Global, regional and national estimates of obstetric fistula prevalence Authors: Saifuddin Ahmed, Rene Genadry, Bridget Asiamah, Mengjia Liang, Vandana Tripathi, Erin Anastasi Published: 2025 ABSTRACT AbstractIntroductionThe UN General Assembly in 2018 passed a resolution to eliminate obstetric fistula within a decade. Because the prevalence of obstetric fistula is not known, it is difficult to develop appropriate national and global health strategies to prevent, treat and eliminate the condition. We estimate the prevalence of obstetric fistula in 55 countries targeted by the UNFPA-led global Campaign to End Fistula Programme.MethodsWe used data on reported symptoms of obstetric fistula from 31 Demographic and Health Surveys conducted in 22 countries and applied a predictive model to estimate the global, regional and country-level prevalences of current obstetric fistula cases in target countries. Self-reported fistula symptoms are subject to over-reporting and misclassification biases. Using known estimates of clinically diagnostic validation from large population studies, we additionally corrected self-reported obstetric fistula estimates for misclassification errors.ResultsWe estimate the global fistula prevalence of about 457 000 (95% CI 303 000 to 709 000) among women aged 15–64 and about 386 000 (95% CI 256 000 to 599 000) among women of reproductive age (15–49). The obstetric fistula prevalence rate was almost two times higher in the sub-Saharan region (71 per 100 000 women) compared with Asia (36 per 100 000 women).ConclusionThe estimated prevalence of obstetric fistula is substantially lower than the previously cited numbers of 2 million women. The prevalence and incidence rates of obstetric fistula have likely declined over time due to a number of factors. Skilled birth attendance, access to emergency obstetric care and caesarean section deliveries have substantially increased in low- and middle-income countries with a concurrent decline in maternal mortality during the last two decades. The global Campaign to End Fistula, launched in 2003, together with other international efforts, has also targeted improving prevention and treatment of obstetric fistula during this period. This is a preventable and treatable condition, and the global community must mobilise to eliminate this devastating and tragic condition by ensuring that women have access to timely and quality emergency obstetric care, including safe surgical care for caesarean section. WHAT IS ALREADY KNOWN ON THIS TOPIC Literature on obstetric fistula suggests that around 2 million women globally suffer from the condition, though some studies estimate the number to be between 1 and 3 million. Overall, it is considered that the true prevalence level of obstetric fistula is not known. Literature on obstetric fistula suggests that around 2 million women globally suffer from the condition, though some studies estimate the number to be between 1 and 3 million. Overall, it is considered that the true prevalence level of obstetric fistula is not known. WHAT THIS STUDY ADDS This study attempts to systematically estimate obstetric fistula prevalence at the global, regional and national levels for the countries known to have a notable number of cases. This study attempts to systematically estimate obstetric fistula prevalence at the global, regional and national levels for the countries known to have a notable number of cases. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY The estimated global prevalence of obstetric fistula is significantly lower than the previously cited numbers. The estimated global prevalence of obstetric fistula is significantly lower than the previously cited numbers. Nevertheless, the current burden of cases poses a huge challenge to ensuring surgical access to these women during their lifetime. Nevertheless, the current burden of cases poses a huge challenge to ensuring surgical access to these women during their lifetime. Introduction Substantial interest has emerged to estimate maternal morbidities in low- and middle-income countries (LMICs), recognising that many of them are also the causes of maternal and perinatal mortality.13 The WHO defines maternal morbidity as “any health condition attributed to and/or aggravated by pregnancy and childbirth that has a negative impact on the woman’s wellbeing”.4 Among all maternal morbidity conditions, obstetric fistula is the most devastating and debilitating5 and is considered a public health and human rights problem.6 Obstetric fistula, often referred to as female genital fistula, develops after prolonged and obstructed labour. In this condition, an abnormal opening develops between the vagina and the urinary bladder and/or rectum, clinically known as vesicovaginal fistula (VVF) and rectovaginal fistula, respectively, leading to continuous urinary and/or faecal incontinence. When the progression of labour is protracted and becomes obstructed, the soft tissues of a woman’s bladder, vagina and rectum are compressed between the fetal head and maternal pelvic bones due to the sustained contractions of the