Obstetric fistulae: a study of women managed at the Monze Mission Hospital, Zambia


  • A Holme,

    Corresponding author
    1. Department of Public Health and Epidemiology, The Medical School, University of Birmingham, Edgbaston, Birmingham, UK
      A Holme, The Medical School, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK. Email arh362@bham.ac.uk
    Search for more papers by this author
  • M Breen,

    1. Maternity Unit, Monze Mission Hospital, Monze, Southern Province, Zambia
    Search for more papers by this author
  • C MacArthur

    1. Department of Public Health and Epidemiology, The Medical School, University of Birmingham, Edgbaston, Birmingham, UK
    Search for more papers by this author

A Holme, The Medical School, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK. Email arh362@bham.ac.uk


Objective  The objective of this study was to describe and compare characteristics of women with obstetric fistula.

Design  Retrospective cross-sectional study.

Setting  Zambia’s primary fistula repair centre, Monze Mission Hospital.

Sample  All women, August 2003 to December 2005.

Method  Review of case notes to obtain data on socio-demographic and obstetric characteristics, causative pregnancy, clinical details, and treatment. Comparison of characteristics with national data was undertaken.

Results  Of 259 women, 239 had socio-demographic and obstetric records and 254 had surgical records. Educational status and height of women were significantly below the national averages, while antenatal care uptake (97.5%) and proportion from the Northern Province were significantly above. Most women (77.9%) weighed ≤50 kg. Median age at marriage was 18 and at development of fistula was 22 years. 15.1% of women were divorced, 49.0% were primiparous, and 27.6% were parity four +. 67.5% of women had spent 2 days or longer in labour. Delays in receiving emergency obstetric care (EmOC) were experienced at home (67.5%) and at clinics (49.4%), usually due to transport difficulties. 89.1% delivered in a health facility, 50.2% of deliveries were by caesarean section, and 78.1% of babies were stillborn. 72.9% of repairs were successful, 17.3% resulted in residual stress incontinence, and 9.8% failed. Failure was significantly associated with previous repair.

Conclusion  More obstetric fistulae occur in areas where early marriage and pregnancy before pelvic maturity is attained is common and where obstetric care is inaccessible. In this study, age at marriage and fistula development was older than usually found, which may indicate that poor access to EmOC contributes more to this problem within Zambia.


Obstetric fistula is no longer seen in affluent nations, yet it continues to cause untold suffering among millions of women in the developing world.1 This preventable and treatable condition2 primarily occurs in sub-Saharan Africa and South Asia, with up to 130 000 new cases each year in Africa alone,3 although the true figure may be higher.1,2 The cause of obstetric fistula is unrelieved obstructed labour, whereby the entrapped fetal head impacts the soft tissues of the vagina, bladder, and rectum against the pelvis, causing widespread pressure necrosis and leading to hole formation between adjacent organs.4 This leaves the woman leaking urine and sometimes faeces uncontrollably through the vagina,5 with possible additional consequences including total urethral loss and neurological damage to the lower limbs.5

Obstetric fistula occurs in the context of some of the highest maternal mortality rates in the world. Of the 529 000 maternal deaths in 2000, 99% occurred in the developing world,6 and obstructed labour was one of the five main causes.7 For every woman who dies, 30 others suffer chronic illness or disability at the height of their productivity and family responsibilities.7,8 In Zambia, this dire picture of reproductive health occurs against a backdrop of 83% of the rural and 56% of the urban population living in poverty, high prevalence of HIV/AIDS, corruption, poor education, and average life expectancy of 37 years.9

Research on obstetric fistula has been highlighted as ‘woefully inadequate’,3 and of those studies that have been conducted,10–24 aside from a few in Ethiopia15–17 and Nigeria,14,19,20 most are small or have limited data. Research has shown that most fistulae are vesicovaginal and occur due to obstructed labour10–24 and poorly supervised delivery.11 Early intervention normally prevents fistula,1 but women often suffer for several days13,15,16 without emergency obstetric care (EmOC), as many come from rural areas with poor access to health care.13,19

