The prevalence rates were 16.3% for bacterial vaginosis, 15.5% for candidiasis, 7.1% for trichomoniasis, 5.5% for Chlamidia trachomatis, 0.9% for Neisseria gonorrhoeae, 0.5% for Treponema pallidum and 0% for human immunodeficiency virus.
There were 154 symptomatic women and 500 asymptomatic women.
Of the 40 women with cervical infection, 40% presented with any of these symptoms, compared with 22% of women without a diagnosis.
In asymptomatic women, the LRs for signs were: 3.4 (95% CI: 1.3 - 9.2) for yellow, green, or bloody cervical or vaginal discharge; 1.2 (95% CI: 0.6 - 2.3) for ectopy; 1.6 (95% CI: 1.2 - 2.2) for friability; and 1.4 (95% CI: 1.1 - 1.9) for erosion.
In asymptomatic women, the LRs for clinic-based tests were: 1.6 (95% CI: 1.0 - 2.5) for more than 30 cervical polymorphonuclear lymphocytes (PMNs); 1.8 (95% CI: 1.1 - 2.8) for cervical leukocyte esterase (LE) of at least 1; 1.4 (95% CI: 1.0 - 1.8) for vaginal LE of at least 2; 1.5 (95% CI: 1.2 - 1.9) for vaginal pH of at least 4.7; and 1.4 (95% CI: 0.9 - 2.2) for potassium hydroxide (KOH) whiff test.
In symptomatic women, the LRs for symptoms were: 0.9 (95% CI: 0.6 - 1.3) for yellow, green, or bloody vaginal discharge; 1.7 (95% CI: 0.4 - 7.2) for vaginal bleeding; and 0.8 (95% CI: 0.3 - 2.0) for lower abdominal pain.
In symptomatic women, the LRs for signs were: 1.6 (95% CI: 0.4 - 6.4) for yellow, green, or bloody cervical or vaginal discharge; 0.8 (95% CI: 0.3 - 1.9) for ectopy; 0.9 (95% CI: 0.5 - 1.8) for friability; and 1.0 (95% CI: 0.6 - 1.6) for erosion.
In symptomatic women, the LRs for clinic-based tests were: 1.3 (95% CI: 0.7 - 2.4) for more than 30 cervical PMNs; 2.0 (95% CI: 1.2 - 3.2) for cervical LE of at least 1; 1.3 (95% CI: 0.9 - 1.8) for vaginal LE of at least 2; 1.6 (95% CI: 1.2 - 2.1) for vaginal pH of at least 4.7; and 1.5 (95% CI: 0.7 - 3.0) for KOH whiff test.
The logistic regression analysis revealed that more than a high school education, age of 25 years or older, and partner with symptoms were significant predictors of cervical infection. These were included in the risk-based algorithm.
The analysis of complete data suggested that pH testing permitted the identification of a sub-group of women who could benefit from prophylaxis.
When considering chlamydial or gonococcal infection (40 cases in the sample of 655 women), the results were as follows.
With WHO risk assessment, the sensitivity was 0.3, the specificity was 0.790, the PPV was 0.085, the total number of women treated was 141, and the number of cases detected and treated was 12.
With WHO examination, the sensitivity was 0.55, the specificity was 0.566, the PPV was 0.76, the total number of women treated was 289, and the number of cases detected and treated was 22.
With WHO microscopy, the sensitivity was 0.55, the specificity was 0.566, the PPV was 0.076, the total number of women treated was 289, and the number of cases detected and treated was 22.
With the risk-based algorithm, the sensitivity was 0.7, the specificity was 0.706, the PPV was 0.134, the total number of women treated was 209, and the number of cases detected and treated was 28.
With universal pH screening, the sensitivity was 0.8, the specificity was 0.475, the PPV was 0.09, the total number of women treated was 355, and the number of cases detected and treated was 32.
With universal prophylaxis, the sensitivity was 1, the specificity was 0, the PPV was 0.061, the total number of women treated was 655, and the number of cases detected and treated was 40.
With universal etiologic screening, the sensitivity was 1, the specificity was 1, the PPV was 1, the total number of women treated was 40, and the number of cases detected and treated was 40.
When considering bacterial vaginosis (95 cases in the sample of 595 women), the results were as follows.
With WHO risk assessment, the sensitivity was 0.463, the specificity was 0.514, the PPV was 0.153, the total number of women treated was 287, and the number of cases detected and treated was 44.
With WHO examination, the sensitivity was 0.463, the specificity was 0.514, the PPV was 0.153, the total number of women treated was 287, and the number of cases detected and treated was 44.
With WHO microscopy, the sensitivity was 0.168, the specificity was 0.912, the PPV was 0.267, the total number of women treated was 60, and the number of cases detected and treated was 16.
With the risk-based algorithm, the sensitivity was 0.926, the specificity was 0.536, the PPV was 0.275, the total number of women treated was 320, and the number of cases detected and treated was 88.
With universal pH screening, the sensitivity was 0.926, the specificity was 0.536, the PPV was 0.275, the total number of women treated was 320, and the number of cases detected and treated was 88.
With universal prophylaxis, the sensitivity was 1, the specificity was 0, the PPV was 0.160, the total number of women treated was 595, and the number of cases detected and treated was 95.
With universal etiologic screening, the sensitivity was 1, the specificity was 1, the PPV was 1, the total number of women treated was 95, and the number of cases detected and treated was 95.
When considering bacterial trichomoniasis (47 cases in the sample of 655 women), the results were as follows.
With WHO risk assessment, the sensitivity was 0.830, the specificity was 0.541, the PPV was 0.123, the total number of women treated was 318, and the number of cases detected and treated was 39.
With WHO examination, the sensitivity was 0.830, the specificity was 0.541, the PPV was 0.123, the total number of women treated was 318, and the number of cases detected and treated was 39.
With WHO microscopy, the sensitivity was 0.745, the specificity was 1, the PPV was 1, the total number of women treated was 35, and the number of cases detected and treated was 35.
With the risk-based algorithm, the sensitivity was 0.872, the specificity was 0.483, the PPV was 0.115, the total number of women treated was 355, and the number of cases detected and treated was 41.
With universal pH screening, the sensitivity was 0.872, the specificity was 0.483, the PPV was 0.115, the total number of women treated was 355, and the number of cases detected and treated was 41.
With universal prophylaxis, the sensitivity was 1, the specificity was 0, the PPV was 0.072, the total number of women treated was 655, and the number of cases detected and treated was 47.
With universal etiologic screening, the sensitivity was 1, the specificity was 1, the PPV was 1, the total number of women treated was 47, and the number of cases detected and treated was 47.