uterus. Consequently, the blood supply to the woman’s genitourinary organs is progressively reduced, resulting in ischaemia and tissue necrosis. Almost 90% of such deliveries result in fetal deaths.7 Often, obstetric fistula results in extensive tissue loss and dense scarring, which makes the surgical operation difficult, often requiring multiple surgeries. Vaginal fistula may also occur from accidental injury during pelvic surgery (iatrogenic fistula), including caesarean section (CS) and hysterectomy, or by violence (traumatic fistula). For the purpose of this study, we include fistula developed after childbirth—a fistula caused by pressure necrosis and could be an iatrogenic fistula—within the definition of obstetric fistula. Women and girls suffering from obstetric fistula are highly stigmatised and often become social outcasts. Arrowsmith and colleagues coined the term ‘obstructed labour injury complex’ to encompass the extent of physical and social injuries caused by fistula.8 In numerous ways, fistula adversely affects women and their families socially, emotionally and economically.9 10 In 2003, the UNFPA and its partners launched the ‘Campaign to End Fistula’ to make obstetric fistula as rare in LMICs as it is in the high-income world and in 2022, the UN Secretary-General called for greater investment to accelerate the action to end fistula within a decade.11 Since then, significant interest has emerged in assessing the burden of obstetric fistula globally and at the country level. It has been estimated that 1–3 million women and girls are suffering from obstetric fistula globally.6 12 However, the true prevalence level is not known.1315 The most commonly quoted estimates of obstetric fistula globally are 2 million prevalent cases and 50 000–100 000 incident cases annually.13 These estimates are often referred to as the WHO estimation of obstetric fistula. However, the original citation for these estimates is a 1993 paper by Waaldijk and Armiya’u in which it is stated: From the authors’ experience in Northern Nigeria, where even 700 repairs a year are not sufficient to cope with all the VVF patients presenting, the incidence rate can be calculated at 1–2 per 1000 deliveries where the mother survives in situations where there is no easy access to a functioning obstetric unit.16 Clearly, the basis of the estimation was not empirical data but, rather, anecdotal. Wall subsequently suggested that about 3 million women are suffering from obstetric fistula; however, no data source or basis was provided for the statistics.6 Overall, a systematic literature review concluded that obstetric fistula prevalence and incidence are understudied and biased.17 Using meta-analytic techniques, Adler and her colleagues estimated a pooled obstetric fistula prevalence rate of 0.29 (95% CI 0.0 to 1.07) per 1000 women of reproductive age globally; 1.60 (95% CI 1.16 to 2.10) per 1000 women aged 15–49 in sub-Saharan Africa and 1.20 per 1000 women aged 15–49 in South Asia.12 The study pooled the results from 10 community-based studies and three hospital studies. Of the 10 community-based studies, 6 studies had estimated a prevalence level of 0%. All these studies were conducted in the Arab States region and relatively economically advantaged countries (Iran, Turkey, Jordan, Lebanon, Oman and Egypt). The study estimates that just over one million women may have an obstetric fistula in sub-Saharan Africa and South Asia. The Demographic and Health Surveys (DHS) began to collect data on incontinence symptoms in 2005 in selected countries of high risk of obstetric fistula, predominantly in sub-Saharan Africa. However, the DHS questionnaire on fistula was never validated for the true diagnosis of obstetric fistula. The reported incontinence rates ranged from a high of about 5% to a low of 0%. There were significant concerns that the tool had low sensitivity and specificity and was, therefore, likely to grossly over-report the prevalence of obstetric fistula. Maheu-Giroux and colleagues re-estimated the prevalence of obstetric fistula in a number of sub-Saharan African (SSA) countries after correcting for misclassification in self-reported symptoms in DHS and Multiple Indicators Cluster Surveys, using priors of a set of ranges in sensitivity and specificity values through a Bayesian analytical method.18 The study estimated a lifetime obstetric fistula prevalence rate of three cases per 1000 women of reproductive age and the current prevalence rate of one case per 1000 women of reproductive age. There are, however, some concerns that the prevalence rates from DHS are higher than the local population level study suggests, and the rate may reflect the estimate of urinary incontinence symptoms rather than the true fistula prevalence rate.19 Another limitation of the study was that the estimates were limited to women of reproductive age. Many women remain untreated for years and even decades, and several studies suggest the presence of a high number of older women as obstetric fistula patients. Moreover, the estimates are limited to 19 countries only; no information was available