Women with fistula are often small and short,10,12,13,16,19,23 an indication of pelvic immaturity or cephalopelvic disproportion (CPD). They are also usually young,11,12,14,15,17,24 illiterate,11,15 and poorly educated.14,19,23 Marriage and causative pregnancy mostly occur before the age of 20,10,12,13,16,19,23,24 although studies in Nigeria and Ethiopia report ages there as even younger (81.7% married by 15 years,24 mean ages at marriage of 14.716 and 15.5,19 and mean age at causative pregnancy of 17.816).

Given that women with fistula are often primiparous,11,14–19 babies usually stillborn,13,15,19,20,22 and secondary infertility common, the childlessness that is so devastating in a culture where a woman’s status is largely determined by her reproductive functioning,4 means often more than 50%5,16,17 are divorced by their husbands.5,10,12,13,15–17,19 Other consequences include severe social stigmatisation5 and loss of support from families and communities.10,11,17,24 Women with fistula often travel long distances to reach repair services and many live with the condition for numerous years.17 It is estimated that 80% never seek treatment,2 yet when repaired by an expert, success rates are reported as high as 87–93%.15,17,18

Only one previous study on fistula in Zambia has been published: a series of 61 women at the University Teaching Hospital, Lusaka, in 1983. It reported one-quarter of women were teenagers, 37.6% were primigravidas, and the cure rate was 59.1%.25 As recommended in the 2002 United Nations Populations Fund (UNFPA) Needs Assessment, gaining a clearer picture of obstetric fistula in Zambia would be valuable as it undoubtedly presents a problem, although may not be as prevalent as in other countries.2 Although knowing morbidity rates is important, a deeper understanding of the characteristics of women with fistula is essential for targeting prevention strategies effectively. The aim of this study was to describe the socio-demographic and obstetric characteristics and the surgical treatment of women with obstetric fistula at the Monze Mission Hospital and to compare characteristics of these women with national data.


The study was a retrospective, cross-sectional investigation. Institutional approval was granted from Monze Mission Hospital and ethical approval from the South Birmingham research ethics committee and the University of Lusaka research ethics committee. Gaining women consent for audit of medical records was not feasible, since it is impossible to contact women once discharged. Monze Mission Hospital in the Southern Province of Zambia is a general hospital, including obstetrics and gynaecology and is the country’s main fistula repair centre. Despite Zambia’s vast size, it is the only hospital to carry out regular fistula repairs and women travel from all over the country for treatment.2

The study sample was all women treated with fistula due to obstetric causes between August 2003 and December 2005. During this period, only six women were nonobstetric, caused by trauma or surgery. Information had been recorded on each woman using a standardised questionnaire. Detailed notes have been kept since UNFPA recommendations that all fistula centres should maintain rigorous record-keeping.2

Data were collected by review of the case notes of women with obstetric fistula and entered into an ACCESS database. This included the women’s age, tribe, home province, marital status, age at marriage, educational level, height, weight, obstetric history, number of sexual partners, antenatal care uptake, most recent delivery, time spent in labour, delays experienced in receiving obstetric care, age at which the fistula developed, associated conditions, consultations at other hospitals, past repairs, operative diagnosis, details of the fistula, level of scarring, and surgical outcome. Outcome of fistula repair was classified as cured, failed, or having stress incontinence. This was determined by examination, with a dye test if in doubt. The diagnosis of stress incontinence was also made if there was a report of persistent incontinence but a negative dye test.

A pilot was undertaken by entering a sample of records and making any necessary adjustments to the database design. National data were taken from the Zambia Demographic and Health Survey 2001–2002,26 a nationally representative sample survey, which includes data on 7658 women aged 15–49 years. When comparing characteristics of women with fistula against national data, only women of a comparable age (15–49 years) were included. The one woman outside this age range and those who did not give their age were excluded from comparison.

