Syndromic management of vaginal discharge and lower abdominal pain in Honduras

Algorithm validation

Abstract

Objectives: 1. to validate the proposed flow charts for the syndromic diagnosis and treatment of vaginal discharge and lower abdominal pain, determining their sensitivity, specificity, positive predictive value and associated socio-demographic, behavioral and clinical factors and comparing them to the etiologic diagnosis; 2. To study the relationship between the presence of vaginal discharge, lower abdominal pain and other STIs and HIV infection in the population studied. Methods: prospective descriptive and analytical study on a sample of 933 sexually active women, pregnant and non-pregnant, presenting spontaneously to eight health centers in the Tegucigalpa Metropolitan Health Region complaining of vaginal discharge and/or lower abdominal pain, between October 1999 and November 2000. The criteria used to validate the proposed methods were sensitivity, specificity and positive predictive value. Sensitivity means the proportion of infections detected by the method (the higher the specificity, the larger the number of infected individuals treated); specificity is the proportion of non-infected cases identified as such; and positive predictive value shows the proportion of truly infected cases in those defined as such by the method (the higher the PPV, the lower the number of individuals unnecessarily treated). Results: Socio-demographic: of the 933 women, 41.8% were under 25 years of age, with a mean age of 28 years and a median age of 26; 6.2% were illiterate, and 16.7% had no stable partner. Obstetric history: 31.4% were pregnant and 46.6% had had more than two previous pregnancies. Sexual history: 13.1% had their first sexual intercourse before 15 years of age; 79.6% had only one sexual partner in the previous 3 months; only 6% used a condom in all sexual relations; 12.8% practiced anal sex and 12.7% were aware that their partners had some genital sign or symptom. Prevalence of genital infections: 4% had specific cervicitis, 53.3% had vulvovaginitis and 4% had pelvic inflammatory disease; 1.2% was HIV-positive. Factors associated to cervicitis: the only significant risk factor for cervicitis was the absence of a stable partner. Associated signs and symptoms were cervical mucopus, cervical erythema and cervical motion tenderness. Validity of the specific clinical diagnosis: Overall the sensitivity, specificity and predictive positive value (PPV) of the clinical diagnosis based solely on clinical experience were low; respectively 18%, 96% and 15% for cervicitis; 11%, 97% and 14% for PID; 63%, 79% and 18% for trichomoniasis. Validity of the proposed flow charts: several flow charts for vaginal discharge proposed for use in Honduras were simulated and validated. Validation of the first flow chart, based solely on risk assessment, showed sensitivity, specificity and PPV for cervicitis of 16%, 94% and 10% respectively. The second flow chart, with the addition of speculum examination to identify the discharge’s origin and characteristics, had sensitivity, specificity and PPV of 49%, 53% and 4% for cervicitis, and 46%, 62% and 38% for trichomoniasis and/or bacterial vaginosis. Sensitivity, specificity and PPV of the third flow chart, which included microscopy, were 57%, 60% and 6% for cervicitis and 90%, 100% and 100% for trichomoniasis. For the fourth flow chart, including some specific cervicitis features and the pH test for vaginitis/vaginosis, sensitivity, specificity and PPV were 67%, 59% and 6% respectively for cervicitis and 96%, 45% and 47% for trichomoniasis and/or bacterial vaginosis. Since the study defined socio-demographic and behavioral risk factors in addition to the associated signs and symptoms, it allowed us to design a new vaginal discharge flow chart, aiming at a balance of validity criteria but always seeking the highest possible level of sensitivity; simulation of this new flow chart showed 85% sensitivity, 38% specificity and 4% PPV. The sensitivity, sensibility and PPV of the flow chart used for lower abdominal pain were 38%, 89% and 21%, respectively. Conclusion: the study demonstrated the poor validity of the specific clinical diagnosis based solely on signs and symptoms, and therefore the need for an alternative method. Although the syndromic management was shown under all comparison criteria to be better than clinical diagnosis, it is necessary to adapt it according to the situation of the health services and particularly to the characteristics of the population. In this sense, the last flow chart proposed will be one of the tools indicated for managing patients with vaginal discharge in those health centers that have the supplies required, such as wet mount microscopy and pH strips. Flow chart 4 is recommended for health centers with no microscopy facilities. The use of the flow chart for lower abdominal pain continues to be recommended: while its sensitivity was only 37.5%, it is higher than that of clinical diagnosis.