about 77 of the participants. Lower abdominal pain and abnormal vaginal discharge were the most frequently reported symptoms of this class. The LCA determined classes along with the WHO syndromic guidelines for STI diagnosis can help physicians to make a more accurate diagnosis. Hence, cost-effectively, only patients who are classified as probably infected need to be referred to medical laboratories for further investigations.