Introduction

Female genital mutilation (FGM), also known as female genital cutting or female circumcision, threatens the health and well-being of millions of girls, women and their children across the globe. The United Nations Sustainable Development Goals call for the elimination of the practice by 2030.1

FGM refers to the partial or total removal of the external female genitalia or other injury to the female genital organs for non-therapeutic reasons. The WHO recognises four distinct types of the practice that range in severity from injuries to the genitalia that do not involve the removal of flesh, such as pricking, piercing or scraping (type 4), to infibulation, in which the vaginal opening is narrowed by cutting and reposition the labia to create a partial cover and may involve stitching the tissues together (type 3). Types 1 and 2, also referred to as clitoridectomy and excision, respectively, involve the partial or complete removal of the clitoris and labia.2

FGM is widely practised in some regions of Africa and the Middle East. Motivations for the practice vary but are often tied to social mores that govern the expression of female sexuality. The practice is believed to reduce women’s desire for sex and therefore limit the likelihood of premarital or extramarital sexual activity.3 It may also be considered a marker of group identity that symbolises a girl’s transition to adulthood. There is strong social pressure to continue the practice among societies that believe cutting is necessary for group identity and marriageability.4

Data and statistical methods

The DHS collects internationally comparable information on FGM that permits examination of the prevalence and severity of the practice across countries. Women between 15 and 49 years of age are first asked whether they know of FGM. Those who are familiar are then asked if they have been cut themselves. Women who indicate that they have been cut are asked whether any flesh was removed from their genitals as part of the procedure and, if so, whether their genital area was sewn closed. Those who report that no flesh was removed are asked whether their genitals were ‘nicked’ without flesh removal. These questions permit a rough distinction between the types of FGM recognised by the WHO. Women who reported that flesh was removed from their genitals likely experienced clitoridectomy or excision (type 1 or 2). Those whose genitals were sewn likely underwent infibulation (type 3). Women who reported that they were cut but no flesh was removed or that their genitals were ‘nicked’ likely experienced type 4.

We obtained DHS data from 22 countries that had included questions on FGM in at least one survey since 1990 and whose data were publicly available. Twenty of these countries are located in sub-Saharan African and two in the Middle East. We pooled data from all of the available surveys within each country, as listed in table 1. Although data from each DHS wave are cross-sectional, they include respondents from a wide range of birth cohorts, which permits the measurement of trends over time. We limited our sample to women born between 1965 and 2000 and examined trends over 3-year birth cohorts in each country.

We examined five measures of FGM: the prevalence of all forms of cutting, forms that involve flesh removal, infibulation and ‘nicking’. We found that in some countries a substantial proportion of women reported that they did not know whether flesh had been removed from their genitals and so we also measured trends in the prevalence of ‘don’t know’ responses. We assumed women who reported that they had not heard of FGM had not been cut to any extent.

We estimated trends in the prevalence of these outcomes by calculating the predicted (marginal) probability of each outcome over birth cohorts using logistic regression. Within each country, we regressed dummy variables for each birth cohort on a binary indicator of each outcome and then used the coefficients associated with the dummy variables to predict the probability of each outcome in each birth cohort. We graphed our estimates over birth cohorts between 1965–1967 and 1995–1997, which allowed us to observe trends over the same time period in all of the countries studied, unlike earlier work which presented trends over survey years that differed across countries. We were able to obtain prevalence estimates for each birth cohort in countries that conducted surveys at approximately 5-year intervals through recent history, but were more restricted in countries that fielded surveys at irregular intervals or that had not conducted one recently. In a few countries, we were able to obtain an estimate for the 1998–2000 cohort as well. All of the individual birth years within each 3-year cohort are represented by at least 50 women in order to ensure that our estimates are reasonably representative of the entire 3-year cohort. We quantified the magnitude of change in each of the outcomes over time by subtracting the prevalence among women born in 1965–1967 from that among women born in the most recent cohort for which data were available. All estimates are weighted using denormalised sampling weights following guidelines for the use of pooled data included in the DHS Sampling and Household Listing Manual.16

All of our outcomes are self-reported. Ongoing legislative and programmatic efforts to eliminate FGM may influence women’s reporting behaviour and could lead to biased trend estimates. In countries that have banned the practice, women may be hesitant to report their status or the severity of the procedure. For example, women in northern Ghana who were interviewed about their FGM status in 1995 and again in 2000 gave inconsistent reports: 13% of women who reported they had been cut in 1995 reported that they had not been 5 years later after the country outlawed the practice.17 We looked for evidence of systematic reporting bias in six countries in which the prevalence of all forms of FGM had declined by 15 percentage points or more over a 30-year period. To do this, we compared prevalence estimates for the same birth cohorts obtained from successive surveys using the method described above. If bias affected our estimates, we would expect to see inconsistent measures of prevalence across survey waves, with lower estimates among more recent surveys.

