Bulletin of the World Health Organization

Female genital cutting: current practices and beliefs in western Africa

Heather L Sipsma a, Peggy G Chen a, Angela Ofori-Atta b, Ukwuoma O Ilozumba c, Kapouné Karfo d & Elizabeth H Bradley a

a. Department of Health Policy and Administration, School of Public Health, Yale University, 2 Church Street South, New Haven, CT 06519, United States of America (USA).
b. Department of Psychiatry, University of Ghana Medical School, Accra, Ghana.
c. Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, USA.
d. Department of Psychiatry, Ouagadougou University School of Medicine, Ouagadougou, Burkina Faso.

Correspondence to Heather L Sipsma (e-mail: heather.sipsma@yale.edu).

(Submitted: 26 May 2011 – Revised version received: 11 October 2011 – Accepted: 12 October 2011 – Published online: 11 November 2011.)

Bulletin of the World Health Organization 2012;90:120-127F. doi: 10.2471/BLT.11.090886

Introduction

More than 100 million girls and women have undergone female genital cutting (FGC, also known as “female circumcision”), and more than 3 million female infants and children are at risk for this procedure annually.1–3 FGC is seen as a rite of passage for young girls in some communities and is most often performed between the ages of 4 and 10.1,4 Reasons for this practice include beliefs that it enhances fertility, promotes purity, increases marriage opportunities and prevents stillbirths. These beliefs are strongly rooted in tradition, culture and religion, but none carries a scientific basis.5–8 Young girls and women who undergo FGC are subjected to extreme pain, since the procedure is often conducted without anaesthesia and under non-sterile conditions. When accompanied by excessive bleeding, it can even lead to death.9,10 Additional complications can include localized infection and abscess formation, pelvic infection, sepsis, tetanus, urinary retention and chronic urinary tract infection, hepatitis and human immunodeficiency virus (HIV) infection.11 Other reproductive health complications can include obstructed menstruation, difficulty conceiving, prolonged labour, tearing of tissues during delivery and neonatal death.8,11–16 Psychiatric sequelae such as flashbacks to the event, affective disorders and post-traumatic stress disorders can also affect women throughout their lives.8,12,13

Despite the risks associated with FGC, the peer-reviewed literature on the prevalence and predictors of FGC is sparse. Studies from Egypt and Ethiopia give FGC prevalences of approximately 85% and 70%, respectively,17,18 and several qualitative and anthropological studies have examined the underlying context for FGC.6,19–22 No studies, however, have reported cross-national comparisons of the prevalence of FGC and of the factors with which it is associated. Such data would be helpful for understanding the variation in the frequency of FGC, particularly in western Africa, where the practice is known to be common despite legislation and other efforts to curb its prevalence. Western Africa is also particularly well suited for cross-national comparisons because substantial differences exist between countries in prevalence rates and in the approaches used to eliminate this practice. These differences can bring to light potential strategies that may be useful in similar settings.

Accordingly, we sought to estimate the prevalence of FGC practices and beliefs across all western African countries for which national data were available from the most recent round of the Multiple Indicator Cluster Surveys (MICS). We also aimed to identify correlates of these practices and beliefs to identify high-risk subpopulations. This evidence may be useful to help better understand country-level variation in this persistent but widely criticized practice1,23,24 and to target efforts to rid future generations of the practice of FGC in western Africa.

Methods

Study design and sample

We conducted a cross-sectional study of 10 countries in western Africa using self-reported data collected between the years 2005 and 2007 during the third round of the Multiple Indicator Cluster Surveys (MICS). Data were available for Burkina Faso, Côte d’Ivoire, Gambia, Ghana, Guinea-Bissau, Mauritania, Niger, Nigeria, Sierra Leone and Togo. Surveys were conducted in French in all countries except Gambia, Ghana, Nigeria and Sierra Leone, where they were conducted in English.