Statistical analysis was carried out using SPSS (SPSS Inc., Chicago, IL, USA) and MINITAB (Minitab Ltd, Coventry, UK). This comprised descriptive analysis using means, medians, ranges, SD, and proportions. Relationships between variables in the dataset were examined using Fisher’s exact test, chi-square test, and Spearman correlation. Comparison of the characteristics of women with fistula (aged 15–49 years) with national data was carried out using the test of one proportion, chi-square test, Wilcoxon signed rank test, and independent samples t test.


There were 259 women with obstetric fistula admitted to the hospital over the study period, of whom 239 had socio-demographic and obstetric records and 254 had surgical records available. Among these women, the proportion of missing data items ranged from 0 to 33%. For simplicity, available data are reported and the proportion of missing data only mentioned if it exceeds 10%.

One-quarter of women were 20 years or younger, and the median age was 25 years (range 15–59, SD 8.01). Twenty-five women did not have age data. Over half of women (56.5%, 135/239) came from the Northern Province and 19.2% (46/239) from the Southern Province. The proportion of women who came from the Northern Province (54.5% of women with fistula aged 15–49 years) differs significantly (P < 0.001) from the national proportion of 13%.26 The tribal affiliation of the women correlates to their geographical distribution: 52.3% were Bemba, the predominant tribe in the North, while 17.6% were Tonga, the predominate tribe in the South.27 Tribal affiliation differs significantly from national proportions (P < 0.001), where only 36% of women are Bemba26 (49.8% of women with fistula aged 15–49 years).

The educational level of the women with fistula aged 15–49 years was significantly lower than the national levels (Table 1), and the median years of schooling (4.5 years) was significantly below the national average of six.26 Women with no education were significantly less likely to give their current age (χ2= 12.7, df= 1, P < 0.001), age at marriage (χ2= 4.6, df= 1, P < 0.05), age at first pregnancy (χ2= 19.5, df= 1, P < 0.001), and age at which the fistula developed (χ2= 14.7, df= 1, P < 0.001).

Table 1.  Highest educational level achieved from the women with fistula aged 14–49 years, compared with the expected levels according to national data
 Observed educational level of women with fistula aged 15–49 years, n (%)Expected educational level of women with fistula aged 15–49 years according to national data, n (%)
  1. n= 213 women.

Educational level
No education42 (19.7)26 (12.1)
Some primary education104 (48.8)81 (38.3)
Completed primary education43 (20.2)42 (19.7)
Some secondary education21 (9.9)49 (23.1)
Completed secondary education3 (1.4)8 (3.8)
Higher education0 (0.0)7 (3.1)

The women with fistula were generally small and short, with their median height being 148.0 cm (SD 8.36, range 115–163 cm). The mean height of women with fistula aged 15–49 years (146.5 cm) was significantly below the national female average of 158 cm26 (t208= 19.818, P < 0.001). The median weight of women was 46.0 kg (SD 7.77, range 25–92 kg) and 77.9% weighed ≤50 kg. The median body mass index (BMI) was 21.2 (SD 3.53), and 41.9% (100/239) of women had a BMI of 20 kg/m2 or below.

Three-quarters of women with fistula were married, 15.1% divorced, 7.5% single, and 1.7% widowed. Of the 45 women who were no longer living with their husbands, 31 (67%) stated that this was due to their fistula. For the 154 women who gave their age at marriage, 81.8% married at 16 years or older, and the median age was 18 (SD 2.91). Although questions on economic status were not asked, 64.0% (153/239) of women received help in paying travel costs to Monze Hospital, suggesting that many came from impoverished backgrounds.

