unlimited access with print and download

Free

Continue searching

The social epidemiology of maternal obesity in Egypt

Dissertation
Author: Petra Nahmias
Abstract:
Obesity is emerging as one of the leading public health challenges in low and middle income countries. In particular, women of reproductive age are vulnerable to many compromised reproductive health outcomes associated with obesity. Egypt is an especially interesting country to study having experienced a rapid rise in obesity, with nearly half of women of reproductive age obese in 2005, exceeding levels of obesity seen in many high income countries. Despite the importance of obesity and its implications for health in developing countries, the subject has not received sufficient research interest; this dissertation contributes to addressing this deficiency. The dissertation is comprised of three empirical chapters all using Egyptian Demographic and Health Surveys from 1992 to 2005. The first uses factor and multilevel analysis to analyze the variables used to measure female empowerment. The findings highlight the difficulty in measuring female empowerment in a meaningful way, with questions around both the reliability and the validity of the data. In the second chapter, I conduct an analysis of the temporal changes in the relationship between maternal obesity and social determinants, using both recursive partitioning and logistic regression. The findings show that not only are Egyptian women becoming more obese but that the increase in obesity has disproportionately affected the most deprived: those with the least education, the poorest, the rural population, and those living in Upper Egypt. Finally, I look at the relationship between maternal obesity and maternal and child health outcomes, and at the mediating effect of socioeconomic status, using Cox proportional hazards and logistic regression models. The findings show that for some outcomes, there is a mediating effect of SES and that this relationship is also changing over time.

Table of Contents

Abstract ........................................................................................................................................................ i Table of Contents ........................................................................................................................................ ii Tables ......................................................................................................................................................... vi Charts ........................................................................................................................................................ vii Chapter One: Introduction .......................................................................................................................... 1 Chapter Two: Data ...................................................................................................................................... 5 Sample restrictions and concerns ........................................................................................................ 6 Sample characteristics ......................................................................................................................... 7 Using Body Mass Index ...................................................................................................................... 8 Chapter Three: Background ........................................................................................................................ 9 Introduction ............................................................................................................................................. 9 Changes in overweight and obesity between 1992 and 2005. .............................................................. 10 Age and cohort effects .......................................................................................................................... 11 Geographic difference ........................................................................................................................... 12 Socioeconomic characteristics .............................................................................................................. 15 Conclusion ............................................................................................................................................ 17 Chapter Four: Using DHS to measure female autonomy and empowerment in Egypt ............................ 19 Introduction ........................................................................................................................................... 19 Defining Female Autonomy ................................................................................................................. 21 Measuring Autonomy: problems and challenges .................................................................................. 22 The Situation in Egypt .......................................................................................................................... 25 Measuring Female Autonomy in Egypt ................................................................................................ 27 Validity of Female Autonomy .............................................................................................................. 33 Bodily integrity autonomy ................................................................................................................ 35 Decision making autonomy .............................................................................................................. 38 Knowledge autonomy ....................................................................................................................... 40 Socioeconomic autonomy ................................................................................................................. 42 Emotional autonomy ......................................................................................................................... 43 Socioeconomic and emotional autonomy ......................................................................................... 44 Physical autonomy ............................................................................................................................ 45 Reactivity of responses ......................................................................................................................... 50 Discussion and conclusion .................................................................................................................... 57 Chapter Five: The changing relationship between maternal obesity and social determinants in Egypt between 1992 and 2005 ............................................................................................................................ 60 Introduction ........................................................................................................................................... 60 Obesity: a challenge for reproductive health ........................................................................................ 61 Obesity and Social Determinants .......................................................................................................... 64 Female obesity in the Middle East ........................................................................................................ 65 Obesity in Egypt ................................................................................................................................... 67 Conceptual framework and Hypotheses ............................................................................................... 72 Methods................................................................................................................................................. 74 Results ................................................................................................................................................... 75 Discussion ............................................................................................................................................. 91 Conclusions ........................................................................................................................................... 96 Chapter Six: Obesity, Socioeconomic Status, and Maternal and Child Health in Egypt ....................... 103 Introduction ......................................................................................................................................... 103

v

Background ......................................................................................................................................... 104

