|
by Tracy E. Slanger , Rachel C. Snow , Friday E. Okonofua The focus of attention in recent decades on female genital cutting as a public health concern has been more political than scientific, with advocacy literature greatly outpacing scientific research (WHO 1998; Obermeyer 1999). This imbalance is particularly true in the case of the more moderate forms of cutting, which account for an estimated 80 percent of the practice worldwide (WHO 2000b). Although case reports, speculation, and anecdotal reports describe extreme morbidity, and in some cases mortality, resulting from short or long-term complications of cutting, these data tend to be hospital-based, they generally lack comparison with a control group, or they come from regions where the most extreme form of cutting, infibulation, is practiced (Obermeyer 1999). Although these reports may be helpful in understanding the potential clinical consequences of the practice, and, where described, how these complications were dealt with, from a public health perspective, these data fail to provide population-relevant m easurements of risk. The dearth of research on types I and II (clitorectomy and excision, respectively) has led to a biased perception of the public health consequences of the practice--that regardless of type, female genital cutting invariably leads to severe morbidity. Too much of review and advocacy literature is based on generalizations but presented as empirical fact. In its widely distributed literature, the advocacy group INTACT, for example, states that, "many girls die shortly after the operation...." (INTACT 2000:2). Popular media and bestselling books, such as Alice Walker's Possessing the Secret of Joy (1992) or Waris Dine's Desert Flower (1998) present the public with a portrait of genital cutting skewed toward the most extreme, and least prevalent, forms of the practice. Even peer-reviewed articles are not exempt from these generalizations and often discuss the obstetric health dangers of cutting without specifying by type. A review published in the International Journal of Gynecology and Obstetrics reports that "mu tilated women are reported to be at greater risk of hemorrhage in childbirth because of obstructed labor and tearing of vaginal and perineal scar tissue" (Kun 1997: 153). A second article from the same journal emphasizes that "women who have undergone FGM [female genital mutilation] suffer from its complications for life ... especially ... during pregnancy, childbirth and the postpartum period" (Rushwan 2000: 104). Although such information is sometimes qualified with the statement that health consequences vary by type and severity, generally the milder forms of genital cutting are not considered in such rhetoric. An epidemiological approach to research on the practice is necessary if we are to understand more fully the contribution of genital cutting to women's sexual and reproductive health problems in the context of what may already be high rates of morbidity and mortality resulting from other causes (Snow 2001). Developing countries suffer the highest rates of maternal mortality worldwide, and for every death, numerous injuries occur (Liskin 1992; Prual et al. 1998; Donnay 2000). The primary causes of maternal mortality include postpartum hemorrhage and obstructed labor (Selo-Ojeme and Okonofua 1997). In populations where women experience multiple child-births, as is the case for those countries where female genital cutting has been documented, the health implications of the practice, particularly as they relate to obstetric morbidity, are potentially large and warrant thorough documentation (Snow 2001). Early epidemiological studies examining the effects of genital cutting on obstetric health have focused on the most severe forms of cutting (Berardi et al. 1985; De Silva 1989). Where more moderate types of cutting have been studied, findings were not stratified by type, and social correlates were not considered in the analysis (Adinma 1997). The most recent efforts to measure the association between cutting and obstetric morbidity using an epidemiological design have included women who underwent moderate forms of cutting. Findings have been mixed, however, with data from Mali showing no association between cutting status and delivery complications among women presenting for delivery and research conducted in Burkina Faso showing a positive association between cutting status and reported complications at any delivery (Jones et al. 1999). Although both studies include several social confounders of genital cutting and obstetric morbidity in their analyses, such as age, education, religion, marital status, and r esidence, the Burkina Faso analysis did not include data on delivery conditions, such as place of delivery (that is, home, hospital, or elsewhere) or the type of assistant who attended the birth. Moreover, neither study distinguished between delivery "complications" and delivery "procedures." The present study shows that delivery conditions, such as place or assistant, potentially can confound associations between genital cutting and obstetric morbidity, and secondly, that this confounding is strongly related to the specific complication or procedure involved. The association between delivery complications and procedures and female genital cutting is examined here in the context of social and demographic factors in southwest Nigeria.
Background Female genital cutting, often referred to as female circumcision, female genital mutilation, or female genital surgeries, has been described as all procedures involving partial or total removal of the external female genitalia and/or other injury to the female genital organs for cultural or other nontherapeutic reasons (WHO 1997). It has been classified by the World Heath Organization, based on its severity, into four major types, (1) from removal of the clitoral foreskin and/or clitoris (type I) to complete excision of the external genitalia and stitching together of the vulva (infibulation) (type III). Although it has been documented historically on all continents, today the practice appears limited to 28 African countries forming a continuous belt across the northern sub-Saharan region, from Sudan to Senegal and along the Nile valley from Egypt down to East Africa (Toubia 1998). It has also been reported in pockets of Malaysia, Saudi Arabia, and Yemen (Isa et al. 1999). Estimates of the number of girls and women currently living with all types of genital cutting range from 100-130 million, with two million girls and women at risk each year (Toubia 1998). Depending on where it is practiced, the procedure is justified on the basis of tradition, religion, rite of passage, health, assurance of virginity before marriage, and marital fidelity. Data on location and practitioners indicate that the surgery is performed most often outside a clinical setting, (2) and therefore, is subject to nonsterile conditions. The artery and dense nerve endings of the clitoris make any surgery, even type I cutting, a painful procedure, carrying a risk of excessive bleeding and infection (Fox et al. 1997). Study Setting In 1999, Nigeria, the most populous country in sub-Saharan Africa, had an estimated total population of 108,945,000 (WHO 2000a). In the southwestern region of the country, the 1999 total fertility rate was 4.5 children per woman of reproductive age. The 2001 estimated maternal mortality rate in the southwest ranged from 165 deaths per 100,000 live births to 1,549 deaths in the northeast (World Bank 2001). The self-reported prevalence of female genital cutting is 25 percent for the whole country and ranges from 2 percent in the northeast to 48 percent in the southwest (National Population Commission 2000). Although infibulation represents half of all cases of cutting in the northeast region, types I (84 percent) and II (5 percent) make up nearly 90 percent of cutting in southwest Nigeria (National Population Commission 2000). This study represents a collaboration between the Department of Tropical Hygiene and Public Health at the University of Heidelberg, the nongovernmental organization Women's Health and Action Research Centre (WHARC) in Benin City, and the University of Heidelberg's Women's Clinic. Benin City (estimated 1991 population: 202,800) is the largest city in Edo State (1991 population: 2.15 million). The three participating hospitals included University of Benin Teaching Hospital (UBTH), Central Hospital, and Specialist Teaching Hospital. Both UBTH and Central are located in Benin City; Specialist is located in Irrua, a large Edo State town about 70 kilometers (44 miles) north of Benin City. UBTH and Central serve a relatively urban population, whereas Specialist has a periurban clientele. As large referral hospitals, all