for countries where DHS data are not available. No reliable estimates are currently available on obstetric fistula at the global and country levels. To address the substantial data gaps in obstetric fistula, we estimated its global, regional and national prevalences using empirical sources. Methods We have re-examined the utility of the DHS data and exploited the availability of auxiliary information in the survey on the possible cause of fistula, timing of fistula formation in relation to pregnancy or adverse birth outcomes, the lag period between delivery and fistula symptoms development, results of treatment (if any) and current incontinence status to correct self-reporting bias in obstetric fistula prevalence estimation. In addition, we have used the recently available empirical data for correcting misclassification bias in survey instruments on the diagnostic validity of obstetric fistula reporting that used confirmation of cases through clinical examination by expert physicians.20 21 Finally, we used predictive modelling for estimating obstetric fistula in all 55 countries in the world that are considered to have non-negligible levels of obstetric fistula to provide national, regional and global estimates of obstetric fistula following a childbirth. Although vaginal fistula may occur from accidental injury during pelvic surgery or by violence, we focused on obstetric fistula only to reduce misclassification errors. The countries selected were identified and supported by the UNFPA-led global Campaign to End Fistula. Data The study uses data from four sources. First, we used 31 DHS data from 22 countries that collected self-reported information on the experience of urinary and/or faecal incontinence from women of reproductive age (https://dhsprogram.com/). Second, we used the WHO/Maternal Mortality Estimation Inter-Agency Group’s (MMEIG’s) 2020 maternal mortality estimation database which estimates global, regional and country-level maternal mortality.22 The database was compiled by the MMEIG. The data on maternal mortality ratio (MMR) were used in the predictive model for estimating obstetric fistula for the countries where DHS data were not available. Third, we used two published reports of validation of the self-reported obstetric fistula symptoms with clinical examinations that used population-level sampling. These two studies were conducted in Bangladesh21 and Ethiopia.20 Bangladesh Maternal Morbidity Validation Survey 2016 was conducted among 65 740 women in two geographical areas where experienced female physicians conducted a clinical examination and dye test with methylene blue to confirm fistula.21 The study in Ethiopia was based on a randomly selected 23 023 women of reproductive age (15–49 years) from 113 villages. The reported suspected cases based on self-reporting of incontinence symptoms were confirmed by a clinical examination and dye test with methylene blue by a clinical member of the research team or a clinician in the health centre. Diagnostic measures of accuracy (sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)) from these two studies were used to correct misclassification of self-reporting of obstetric fistula symptoms. A few other studies in Nepal,23 Rwanda,24 Pakistan25 and Nigeria26 also examined the validation of self-reporting with a clinical examination. However, these studies were based on a case-control design or small sample sizes or the patients were identified through outreach efforts, and they were not suitable for population-level diagnostic accuracy estimates. Fourth, we used data from the female population size aged 15–64 years from the United Nations World Population Prospects database, 2020 (the corresponding year of the maternal mortality estimates used for the analysis (https://population.un.org/wpp/). Statistical analysis method We estimated the current prevalence of obstetric fistula in five steps. First, we examined the self-reported positive response to urinary and/or faecal incontinence symptoms as reported by the respondents in the DHS. Second, we attempted to correct misclassification in self-reporting from other auxiliary information available in the DHS. We consider valid obstetric fistula cases to be those who reported the symptoms to start after a child delivery or difficult labour or after a stillbirth. An obstetric fistula usually starts within 10 days of delivery (after hypoxic necrotic tissue starts sloughing). In a few countries, an additional question was added on the timing of fistula development after the delivery. Any report that suggested fistula developed after 30 days was considered a misclassification; we conservatively used 30 days as the cut-off period, considering the low literacy of the study respondents and concerns of recall bias. Third, for the estimation of the current prevalence of obstetric fistula, we considered the women who reported that they had received a (surgical) treatment and their condition had been cured as non-cases. Fourth, we used a predictive model to estimate the obstetric fistula prevalence rate, primarily because several countries have