Maternal obesity and complications during labor and delivery ...................................................... 104 Maternal obesity and Cesareans ...................................................................................................... 105 Maternal obesity and infant mortality ............................................................................................. 105 Infant mortality and SES ................................................................................................................. 106 Cesareans in Egypt .......................................................................................................................... 108 Labor and delivery complications and SES .................................................................................... 110 Obesity and SES ............................................................................................................................. 112 Hypotheses .......................................................................................................................................... 113 Methods............................................................................................................................................... 114 Outcome variables .......................................................................................................................... 114 Analysis........................................................................................................................................... 115 Results ................................................................................................................................................. 116 Discussion ........................................................................................................................................... 140 Conclusion .......................................................................................................................................... 144 Chapter Seven: Conclusion ..................................................................................................................... 147 Summary ......................................................................................................................................... 147 Further research .............................................................................................................................. 148 References ............................................................................................................................................... 150 Appendix ................................................................................................................................................. 160

vi

Tables Table 2.1: Survey details, 1992-2005 ......................................................................................................... 5

Table 2.2: Sample characteristics, mothers of young children, 1992-2005, percentages ........................... 7 Table 3.1: Mean BMI by cohort and age group ........................................................................................ 12 Table 3.2: Prevalence of obesity by socio-demographic characteristics (weighted), percentages (95% confidence interval in brackets) ................................................................................................................ 17 Table 4.1: Variables used to measure female autonomy and dimension measured. ................................. 29 Table 4.2: Justification of proxy measurements of female autonomy ...................................................... 30 Table 4.3: Weighted distribution of variables used to measure female autonomy ................................... 31 Table 4.4: Variables used .......................................................................................................................... 46 Table 4.5: Summary of results of one dimensional models ...................................................................... 48 Table 4.6: Intraclass coefficients (percentages) ........................................................................................ 52 Table 4.7: Coefficients of interviewer determined level of cooperation, random effects regression analysis. ..................................................................................................................................................... 54 Table 4.8: Coefficients of presence of people during interview, random effects regression analysis, EDHS 2005 ............................................................................................................................................... 56 Table 5.1: Mean BMI of ever married women in Egypt by age group, 1992-2005 ................................. 69 Table 5.2: Descriptive characteristics (unweighted), percentages unless otherwise stated* .................... 77 Table 5.3: Bivariate analysis ..................................................................................................................... 78 Table 5.4: Logistic regression nested models, odds ratio of obesity (SE in brackets) ............................. 80 Table 5.5: Two way interaction models .................................................................................................... 82 Table 5.6: CHAID terminal nodes ............................................................................................................ 87 Table 6.1: Percentage of obese mothers by socioeconomic characteristics and year (95% confidence interval in parentheses, mothers with birth in past five years) ............................................................... 113 Table 6.2: Descriptive statistics (percentages) ........................................................................................ 116 Table 6.3: Percentage of women with births in past five years with poor maternal and infant health outcomes by BMI category ..................................................................................................................... 117 Table 6.4: Cox proportional hazards, risk of infant death in first year ................................................... 118 Table 6.5: Cox proportional hazards, risk of infant death in first year controlling for infant‘s birthweight, 1995 only ................................................................................................................................................ 120 Table 6.6: Cox proportional hazards, risk of infant death in first year controlling for infant‘s birthweight, 1995 only ................................................................................................................................................ 121 Table 6.7: Cox proportional hazards, risk of infant death in first month, all years and 2005 only ........ 122 Table 6.8: Logistic regression, odds ratio that last birth was a Cesarean ............................................... 123 Table 6.9: Logistic regression, odds ratio that last birth was a Cesarean by birth order ........................ 127 Table 6.10: Logistic regression, odds ratio that respondent experienced prolonged labor at last birth .. 130 Table 6.11: Logistic regression, odds ratio that respondent experienced excessive bleeding at last birth ................................................................................................................................................................. 132 Table 6.12: Logistic regression, odds ratio that respondent experienced fever or vaginal infection at last birth ......................................................................................................................................................... 134 Table 6.13: Summary of findings and support for hypotheses ............................................................... 140 Table A1: Definitions of autonomy for interviewer and presence of other people effect ...................... 160

