Obstetric fistula in Bangladesh: estimates from a national survey with clinical validation correction

Background: Obstetric fistula, which develops after a prolonged or obstructed labour, is preventable and treatable. However, many women are still afflicted with the condition and remain untreated in low-income and middle-income countries. Concerns have also been raised that an increasing trend of caesarean sections is increasing the risk and share of iatrogenic obstetric fistula in these countries. The true prevalence of this condition is not known, which makes it difficult for health planners and policy makers to develop appropriate national health strategies to address the problem. The estimation of obstetric fistula with surveys is difficult because self-reporting of incontinence symptoms is subject to misclassification bias. In this study, we aimed to estimate the prevalence and burden of obstetric fistula in Bangladesh. Methods: For a valid estimation addressing misclassification bias, we implemented the study in two steps. First, we did the Maternal Morbidity Validation Study (MMVS) among a population of 65 740 women in Sylhet, Bangladesh, to assess the sensitivity, specificity, positive predictive values (PPVs), and negative predictive values of the survey questions. This was done through confirmation of the diagnosis with clinical examinations of suspected cases by female physicians; a sample of women who screened positive for pelvic organ prolapse and other urinary incontinence symptoms were also examined and used as controls for clinical diagnosis confirmation. Second, we used the estimated diagnostic test values, after correcting for verification bias, to adjust the reported prevalence in the nationally representative Bangladesh Maternal Mortality and Health Care Survey 2016 for the unbiased estimation of obstetric fistula prevalence in Bangladesh. Findings: The MMVS, done from Aug 3 to Dec 9, 2016, identified 67 potential cases of obstetric fistula; of them, 57 (85%) women completed the clinical examination, and 19 were confirmed as obstetric fistula cases. The adjusted sensitivity of the self-reports of obstetric fistula was 100% (95% uncertainty interval [UI] 99·8–100) and the observed specificity was 99·9% (95% UI 99·9–100) among women aged 15–49 years. However, the PPV was low, at 31·6% (95% UI 19·2–46·2), suggesting that almost two thirds of the self-reported cases were not true obstetric fistula cases. We estimated an adjusted obstetric fistula prevalence rate of 38 (90% UI 25–58) per 100 000 women aged 15–49 years in Bangladesh. Nationally, we estimated about 13 376 (90% UI 8686–20 112) women of reproductive age living with obstetric fistula. Additionally, we estimated 4081 (1773–8790) women aged 50–64 years to be living with obstetric fistula in Bangladesh; overall, we estimated that there are 17 457 (10 459–28 902) women aged 15–64 years in Bangladesh with obstetric fistula. Interpretation: The burden of obstetric fistula is still high in Bangladesh. Prevention and provision of surgical treatment to so many women will need coordinated efforts, planning, allocation of resources, and training of surgeons. Funding: US Agency for International Development, Government of Bangladesh, and UKAid.