Development and validation of a predictive model for the risk of developing trichomonas vaginitis in women

Abstract

Trichomonas vaginitis (TV) is the most common non-viral sexually transmitted infection (STI) worldwide. The high prevalence of TV combined with mild or asymptomatic early symptoms leads to clinical vulnerability from delayed diagnosis. Latent infection can increase the incidence of pelvic infections, infertility, and adverse pregnancy complications. Data from 898 women who underwent vaginal flora testing from June 2014 to December 2014 were used to create a nomogram to assess the risk of TV in women in order to guide TV prevention and clinical intervention. The prediction model was evaluated in terms of identification, calibration, and clinical utility using the C-index, calibration plots, decision curve analysis, and internal validation. Predictors in the TV nomogram included age, occupation, yearly income, tea drinking, bathing frequency, menopause, spontaneous abortion, use of contraceptives, history of gynecological surgery, and HPV infection. The C-index of the TV risk prediction model was 0.732 (95% confidence interval: 0.695–0.768). It showed good discriminatory and predictive power. Decision curve analysis indicated that the nomogram had a good net benefit when the threshold probability of TV in women was 2–80%. The established TV prediction model easily, accurately, and quickly predicts the risk of TV onset.