three also draw clients from the adjoining rural communities. Methods The data presented in this report were gathered as part of a larger study examining the association between female genital cutting and a number of aspects of both reproductive and sexual health. Therefore, study participants had to be recruited from a population for which current sexual activity was established. For this reason, users of family planning and antenatal services were targeted, providing a sample of young, reproductively active women. This facility-based study had a cross-sectional design and relied on women's self-reporting of obstetric experiences at first delivery. A pretest of the survey questionnaire and extensive consultations with both interviewers and clinicians concerning language and content formed the basis for the final survey tool. Between August 1998 and March 1999, a total of 1,861 women aged 15 to 49 were invited to participate in the study while attending family planning or antenatal services at one of the three Edo State hospitals mentioned above. Women were informed of the study and their consent was obtained. The participation rate was 99.5 percent, with 0.5 percent (n = 10) declining to take part. The survey was administered by a nurse or nurse/midwife in a private office. The questionnaire covered a wide range of issues, including obstetric and gynecologic history, sexual practices, and personal experience of genital cutting and attitudes toward the practice. The interview was followed by a clinical exam to determine the presence and extent of genital cutting. A total of 1,709 questionnaires with information about respondent's age remained in the data set after cleaning of the data. Early in the analysis, age was found to be an important determinant of genital cutting status. Therefore, only those women were selected who gave this information (either "age now" or "year born"), reducing the data set from 1,861 to 1,717. An additional eight women were dropped from the sample because of large discrepancies in their responses between "year born" and "age now" and because no further information was available for ascertaining their ages, leaving a data set of 1,709. The analysis presented in this report is based only on information from those women who reported ever having had a live (3) delivery. Of the 1,709 women in the study sample, 65 percent reported at least one delivery, leaving a subset of 1,107 women. Because of the ethical considerations involved in following a cohort of girls and women into genital cutting ceremonies, and the often lengthy period between circumcision and first delivery, prospective studies of the obstetric health consequences of the practice would have been difficult to conduct. The only practical options were either a cross-sectional study design, based on the hospitals' records or women's recall of delivery experience, or a case-control or cohort observation at time of delivery. The latter choices might have been preferable had the sole objective been to look at delivery experience. A cross-sectional design targeting antenatal and family planning clients was deemed the most appropriate to obtain the range of information sought, in light of the finding that the proportion of women attending antenatal services in southwest Nigeria is higher than the proportion of women delivering at a health-care facility (89 percent and 67 percent, respectively) (National Population Commission 2000). In deed, the importance to the study of securing data from women who were undergoing a routine vaginal examination argued for this design, whereby the trade-off of selection bias was acceptable. Outcome Variables Sixty-four percent of women who had experienced genital cutting reported knowing their age at circumcision. The mean age reported was 4.8 years (range: 0-27). Although 95 percent of women reported they had undergone cutting in the calendar year that preceded the calendar year of their first delivery, in most cases by many years, 19 women (5 percent) reported that cutting and first delivery occurred in the same year. Reports describing the timing of circumcision in this region indicate that for those women cut in the same year as their first delivery, the surgery typically occurs either prior to, or during the first pregnancy, but not during or immediately following delivery (Myers et al. 1985; Caldwell et al. 1997; Babatunde 1998). The analysis of delivery experience is limited to direct complications of delivery, ensuing morbidity in the immediate postpartum period, and clinical procedures at delivery, and does not include fetal distress or complications that occurred during pregnancy and ended before delivery. The clinical procedures include episiotomy, cesarean section, and use of instruments. Episiotomy, depending on its indications, is generally a preventive procedure to treat obstruction or to avoid perineal tearing. In some health-care facilities, however, the procedure may be routine, even when not indicated. A poll of Nigerian midwives attending courses in England noted that health professionals performed episiotomies routinely on primigravidas to prevent third-degree perineal tears, and midwives reported that some episiotomies were performed to allow midwifery and medical students an opportunity to practice the procedure (Maduma-Butshe et al. 1998). At one teaching hospital included in this study, the knowledge that episiotomi es were performed routinely for first deliveries was widespread (Otoide et al. 2000). Study participants were asked, pregnancy by pregnancy, if they had experienced any of five complications: strong and regular pains lasting longer than 24 hours (obstructed labor); a large tear (perineal tear); convulsion or seizure; fever; or so much bleeding that they feared they would die (hemorrhage). (4) Complications of first delivery were based on women's self-reports, and because only their first delivery experiences were analyzed, in some cases women were reporting on events that had occurred more than 30 years before the interview. Indeed, the two studies published to date on recall of obstetric complications found that reliance on women's self-reports may lead both to over-and underreporting of the incidence of complications, such as heavy bleeding or dysfunctional labor (Stewart and Festin 1995; Ronsmans et al. 1997). In order for recall bias to have had an impact upon the major findings presented here, however, systematic deviation from the true delivery experience would have had to occur, based o n a woman's genital cutting status. No basis is found upon which to speculate such a nonrandom deviation in recall bias. Because all women were asked the same questions in the same manner, no reason is found to believe that women who had experienced genital cutting would have responded any differently to the questions on delivery complications from those who had not. Demographic information and data on place of each delivery and on who attended each birth were collected as well. Statistical Methods Statistical analyses were conducted using SAS software (version 6.0). Univariate statistical associations were measured using student's t-tests, chi-square, and Mantel-Haenszel chi-square tests. Logistic regression models were used to measure both univariate and multivariate associations and to calculate odds ratios, 95 percent confidence intervals (CI), and p-values. The change in chi-square for covariate between logistic regression models measured the relative contribution of an individual variable to a multiple regression model. Findings A demographic breakdown of the 1,107 women in our sample (see Table 1) shows that 57 percent of women were younger than 35, with a mean age of 33.7 years. Ethnic representation was highest among Bini (33 percent), followed by Esan (23 percent). The population was predominantly Pentecostal, followed by Catholic and Protestant (36 percent, 30 percent, and 16 percent, respectively), with Muslim women making up only 6 percent of the sample. Educational levels were high, with 37 percent of women reporting some post-secondary education, another 31 percent some secondary education, 26 percent some primary education, and only 6 percent reporting no formal education. Slightly more than 56 percent of the sample had undergone genital cutting, as determined by clinical examination. The breakdown by type showed that of those, 72 percent had undergone type I cutting, 24 percent type II, and 4 percent type III or IV. With the exception of age at menarche, women who had been cut and those who had not differed significantly in characteristics relating to delivery and sociodemographics. Women in the former category were nearly two years older at the time of interview than those in the latter, and also were one