Results

All forms of cutting

The proportion of women who experienced any type of genital cutting has fallen over time in almost all of the countries we examined, but it remains a pervasive problem in many of them (figures 1–4, table 2). Despite small declines, the practice is still nearly universal in Guinea, Mali and Egypt, where 90% or more of women born between 1995 and 1997 reported undergoing some form of the procedure. More than half of women continue to be cut in Burkina Faso, Sierra Leone and the Gambia, where rates of cutting have not changed in 30 years (table 2). In contrast, fewer than 5% of women born in the most recent cohort reported being cut in Benin, Niger, Tanzania, Togo and Uganda.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Trends in the prevalence and severity of female genital mutilation in Western African countries.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Trends in the prevalence and severity of female genital mutilation in Eastern African countries.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Trends in the prevalence and severity of female genital mutilation in Central and Southern African countries.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Trends in the prevalence of female genital mutilation in Egypt and Yemen.

View this table:

• View inline
• View popup

Table 2

Changes in the prevalence of various forms of FGM  over birth cohorts

Discussion

The prevalence of FGM is falling in most of the countries that we assessed. The practice has been declining slowly and steadily for 30 years in many countries including Cote d’Ivoire, Nigeria, Ethiopia and Kenya, while reductions appear more recent in Chad and Sierra Leone. However, progress has not been uniform: the prevalence of FGM has remained high and stable in Mali and the Gambia for 30 years.

We found little evidence of a substantial shift in the severity of the procedures performed, although symbolic forms of FGM appear to be making small inroads in a handful of countries. The vast majority of women who are cut likely experience partial or total removal of the clitoris and labia, corresponding to WHO types 1 or 2. Infibulation continues to be practised in many countries and rates may be rising in Mali and Sierra Leone. These are serious injuries that hinder women’s health and sexual functioning throughout their lives. Estimates of flesh removal should be treated as lower bounds given the substantial proportion of women who reported that they did not know whether flesh had been removed from their genitals.

Our estimates assume that women correctly report their FGM status and are aware of the extent of their injuries. Validation studies have been conducted in clinical settings but may capture highly selected samples in countries where healthcare is not equally accessible across all population groups. However, validation studies outside of clinical settings would be ethically challenging to conduct as they would require the direct examination of women’s genitalia, and are therefore unlikely to be pursued. One study conducted in Sudan compared women’s self-reports of the extent of their cutting with classification based on clinical exams and found that as many as 25% of women did not accurately report the type of FGM they had experienced.18 The women were slightly more likely to under-report the extent of the injuries. A tendency to under-report FGM status was also found in Tanzania, providing further support for treating estimates of flesh removal and infibulation as lower bounds.19 The vast majority of women in a Nigerian study accurately reported their FGM status.20

The specific drivers of the decline in FGM are uncertain. Despite decades of efforts to stem the practice, there is very little high-quality evidence on the types of interventions that may be effective at preventing it. Many countries have instituted bans on FGM, but such policies may be difficult to enforce where the practice continues to have wide social support.21 This appears to be the case in Egypt, where multiple attempts to ban the practice have resulted in little progress towards reducing its prevalence.22 Aside from legislative efforts, a wide variety of programmes that encourage communities to abandon FGM traditions have been fielded. Unfortunately, few have been evaluated. A 2006 report by the Population Reference Bureau catalogued 92 projects aimed at abandoning FGM in 19 countries. Only four had been evaluated using controlled before-and-after designs.23 A more recent systematic review identified a total of eight impact evaluation studies conducted in sub-Saharan Africa.24 Of these, only two measured the effect of interventions on the prevalence of FGM. Both studies measured the effect of community-based educational programmes for adults and reported a reduction in the number of young girls who had been cut.24 25 Population-level improvements in educational attainment and increasing urbanisation may also contribute to changing attitudes towards these practices, although these relationships may vary across and within countries. Further investigation is warranted to better understand what works to prevent FGM and how to mitigate the harms associated with it.

We present national-level estimates of the prevalence of FGM in 22 countries but recognise that both prevalence and trends vary substantially across subgroups within many countries. In some, it is practised primarily within certain ethnic groups. Estimates may also differ according to educational attainment, religious identity or area of residence.14 A full examination of trends among each of these categories in every country was beyond the scope of this analysis but may assist in the targeting of programmes related to FGM. Research is also needed to understand the extent of FGM in countries that have not collected nationally representative data on the topic. There is evidence of FGM customs in Indonesia, Malaysia and India, among others.26–28 Immigrant communities may also continue FGM in countries where it has not been traditionally practised. For example, two recent criminal prosecutions have drawn attention to the practice within the USA.29 30

Slow progress is being made toward reducing the prevalence of FGM but the practice remains nearly universal in some countries, irreparably damaging the health and well-being of hundreds of millions of women. There is little high-quality evidence of what works to prevent girls from undergoing these procedures. Studies suggest that community-based educational interventions may shift attitudes toward the practice, but it remains unclear whether changes in attitudes translate into reductions in FGM or whether such programmes could be scaled up to affect change on a national level. Well-planned and documented interventions and rigorous evaluations are necessary to advance efforts to curtail FGM and the harms associated with it, especially as the 2030 deadline for the Sustainable Development Goals draws nearer.