Modeling of sti prevalence among hiv-infected adults in hiv care programs in Kenya using logistic regression

Abstract

Objectives and methods: This paper investigates risk factors associated with prevalence of sexually transmitted infections. Logistics regression is employed to determine the risk factors. Furthermore, the study validates the syndromic diagnosis of STI using etiological diagnosis as the gold standard. Sensitivity, specificity and positive predictive values is used in this in this validation process. Results: 9.81% of HIV-infected participant had an etiologic diagnosis with majority being trichomoniasis (8.04% overall prevalence). There was much lower prevalence of gonorrhea (1.33%), chlamydia (0.36%), and syphilis (0.48%). 69% of the participants had positive serology tests for HSV-2. Among women who participated in the study, 18.6% were diagnosed with bacterial vaginosis. Sensitivity, specificity and PPV of genital ulcer were 0.00%. Sensitivity of urethral discharge for predicting Gonorrhea or chlamydia low (20%), specificity was high at (95%) while PPV was 4%. Sensitivity, specificity and PPV of virginal discharge for predicting Gonorrhea, chlamydia or trichomoniasis in females are 36%, 76%, and 18% respectively. The odds of having STI for male was 0.33 (95% CI : 0.20- 0.53) compared to women. The odds of having STI for the participants aged 45 years and above was found to be 0.33 (95% CI : 0.15 - 0 .71) as compared to those of age between 18 to 24 years old. In addition, those with vocational training/secondary education and post secondary education and above had odds ratio of 0.56 (95% CI : 0.39- 0.82) and 0.05 (95% CI : 0.01-0.38). Those with more than one sexual partner was significantly associated with having STI, odds ration of 3.45 (95% CI: 1.43-8.30). Conclusion: Prevalence of any STI was relatively low compared to results from other studies. Trichomoniasis was overwhelmingly the most common genital nonviral infection, while HSV-2 had the highest overall prevalence. Being a femaleyounger age, less education, and more recent sexual partners were all associated with an increased risk of STI. Syndromic diagnosis of DD, VD and CD were insensitive, while their specificity were relatively good.