Obstetric characteristics are shown in Table 2. Of the 184 women who could give their age at menarche, the median age was 15 years (SD 1.56). There were 91.2% of women (196/215) who had never used birth control, and 88.9% (208/234) had only one sexual partner (range 1–3). Almost half of women were primiparous, although parity ranged up to 12. Of the women, 24.3% could not state their age of first pregnancy, but of those with data, the median age was 18 (SD 2.73). Of the total of 661 babies born to these women, only 305 (46.1%) were still alive. There were 55.6% (133/239) of women who were childless, 83.3% (199/239) who had suffered at least one stillbirth (range 0–5), and 31.0% (74/239) who had at least one dead child (range 0–4).

Table 2.  Obstetric characteristics of the women with fistula
Obstetric characteristicsn (%)
  1. n= 239 women.

1117 (49.0)
2–356 (23.4)
4+66 (27.6)
Age at first pregnancy (years)
14 or younger6 (3.3)
15–1629 (16.0)
17–1859 (32.6)
19–2052 (28.7)
21–2223 (12.7)
23–247 (3.9)
25 or older5 (2.8)
Age at first pregnancy not given58
Antenatal care in most recent pregnancy
Did attend233 (97.5)
Did not attend6 (2.5)
Place of most recent delivery
Home23 (9.6)
Clinic30 (12.6)
Hospital183 (76.6)
Other3 (1.3)
Mode of most recent delivery
Caesarean section119 (50.2)
Spontaneous vaginal93 (39.2)
Vacuum extraction/forceps/destruction25 (10.5)
Mode of delivery not given2
Outcome of most recent delivery
Stillbirth185 (78.1)
Live birth52 (21.9)
Outcome of birth not given2
Duration of labour
Less than 1 day10 (4.3)
1–3 days197 (84.5)
Over 3 days26 (11.2)
Duration of labour not given6

Antenatal care attendance during the most recent pregnancy was significantly associated with having some education (Fisher’s exact test, P < 0.05). For women with fistula aged 15–49 years, antenatal care uptake (98.1%) was significantly higher than the national figure of 93.4%26 (P < 0.01). Of the 153 women with data on the number of antenatal visits, 76.5% (117/153) attended four or more (range 2–10).

National data suggest that 44% of women in Zambia had their most recent delivery at a health facility and 56% at home.26 In contrast, nine of ten women with fistula had their most recent delivery at a health facility; although it is likely most began their labour at home with transfer required because of their obstructed labour. Delivery in a health facility was significantly associated with the women having some education (χ2= 19.1, df= 1, P < 0.001). The proportion of women attended by a doctor during delivery was similarly very high: 65.9% (153/232) compared with 3.2% nationally. For half of the women, their most recent delivery was by caesarean section, and since only one in five babies were born alive, a substantial proportion of these caesarean sections probably occurred after the fetus had died.

Many women experienced prolonged obstructed labour, with two-thirds spending 2 days or more in labour and some up to 7 days. In the most recent delivery, the most widely experienced delay in accessing EmOC was at home, with 67.5% (158/234) of women delayed there. Of those women, 84.4% (130/154) spent at least 1 day in labour at home and 52.6% (86/154) spent 2 days or more. The most common reason for this was transport difficulties, either because no transport was available (48/144, 33.3%) or because there was no money for transport (39/144, 27.1%). Of those women who stated that they experienced a delay at home, 31.3% (45/144) said it was because they wanted to have a home delivery. The next most common delay was in the clinics, with almost half (116/235) of women delayed there. There were 54.9% (62/113) of women who spent a day or more in labour at a clinic and 26.5% (30/113) spent 2 days or more. Again, the most common reason for the delay was transport difficulties (56/120, 46.7%). In 26.7% (32/120) of cases, the women were not aware of why the delay occurred, and in 19.2% (23/120) of cases, the delay was in referral due to staff stating that the women would deliver normally at the clinic. The least common delay occurred after reaching a hospital, with only 43 (18.3%) women experiencing a delay there. However, of those women, 11 still spent a day or more in labour at hospital. Most women (29/43) were not told or were not aware of why the delay occurred.