no data available on obstetric fistula. For setting the predictive model, we hypothesise that the prevalence of fistula rate in a country is proportional to the level of MMR, skilled attendance at birth (SAB) rate, general fertility rate (GFR) and CS rate: MMR reflects the country’s mortality risk per birth, and GFR reflects each woman’s exposure to such risks in a calendar period. Higher obstetric fistula is expected in countries where MMR and GFR are high. In contrast, where SAB and CS rates are high, at least theoretically, obstetric fistula rates are expected to be low. However, fitting such a model empirically is challenging. The MMRs for LMICs are model-based estimates from GFR and SAB (including per-capita gross domestic product to capture developmental status) and are likely to be highly correlated. Despite the WHO guidance that CS levels of about 15% are not associated with further reductions in maternal mortality, the exact relationship between the level of CS and maternal morbidity is not known. Moreover, in countries with low surgical competency, CS may increase the risk of iatrogenic fistula and it is incredibly challenging to differentiate the cause due to iatrogenic or obstructed labour among women receiving CS from survey data and/or self-report alone. Another challenge is the very low sample size of the countries with data on obstetric fistula that may not allow them to fit multiple covariates in a regression-based prediction model. Because MMR captures overall risks of mortality from maternal causes and is estimated from the level of GFR and SAB, we used the MMR variable in the prediction model, including regional classification. As a part of sensitivity analysis, we also included SAB in the model but found the model fit deteriorated based on Akaike Information Criterion and Bayesian Information Criterion statistics. Because obstetric fistula rates are expected to vary across geographical regions, we used the UN Population subregion classification of the countries as a controlling variable in the prediction model. We used a fractional logistic model to predict the (expected) proportion of obstetric fistula in each country. This model specification prevents predicting negative values, which may occur with a linear regression model, specifically with very low prevalence level data. Because we used the data from the WHO/MMEIG 2020 maternal mortality estimates, the results in this analysis represent the obstetric fistula estimates for 2020. All DHS that included the fistula module were conducted prior to the publication of the MMEIG’s 2020 MMR estimates. The results are based on the predicted estimates from the model, which allowed us to present the estimates for countries with and without DHS data. Finally, we applied a statistical method to correct the misclassification of the obstetric fistula survey instrument. A number of studies have suggested that self-reporting of obstetric fistula overestimates the obstetric fistula prevalence rate. A large study of 23 023 women in Ethiopia found 35 women reported fistula symptoms, but a clinical examination confirmed only 13 women as true obstetric fistula cases (37.1%).20 Another study of 17 893 women in Nepal found only two confirmed cases of fistula among 55 symptomatic patients.23 A study of 65 740 women in Bangladesh found only 19 confirmed cases among 57 (33.3%) symptomatic patients.21 None of these studies found false negative cases. It is possible to use sensitivity (Se) and specificity (Sp) information of a diagnostic/screening test to correct (self-reported) prevalence level (Pobs) for an unbiased estimate (Punb) by an epidemiological method proposed by Rogan and Gladen,27 where: There are two problems with using the Rogan and Gladen method for correcting the misclassification of obstetric fistula based on reported symptoms. If the observed prevalence rate is lower than the false positive rate (1–specificity) of the diagnostic test, the estimated unbiased prevalence rate becomes negative, which does not make any sense.28 Another problem is verification bias. For ethical and logistical reasons, it is a common practice to conduct confirmatory diagnostic tests on positive symptomatic subjects only, yet this may miss cases among non-symptomatic cases. A major problem of this practice is verification bias that overestimates Se and underestimates Sp, which consequently biases prevalence estimates. It is, however, possible to estimate an unbiased prevalence rate from the PPV and NPV of the diagnostic test under the assumption of conditional independence of missing confirmed diagnosis of unverified cases, which is often referred to as missing at random for the unobserved values of the test results.29 In this case, where P(T=1) and P(T=0) are the probabilities of self-reporting positive and negative symptoms, respectively We use PPV and NPV estimates from two published studies of validation of the self-reported obstetric fistula symptoms with clinical examinations that used population-level sampling. These two studies were conducted in Bangladesh21 and Ethiopia.20 We corrected