vii

Charts

Chart 3.1: Distribution of BMI categories, 1992-2005 ............................................................................. 11 Chart 3.2: Changes in obesity by governorate, 1992-2005 ....................................................................... 13 Chart 3.3: Prevalence of obesity by region, 1992-2005............................................................................ 14 Chart 3.4: Prevalence of obesity by type of place of residence, 1992-2005 ............................................. 15 Chart 4.1: Models of bodily integrity autonomy ...................................................................................... 37 Chart 4.2: Models of decision making autonomy ..................................................................................... 39 Chart 4.3: Models of knowledge autonomy .............................................................................................. 41 Chart 4.4: Models of socioeconomic autonomy ....................................................................................... 42 Chart 4.5: Models of emotional autonomy ............................................................................................... 43 Chart 4.6: Models of socioeconomic and emotional autonomy ............................................................... 44 Chart 4.7: Models of physical autonomy .................................................................................................. 45 Chart 5.1: Pathways through which maternal obesity is associated with poor maternal and infant health outcomes ................................................................................................................................................... 63 Chart 5.2: Weight distribution of women of reproductive age (15-49), 1992 and 2005 .......................... 68 Chart 5.3: GDP per capita, 1980-2005...................................................................................................... 70 Chart 5.4: Conceptual framework ............................................................................................................. 73 Chart 5.5: Chaid Analysis Pathways ......................................................................................................... 99 Chart 5.6: Predicted probability of obesity by wealth, age and education ............................................. 100 Chart 5.7: Predicted probability of obesity by wealth, year and region ................................................. 101 Chart 5.8 Predicted probability of obesity by wealth, year and education ............................................. 102 Chart 6.1: Infant mortality by year and rural/urban residence in Egypt, 1992-2005 .............................. 108 Chart 6.2: Percentage of last births by Cesarean in Egypt, 1992-2005 .................................................. 110 Chart 6.3: Percentage of last births with complications of labor and delivery in Egypt, 1995 and 2000111 Chart 6.4 Predicted probability of Cesarean section by year and maternal weight ................................ 128 Chart 6.5: Predicted probability of Cesarean (with 95% confidence intervals) by maternal weight, year and place of residence ............................................................................................................................. 129 Chart 6.6: Predicted probability of excessive bleeding by BMI and region of residence ...................... 137 Chart 6.7: Predicted probability of excessive bleeding by BMI and education level ............................. 137 Chart 6.8: Predicted probability of prolonged labor by BMI and region of residence ........................... 138 Chart 6.9: Predicted probability of prolonged labor by BMI and education .......................................... 139

Chapter One: Introduction

Initially, I intended to write a dissertation about maternal health in general in Egypt, linking it with women‘s status and female empowerment. For a number of reasons, my dissertation has taken a different path. Originally, obesity was only one of a number of maternal health issues that I wished to study. However, once the extent, the dramatic increase and the importance of maternal obesity in Egypt became apparent, my committee encouraged me to focus only on this one issue. Additionally, when analyzing the operationalization, validity and reliability of measurements of female empowerment, I discovered that measuring female empowerment in a meaningful way would require more sophisticated and better quality data than available to me which led to me broaden the scope of the determinants of obesity.

Obesity is important because it has many deleterious effects for women of reproductive age. In the first place, obese women are more likely to encounter problems becoming pregnant (Hartz et al. 1979; Normal and Clark 1998; Clark et al. 1998) and obese women are more likely to miscarry during early pregnancy (Lashen 2004). There is a large literature demonstrating women who are obese are at greater risk of developing pregnancy complications and problems associated with labor and delivery. Finally, obese women are more at risk of postpartum complications such as infections, hemorrhage and embolisms. Taken all together, maternal mortality and morbidity is significantly elevated for obese women (Cedergen et al. 2004; Pathi et al. 2006; Robinson et al. 2005). Maternal obesity is also dangerous for the fetus, who is more likely to have congenital anomalies, and the newborn who has a significantly higher perinatal morbidity rate (Fretts 2005; Ramsay et al. 2006). The effects of maternal

2

obesity do not stop at birth - babies born to obese mothers are at greater risk of later developing diabetes, cardiovascular disease and obesity themselves (Ramsay et al. 2006).