Of the 210 women who stated their age at the time of fistula development, the median was 22 (SD 7.17). However, the distribution of these ages (Figure 1) shows two peaks: a younger primiparous population (mean 19.2, SD 2.6) and an older multiparous population (mean 29.3, SD 7.1). Of the women, 39.9% (93/233) presented to Monze Hospital within a year of developing a fistula, but for 23.2% (54/233), it was 4 or more years, and four had fistula for over 20 years. On initial examination, 90.9% (209/230) of women had a vesicovaginal fistula (VVF).

Figure 1.

Bar chart showing the age at which the women developed obstetric fistula (n= 210 women).

Nerve damage in obstructed labour meant 43.5% (103/237) of women experienced difficulty walking since the delivery. Only 54.0% (128/237) of women had menstrual periods. There were 16.9% (40/237) of women who no longer menstruated, and in 29.1% (69/237), it was too soon after delivery to tell if menstruation had been affected. Almost half (117/237) of the women had been seen in other hospitals before reaching Monze. Of these, 53.2% (59/111) had been referred, 28.8% (32/111) had experienced attempted surgical repair, 8.1% (9/111) had only been examined or treated nonsurgically, and in 9.9% (11/111), nothing had been performed. In total, 93% (223/239) of women were recommended to come to Monze Hospital by a health worker.

The surgical dataset included details on 277 operations; 232 of the 254 women had one operation, 20 had two, and two had three operations. However, these were not necessarily the women’s first, second, and third repair overall. Results are reported in terms of the proportion of operations and not patients.

In 72.6% (199/274) of operations, it was the first repair attempt anywhere, 19.3% (53/274) were the second, and 8.1% (22/274) were the third or more (up to five) attempts to repair. However, 91.6% (251/274) of operations were the first repair attempt at Monze Mission Hospital, and 8.4% (23/274) were the second or third attempt. The operative diagnoses for all 277 operations were 247 VVF repairs (18 included additional rectovaginal fistula (RVF) repairs and five additional ureteric fistula repairs), four RVF repairs, and 26 ureteric fistula repairs. Of the operations, 10.5% (26/248) involved multiple fistulae, and 85.9% (238/277) were carried out under spinal anaesthesia alone.

The most frequent VVF site was juxtaurethral (82/259, 31.7%), commonly associated with obstructed labour.25 In women with a VVF, the level of scarring was slight in 66.8% (169/253) of cases and moderate or severe in 26.1% (66/253). Women with multiple previous repairs were significantly more likely to have a worse level of scarring (Spearman correlation, r= 0.211, P= 0.001). In 56.7% (144/254) of VVF women, the urethra was partially damaged, and in 1.6% (4/254), it was completely destroyed.

Of all the operations, 72.9% (186/255) were successful, 17.3% (44/255) resulted in residual stress incontinence, and 9.8% (25/255) failed. The failure rate of surgery increased three-fold if the women had experienced one or more previous repair attempts (Table 3). The outcome of surgery was significantly associated with the level of scarring, with those women with less scarring more likely to have a successful surgical outcome (Spearman correlation, r= 0.412, P < 0.001).

Table 3.  Cross-tabulation of outcome of surgery by repair attempt
  Outcome of surgery by repair attempt
CuredFailedStress incontinence
CountRow (%)CountRow (%)CountRow (%)
  1. n= 252 operations.