estimates of self-reported obstetric fistula prevalence for misclassification with three methods: Method 1: applied Bangladesh validation estimates of PPV and NPV to all countries. Method 1: applied Bangladesh validation estimates of PPV and NPV to all countries. Method 2: applied the PPV and NPV from the combined data of Bangladesh and Ethiopia studies. Method 2: applied the PPV and NPV from the combined data of Bangladesh and Ethiopia studies. Method 3: applied PPV/NPV estimates from Ethiopia data to SSA countries and Bangladesh estimates for Asian and other region countries. Method 3: applied PPV/NPV estimates from Ethiopia data to SSA countries and Bangladesh estimates for Asian and other region countries. PPV was 31.8% in the Bangladesh study, 37.1% in the Ethiopia study and 34.2% in the combined study estimate. In both countries, NPV was 100%. Women develop frank incontinence in the case of obstetric fistula, and it is unlikely not to report incontinence if present. Conservatively, however, we considered NPV=0.99999, as a few ostracised women with obstetric fistula may be excluded in the household surveys and not counted. This assumption is likely to slightly increase the prevalence estimation. Patient and public involvement It was not appropriate or possible to involve patients or the public in the design, or conduct, or reporting, or dissemination plans of our research. Results Prevalence of incontinence symptoms Figure 1 shows the reported level of self-reported urinary and/or faecal incontinence symptoms in 22 DHS countries. The self-reported prevalence levels of incontinence symptoms range from a low of 1 to a high of 41 per 1000 women of reproductive age (none in Gambia DHS 2019–2020 has reported incontinence symptoms and was not included in the analysis). The average reported prevalence level was 7.20 (95% CI 6.10 to 8.30) per 1000 women. After correcting for the obstetric origin of symptoms, the overall average prevalence level of incontinence following childbirth was reduced to 3.1 per 1000 women (figure 2). The reported prevalence of obstetric fistula incontinence symptoms in the Demographic and Health Surveys. The prevalence of obstetric fistula incontinence symptoms in Demographic and Health Surveys after correction for obstetric origin (eg, developed after childbirth within 30 days). The overall current prevalence level of incontinence symptoms, after excluding the cases who reported cured, was 1.95 per 1000 (figure 3) in 22 DHS countries. These estimates were used in the fractional logistic regression models for predicting the expected prevalence of self-reported obstetric fistula in 55 countries based on the hypothesised relationship between MMR and obstetric fistula symptoms. The current prevalence of obstetric fistula incontinence symptoms in Demographic and Health Surveys after excluding the cases who were cured. Prevalence of obstetric fistula We used PPV and NPV values for correcting the estimated self-reported symptoms and estimated the current prevalence rates of obstetric fistula using three methods. The results are comparable and we present the model 3 results, which applied the regional-specific diagnostic correction factors, in the main text, and the model 1 and 2 results are presented in online supplemental web appendix table S1. Table 1 shows the number of obstetric fistula cases by countries among women of reproductive age (15–49) and among women aged 15–64, along with the corresponding MMR in 2020. Nigeria has the highest number of obstetric fistula cases in sub-Saharan Africa, estimated at 45 166 (95% CI 24.275 to 84 311) women of reproductive age and 52 004 (27 950 to 97 075) women aged 15–64, followed by Ethiopia (19 367 (12 742 to 29 507) women 15–49 and 22 223 (14 621 to 33 858) women 15–64) and DR Congo (11 701 (9484 to 14 447) women 15–49 and 13 477 (10 923 to 16 640) women aged 15–64). Table 1The number of obstetric fistula cases by countriesCountriesMMRSelf-reportedOF estimatesStatistically corrected OF estimatesCases aged 15–49 (n)Cases aged 15–64 (n)Cases (n)95% CIlower limit95% CIupper limitCases (n)95% CIlower limit95% CIupper limitEast and Southern Africa Angola22212046350918466734404421277759 Botswana18616261427244752503287885 Burundi49420878218914333350245416073756 Democratic Republic of the Congo5471505711 701948414 44713 47710 92316 640 Eritrea322186706034128837094851038 Eswatini (formerly Swaziland)24017365199126314234149369 Ethiopia2671786719 36712 74229 50722 22314 62133 858 Kenya5302117910 641681016 66512 135776619 004 Lesotho56621581477298765563352904 Madagascar39219674512335167475593640748662 Malawi38119473356124505186394927175751 Mozambique127148564180196489844792225210 301 Namibia21516864415252684491299809 Rwanda25917767226414773481260717004008 South Africa127148568714409118 75310 942513623 547 South Sudan12232619823179775527274111566539 Uganda284181687285487410 9138118543212 161 United Republic of Tanzania238173659643611415 25410 938693517 303 Zambia13515057265712885532297614436195 Zimbabwe35719172304721004431341323514962West and Central Africa Benin52321179235514803755272317124342 