year younger at their first marriage. Women who had been cut were one year younger than those who had not at the time of their first delivery, and this delivery took place nearly three calendar years earlier for them than did that for women who had not been cut (mid-1986 versus 1989) (all chi-square test p-values <0.0001). Women who had been cut were predominantly Bini, more likely to be Pentecostal, and more likely to have received only a primary education (women with type II+ cutting were more likely to have received a secondary education); those who had not tended to be Esan, Catholic, and to have had some post-secondary education. A comparison of sociodemographic characteristics by type of cutting (type I versus type II+) revealed that women with type II+ cutting were more frequently represented among those with higher education (Mantel-Haenszel chi-square p = 0.001). In addition, type of cutting was associated with ethnicity; Bini and Esan women were more likely than Igbo, Urhobo, and Yoruba women to have undergone only type I cutting. Neither age nor religion was associated with the severity of genital cutting. Delivery Complications and Procedures by Cutting Status As shown in Table 2, slightly fewer than half of all women (48 percent, n = 535) reported a procedure or complication at first live delivery. (5) Of the 1,107 women in the study sample, 31 percent reported episiotomy and 7 percent reported cesarean section, while 4 percent reported a perineal tear, 3 percent reported obstructed labor, and 2 percent reported hemorrhage. (6) Hemorrhage, episiotomy, and cesarean section were all crudely associated with cutting status. Women who had been cut were only 57 percent as likely to report a cesarean section, only 60 percent as likely to report an episiotomy, and nearly three times more likely than were women who had not been cut to report hemorrhage at first delivery. Although the frequency distributions for obstructed labor and perineal tear also suggest a possible association with cutting status, the odds ratios are not statistically significant. Crude Predictors of First-delivery Complications and Procedures After measuring their crude association with cutting status, (7) each of the complications and procedures examined above was tested for potential confounding with a variety of sociodemographic data (age at first marriage, age at first delivery, year of first delivery, ethnicity, educational level, and religion), and with delivery-setting variables (place of delivery and assistant at delivery). Educational level was inversely related to reporting of overall delivery complications, that is, combined reports of hemorrhage, obstructed labor, and tearing were more frequent among those with less education (Mantel-Haenszel chi-square p = 0.001) (data not shown). Hence, the data would not suggest differential recall by cutting status based on educational level. As Table 3 makes clear, the frequency distribution of place of first delivery by first-delivery complications and procedures indicates a strong association. As expected, deliveries in which surgery or instruments were involved occurred exclusively at institutions (8) (with the exception of two episiotomies reported at home and at a traditional healer's or traditional birth attendant's workplace). Conversely, hemorrhage, obstructed labor, and perineal tears occurred proportionately more often in noninstitutional locations. Like delivery place, delivery assistant was also associated with first-delivery experiences; episiotomies occurred more often with a professional assistant (9) than with a nonprofessional assistant, cesarean section occurred solely with a professional, and hemorrhage, obstructed labor, and perineal tear were more likely to occur where a nonprofessional was assisting with the delivery. (10) As shown in Tables 4 and 5, nearly all complications and procedures were associated with at least one of the following sociodemographic variables: Age at first marriage, age at first delivery, year of first delivery, ethnicity, educational level, and religion. Cesarean section was associated with educational level, ethnicity, delivery year, age at first marriage, age at first delivery, and religion; episiotomy with educational level, ethnicity, and religion; obstructed labor with age at first marriage, and educational level. Hemorrhage was associated only with religion. Perineal tear was not associated with any of the six sociodemographic characteristics. Delivery Place, Birth Assistant, and FGC In addition to age, ethnicity, religion, educational level, age at first marriage, age at first delivery, and year of first delivery, cutting status was also associated with delivery place and delivery assistant. Although the overall majority of deliveries (59 percent) occurred at private, teaching, or referral hospitals, and only 15 percent took place in a noninstitutional setting, women who experienced genital cutting were more than two times more likely than women who did not to report that their first delivery took place at a traditional healer's or a traditional birth attendant's workplace (5 percent versus 2 percent) and two times more likely to report that it occurred at home (14 percent versus 7 percent). Both groups of women were most often assisted in their first delivery by a health professional (85 percent), and although 38 percent of the entire sample reported that a midwife assisted at their first delivery, women who had been cut were one-third more likely than those who had not to have been ass isted by a midwife (43 percent versus 31 percent). In addition, women who had been cut were more than two times more likely than those who had not to report that a nonprofessional assisted at their first delivery (21 percent versus 9 percent; OR 2.51; 95 percent CI 1.74-3.61). (11) In summary, women who had undergone cutting were proportionately more likely than those who had not to report that their first delivery took place in and was attended by less sophisticated services and assistants. First-delivery Complications and Procedures A multivariate logistic regression model was used to examine the relative contribution of genital cutting to first-delivery complications and procedures, as well as to assess which sociodemographic variables predict most strongly each first-delivery complication or procedure, while simultaneously controlling for the effects of other variables identified in the preceding analysis. Hemorrhage In the multivariate model, the variables genital cutting, delivery place, and delivery assistant all lost their statistically significant association with hemorrhage at first delivery. Once all relevant sociodemographic correlates were taken into account, only religion remained significantly associated with the occurrence of hemorrhage (see Table 4). Obstructed Labor As with hemorrhage, in the multivariate model, obstructed labor lost its association with both delivery place and delivery assistant. Age at first marriage also lost any statistical association, as did educational level. Genital cutting bore no association to obstructed labor, and no other variable in the model retained an association with this outcome. Perineal Tear In the multivariate analysis, the strongest determinant of perineal tear at first delivery was place of delivery. A woman delivering at home was nearly three times more likely than a woman delivering in any other setting to report tearing. Genital cutting was not significantly associated with perineal tear. Cesarean Section Among delivery procedures, the strongest predictor of cesarean section in the multivariate model was delivery assistant (change in chi-square for covariates p-value <0.0001). In addition to delivery assistant, delivery year before 1990 also retained an independent association with cesarean section, as shown in Table 5. Genital cutting status, delivery place, educational level, ethnicity, age at first marriage, age at first delivery, and religion had no significant association with cesarean section when other social variables were controlled. Episiotomy After all previously established predictors of episiotomy were controlled simultaneously, a woman's educational level was the strongest predictor of whether or not she reported an episiotomy at first delivery (change in chisquare for covariates p-value <0.0001). Whether or not a woman delivered at an institution was the next strongest predictor. A woman delivering at an institution was 13 times more likely to report an episiotomy compared with a woman who delivered elsewhere. Genital cutting status, delivery assistant, and religion all remained significantly associated with episiotomy, even when controlling for other factors. The association with cutting went in an unexpected direction, however, and