Egypt, and the countries of the Middle East in general, are typical of many middle income developing countries in experiencing a rapid rise in the prevalence of obesity. Obesity in the Eastern Mediterranean Region has reached ‗alarming levels‘ according to the World Health Organization (Musaiger 2004). The same W.H.O. study pointed out that, unlike Europe and North America, obesity is more prevalent among women and in urban areas. Data from the Demographic and Health Surveys show in 1992 mothers with young children had a mean body mass index (BMI) of 26.9. By 2005, obesity levels had risen to a mean BMI of 30.1, with nearly half of Egyptian women of reproductive age classified as obese. Clearly, Egypt is facing extraordinary changes in the prevalence of obesity in a comparatively short period of time.

Most studies focus on the experience of obesity in developed countries, with the emergence of obesity in developing countries a relatively new and poorly understood phenomenon. Despite the challenges facing many middle income developing countries such as Egypt, research on obesity is lacking, partly reflecting the recent development of the obesity problem in these countries. The International Journal of Epidemiology in 2006 devoted an entire issue to the global pandemic of obesity, with many authors lamenting the lack of attention being accorded to the problem of obesity in the developing world (Prentice, 2006; Kim and Popkin, 2006). Clearly research on the sociology of obesity in the developing world in general and Egypt in particular is needed and this dissertation seeks to, at least partly, address this deficiency

3

Following this introduction, Chapter Two describes the Egypt Demographic and Health Survey data used in the dissertation. Chapter Three lays out the background, describing the socio-demographic, temporal and geographic nature of maternal obesity in Egypt. In Chapter Three, I demonstrate the extraordinary increases in maternal obesity that Egypt has experienced over a relatively short period of time, with different groups responding in different ways.

Chapter Four analyzes the variables used to measure female empowerment – a concept I had initially intended to study in relation to maternal obesity. The findings highlight the difficulty in measuring female autonomy in a meaningful way, with questions around both the reliability and the validity of the data. Although I specifically looked at Demographic and Health surveys in Egypt, the findings suggest that similar problems could be found in other surveys and settings.

In Chapter Five, I conduct an analysis of the temporal changes in the relationship between female obesity and social determinants. Social factors are crucial when studying obesity and important for targeting effective interventions as obesity, operating through both food intake and energy expenditure, is socially influenced. However, the relationship between obesity and society will vary depending on the type of society and the level of obesity. This relationship can also be expected to vary over time as both obesity rates change and societies evolve. Egypt is a particularly interesting country for studying this relationship, having experienced a sharp rise in obesity, coupled with rapid social change. I hypothesize that the relationship between female obesity and social determinants has changed over time concurrent with changing social conditions and increasing obesity. The main question addressed in this chapter is: as obesity increases among women in Egypt, how is the sociology of obesity changing? Chapter Four shows that not only are Egyptian women becoming more obese but that the increase in obesity has

4

disproportionately affected the most deprived: those with the least education, the poorest, the rural population, and those living in Upper Egypt.

Finally, Chapter Six looks at the relationship between maternal obesity and maternal and child health outcomes, and at the mediating effect of socioeconomic status (SES). In this chapter, I analyze the relationship between maternal obesity, SES and infant mortality, Cesarean section and pregnancy complications. Egypt is currently facing many challenges in addressing maternal and child health, and there is much need that is not being currently met (Stanton and Langsten 2000). For example, in 2005, a third of births in rural areas were not attended by trained medical personnel (El-Zanaty and Way 2006). Increasing maternal obesity will increase the burden on the health services – as has already been shown in developed countries where obese pregnant women use more health care and incur more costs than non obese women (Chu et al. 2008; Heslehurst et al. 2007; Pathi et al. 2006): The questions addressed in this chapter are: who are the women who will bear the burden of increasing obesity levels? Will obesity affect all mothers equally? In short, the chapter tackles a previously unaddressed issue: whether the effect of maternal obesity on maternal and infant health outcomes is the same for women of different SES. Chapter Six demonstrates that for some outcomes, there is a mediating effect of SES and that this relationship is also changing over time.