The MICS is a household survey developed by the United Nations Children’s Fund (UNICEF) to provide national estimates of health indicators for women, men and children. The third round of the MICS used a two-stage stratified sample design in which enumeration areas were selected first, stratified by administrative region and area type (urban/rural). Households were systematically sampled from each enumeration area. Total sample size varied by country. The MICS used three sets of questionnaires, including a women’s self-reporting questionnaire administered face to face to all women in each household between the ages of 15 and 49 years.25

Measures

We examined three primary outcome measures. First, participants were asked “Have you yourself ever been circumcised?” (yes/no). Second, participants with at least one living daughter were asked “Have any of your daughters been circumcised? If yes, how many?” Responses indicated the number of daughters circumcised. For analysis, this variable was dichotomized into yes (at least 1 daughter circumcised) versus no (no daughters circumcised). Participants who had not heard of FGC or female circumcision (women who responded no to both “Have you ever heard of female circumcision?” and “In a number of countries, there is a practice in which a girl may have part of her genitals cut. Have you ever heard about this practice?”) were assumed to have not been circumcised and to have not had their daughters circumcised and were coded accordingly. Third, participants were asked “Do you think this practice should be continued?” (yes/depends/no). For analysis, this third variable was collapsed into two categories: yes/depends versus no. Participants who had not heard of FGC were coded as missing for this outcome.

Our independent variables included basic sociodemographic characteristics that were available across all countries, including age (5-year age groups), educational level (none, primary, above primary), marital status (currently married, formerly married, never married), wealth quintile and religion (Muslim versus non-Muslim). Subgroups of non-Muslims were examined for differential effects. Wealth quintiles were derived by the MICS using a combination of reported household assets and utility services.26

Statistical analysis

Analyses were performed separately for each country. We first generated weighted frequencies to determine the prevalence of the three outcomes and to describe the characteristics of the sample populations. We then examined the potential for multicollinearity among the independent variables by using an average threshold correlation coefficient of 0.50. We constructed logistic regression models for each outcome, including all sociodemographic characteristics, whether or not they significantly contributed to the model. This approach was used to ensure the comparability of the models across countries. Because not all circumcised women are aware of having undergone circumcision, particularly if they have smaller incisions or were circumcised in infancy, we conducted a sensitivity analysis by repeating all analyses without including those respondents who reported never having heard of FGC, and we compared these results with our other findings. In all models we accounted for sample weighting and complex survey designs by adjusting for strata and cluster membership. Cases with missing data (< 5%) were excluded from the analyses, which were completed with SAS version 9.2 (SAS Institute, Cary, United States of America).

Results

Sample characteristics

On average, slightly more than half of the women in each sample were between the ages 15 and 29 (Table 1). More than half of the women in Burkina Faso, Côte d’ Ivoire, Gambia, Guinea-Bissau, Niger and Sierra Leone had had no formal education. In most countries, the majority of women were currently married, and the proportion of Muslims per country ranged from 14% to 99%.

Country prevalence

The prevalence of FGC was high overall but varied substantially between countries. In Sierra Leone, Gambia, Burkina Faso and Mauritania, prevalence rates for FGC were 94%, 79%, 74% and 72%, respectively (Table 2). In contrast, fewer than 6% of women had been circumcised in Ghana, Niger and Togo. Gambia and Mauritania had the highest percentage of daughters circumcised (64%) and Togo, Ghana and Niger had the lowest (1%). In three countries, more than half of the women believed that the practice of FGC should continue (Sierra Leone, 88%, Gambia, 77% and Mauritania, 59%). The lowest percentages of women believing the practice should continue were found in countries with the lowest reported rates of FGC: Ghana (4%) and Niger (7%). Prevalence estimates after excluding women who had never heard of FGC remained largely unchanged.

Associated sociodemographic characteristics

In general, being older, having less education, and being currently or formerly married as opposed to never married were all associated with increased odds of having been circumcised (Table 3, available at: http://www.who.int/bulletin/volumes/90/2/11-090886). However, the opposite effects were seen in Gambia, where being older was associated with lower odds of having been circumcised and no association was noted between having been circumcised and educational level. Furthermore, being Muslim was generally associated with increased odds of having been circumcised. Effects across non-Muslim subgroups were largely similar, and the non-Muslim reference category was therefore maintained for analysis. The association between wealth and FGC varied across samples. In five countries greater wealth was associated with increased odds of having been circumcised; in the other five, less wealth was associated with increased odds of having been circumcised.