results showed that a woman who had undergone circumcision was less likely (only 69 percent as likely) to report episiotomy at first delivery than was a woman who had not. A Closer Look at Genital Cutting and Episiotomy The highly significant negative association between genital cutting and episiotomy, remaining even after other variables were controlled, warrants closer examination. Although few cases of noninstitutional episiotomies were found, the number of reported cases performed at institutions was large enough to allow for stratification among different institutional settings: private hospital, teaching! referral hospital, state or district hospital, and a health center/primary health center. The stratified analysis revealed that only at the private hospital level did a statistically significant association exist between reported episiotomy and genital cutting status (data not shown). At a private hospital, a woman who had been cut was less likely (only 45 percent as likely) than a woman who had not to report an episiotomy at first delivery. In a logistic regression model consisting only of women delivering at a private hospital and adjusting for assistant at delivery, (12) the type of assistant remained highly predictive of whether or not a woman reported episiotomy. A woman was nearly two times more likely to report an episiotomy if she had been assisted by a midwife (p = 0.008) than if she had been assisted by a private doctor. Genital cutting status remained the most predictive variable, however, and retained its independent effect on episiotomy in the multivariate model. At a private hospital, a woman who had been cut was only 39 percent as likely as one who had not to report an episiotomy at first delivery, even when controlling for delivery assistant, as shown in Table 6. Discussion Data from the current study reveal four important findings: First, in this population, where type I and II are the predominant forms of female genital cutting (72 percent and 24 percent, respectively), women who have been cut are no more likely than those who have not to report either complications or surgical procedures at first delivery when known social correlates and delivery conditions are taken into account. Second, women who have been cut are proportionately more likely to deliver in settings where delivery complications are more common, whereas those who have not are more likely to deliver in settings where surgical procedures are more common. Third, in multivariate analyses, delivery place and assistant are the strongest predictors of perineal tear, cesarean section, and episiotomy, and these predictors eclipse any contribution of genital cutting. Finally, in a private hospital setting, women who have been cut are less likely than others to undergo an episiotomy at first delivery. These findings are provocative, insomuch as they present a picture of genital cutting as failing to contribute independently to poor first-delivery experiences, a finding that deviates from that portrayed in the mainstream discourse on the consequences of the procedure. Select findings of the present study are consistent, however, with the few similar studies that have carried out epidemiologic assessments. Our findings on cesarean section, perineal tear, hemorrhage, and episiotomy (in a public hospital setting) are consistent with findings from both a 1985 study of West African immigrants and with those of a 1998 study conducted in Mali. Among West African immigrants to France, no difference was seen between circumcised and noncircumcised women in the observed rates of cesarean section (Berardi et al. 1985). The researchers in the Mali study found that the severity of the cutting, and not its occurrence, predicted a woman's likelihood of experiencing a perineal tear, episiotomy, or hemorrhage (Jones et al. 1 999). Our findings on episiotomy performed outside of a private hospital setting were also consistent with a 1994 study conducted in southeast Nigeria that found that the incidence of self-reported episiotomy at first delivery was the same between circumcised and uncircumcised women (Adinma 1997). Our study results differ, however, from certain findings of earlier studies. The authors of a 1989 study of Sudanese immigrants in Saudi Arabia, for example, observed an increased rate of episiotomy and hemorrhage (due to episiotomy), and a significantly longer second stage of labor among primiparous women who had been cut compared with primiparous uncircumcised women (De Silva 1989). The Berardi study mentioned above also found that primiparous circumcised women had an increased likelihood of experiencing an episiotomy or a perineal tear, compared with those who had not been cut (Berardi et al. 1985). The likely explanation for the discrepancy with our findings lies in the differences in types of cutting. In the study conducted in Saudi Arabia, the distribution of types of cutting was skewed toward the more severe forms of the practice (45 percent of cutting was type II and 45 percent was type III), whereas the French researchers observed a population of women in which all cutting was type II. As noted above , the sub-sample in our study who had been cut, in contrast, had a distribution of 72 percent: type I; 24 percent: type II; and 4 percent: type III or IV. The De Silva study neglected to discuss type of cutting in relation to its findings. The most intriguing comparison with our study results comes from the 1998 Burkina Faso study, which is similar in design to ours (Jones et al. 1999). Like ours, the Burkina Faso survey was cross-sectional, relied on self-reporting, was conducted among women presenting for OB/GYN services, included a clinical exam, and was made up of a study sample of women with proportionally more type I than type II or type III cutting (56 percent, 39 percent, and 5 percent, respectively). Overall, individual delivery complication rates were similar to those reported in our study, with episiotomy being the most commonly reported complication. However, a logistic regression analysis of the Burkina Faso data comparing circumcised and noncircumcised women's likelihood of reporting any delivery complications while simultaneously controlling for type of cutting, age, parity, education, religion, marital status, and type of consultation found that circumcised women were significantly more likely than others to report any delivery complications. These results included the finding that noncut women were only 32 percent as likely as women with type I cutting to report a past delivery complication. This discrepancy between the Burkina Faso study's findings and our findings most likely reflects three points. First, the Burkina Faso study did not take into account the delivery assistant and location of deliveries. As the present study shows, delivery complications and procedures are strongly associated with delivery place and assistant, a finding supported by many studies, including recent research conducted in southeast Nigeria reporting that perineal tear, hemorrhage, and prolonged labor occur more often among women delivering outside formal health facilities (Etuk et al. 2000). Delivery place and assistant continue to be significant predictors of perineal tear, cesarean section, and episiotomy, even when they are included in a multivariate model, and they seem to be confounders in the apparent association between genital cutting and delivery experience. Because cutting status is also shown in this study to be strongly associated with delivery place and assistant, neglecting these variables, even while considering other sociodemographic correlates, clearly may lead to an incomplete analysis of the association between genital cutting and the experience of childbirth. Second, distinguishing delivery complications from procedures may be important, because reports of procedures logically imply women's access to higher-grade clinical services, and these may be differentially accessible to women who have been cut and those who have not. In our study, we found that cutting status predicted the setting where a delivery took place and suggested lower economic or social status among circumcised women. Such a social-class gradient with genital cutting may be entirely absent in other settings, but it warrants closer inspection where this procedure is not universally practiced. The Burkina Faso study compounded complications and procedures into a single category, muting the prospects for such an analysis. Third, the high prevalence of genital cutting in the Burkina Faso sample population (93 percent) meant that the size of the noncircumcised portion of the sample was small. Our study's distribution of circumcised to noncircumcised women was nearly equal, allowing for a more