5

Chapter Two: Data

The data used in the dissertation are from the Demographic and Health Surveys conducted in Egypt in 1992, 1995, 2000 and 2005. These are large surveys carried out periodically and intended to be uniform to facilitate cross-national comparisons; they are generally representative at the regional level. The surveys provide estimates for key indicators such as fertility, contraceptive use, infant and child mortality, immunization levels, coverage of antenatal and delivery care, nutrition, and anthropometric measurements.

The methodology is similar in each survey, with a three stage sampling process randomly selecting households to be interviewed for the household questionnaire. All ever-married women aged 15-49 who are usual residents or who were present in the sampled households on the night before the interview were eligible for the women‘s questionnaire. The EDHS have very high response rates. The refusal rate is even lower, since a large proportion of the non-responders were women who were not located by the interviewers Table 2.1 presents details on the surveys used. Table 2.1: Survey details, 1992-2005

Year

1992

1995

2000

2005

Household

Number

10,760

15,567

16,957

21,972

Response rate (%)

98.3

99.2

99.1

98.9

Women

Number

9,864

14,779

15,573

19,474

Response rate (%)

98.9

99.3

99.5

99.5

Overal l response rate

97.2

98.5

98.6

98.4

6

Sample restrictions and concerns The sample used is restricted to women who had had a birth in the past five years since in 1992 and 1995 only mothers with children aged five and under in the household were sampled for anthropometric measurements. Therefore the sample is not representative of all Egyptian women of reproductive age, with younger and older women under-represented. However, comparing the BMI of ever married women and mothers with children aged under five in the household, women in the more inclusive sample have an even higher mean BMI than women with a young child. Mendez, Monteiro and Popkin also identified the restriction to women of young children as a limitation to the use of DHS for obesity studies (Mendez et al. 2005). They showed that while parity is associated with weight gain, this is only significant for women who were overweight pre-parity and since mean BMI is higher for all ever- married women in 2000 and 2005, the results should be conservative although care should be taken if extrapolating to the general Egyptian population.

Another concern is that the 1992 sample did not sample in the frontier governorates (Red Sea, North Sinai, South Sinai, Matruh and New Valley). While these governorates are sparsely populated, they have seen the most extreme changes in the prevalence of obesity and overweight. Although studying the changes in these regions is valuable due to the tremendous increase in obesity, the population of these regions is only around one percent of the weighted sample population. Finally, while marriage is still nearly universal in Egypt, the age of marriage is increasing. The increase in age of marriage means younger married women are becoming a more select group, so changes in obesity for young women over time need to be considered in that context.

7

Sample characteristics The sample characteristics between 1992 and 2005 are surprisingly similar. As expected, there was an increase in the educational level of Egyptian women. In 1992, 47 percent of mothers of young children had no education. By 2005, this had decreased to 31 percent. At the same time, the proportion with higher education doubled from nearly five percent to more than ten percent. There is a slight trend to a reduction in the proportion of older mothers of young children. This trend is consistent with fertility decline in the Middle East which has tended to be led by older women stopping childbearing (Fargues 2000; Khawaja 2000). The proportion of women aged 35 and older has declined from 27 percent in 1992 to 22 percent in 2005. Table 2.2: Sample characteristics, mothers of young children, 1992-2005, percentages