Overall repair attemptFirst overall repair13975.5126.53317.9
Women has had one or more previous repair4464.71319.11116.2
Total 18372.6259.94417.5


This study has shown that, in keeping with previous literature, most women with fistula were shorter than the national average. A small, short physique is a predisposing factor to CPD, which increases the risk of obstructed labour.28 The low weight of many of the women may also be the result of neglect after development of the fistula or weight loss after obstructed labour, caesarean section, or sepsis. Most women in this study experienced prolonged obstructed labour of at least 2 days, where they were delayed in accessing EmOC due to transport difficulties and where the delivery resulted in a stillbirth. Although a high proportion ended up delivering at a health facility with skilled attendance, most began their labour at home, only reaching a health facility after the ischaemic damage of prolonged obstructed labour had occurred. Although half of the women were primiparous, there were also numerous grand multiparae, an at-risk population that should not be forgotten when addressing obstetric fistula. In keeping with previous reports, the educational status of the women was significantly lower than the national averages. The median age at marriage of 18 years suggests that it was not that women with fistula gave up their education to get married, so poverty may be the linking factor. Both education and access to health care cost money, and poor families may be unable to afford either.

Significantly more women with fistula than expected came from the Northern Province of Zambia, despite Monze Mission Hospital being in the South. This may be because the Northern Province is a large, sparsely populated rural area, with few hospitals, limited transportation, difficult communication, and thus poorer access to health care than other regions.29 However, the specific arrangement whereby women with fistula are transported from health facilities in the North down to Monze Mission Hospital for repair may also contribute to this occurrence.

Antenatal care uptake among the women with fistula was significantly higher than that in national data, and the average number of visits was within international standards.30 This questions the effectiveness of such antenatal care since it is possible to assess for the risk of obstructed labour, and if appropriate, refers to a health facility for delivery, especially if the woman lives in a remote area. Education of women on when to seek help in labour is good practice, but it is not known the extent to which this occurs in antenatal care within this area.

The small number of women at Monze who were divorced differs to previous literature, which often shows that women with fistula are rejected by their husbands. The large proportion still married in this study is difficult to explain, but may be due to differences in the ease or acceptability of divorce in Zambia, compared with other countries. Furthermore, women may have been too ashamed to say they were divorced. One possibility is that in communities that practice polygamy, the husband may take another wife without having to divorce an existing wife with fistula.

The age at marriage and at first pregnancy in this study is generally older than that in previous reports. The age at which the fistula developed (median 22 years) also differs from previous literature, which usually shows the injury is sustained at a younger age. Most high-quality research on fistula comes from Nigeria and Ethiopia, both countries with high occurrence of arranged marriage of young girls often as soon as or even before menarche.16,24 Data suggest that in Nigeria 20% of girls are married by age 15, rising to 48% in some regions,31 while in Ethiopia, 19% of girls are married by age 15, rising to 50% in certain areas.32 Obviously, earlier marriage means exposure to pregnancy at an age when mature pelvic capacity is not yet attained,19 with greater risk of complications, such as CPD, predisposing to obstructed labour, and fistula. In contrast, in Zambia, only 8% of girls are married by age 15, and the average age of marriage among the women with fistula was similar to the national average of approximately 18 years.26 Since women in Zambia are generally older at marriage and at first pregnancy, they are at reduced risk of such complications, thus possibly explaining why the UNFPA Needs Assessment estimated the occurrence of fistula might be lower than in other developing countries.2 However, the maternal mortality ratio (MMR) in Zambia is extremely high; 750 per 100 000 live births33 (as opposed to 20 per 100 000 in developed countries6), with rural Zambia having one of the highest MMR’s in Africa.29,34 Therefore, it may be that women suffering obstructed labour are dying, rather than surviving and presenting with fistula, or there may be huge number of women with fistula, but they cannot access treatment.2

The success of fistula repair at Monze was high. However, there are limitations in defining the success of surgical outcome on whether women are dry, have residual stress incontinence, or are still leaking. These physical measures do not take into account the complex nature of the injury, and the holistic approach needed for complete rehabilitation. Difficulties in follow up also mean that it is hard to ascertain women’s long-term outcomes, in terms of the surgical repair, associated conditions such as foot drop and secondary infertility, and whether there has been successful re-integration into the community and psychological recovery.