robust statistical analysis. Nonetheless, an optimal study would include the features described above plus a larger sample size to allow for a stronger analysis of social and clinical predictors of obstetric complications. Although the prevalence of postpartum hemorrhage in our study is similar to the 2 percent-6 percent reported in facility-based studies from the same region of Nigeria (Adetoro 1992; Opaneye 1998), the wide confidence intervals in our multivariate models of the predictors of delivery complications underscore the need for larger and more robust studies of the association between female genital cutting and clinical outcomes. The most unexpected of our study results relates to episiotomy. We found that despite controlling for a number of confounders, circumcised women were less likely than noncircumcised women to report episiotomy at first delivery if delivery took place at a private hospital. Whether this finding is due to chance or to the failure to include in the analysis a variable that is as yet undetermined is not clear. Because private hospitals charge a fee for each extra surgical procedure (episiotomy is considered an additional surgery), circumcised women may be refusing episiotomy to avoid additional cost. If that were the case, we would have expected to see a significant difference at the private hospital level between the socioeconomic status of women reporting that they had undergone an episiotomy and that of women who did not (that is, women reporting episiotomy should have higher socioeconomic status). No difference in such status was found, however. An additional possibility is that private hospital personnel disc riminate against women who have been cut, a situation that would lead to systematic differences in delivery practices. Our study did not allow further inquiry into such possibilities. The generalizability of our study was limited by a number of factors. As noted above, reliance on women's self-reporting of delivery experiences creates an opportunity for recall bias. Research to date on the association between educational level and recall shows no significant association between schooling and the accuracy of reporting past obstetric complications (Ronsmans et al. 1997). Indeed, our findings reveal that women in our sample with low educational attainment were more likely to report obstetric complications than were women with higher education, thus reducing the likelihood that circumcised women may have underreported obstetric complications because they lacked schooling. Selection bias also limits the extent to which our findings can be generalized. Approaching women at antenatal and family planning services limited our sample to women who were able to afford and chose to seek such services. This approach causes the sample population to be skewed toward women with relatively higher educational attainment and relatively low parity. Finally, our study results are based on first live delivery only, and therefore they cannot be generalized to later childbirth experiences. These findings highlight the importance of closer analyses of socioeconomic differences between women who have undergone genital cutting and those who have not when an attempt is being made to differentiate the role of genital cutting on health outcomes. For a more thorough understanding of the association between female genital cutting and obstetric morbidity, future analyses should look closely into the women's relative access to birth services and also to both policy and practice in different delivery settings. While contributing to the growing body of literature calling for improved scientific data on the health effects of genital cutting, this study furthers arguments that the generalization of health risks from infibulation to all forms of female genital cutting should be avoided (Obermeyer 1999; Larsen and Yan 2000; Shell-Duncan 2001; Snow 2001). Finally, the study underscores that although severe cutting (that is, infibulation) may have grave consequences in terms of obstetric morbidity, in settings w here types I and II predominate, the impact of genital cutting on birth outcomes is eclipsed by women's basic necessities, including access to trained birth attendants and adequate clinical facilities. In places like southwest Nigeria, the attention focused on female genital cutting may act as a smokescreen that distracts from more widespread issues threatening the health of women (Gruenbaum 1996). | |
Table 1
Percentage distribution of women surveyed, by selected demographic
characteristics and delivery setting, according to type of genital
cutting they had undergone, Edo State, Nigeria, 1999
Type of cutting
Characteristic/ Types
setting Type I II, III, IV Not cut Total
Current age
15-19 0.2 0.0 0.8 0.5
20-24 2.2 2.3 4.3 3.2
25-29 18.6 21.7 23.7 21.3
30-34 29.8 32.0 33.3 31.7
35-39 25.3 27.4 27.8 26.7
40-44 17.9 8.0 8.0 12.0
45-49 5.8 8.6 2.1 4.6
Ethnicity
Bini 49.8 36.5 16.3 33.0
Esan 17.7 20.6 28.5 22.9
Igbo 11.9 14.3 14.4 13.4
Urhobo 8.1 12.0 6.6 8.0
Yoruba 6.0 4.0 13.4 9.0
Other 6.5 12.6 20.8 13.7
Religion
Catholic 21.8 26.6 39.3 30.2
Muslim 2.5 2.9 11.1 6.3
Pentecostal 44.4 41.6 25.1 35.5
Protestant 14.8 16.8 17.5 16.3
Traditional 6.0 5.8 2.3 4.3
None 1.3 0.0 0.6 0.9
Other 9.2 6.3 4.1 6.5
Educational level
None 7.2 4.0 6.0 6.2
Primary 37.2 24.6 15.0 25.5
Secondary 29.6 38.3 29.7 31.0
Tertiary 26.0 33.1 49.3 37.3
Place of first delivery
Private hospital 29.3 29.1 36.3 32.3
Teaching/referral
hospital 21.0 23.8 33.0 26.7
State/district hospital 21.4 20.4 16.2 18.9
Health center/PHC 9.6 6.4 5.6 7.4
Traditional healer's/
traditional birth
attendant's workplace 4.5 6.4 2.1 3.7
Home 14.2 13.9 6.8 10.9
Assistant at first
delivery
Private doctor 14.0 16.1 25.5 19.4
Public doctor/nurse 23.6 17.2 34.8 27.5
Midwife (a) 41.7 47.1 30.5 37.6
Traditional or faith
healer/traditional
birth attendant 13.3 9.8 5.1 9.2
Self/female relative 7.4 9.8 4.1 6.3
Total prevalence 40.3 15.8 43.9 100.00
(N) (446) (175) (486) (1,107) (b)
PHC = Primary health center.
(a)The WHARC survey questionnaire did not differentiate between trained
and untrained midwives.
(b)Numbers (N) are valid only for age category. For other categories
they may vary because of "no response" answers.
Table 2
Among 1,107 parous women interviewed, first-delivery procedures and
complications, by genital cutting status, frequency distribution, and
crude odds ratio, Edo State, Nigeria, 1999
Not
Procedure/ Cut cut Total
complication (n) (n) (Percent) (n)
Procedure
Episiotomy (162) (180) 30.9 (342)
Cesarean section (32) (42) 6.7 (74)
Instruments (3) (1) 0.4 (4)
Complication
Perineal tear (25) (17) 3.8 (42)
Obstructed labor (21) (10) 2.8 (31)
Hemorrhage (18) (5) 2.1 (23)
Fever (3) (3) 0.5 (6)
Convulsion/seizure (4) (0) 0.4 (4)
Other (5) (4) 0.8 (9)
None (348) (224) 51.7 (572)
Total (N) (621) (486) (1,107)
Percent 56.1 43.9 100.0
Crude odds ratio Not
Procedure/ (95 percent CI) (a) cut
complication Cut (r)
Procedure
Episiotomy 0.60 (0.46-0.78) *** 1.00
Cesarean section 0.57 (0.36-0.93) * 1.00
Instruments -- (b)
Complication
Perineal tear 1.16 (0.62-2.17) 1.00
Obstructed labor 1.67 (0.78-3.57) 1.00
Hemorrhage 2.87 (1.06-7.79) *
Fever -- (b)
Convulsion/seizure -- (b)
Other -- (b)
None 1.49 (1.18-1.89) ** 1.00
Total (N)
Percent
* Significant at p <0.05
** p <0.01
*** p <0.001.
CI = Confidence interval.
(a)Odds of reporting a delivery complication or procedure among women
who have been cut relative to those who have not.
(b)Odds ratio not calculated for complications or procedures where
(n) <10.
Table 3
Occurrence of selected obstetric complications or procedures, by
delivery location and assistant, and crude odds ratios (OR), Edo State,
Nigeria, 1999
Delivery place/ Cesarean section Episiotomy
assistant OR (95% percent CI) OR (95 percent CI)
Deliver place
Institution (r) 1.00 1.00
Other (a) No occurrence 0.02 (0.01-0.09) ***
Delivery assistant
Professional (r) 1.00 1.00
Nonprofessional (b) No occurrence 0.04 (0.02-0.12) ***
Delivery place/ Hemorrhage Obstructed labor
assistant OR (95 percent CI) OR (95 percent CI)
Deliver place
Institution (r) 1.00 1.00
Other (a) 4.21 (1.77-10.02) *** 5.91 (2.86-12.22) ***
Delivery assistant
Professional (r) 1.00 1.00
Nonprofessional (b) 4.77 (2.03-11.22) *** 5.51 (2.67-11.37) ***
Delivery place/ Perineal tear
assistant OR (95 percent CI)
Deliver place
Institution (r) 1.00
Other (a) 2.80 (1.40-5.41) *
Delivery assistant
Professional (r) 1.00 **
Nonprofessional (b) 2.07 (1.02-4.21)
* Significant at p <0.05
** p <0.01
*** p <0.001.
CI = confidence interval.
(r) = reference category.
(a)Includes at home, at a traditional healer's workplace, or at a
traditional birth attendant's workplace.
(b)Includes a traditional healer, faith healer, traditional birth
attendant, self, or female relative.