Survey year

1992

1995

2000

2005

Religion

Muslim

94.4

95.2

95.1

Christian

5.6

4.8

4.9

Age

15 - 19

3 .5

3.9

3.2

3.5

20 - 24

17.0

18.8

19.8

21.4

25 - 29

29.0

28.2

29.0

30.2

30 - 34

23.3

23.0

23.5

22.5

35 - 39

16.4

16.1

16.3

14.8

40 - 44

8.6

7.7

6.3

6.3

45 - 49

2.3

2.4

1.9

1.4

Education

None

47.0

43.2

38.6

31.0

Primary

25. 0

21.9

16.2

13.1

Secondary

23.1

28.8

36.4

45.5

Higher

4.9

6.0

8.9

10.4

Children ever born

1 - 2

33.9

37.2

42.6

48.1

3 - 4

31.6

31.8

33.0

34.5

5+

34.5

31.1

24.4

17.5

Age at first marriage

14 and under

13.0

13.6

9.5

6.9

15 - 16

19.0

20.5

17.6

14.5

17 - 18

22.1

20.5

20.9

21.9

19 - 20

17.6

16.1

20.3

20.7

21+

28.4

29.2

31.6

35.9

Currently working

21.6

18.1

15.3

18.0

Electricity

93.1

95.4

97.7

Radio

62.9

65.8

82.2

8

TV

52.5

52.3

63.5

Region

Urb an governorates

20.8

20.1

17.3

15.1

Lower Egypt Urban

11.4

10.6

12.2

10.2

Lower Egypt Rural

30.1

29.0

31.1

31.4

Upper Egypt Urban

10.4

11.4

11.4

12.6

Upper Egypt Rural

27.3

27.9

26.5

29.6

Frontier

0.9

1.5

1.2

Using Body Mass Index A key variable in the dissertation, both as an outcome and an explanatory variable, is maternal Body Mass Index (BMI), measured at the time of the survey. The BMI is calculated by the mass (in kilograms) divided by the square of height (in centimeters). While BMI is widely used as a diagnostic tool for obesity, it is not a perfect measure. In particular, BMI is not sensitive to weight due to adiposity and weight due to muscle mass. Further, it takes no account of frame size, body shape and where on the body fat is accumulated – an important determinant in the effect of obesity on health. These problems pertain more to BMI as a medical diagnostic tool for individuals. As a statistical categorization, it is still useful for determining obesity since there is a strong correlation between BMI and other, more precise, measures of obesity. It is also readily available and the measurements are reliable and easy to take, based solely on height and weight.

Although I am interested in pre-pregnancy BMI, it is fair to assume that post-pregnancy BMI will be strongly associated with pre-pregnancy BMI. The most significant issue is that obese women are more likely to gain more weight during pregnancy. Both weight gain during pregnancy, as well as pre- pregnancy weight, are independently and positively associated with adverse pregnancy outcomes (Jain et al. 2007; Cedergren 2006), so the effects may be conflated. However, other research among a non- European (Japanese) population found that excess gestational weight gain was not associated with poor outcomes (Murakami et al. 2005). It is likely that there is some exaggeration of the effect of BMI on maternal outcomes but the variable is still useful in this dissertation in highlighting areas of concern that require further research.

9

Chapter Three: Background

INTRODUCTION Obesity has leads to various compromised health outcomes for women of reproductive age. Firstly, obese women are more likely to experience infertility (Hartz 1979; Norman and Clark 1998; Clark et al. 1998) , and obese women are also more likely to miscarry during early pregnancy (Lashen et al. 2004). It has been established that women who are obese are at greater risk of developing pregnancy complications and problems associated with labor and delivery. Finally, obese women are more at risk of postpartum complications such as infections, hemorrhage and embolisms. Looking at increased risk across pregnancy, childbirth and postpartum, maternal mortality and morbidity is significantly elevated for obese women (Cedergen 2004; Pathi et al. 2006; Robinson et al. 2005). Maternal obesity is also dangerous for the fetus, who is more likely to have congenital anomalies, and the newborn who has a significantly higher perinatal morbidity rate (Fretts 2005; Ramsay et al. 2006). The effects of maternal obesity do not stop at birth - babies born to obese mothers are at greater risk of later developing diabetes, cardiovascular disease and obesity themselves (Ramsay et al. 2006).

Egypt, and the countries of the Middle East in general, are typical of many middle income developing countries in experiencing a rapid rise in the prevalence of obesity. Obesity in the Eastern Mediterranean Region has reached ‗alarming levels‘ according to the World Health Organization (Musaiger 2004). The same W.H.O. study pointed out that, unlike Europe and North America, obesity is more prevalent among women and in urban areas. Data from the Demographic and Health Surveys show in 1992 mothers with young children had a mean BMI of 26.9. By 2005, obesity levels had risen to a mean BMI

10

of 30.1, with nearly half of Egyptian women of reproductive age classified as obese. Clearly, Egypt is facing extraordinary changes in the prevalence of obesity in a comparatively short period of time. Most studies focus on the experience of obesity in developed countries, with the emergence of obesity in developing countries a relatively new and poorly understood phenomenon. This paper addresses the lack of research on obesity in middle-income developing countries, and describes the socio-demographic, temporal and geographic nature of obesity in Egypt.