Burkina Faso26417867337517536536386520087486 Central African Republic835167637235329848386161140 Chad106317867246716003813280118174330 Côte d'Ivoire48020678507431698146570535639160 Equatorial Guinea212119451608231618896370 Gabon22712146260138492302161573 Ghana263178675439282210 5476403332212 415 Guinea55321480256616144087297018694731 Guinea-Bissau72522986435265716499304822 Liberia652223841030640166311887381918 Mali44020276354521855765397424506464 Mauritania4642057785152913739836111585 Niger44120276386223816279441827247183 Nigeria10472519445 16624 27584 31152 00427 95097 075 Republic of Cameroon43814254344726164547391529725165 Republic of Congo2821274865939910967704651279 Senegal26117767269113915239310116036036 Sierra Leone4432027615429522505178611022902 The Gambia45820477470291759539334870 Togo3991977415129122514175910612923Arab States Djibouti23417265193122307230145366 Somalia62122083300018035003343120625722 Sudan270178677169473510 8798359552112 686 Yemen1833501139175809810 39610 273906711 640Asia and the Pacific Afghanistan62045314513 36111 57515 42215 11313 09317 444 Bangladesh1231073516 57212 36722 26120 03314 95026 911 Cambodia21816955232714283809295318114833 India10310233121 93581 346183 695151 402101 004228 087 Lao People’s Democratic Republic12614748940441202711195252412 Nepal17411738324030083489383635624132 Pakistan1541133720 63517 73024 03024 39920 96428 414 Timor-Leste20416654175104294202120340Latin America and the Caribbean Guyana11214347994322812455286 Haiti35019062186412852709220915233211MMR, maternal mortality ratio; OF, obstetric fistula. The number of obstetric fistula cases by countries MMR, maternal mortality ratio; OF, obstetric fistula. We present the corrected prevalence rates per 100 000 women, as these estimates are reasonably low and easier to express in integer numbers. India, the second largest populous country in the world with 1.3 billion population, has an obstetric fistula prevalence rate of 33 per 100 000 women. Although the country has a much lower obstetric fistula prevalence rate compared with the global rate of 50 per 100 000 women, India represents the highest burden of obstetric fistula cases globally because of its enormous population size: 121 720 (81 346 to 183 695) women of reproductive age and 151 402 (101 004 to 228 087) women aged 15–64. Afghanistan has the highest obstetric fistula prevalence rates in the world (145 per 100 000 women). It is estimated that there are about 13 360 women aged 15–49 and 15 110 women aged 15–64 suffering from obstetric fistula in Afghanistan. Pakistan has the third highest level of obstetric fistula cases globally: 20 635 (17 730 to 24 030) women aged 15–49 and 24 399 (20 964 to 28 414) women aged 15–64. We estimate that globally there are about 386 631 (256 611 to 599 630) obstetric fistula cases aged 15–49 and 457 657 (303 693 to 709 705) cases aged 15–64 (table 2). In SSA countries, there are approximately 196 309 (119 184 to 331 269) obstetric fistula cases of women aged 15–49 and 225 996 (137 018 to 381 997) aged 15–64. Other regional estimates are shown by the UNFPA’s regional designation. In sub-Saharan Africa, West and Central African countries have the highest prevalence rates (81 per 100 000 women) with an estimated prevalence of 178 769. Among all regions, the obstetric fistula prevalence rate is the highest in the Arab region, which includes four countries (Djibouti, Somalia, Sudan, Yemen): 86 per 100 000 women. Although the prevalence rate of obstetric fistula is lowest in Asia and the Pacific region (36 per 100 000 women), because of the larger population size, the absolute number of obstetric fistula cases is high in the region: approximately 179 184 and 219 056 women aged 15–49 and 15–64, respectively. Only two countries in the Latin America and Caribbean region—Guyana and Haiti—are included in the analysis, and it is estimated that there are about 2300 cases in this region. Table 2The number of obstetric fistula cases by regionCountryFemale population(15–49 in ‘000)Female population(15–64 in ‘000)Obstetric fistula prevalence rate 100 000Statistically corrected OF estimatesCases aged 15–49 (n)Cases aged 15–64 (n)Cases (n)95% CIlower limit95% CIupper limitCases (n)95% CIlower limit95% CIupper limitSub-Saharan Africa* 274 659.5 316 505.1 71 196 309 119 184 331 269 225 996 137 018 381 997 East and Southern Africa151 520.3174 795.76598 31962 497159 636113 24571 811184 453 West and Central Africa108 560.3124 782.88187 62850 028155 442100 73057 478178 769 Arab States22 733.926 057.78619 53614 75826 58622 29316 79630 415 Asia and the Pacific491 959.0604 458.636179 184128 000255 029219 056156 031312 572 Latin America and the Caribbean3237.73852.662196313292937233315783497 World 778 011.2 933 947.3 50 386 631 256 611 599 630 457 657 303 693 709 705 *Sub-Saharan Africa includes two African regions and three Arab States (Djibouti, Somalia, Sudan).OF, obstetric fistula. The number of obstetric fistula cases by region Sub-Saharan Africa includes two African regions and three Arab States (Djibouti, Somalia, Sudan). OF, obstetric fistula. Discussion A critical need