Table 4
Complications at first delivery, by selected variables, and crude and
and multivariate odds ratios, Edo State, Nigeria, 1999
Hemorrhage
crude
Variable OR (95 percent CI) *
cutting status
Notcut(r) 1.00
cut 2.87 (1.06-7.79)
Place of first delivery
Institution (r) 1.00
Other 4.21 (1.77-10.02) ***
Delivery site other
than home (r) (a) --
Delivery site at home --
Assistant at delivery
Professional (r) 1.00
Nonprofessional 4.77 (2.03-11.22) ***
Assistant other than
public doctor or nurse (r) (a) --
Public doctor or nurse --
Age at first marriage (b) --
Religion
Pentecostal (r) 1.00
Protestant 1.48 (0.43-5.13)
cathollc 0.79 (0.23-2.73)
Muslim -- (c)
Traditional/other/none 4.32 (1.54-12.14)
Educational level
None and primary (r) --
Secondary + --
Hemorrhage
Multivariate
Variable OR (95 percent CI)
cutting status
Notcut(r) 1.00
cut 2.03 (0.72-5.71)
Place of first delivery
Institution (r) 1.00
Other 0.69 (0.09-5.19)
Delivery site other
than home (r) (a) --
Delivery site at home --
Assistant at delivery
Professional (r) 1.00
Nonprofessional 4.55 (0.61-34.08)
Assistant other than
public doctor or nurse (r) (a) --
Public doctor or nurse --
Age at first marriage (b) --
Religion
Pentecostal (r) 1.00
Protestant 2.00 (0.55-7.25)
cathollc 1.29 (0.35-4.74)
Muslim -- (c)
Traditional/other/none 3.34 (1.08-10.32) *
Educational level
None and primary (r) --
Secondary + --
Obstructed labor
crude
Variable OR (95 percent CI)
cutting status
Notcut(r) 1.00
cut 1.67 (0.78-3.57)
Place of first delivery
Institution (r) 1.00
Other 5.91 (2.86-12.22) ***
Delivery site other
than home (r) (a) --
Delivery site at home --
Assistant at delivery
Professional (r) 1.00
Nonprofessional 5.51 (2.67-11.37) ***
Assistant other than
public doctor or nurse (r) (a) --
Public doctor or nurse --
Age at first marriage (b) 0.90 (0.81-0.99) *
Religion
Pentecostal (r) --
Protestant --
cathollc --
Muslim
Traditional/other/none --
Educational level
None and primary (r) 1.00
Secondary + 0.28 (0.14-0.59) ***
Obstructed labor
Multivarlate
Variable OR (95 percent CI)
cutting status
Notcut(r) 1.00
cut 1.22 (0.54-2.79)
Place of first delivery
Institution (r) 1.00
Other 2.84 (0.38-21,22)
Delivery site other
than home (r) (a) --
Delivery site at home --
Assistant at delivery
Professional (r) 1.00
Nonprofessional 1.54 (0.20-11.80)
Assistant other than
public doctor or nurse (r) (a) --
Public doctor or nurse --
Age at first marriage (b) 0.98 (0.87-1.09)
Religion
Pentecostal (r) --
Protestant --
cathollc --
Muslim
Traditional/other/none --
Educational level
None and primary (r) 1.00
Secondary + 0.47 (0.19-1.18)
Perineal tear
crude
Variable OR (95 percent CI)
cutting status
Notcut(r) 1.00
cut 1.16 (0.62-2.17)
Place of first delivery
Institution (r) --
Other --
Delivery site other
than home (r) (a) 1.00
Delivery site at home 3.08 (1.51-6.31) ***
Assistant at delivery
Professional (r) --
Nonprofessional --
Assistant other than
public doctor or nurse (r) (a) 1.00
Public doctor or nurse 0.36 (0.14-0.91) ***
Age at first marriage (b) --
Religion
Pentecostal (r) --
Protestant --
cathollc --
Muslim
Traditional/other/none --
Educational level
None and primary (r) --
Secondary + --
Perineal tear
Multivarlate
Variable OR (95 percent CI)
cutting status
Notcut(r) 1.00
cut 0.92 (0.48-1.75)
Place of first delivery
Institution (r) --
Other --
Delivery site other
than home (r) (a) 1.00
Delivery site at home 2.70 (1.29-5.68) **
Assistant at delivery
Professional (r) --
Nonprofessional --
Assistant other than
public doctor or nurse (r) (a) 1.00
Public doctor or nurse 0.43 (0.16-1.12)
Age at first marriage (b) --
Religion
Pentecostal (r) --
Protestant --
cathollc --
Muslim
Traditional/other/none --
Educational level
None and primary (r) --
Secondary + --
* Significant at p <0.05
** p <0.01
*** p <0.001.
(r) = Reference category.
CI = Confidence interval.
-- = No statistical association with particular procedure.
(a)Reference category for perineal tear.
(b)Continuous variable, no reference category.
(c)No women in this category had the particular complication.
Table 5
Procedures at first delivery, by selected variables, and crude and
multivariate odds ratios, Edo State, Nigeria, 1999
Cesarean section
Crude
Variable OR (95 percent CI)
Cutting status
Not cut (r) 1.00
Cut 0.57 (0.36-0.93) *
Place of delivery
Private hospital (r) (a) 1.00
Teaching/referral hospital 6.14 (3.41-11.02) ***
State/district hospital 1.09 (0.44-2.69)
Health center/PHC 1.20 (0.34-4.23)
Institution (r) (b) --
Other --
Assistant at delivery
Private doctor (r) (a) 1.00
Public doctor/nurse 4.88 (2.73-8.74) ***
Midwife 0.28 (0.10-0.78) **
Professionat (r) (b) --
Nonprofessional --
Educational level
None (r) (b) -- (d)
Primary (r) (a) 1.00
Secondary 3.04 (1.27-7.29) **
Tertiary 6.31 (2.81-14.15) ***
Ethnic group
Bini (r) 1.00
Esan 1.26 (0.66-2.41)
Igbo 0.94 (0.41-2.17)
Urhobo 0.38 (0.09-1.64)
Yoruba 3.41 (1.72-6.75) ***
Other 0.91 (0.40-2.11)
Year of delivery
After 1990 (r) 1.00
Before 1990 0.19 (0.10-0.34) ***
Age at first marriage (c) 1.22 (1.16-1.29) ***
Age at first delivery (years)
<20 (r) 1.00
20-24 2.98 (0.86-10.09)
25-29 5.75 (1.73-19.08) **
30+ 37.33 (10.53-132.36) ***
Religion
Pentecostal (r) 1.00
Protestant 1.26 (0.57-2.78)
Catholic 2.58 (1.44-4.64) **
Muslim 2.76 (1.15-6.62) *
Traditional religion 0.46 (0.60-3.49)
Other religion 0.30 (0.04-2.29)
None -- (d)
Cesarean section
Multivariate
Variable OR (95 percent CI)
Cutting status
Not cut (r) 1.00
Cut 1.36 (0.72-2.55)
Place of delivery
Private hospital (r) (a) 1.00
Teaching/referral hospital 2.46 (0.72-8.38)
State/district hospital 0.89 (0.21-3.83)
Health center/PHC 1.27 (0.24-6.66)
Institution (r) (b) --
Other --
Assistant at delivery
Private doctor (r) (a) 1.00
Public doctor/nurse 1.83 (0.54-6.23)
Midwife 0.16 (0.04-0.65) *
Professionat (r) (b) --
Nonprofessional --
Educational level
None (r) (b) -- (d)
Primary (r) (a) 1.00
Secondary 1.76 (0.62-5.01)
Tertiary 2.23 (0.76-6.57)
Ethnic group
Bini (r) 1.00
Esan 0.96 (0.43-2.11)
Igbo 1.05 (0.39-2.79)
Urhobo 0.34 (0.07-1.58)
Yoruba 1.43 (0.57-3.60)
Other 0.87 (0.31-2.43)
Year of delivery
After 1990 (r) 1.00
Before 1990 0.43 (0.21-0.87) *
Age at first marriage (c) 1.10 (0.97-1.25)
Age at first delivery (years)
<20 (r) 1.00
20-24 0.75 (0.63-8.96)
25-29 0.93 (0.20-4.40)
30+ 0.74 (0.24-2.33)
Religion
Pentecostal (r) 1.00
Protestant 0.81 (0.33-1.99)
Catholic 1.04 (0.50-2.20)
Muslim 2.52 (0.77-8.23)
Traditional religion 1.19 (0.13-11.35)
Other religion 0.67 (0.08-5.64)
None -- (d)
Episiotomy
Crude
Variable OR (95 percent CI)
Cutting status
Not cut (r) 1.00
Cut 0.60 (0.46-0.78) ***
Place of delivery
Private hospital (r) (a) --
Teaching/referral hospital --
State/district hospital --
Health center/PHC --
Institution (r) (b) 1.00
Other 0.02 (0.01-0.09) ***
Assistant at delivery
Private doctor (r) (a) --
Public doctor/nurse --
Midwife --
Professionat (r) (b) 1.00
Nonprofessional 0.04 (0.02-0.12) ***