CHANGES IN OVERWEIGHT AND OBESITY BETWEEN 1992 AND 2005. In just thirteen years, Egypt has undergone a significant increase in the prevalence of obesity. The average BMI among mothers of young children has increased from 26.9 in 1992 to 28.6 in 2005, bearing in mind the BMI for normal weight is considered 18 to 25. Chart 3.1 shows the weight distribution of women is increasingly shifting toward heavier women. Between 2000 and 2005, most of the increase in BMI was due to increasing numbers of obese women; the proportion of women normal and just overweight remained approximately the same. The proportion of women morbidly obese increased nearly three percentage points from eight percent to eleven percent.

11

Chart 3.1: Distribution of BMI categories, 1992-2005 (weighted) 0% 20% 40% 60% 80% 100% 1992 1995 2000 2005 morbidly obese obese overweight normal

AGE AND COHORT EFFECTS Although the data used are not longitudinal data, it is still possible to follow age cohorts over the thirteen years covered by the surveys, as shown in Table 3.1.

12

Table 3.1: Mean BMI by cohort and age group

Birth year

Age group

15 - 19

20 - 24

25 - 29

30 - 34

35 - 39

40 - 44

45 - 49

1945 - 1949

28.1

1950 - 1954

29.0

28.9

1955 - 1959

27.9

28.9

31.0

1960 - 1964

27.3

28.6

3 1.7

1965 - 1969

26.1

28.3

30.0

1970 - 1974

24.9

27.3

29.3

1975 - 1979

24.0

26.3

28.1

1980 - 1984

25.4

26.9

1985 - 1989

25.5

While there are increases in BMI as a cohort ages, there are also increases in the BMI of an age group over time. This increase is seen most sharply in the oldest age groups. Although the 15-19 age group has seen only an increase in mean BMI of about one point between 1992 and 2005, the 45-49 age group has seen an increase of nearly three points.

GEOGRAPHIC DIFFERENCE The distribution of obesity varies dramatically by geographic region. Chart 3.2 shows the most dramatic changes have been in the frontier governorates of the North and South Sinai, Matruh, New Valley and Red Sea. Chart 3.3 shows the percent overweight by region and urban/rural. A one way analysis of variance found that, in all years, there were significant differences according to the primary sampling unit.

Full document contains 170 pages
Abstract: Obesity is emerging as one of the leading public health challenges in low and middle income countries. In particular, women of reproductive age are vulnerable to many compromised reproductive health outcomes associated with obesity. Egypt is an especially interesting country to study having experienced a rapid rise in obesity, with nearly half of women of reproductive age obese in 2005, exceeding levels of obesity seen in many high income countries. Despite the importance of obesity and its implications for health in developing countries, the subject has not received sufficient research interest; this dissertation contributes to addressing this deficiency. The dissertation is comprised of three empirical chapters all using Egyptian Demographic and Health Surveys from 1992 to 2005. The first uses factor and multilevel analysis to analyze the variables used to measure female empowerment. The findings highlight the difficulty in measuring female empowerment in a meaningful way, with questions around both the reliability and the validity of the data. In the second chapter, I conduct an analysis of the temporal changes in the relationship between maternal obesity and social determinants, using both recursive partitioning and logistic regression. The findings show that not only are Egyptian women becoming more obese but that the increase in obesity has disproportionately affected the most deprived: those with the least education, the poorest, the rural population, and those living in Upper Egypt. Finally, I look at the relationship between maternal obesity and maternal and child health outcomes, and at the mediating effect of socioeconomic status, using Cox proportional hazards and logistic regression models. The findings show that for some outcomes, there is a mediating effect of SES and that this relationship is also changing over time.