for reliable data on the regional and global burden of obstetric fistula led us to use a novel methodology using the best available national-level empirical data to date. We believe that this study represents the optimal utilisation of self-reported national data on obstetric fistula and statistical methodology to provide acceptable national, regional and global estimates. The estimated numbers are substantially lower than the previously cited numbers of 2 million, which were first suggested in the early 1990s, and are still frequently cited. We recognise several limitations of our study. In the absence of validated data on obstetric fistula prevalence at the national level, we used self-reported data, which are subject to gross misclassification. We have used the validation data that employed confirmation of cases through clinical examination by expert physicians or clinical professionals to correct any misclassification bias in survey instruments on the diagnostic validity of obstetric fistula. The DHS began to collect data on incontinence symptoms in 2005. There are two major challenges to using the DHS data for reliably estimating and tracking changes in obstetric fistula prevalence. The DHS questionnaires were not consistent across all the survey rounds and countries. They have often been modified over the DHS cycles and were never validated for checking self-reported responses to correct clinical diagnosis. Such information is critical for true estimation of the obstetric fistula prevalence and correcting misclassification in reporting. Maheu-Giroux and colleagues have used the DHS data for estimating obstetric fistula in 19 SSA countries, which used theoretical sensitivity and specificity levels for correcting the prevalence level.18 We rather used empirical data from two large validation studies from two countries, Bangladesh and Ethiopia, to correct the self-reported DHS data and used auxiliary data on obstetric contexts of the index pregnancy for addressing reporting biases. We recognise the limitations of this approach. The two countries may not represent the two distinct geographical regions. Nevertheless, we consider that our model-based results are comparable to many published results. The study in rural Ethiopia, based on data collected in 2014, estimated a prevalence rate of 60 (95% CI 30 to 100) per 100 000 women, and the current study for Ethiopia estimates 67 (95% CI 44 to 103) per 100 000 women. Bangladesh Mortality and Health Care Survey 2016 estimated a prevalence rate of 37 per 1 000 000 women; the current study estimated a prevalence rate of 35 (95% CI 26 to 47) per 100 000 women. We tried three different methods of estimation to check the robustness of the results under different assumptions, and the results from these methods are comparable. Another problem of our study was that due to the small sample size of countries for which DHS data were available on obstetric fistula, we used only two covariates in the predictive models. The small sample size and large variability have likely contributed to wide uncertainty in our estimates. We used the WHO/MMEIG estimated MMR to predict obstetric fistula for the countries. However, MMR estimates are also predicted from model-based methods and are subject to large uncertainty. Our results are affected by the accuracy level of MMR estimates. It is likely that the uncertainty estimates of obstetric fistula shown in this study are in the low range. The DHS implemented the fistula module questions predominantly in SSA countries, with only two in Southeast Asia and one in North Africa. Disproportionate representations of the countries in the dataset may also affect our estimates. Moreover, this study focused on obstetric fistula estimation and did not capture urogenital fistulas unrelated to childbirth. Most studies on obstetric fistula prevalence are based on self-reported symptoms of urinary and faecal incontinence. Our results may seem lower than expected. Because these symptoms, especially urinary incontinence, may occur for other reasons, including postpartum incontinence, weakened pelvic floor muscles and prolapse, self-reported symptoms may be misclassified as obstetric fistula. Although not generalisable, situational analysis in southern Sudan using the key informant method with clinical examination revealed lower estimates with the prevalence of 30/100 000 women aged 15–49, validating the use of the key informant method and the benefit of clinical assessment.30 Indeed, in studies where a clinical examination was carried out, the frequency of obstetric fistula was substantially lower than from self-report estimates; in a Pakistan study, only 4.3% were confirmed obstetric fistula cases among the women who reported urinary incontinence.25 In a systematic review of obstetric fistula in low-resource countries published between 1995 and 2014, Cowgill et al also noted the wide variability of the incidence and prevalence estimates of obstetric fistula in South Asia and Africa, respectively, from 0 to 4.09 obstetric fistula cases/1000 deliveries and 0–81 cases/1000 deliveries.17 