Educational level
None (r) (b) 1.00
Primary (r) (a) 2.52 (1.15-5.52) *
Secondary 3.14 (1.45-6.80) **
Tertiary 5.02 (2.34-10.77) ***
Ethnic group
Bini (r) 1.00
Esan 1.49 (1.05-2.11) *
Igbo 1.95 (1.30-2.92) **
Urhobo 1.08 (0.64-1.83)
Yoruba 1.34 (0.82-2.18)
Other 1.44 (0.95-2.17)
Year of delivery
After 1990 (r) --
Before 1990 --
Age at first marriage (c) --
Age at first delivery (years)
<20 (r) --
20-24 --
25-29 --
30+ --
Religion
Pentecostal (r) 1.00
Protestant 0.98 (0.68-1.43)
Catholic 0.82 (0.60-1.12)
Muslim 0.57 (0.31-1.03)
Traditional religion 0.27 (0.11-0.66) **
Other religion 0.84 (0.49-1.45)
None 0.55 (0.11-2.68)
Episiotomy
Multivariate
Variable OR (95 percent CI)
Cutting status
Not cut (r) 1.00
Cut 0.69 (0.51-0.93) **
Place of delivery
Private hospital (r) (a) --
Teaching/referral hospital --
State/district hospital --
Health center/PHC --
Institution (r) (b) 1.00
Other 0.07 (0.01-0.42) **
Assistant at delivery
Private doctor (r) (a) --
Public doctor/nurse --
Midwife --
Professionat (r) (b) 1.00
Nonprofessional 0.27 (0.08-0.99) *
Educational level
None (r) (b) 1.00
Primary (r) (a) 1.87 (0.80-4.38)
Secondary 1.59 (0.69-3.69)
Tertiary 2.56 (1.11-5.91) *
Ethnic group
Bini (r) 1.00
Esan 1.33 (0.90-1.96)
Igbo 1.73 (1.11-2.69) *
Urhobo 1.12 (0.64-1.98)
Yoruba 1.27 (0.73-2.21)
Other 1.35 (0.84-2.19)
Year of delivery
After 1990 (r) --
Before 1990 --
Age at first marriage (c) --
Age at first delivery (years)
<20 (r) --
20-24 --
25-29 --
30+ --
Religion
Pentecostal (r) 1.00
Protestant 0.79 (0.53-1.17)
Catholic 0.53 (0.37-0.75) ***
Muslim 0.49 (0.25-0.98) *
Traditional religion 1.43 (0.48-4.23)
Other religion 1.07 (0.59-1.94)
None 1.79 (0.25-12.71)
* Significant at p <0.05
** p <0.01
*** p <0.001.
(r) = reference category.
CI = confidence interval.
PHC = Public health center.
-- = No statistical association with particular procedure.
(a)Reference category for cesarean section.
(b)Reference category for episiotomy.
(c)Continuous variable, no reference category.
(d)No women in this category underwent the particular procedure.
Table 6
Among 355 women reporting first delivery at a private hospital,
multivariate odds ratios for those who reported undergoing episiotomy,
by genital cutting status and type of assistant at delivery, Edo State,
Nigeria, 1999
Cutting status/ Episiotomy
assistant OR (95 percent CI)
Cutting status
Not cut (r) 1.00
Cut 0.39 (0.25-0.62) ***
Delivery assistant
Private doctor (r) 1.00
Public doctor/nurse 0.37 (0.10-1.31)
Midwife 1.92 (1.18-3.12) **
Self/relative 0.86 (0.08-8.87)
** Significant at p <0.01
*** p <0.001.
(r) = Reference category.
CI = confidence interval.
| | Acknowledgments Funding for this research was provided by a grant from the Bundesministerium fur Bildung und Forschung (BMBF) to Rachel Snow and Juergen Wacker, as part of the research program TropMed-Heidelberg. We are grateful to Heiko Becher for advice and support regarding the statistical analysis. We gratefully acknowledge the participation of colleagues in each of the collaborating institutions: University of Benin Teaching Hospital and Central Hospital in Benin City and Specialist Teaching Hospital in Irrua. Finally, we thank all the women who took part in the study and shared their experiences. Notes (1.) Type I includes excision of the prepuce with or without removal of part or all of the clitoris; type II refers to excision of the prepuce and clitoris with partial or total excision of the labia minora; type III includes excision of all or part of the external genitalia with stitching or narrowing of the vaginal opening (infibulation); and type IV involves pricking, piercing, or incision of the clitoris and/or labia minora, stretching of the clitoris and/or labia, cauterization by burning of the clitoris and surrounding tissues, scraping (augurya cuts) of the vaginal orifice or cutting (gishiri cuts) of the vagina, introduction of corrosive substances into the vagina to cause bleeding or of herbs into the vagina with the aim of tightening or narrowing the vagina, and any other procedure that falls under the definition of female genital mutilation given above (WHO 1997: 3). (2.) Depending on the region, the choice of setting may be a result of cost, tradition, or legislation against health practitioners' performing the cutting in a hospital setting. (3.) Stillbirths were not included in the analysis. Examined separately, a crude logistic regression analysis showed that, for the 51 women whose first delivery was a stillbirth, a slight but not statistically significant association existed with genital cutting status (OR 1.6;95 percent CI 0.88-2.90, p = 0.12). (4.) Although interviewers asked women to report multiple problems at delivery, each woman gave only one response. (5.) By confining the analysis to first live delivery only, it was possible to (1) include the largest number of women, because the population was young, and the majority of parous women had had only one live birth; (2) control for parity; and (3) control for complications specifically associated with stillbirths. (6.) Use of instruments, incidence of convulsion or seizure, and incidence of fever, each having a frequency of fewer than ten, were not taken further in the analysis. (7.) First-delivery complications and procedures did not prove statistically significantly associated with circumcision type, with the exception of episiotomy and type III cutting. Among women who had been cut, women with type III cutting were 2.5 times more likely to report episiotomy at first delivery, compared with women who had undergone type I cutting (p = 0.05). This analysis did not include women who are not cut and did not control for socio-demographic confounders, however. (8.) The terms used here to distinguish between quality of delivery services are: "institution" (private hospital, teaching/referral hospital, state/district hospital, health center/primary health center [PHC]), and "other" (traditional healer's or traditional birth attendant's workplace and home). (9.) A doctor, nurse, or midwife is considered a professional assistant, while a traditional healer, faith healer, traditional birth attendant, self, or a female relative is considered a nonprofessional assistant. (10.) The number of women reporting "no complications" at first delivery was proportionately higher among those who delivered at noninstitutional sites (OR 2.9; 95 percent CI 2.01-4.22), compared with those who gave birth in institutions. This result likely reflects that institutions are a magnet for women who anticipate delivery complications (and can afford the cost). (11.) As anticipated, the relationship between place and assistant was strong. Private doctors delivered at private hospitals (in 96 percent of cases), and women who delivered by themselves did so at home, with a few exceptions. Midwives were the least likely to assist