Cardiovascular disease has become the leading cause of death in sub-Saharan Africa (1). Increasing urbanization and associated lifestyle changes may induce risk factors, such as obesity, for cardiovascular disease (2,3). In sub-Saharan Africa, the rise in the prevalence of obesity in urban areas has coincided with the growth of an increasingly educated and wealthier middle class that engages in less physical activity and consumes greater amounts of calorie-dense foods than the poorer segments of the population (3,4). However, the change in lifestyle in urban sub-Saharan Africa also profoundly affects the urban poor, and the prevalence of risk factors for cardiovascular disease is high in resource-deprived city slums (4,5). Although studies show that obesity is increasing among the urban poor (4), data are limited on the extent of the problem among city slum populations in different countries in sub-Saharan Africa.

Cultural ideals influence how people assess their body image and body weight (6,7). Studies in sub-Saharan Africa suggest that the poor tend to perceive body size and its health implications differently from people in the same cities with greater wealth and education (8). In some societies in sub-Saharan Africa, a larger body size is commonly assumed to reflect good health and higher social status and may thus be considered desirable (8,9). In addition to examining culturally desired body sizes among the poor in sub-Saharan Africa, it is also relevant to learn how people perceive their own body size and weight, because underestimation of body mass index (BMI) is prevalent in low-income settings (10) and may predict overweight or obesity (11). Although some of the factors that influence perception of body image have been studied in developed countries, this topic remains largely unexplored in sub-Saharan Africa. Body image perceptions and preferences differ by sex in the United States (12,13), but the effects of age and marital status are unclear (14,15). The objective of this study was to determine the prevalence of overweight and obesity and examine perceptions of body size differentiated by sex and other determinants among slum dwellers in Nairobi, Kenya.

Of the 5,190 people recruited into the main study, 4,934 people (comprising 2,669 [54%] men and 2,265 [46%] women) had the data for the required variables and were included in the analysis. The mean age of the respondents was 42 years: 45% of respondents were younger than 40 years, and 55% were aged 40 years or older. The proportion of the respondents who had primary education or less was 75%, and approximately 2.5% had postsecondary education.

Overall, 43.4% of women and 17.3% of men in the study population were overweight or obese (Table 1). Obesity was more prevalent among women (15.5%) than among men (2.3%). Approximately half (50.5%) of women rated their current body image as normal weight; although 14.2% rated their current body image as underweight, only 5.1% were underweight according to BMI. Among men, 52.7% rated their current body image as normal weight; according to BMI, 72.9% were normal weight. The proportions of men who perceived themselves to be overweight (20.8%) or obese (13.4%) were greater than the proportions who were overweight or obese according to BMI (15.0% and 2.3%, respectively). Among women, the distribution of categories for ideal body image was not substantially different from the distribution of categories for BMI. For example, 14.8% of women chose an obese image as ideal, and 15.5% of women were obese. These distributions, however, were different among men: 20.6% of men chose an obese image as ideal, whereas only 2.3% of men were obese according to BMI.

In all BMI categories, more than one-third of women preferred body images classified as overweight or obese, and the proportion of men with a preference for overweight or obese body images was greater than that of women (Figure 2). More than half of women and men classified as overweight or obese indicated a preference for overweight or obesity.

Among women, 18.2% overestimated their BMI; 33.0% of men overestimated their BMI (Table 2).Underestimation was more prevalent among women (34.6%) than among men (16.9%). Obese women accounted for 28.8% of total underestimation among women, whereas obese men accounted for only 7.8% of total underestimation among men.

Of 1,444 participants classified as overweight or obese, 776 underestimated their BMI. Of the key sociodemographic characteristics of the participants, sex was a significant predictor of underestimation of BMI among normal-weight participants, whereas age and sex were significant predictors among overweight or obese participants (Table 3). Overweight or obese participants aged 50 years or older were less likely than their younger counterparts to underestimate their BMI. Similarly, men were less likely than women to underestimate their BMI in both age categories. Marital status and education were not significantly associated with the likelihood of BMI underestimation.

Our study was designed to investigate the prevalence of overweight and obesity, the extent of disagreement between perceived body image and BMI, and the patterns of preferred body size in 2 slums in Nairobi. The prevalence of overweight or obesity among women in these urban slums was similar to the prevalence reported in a national survey of urban women in Kenya (4), and the sex pattern shown in our study confirms the patterns reported in other studies in Africa (21,22). In the context of poverty and resource deprivation in slums, overweight and obesity are a challenge and help to drive the increasing rates of morbidity associated with cardiovascular disease in these urban settlements (23–25).

A major issue associated with overweight and obesity is the perception of body size in different cultural settings (26,27). Our study used body images or silhouettes to estimate self-reported body size. Differences between objective measures of body size (ie, BMI) and individual perception of current body image suggest individual or societal definitions of desirable or “normal” body size that are not based on medical facts. Our study found that more than half of the population who were overweight or obese underestimated their weight, suggesting that most of these people do not acknowledge their extra weight and the related health risks.

We also determined the ideal body image for each survey participant and how it related to their actual BMI category. Regardless of BMI category, substantial proportions of women and men indicated a desire to have a body size that was larger than their actual size. More than half of the participants who were overweight or obese chose an ideal body image that was either overweight or obese. Families, communities, and society strongly influence perceptions of body image, and excess body weight has generally been associated with wealth and health and considered a desirable attribute in many parts of sub-Saharan Africa (6,8,22). The implications of excess body weight as a risk factor for noncommunicable diseases are not readily apparent to most residents of poor settlements because of low levels of education and the lack of emphasis on noncommunicable diseases in traditional health education and counseling at the community level in sub-Saharan Africa.

In our focus on overweight and obese participants, we sought to identify sociodemographic characteristics associated with underestimation of BMI. More than half (54%) of this subgroup underestimated their BMI, and age and sex were the only significant predictors of underestimation. These results are consistent with reports of greater dissatisfaction with body image among younger adults and women (28). Although studies have found that education and income are associated with the likelihood of obesity in urban sub-Saharan Africa (4), we could not find reports on the determinants of underestimation of BMI in this high-risk subgroup.

This study has several limitations. One is the use of a discrete number of silhouettes to assess body image. By using this method, body size is considered a continuous variable measured from a finite number of images (29). The amount of information lost by using this method can be reduced by increasing the precision of the instrument. In our study, we attempted to reduce the coarseness of the scale by using an 18-silhouette scale developed and validated in a previous study (30). Other limitations of the study include the challenge of establishing the influence of cultural preferences and beliefs on body image in slum settlements comprising diverse ethnic groups and the extent to which these 2 slums are representative of other slums in Nairobi. A strength of the study is the participation of a large number of residents of an informal urban settlement in Kenya. Although a large proportion of the urban population in sub-Saharan Africa lives in slums, data are limited on the increasingly relevant topic of obesity and weight-related perceptions among this growing group.

Our study has important implications for the effort to reduce the prevalence of risk factors for cardiovascular disease in poor urban settings in Kenya and the rest of sub-Saharan Africa. As the epidemiological transition continues among low- and middle-income countries, the prevalence of obesity also increases, even among low-income groups. The high levels of overweight and obesity reported in our study — despite the high levels of poverty found in the slums we studied — suggest that effective dietary and lifestyle interventions are needed to address this issue in slums. Innovative strategies to ensure better appreciation and understanding of the health consequences of overweight and obesity by slum residents are required if efforts to reverse the greater preference for larger body size are to be successful.