As the published estimates are unreliable and unsuitable for a useful meta-analysis, Tuncalp et al suggested a novel approach involving community outreach followed by physical examinations by nurse-midwives at lower-level facilities closer to where the women live.26 The authors reported a PPV of 47% and a 98% NPV of self-report of fistula-like symptoms based on a DHS fistula questionnaire module. However, the authors opined that these estimates of positive and negative predictive value do not represent the validity of the DHS fistula module questions, as the sample selection was based on self-selected screened patients and not a population sample of women of reproductive age. Using self-reports of constant urine leakage with clinical assessment for confirmation, Mocumbi et al report an incidence of 1.1/1000 recently pregnant women in a population-based study in a high-density health facility area in southern Mozambique.31 Except for the Bangladesh study, reliable national-level obstetric fistula estimation is currently not available. This study attempts to address the data gaps. Our study suggests that it is likely that the prevalence of obstetric fistula has substantially declined from the earlier prevalence level. The frequently cited figure of 2 million obstetric fistula cases emerged in the early 1990s. Maternal mortality declined by about 45% between 1990 and 2020 and CS has increased from 7% in 1990 to 21% by 2018 globally.32 Studies also suggest that the maternal deaths due to obstructed labour, the primary cause of obstetric fistula, have reduced from 8% in the 1990s33 to 1.9% (80% UI 1.4 to 2.9)34 by 2009–2020. Nevertheless, many may find it surprising that fewer than 500 000 cases of obstetric fistula are prevalent globally. Consider the case of India, where obstetric fistula prevalence is estimated to be the highest in the world, MMR has declined about 81.6% (560–103 between 1990 and 2020) and CS increased substantially from 2.5% in the 1990s to 8.5% in 2005–2006 to 21.5% from 2019 to 2021. Skilled birth attendance increased from 31.5% to ~90% between 1990 and 2019–2020. Our analysis suggests India has the highest estimated number of obstetric fistula cases (~150 000, almost one-third of global obstetric fistula estimates). So, many may consider the present analysis overestimating India’s obstetric fistula prevalence level. Overall, we consider that our estimates of less than 500 000 cases globally seem reasonable and may be used by country officials, planners and developmental partners as a guide for planning and developing strategies to eliminate obstetric fistula. A large number of women received surgical treatment for obstetric fistula repair through the UNFPA-led global Campaign to End Fistula and international development and humanitarian agencies, like USAID, FIGO and Fistula Foundation, during the last two decades. However, there is still an unmet need for treatment in many countries. More women now have access to emergency obstetric and neonatal care, including CS. However, there are concerns that an increasing trend in CS rates in low-resource settings with low surgical capacity may also increase the incidence of iatrogenic fistula. A retrospective review in nine sub-Saharan countries found that 26.8% of women who developed fistula after a CS were caused by surgery rather than pressure necrosis from obstructed labour,35 and a number of individual country studies in sub-Saharan Africa have similar findings.36 37 However, in South Asia, some studies have implicated other surgeries, notably hysterectomy, in contributing to a significant proportion of iatrogenic fistula.38 39 It is critical to address this challenge to end fistula by 2030. This aetiological shift from ‘poor access’ to ‘poor obstetrical and gynaecological care’ will require vigilance, accountability and regionally specific interventions directed at healthcare facilities and systems that address optimal training and equipment needs for conducting CSs and pelvic surgery. Conclusion Although the number of obstetric fistula cases is less than previously anticipated, the current prevalence of about 500 000 cases still poses a huge challenge to ensure surgical access to these women during their lifetime at the current rate of repair. Obstetric fistula is preventable and treatable: these estimates should guide the development and application of much-needed preventive and curative programmes for the eradication of this unacceptable condition.40 The global community must unite to eliminate this devastating and tragic condition by ensuring that all women have access to timely and high-quality emergency obstetric and safe surgical care through strengthened health systems. This effort must also address the social determinants that impact the well-being of women and girls, foster meaningful community engagement and build robust partnerships. Through evidence-based, country-driven, sustainable, quality maternal/newborn healthcare—including in humanitarian settings—it is possible to achieve the promise of eliminating obstetric fistula within a decade. Supplementary material