delivery at only one particular facility, and were fairly evenly distributed between state or district hospitals (30 percent), teaching or referral hospitals (28 percent), and private hospitals (30 percent), with the remainder delivering at a health center or primary health center (10 percent). (12.) An analysis of episiotomy at a private hospital by religion, delivery year, ethnicity, age at delivery, educational level, delivery assistant, circumcision type, and age at circumcision revealed that the only statistically significant differences between women reporting episiotomy and those not reporting episiotomy at a private hospital were (1) their assistant at delivery (specifically a midwife), and (2) their genital cutting status. References Adetoro, O.O. 1992. "Primary post-partum haemorrhage at a university hospital in Nigeria." West African Journal of Medicine 11(3): 172-178. Adinma, J.I.B. 1997. "Current status of female circumcision among Nigerian Igbos." West African Journal of Medicine 16(4): 227-231. Babatunde, Emmanuel D. 1998. Women's Rites vs. Women's Rights: A Study of Circumcision among the Ketu Yoruba of South Western Nigeria. Trenton, NJ and Asmara, Eritrea: Africa World Press. Berardi, J.C., J.F. Teillet, J. Godard, V. Laloux, P. Allane, and M.H. Franjou. 1985. "Consequences obstetricales de 1'excision feminine." Journal de Gynecologie Obstetrique et Biologie de la Reproduction 14: 743-746. Caldwell, John C.,I.O. Orubuloye, and Pat Caldwell. 1997. "Male and female circumcision in Africa--from a regional to a specific Nigerian examination." Social Science and Medicine 44(8): 1,181-1,193. De Silva, S. 1989. "Obstetric sequelae of female circumcision." European Journal of Obstetrics and Gynecology and Reproductive Biology 32: 233-240. Dirie, Waris. 1998. Desert Flower. New York: William Morrow. Donnay, France. 2000. "Maternal survival in developing countries: What has been done, what can be achieved in the next decade." International Journal of Gynecology and Obstetrics 70:89-97. Etuk, S.J., I.H. Itam, and E.E.J. Asuquo. 2000. "Morbidity and mortality in booked women who deliver outside orthodox health facilities in Calabar, Nigeria." Acta Tropica 75:309-313. Fox, E.F., A. de Ruiter, and J.S. Bingham. 1997. "Female genital mutilation." International Journal of STD and AIDS 8: 599-601. Gruenbaum, Ellen. 1996. "The cultural debate over female circumcision: The Sudanese are arguing this one out for themselves." Medical Anthropology Quarterly 10(4): 455-475. International Action against Female Genital Mutilation (INTACT). Brochure. 2000. Saarbruken, Germany: HAAG Agentur fur Kommunikation. Isa, Ab. Rahman, Rashidah Shuib, and M. Shukri Othman. 1999. "The practice of female circumcision among Muslims in Kelantan, Malaysia." Reproductive Health Matters 7(13): 137-144. Jones, Heidi, Nafissatou Diop, Ian Askew, and Inoussa Kabore. 1999. "Female genital cutting practices in Burkina Paso and Mali and their negative health outcomes." Studies in Family Planning 30(3): 219-229. Kun, K.E. 1997. "Female genital mutilation: The potential for increased risk of HIV infection." International Journal of Gynecology and Obstetrics 59: 153-155. Larsen, Ulla and Sharon Yan. 2000. "Does female circumcision affect infertility and fertility? A study of the Central African Republic, Cote d'Ivoire, and Tanzania." Demography 37(3): 313-321. Liskin, L.S. 1992. "Maternal morbidity in developing countries: A review and comments." International Journal of Gynecology and Obstetrics 37(2): 77-87. Maduma-Butshe, A., Adele Dyall, and Paul Garner. 1998. "Routine episiotomy in developing countries." British Medical Journal 316: 1,179-1,180. Myers, Robert A., Francisca I. Omorodion, Anthony E. Isenalumhe, and Gregory I. Akenzua. 1985. "Circumcision: Its nature and practice among some ethnic groups in southern Nigeria." Social Science and Medicine 21(5): 581-588. National Population Commission, Nigeria (NPC). 2000. Nigeria Demographic and Health Survey 1999. Calverton, MD: NPC and ORC/Macro. Obermeyer, Carla Makhlouf. 1999. "Female genital surgeries: The known, the unknown, and the unknowable." Medical Anthropology Quarterly 13(1): 79-106. Opaneye A.A. 1998. "Traditional medicine in Nigeria and modern obstetric practice: Need for cooperation." Central African Journal of Medicine 44(10): 258-261. Otoide, V.O., S.M. Ogbonmwan, and Friday E. Okonofua. 2000. "Episiotomy in Nigeria." International Journal of Gynecology and Obstetrics 68: 13-17. Prual, Alain, Dominique Huguet, Olivier Garbin, and Gomna Rabe. 1998. "Severe obstetric morbidity of the third trimester, delivery and early puerperium in Niamey (Niger)." African Journal of Reproductive Health 2(1): 10-19. Ronsmans, Canine, Endang Achadi, Surekha Cohen, and Ali Zazri. 1997. "Women's recall of obstetric complications in South Kalimantan, Indonesia." Studies in Family Planning 28(3): 203-214. Rushwan, H. 2000. "Female genital mutilation (FGM) management during pregnancy, childbirth and the postpartum period." International Journal of Gynecology and Obstetrics 70: 99-104. Selo-Ojeme, D.O. and Friday E. Okonofua. 1997. "Risk factors for primary postpartum haemorrhage." Archives of Gynecology and Obstetrics 259: 179-187. Shell-Duncan, Bettina. 2001. "The medicalization of female 'circumcision': Harm reduction or promotion of a dangerous practice?" Social Science and Medicine 52(7): 1,013-1,028. Snow, Rachel C. 2001. "Female genital cutting: Distinguishing the rights from the health agenda." Tropical Medicine and International Health 6(2): 89-91. Stewart, Kathryn S. and Mario Festin. 1995. "Validation study of women's reporting and recall of major obstetric complications treated at the Philippine General Hospital." International Journal of Gynecology and Obstetrics 48 (Supplement): S53-S66. Toubia, Nahid. 1998. "Female circumcision/female genital mutilation." African Journal of Reproductive Health 2(2): 6-7. Walker, Alice. 1992. Possessing the Secret of Joy. New York: Harcourt Brace Jovanovich. World Bank. 2001. World Development Indicators. Washington, DC: World Bank. World Health Organisation (WHO). 1997. Female genital mutilation- a joint WHO/UNICEF/UNFPA statement. Geneva: WHO. -----. 1998. A Systematic Review of Research on Health Complications Following Female Genital Mutilation Including Sequelae in Childbirth. WHO/FRH/WHD/98.21. Geneva: WHO. -----. 2000a. World Health Report 2000. Geneva: WHO. -----. 2000b. "Female Genital Mutilation: Fact Sheet No. 241." Geneva: WHO. Tracy E. Slanger is Research Associate and Rachel C. Snow is Unit Head, University of Heidelberg, Department of Tropical Hygiene and Public Health, Reproductive Health Unit, Im Neuenheimer Feld 324, 69120, Heidelberg, Germany. E-mail: tracy.slanger@urz.uni-heidelberg.de. Friday E. Okonofua is Director, WHARC (Women's Health and Action Research Centre), Benin City, Nigeria. |
|