Being older, having less education and being Muslim were associated with higher odds of having had a daughter circumcised (Table 4, available at: http://www.who.int/bulletin/volumes/90/2/11-090886). Wealth was inconsistently associated with having had a daughter circumcised; these associations were similar to those seen for having been circumcised. Finally, in most countries believing that the practice of FGC should continue was associated with being younger, having less education, being currently married and being poorer (Table 5, available at: http://www.who.int/bulletin/volumes/90/2/11-090886). The associations between sociodemographic characteristics and outcomes were largely unaffected by excluding from the analysis those women who had never heard of FGC.

Comparisons across outcomes

Recognizing that the frequencies of our outcomes could represent changes in the prevalence of FGC over time, we plotted the data sequentially by country (Fig. 1). In all countries, the percentage of women who had their daughters circumcised was lower than the percentage who had themselves been circumcised. The relationship between believing that FGC should continue and having had a daughter circumcised was not consistent across countries. Burkina Faso and Mauritania were the only two countries where the percentage of women who believed that FGC should continue was lower than the percentage that had had their daughters circumcised.

Fig. 1. Graphical depiction of prevalence for each outcome by countrya,b
Fig. 1. Graphical depiction of prevalence for each outcome by country<sup>a,b</sup>
FGC, female genital cutting; MICS, multiple indicator cluster surveys.a Due to the proxy measure used to assess FGC in Sierra Leone, the MICS recommends not interpreting the estimate for believing that this practice should continue.27b Percentages include both yes and depends responses.

Discussion

The estimated prevalence of FGC varied widely across countries, despite their geographic proximity, and this variation probably reflects the differences in political, social and historical contexts in countries where FGC is practiced. In Burkina Faso, Gambia, Mauritania and Sierra Leone more than 70% of women had been circumcised, whereas in Ghana, Niger and Togo less than 6% had been circumcised. Additionally, in four countries at least one third of the women reported having had their daughters circumcised, and in six countries more than 20% believed that FGC practices should continue despite recent public criticism of FGC. These findings are concerning, given the potential for causing a girl severe physical and psychological harm.

We also recognize that the prevalence of FGC can vary substantially within the same country. Our findings show that certain women belonging to certain subgroups based on educational level, wealth and religion have significantly increased rates of FGC. Such rates can also vary within countries depending on ethnicity and geographic region. Although these variables could have enhanced our analysis, they were excluded to maintain comparability across country models. For instance, reports of FGC are common in the southern regions of Nigeria but are substantially less frequent in its northern regions. Recognizing the diversity within countries in western Africa is particularly important for developing interventions and targeting efforts to reduce FGC.

Two countries – Burkina Faso and Mauritania – stand out for having succeeded in reducing FGC and support for this practice. This is evidenced by the fact that the percentages of women who have been circumcised, who report that their daughters have been circumcised, and who believe that FGC should continue have shown steady declines. Burkina Faso, for instance, has established the National Committee to Fight against the Practice of Excision, a government-led entity that seeks to make citizens aware of the dangers of FGC and to ensure that proper law enforcement is in place to convict people who continue the practice. Although several countries have passed legislation banning FGC, Burkina Faso is the only country in which people who break this law are commonly prosecuted.28,29 Mauritania has also experienced consistent declines in FGC and in support for the practice, although less dramatically than Burkina Faso. The government of Mauritania has established the Ministry of Women's Affairs (MCPFEF), which works with governmental groups as well as community and religious leaders to promote FGC awareness campaigns.30 Mauritania also has a law against “harm[ing] the genital organs of a female child". However, prosecution is rare.31 Diligence in prosecution could contribute to the differing rates of decline between Burkina Faso and Mauritania.

Based on the efforts and outcomes in both Burkina Faso and Mauritania, we postulate that four components are necessary for effectively reducing FGC practice and support. These include: (i) community education and awareness, (ii) the use of prominent groups to champion the cause, (iii) the support of FGC practitioners such as nurses, midwives, and traditional healers, and (iv) enforced legislation. First, community education and awareness can enable and facilitate affected communities to promote positive attitudes towards discontinuing the practice.5,32,33 Successful community strategies have been documented, including circumcision-free rite-of-passage ceremonies and collective declarations in which villages pledge not to circumcise their daughters.8 Second, strategies in Burkina Faso and Mauritania have used high-ranking women and a broad array of professionals and other influential people, such as government representatives and religious leaders to champion reform efforts.29,34 Third, the participation of FGC practitioners within each country are central to meaningful implementation of reform strategies by limiting access to FGC services and educating their patients.35,36 Last, enforced legislation is particularly critical; legislation alone, without broad-based political support and enforcement, is likely to be ineffective.37 Such concerted and multifaceted commitments allow for the preservation of a community’s cultural heritage and social values while sustaining attitude shifts away from FGC for the benefit of future generations of women and girls.