Given the impact of fistula on the women’s lives, treatment and prevention are critical public health issues.35 Prevention strategies must focus on factors predisposing to prolonged obstructed labour, including access to obstetric care, early childbearing, malnutrition, and female genital mutilation.36 Indirect risk factors include poor education, poverty, lack of antenatal care or skilled birth attendance, and low status of women, which may limit access to services that could prevent the onset of such conditions.2 This study suggests that in dealing with obstetric fistula in Zambia, it is essential to address the transport problem and ensure all pregnant women have timely access to EmOC. Specifically, clinics must recognise the importance of swift referral of women in obstructed labour, referral systems and transportation to hospital for emergencies must be supported, and healthcare providers must be fully educated in managing prolonged obstructed labour. Additionally, it is important to improve community health education, raise awareness of labour complications, and support campaigns encouraging women to give birth at health facilities. The provision of maternity waiting homes, where women from remote areas can await delivery, avoids the problem of ‘obstructed transport’ should complications arise. Increasing the educational status of women is also essential, for all the interrelated benefits this can bring.

Since complete prevention is a distant goal,1 effective repair centres must be made accessible and affordable5 and training programmes for fistula repair surgeons increased. The length of time some women lived with fistula before treatment also demonstrates the need to raise awareness about repair services. In this study, about half of women had consulted other hospitals before Monze, and many experienced failed repair attempts or incorrect treatments. This reveals the need to inform all maternity care providers on managing women with obstetric fistula. For example, women can be catheterised for 4–5 weeks as small fistulas may heal spontaneously this way,36 or if this is unsuccessful, referred to a designated fistula repair centre. The fact that first repairs had a lower failure rate emphasises the importance of sufficient expertise, since multiple operations increase the risk of failure in subsequent surgery due to increased scar tissue.

The only previous study on fistula in Zambia was a small surgical review from over 20 years ago. The larger present study investigates the country’s primary fistula repair centre and includes detailed social and clinical data, fulfilling the UNFPA recommendation of conducting research to gain a clearer picture of fistula in Zambia. However, there are some limitations. Like most fistula research, the study was institution based, therefore, only accessed and can only be generalised to women who have reached treatment, since those who have not reached treatment may differ in certain ways. Missing datasheets depleted the sample only slightly but may be a source of bias. Also, women with missing data items may differ from those with complete data, for instance, those with no education were significantly less likely to have data on their current age, age at marriage, age at first pregnancy, and age at which the fistula developed. This creates bias because those with no education may be more likely to have married and given birth at a younger age. Furthermore, answers involving numerical quantities, such as times, dates, or distances, were sometimes incomplete. A nominated senior nurse had collected about 90% of the social data, and the fistula surgeon completed all surgical datasheets. Data may therefore be subject to interviewer bias as the women may have worried that their answers could affect their care. It is also likely that sensitive questions, such as the number of sexual partners, may be subject to social desirability bias. Furthermore, since there are over 70 different languages in Zambia, data from women who spoke less well-known languages may have suffered interpreter bias or been missing. However, 70% were Bemba or Tonga and the nurse collecting social data spoke both languages.

Future research to obtain accurate prevalence estimates for fistula in Zambia would require population-based studies. It would also be beneficial to understand the long-term outcomes of women with fistula after they are discharged, although the logistics of such follow-up research presents a challenge.


More obstetric fistulae occur in areas where early marriage and pregnancy before pelvic maturity has been attained are common and where obstetric care is inaccessible. Among women in this study, the age at marriage and age at fistula development were older than those found in previous studies, which may indicate that poor access to EmOC has a greater contribution to the problem of obstetric fistula within Zambia than early childbearing.


We are grateful to Dr John Kelly for his excellent advice on this project. Our thanks also go to the Monze Mission Hospital for enabling the study to go ahead, Dr Roger Holder for assistance with statistics, and to the Arthur Thompson Trust, the Devon Educational Trust, and the University of Birmingham Intercalated BMedSc grant for help with funding.