Our results suggest subpopulations in each country where efforts could be targeted. For instance, older ages were consistently associated with the practice of FGC. The FGC practices of older women may be a result of past societal norms, including pressure from family members or spouses, even if they did not support the practice. Conversely, younger ages were more likely to believe the practice should continue, possibly because many may not yet have had to consider circumcising their own daughter. When women are faced with the decision to circumcise their own daughters or have already done so, their support of FGC may diminish. Furthermore, these younger women may not have been circumcised themselves and are therefore able to support the practice more easily without having experienced the medical and psychological complications that many of the older women had. Being less educated and Muslim were also generally associated with all three of our outcomes, suggesting that these subpopulations may be important groups for intervention approaches in some countries. Correlates, however, were not always consistent across countries. FGC in Niger and Nigeria, for instance, was not associated with Muslim religion. Additionally, in some countries wealth was associated with higher odds of engaging in the practice, and in others it was associated with lower odds of engaging in the practice. These variations reiterate contextual differences and the importance of having country-specific data to effectively tailor approaches for reducing and eliminating the practice of FGC.

Strengths of this study include its role in filling a large and significant gap in the peer-reviewed literature. Given the risks to both the physical and mental health of girls and women who undergo FGC and the desperate need for more effective strategies to eliminate this practice, our research calls attention to the substantial prevalence in several countries and provides information for affected governments and communities. Furthermore, medical and public health professionals in western Africa can use these results to identify patient subpopulations with whom they may need to address FGC practices and beliefs. Through this cross-country comparison, we were also able to highlight potential approaches for effectively reducing and eliminating this complex and deeply rooted practice. We also generated the data needed to identify subgroups of women at high risk for FGC to ensure that these strategies may be effectively implemented.

Our results should be interpreted in light of some limitations. First, we used only sample characteristics that were available in all countries and those that were largely consistent across countries, limiting our ability to describe practicing communities further. Second, it is possible that FGC-related practices and attitudes may have changed since the time of data collection (2005–2007); however, we used the most recent round of data made available by the MICS for this region. Third, the data collected from Sierra Leone may not be as comparable to other data as we may have wished. Questions for Sierra Leone did not refer to female circumcision or genital cutting, but to initiation into the Bondo Society. This alternative phrasing was deemed most relevant to the practicing culture in Sierra Leone,27 although this measurement was markedly different from that used in other countries. Fourth, responses may be subject to social desirability or recall bias, depending on the cultural context and strategies in place to eliminate FGC. Women may have been motivated to underreport circumcision and their support for the practice, particularly in countries in which legislation exists against such behaviour. Although other approaches, such as medical record reviews or examinations,38 may have been more valid for measuring the prevalence of FGC, they were not feasible in these settings due to additional cost and time. As a result, these self-reported data across several countries were best for meeting our research objectives. Last, our analysis did not consider the type of FGC that had been experienced by the women in our samples. The type generally performed in each country, however, was not markedly different. Most of the countries in our sample perform a combination of Type I (clitoridectomy) and Type II (excision) or Type II only. It has been estimated that approximately 90% of all FGC includes these types where the flesh is either nicked or removed. Three countries in our sample (Gambia, Ghana, and Nigeria), however, perform the most severe form of FGC (Type III, infibulation) in at least some parts of the country. Infibulation, or being sewn shut, may describe up to 10% of all FGC.5,34,39 The countries where this type occurs, and these affected women, may warrant additional and increased efforts.

Although action against FGC must be tempered with an understanding of the deeply rooted traditions that have allowed this practice to continue for so many generations, effective approaches for reducing FGC are critical. Despite widespread efforts, prevalence remains high in many countries, putting millions of girls at risk every year. Successful strategies for eliminating FGC are likely to require multi-pronged approaches in which political, legal and cultural elements are choreographed to effect large-scale change. Such concerted societal commitments are necessary for the benefit of future generations of women and girls.

Competing interests:

None declared.

References

Share
Email Twitter Facebook Google Delicious LinkedIn
More...