October 2001
by Alain Marcoux1
Population Programme Service
FAO Women and Population Division
The greater susceptibility of females to malnutrition - hampering their economic, reproductive, educational and social roles - is the subject of much discussion in a policy perspective. The objective obviously is to redress inequities and to end detrimental situations. But what is the extent of that phenomenon and what are the most affected countries or areas - those where efforts need to be concentrated?
This article reviews available evidence regarding gender biases in malnutrition in developing countries and countries in transition. For this purpose, we inventory findings regarding sex differentials for the classical malnutrition indicators. Some 350 nutrition surveys taken since 1985 in developing countries and countries in transition have been reviewed and differences between sex-specific indicators have been checked for statistical significance2. As customary, the conditions of adults and children are dealt with separately.
The nutritional status of adults is assessed using the Body Mass Index (BMI)3. For classifying individuals according to their nutritional status, cut-off levels of the BMI are used. Adults with a BMI less than 18.5 kg/m2 are considered to suffer from chronic energy deficiency (CED).
Survey results have been extracted from FAO Nutrition Country Profiles.4 Thirty of the surveys reviewed contained sex-disaggregated data on population distribution by BMI category.5 Out of these, 11 revealed a greater incidence of CED among women, while one showed the opposite bias. The countries concerned are listed in Table 1 below.6
Since none of these countries has such data for at least two points in time, it is impossible to tell whether these biases are growing or not.
Let us briefly comment and look into additional details, where available, on the countries signalled in this table. In particular, we shall examine possible variations in the pattern of gender biases when one looks specifically at severe chronic energy deficiency (BMI < 16).
In the case of China, the bias was assessed in the rural population only (no significant difference between sexes was found in the urban segment). Of course, this alone represents an issue of great magnitude: suffice it to say that the adult female population of rural China numbers about 280 million persons. No significant variation in the pattern of excess female undernutrition was evident across degrees of CED. As these results are likely to vary in intensity across regions, however, one would wish to see data by Province, such as was collected in 1987 on child nutrition (see below).
Regarding Fiji, two notations may be warranted. Among Fijians, women were also significantly worse off at the level of severe and moderate CED (BMI < 17). And among Indians, although there was no significant gap for CED in general, women were worse off (very close to the significance level) for severe CED.
In India, eight States or Union Territories appear in Table 1. According to the preliminary results of the 2001 Census of India, these eight entities total 142 million inhabitants7 (i.e. 14 % of the estimated 1 027 million total population), with probably close to 40 million adult women.8 Women seemed particularly at risk of severe CED in Bihar and Goa. In Bihar, the gap was also significant for moderate CED (BMI = 17 to 18.5).
We also looked for cases of countries or other entities which do not have a female bias for CED in general but present such a bias for severe CED. One such case is offered by Sri Lanka (1995). Another is Punjab in the 1994 Indian survey.
In some surveys, data were collected on women only, giving no indication on gender gaps but revealing a serious situation regarding women's undernutrition. This was in particular the case for Bangladesh (1996-7), where 52 % of women were affected (36 in urban areas, 54 in rural areas). In Cambodia (1998), 29 % of women were affected by CED (60 % in the forest zone). In Nepal (1996), 28% of women were affected, especially in the Terai (41%). Such cases deserve attention, irrespective of the status of males relative to females.
Finally, let us look at the other end of the malnutrition range, for eventual biases in overweight (BMI = 25-30) and obesity (BMI > 30). In China (1992), women overweight were much more numerous than men, especially in rural areas, but there was no significant gender bias in obesity. In India (1994), a greater prevalence of obesity among females was found in Delhi, Himachal Pradesh, Meghalaya and Punjab, but there was no significant gender bias in overweight. Finally, overweight was found to be significantly more prevalent in females than in males in Chile (1992), Jamaica (1994-5) and the Ha Tay Province of Vietnam (1995). The same applied to obesity in Chile (1992, 1993), Jamaica (1994-5) and in Ankara (1999).9
The nutritional status of pre-school children is assessed using three indices: weight-for-height, height-for-age and. Low weight-for-height (wasting) indicates acute growth disturbance, low height-for-age (stunting) indicates long-term growth faltering, and low weight-for-age (underweight) reflects a mix of the above short- and long-term effects. These disturbances in turn denote malnutrition, usually in conjunction with other health factors. Weights and heights of children are compared with the reference standards (NCHS/CDC/WHO) and the prevalence of deficits is expressed as the percentage of children below a specific cut-off point (here, minus two standard deviations from the median value).
Survey results in respect of these three indicators have been extracted from the WHO Global Database on Child Growth and Malnutrition.10
222 of the surveys reviewed detected no gender bias for any of the indicators.11 Otherwise,
Clearly, boys generally fare worse than girls do by anthropometric indicators. This pattern - empirically familiar to nutritionists - was first confirmed by a United Nations-sponsored study.12 Yet, this pattern stands in contrast with the widespread notion of an anti-female bias in intra-household food allocation. But the basis for this notion is a handful of small-scale studies, the findings of which have been unwarrantedly extrapolated. Sommerfelt and Arnold (op. cit.) note that "despite evidence of discrimination against girls in feeding practices and medical treatment from small-scale studies in countries with a strong preference for sons, those types of discriminatory practices are evidently either not widespread or not sufficiently severe to show up in data on nutritional status at the national level". In addition, those studies refer to situations of the late 1970s or early 1980s, and many things may have changed since those times. At any rate, the comparison with national data on nutritional outcomes is telling, as we shall now see.
Two of those studies dealt with rural Bangladesh;13 but none of seven national surveys taken since 1985 found an anti-female bias for any of the indicators. Another study dealt with Egypt;14 none of nine national surveys there found an anti-female bias for any indicator. Another dealt with Kiribati,15 where the 1985 national survey again did not detect any anti-female bias.16 Two studies dealt with North India and Punjab. The 1994 state-wide survey does not confirm the finding for Punjab.17 Otherwise, the existence of an anti-female bias for wasting in four northern States or Territories (see Table 2), and for underweight in one (Table 4), might lend some support to the earlier finding. But the fact that 45 surveys at State or Territory level between 1988 and 1999 did not find any bias - and that the 1992-3 survey even found an anti-male bias for wasting at national level - suggests that the said finding cannot be generalized.18
Finally, what can be said about trends in the gender biases recorded in Tables 2-4? Two facts seem relevant:
The female bias in undernutrition appears to be limited to the adult population in a few countries, including some areas of the two most populous countries in the world, China and (North) India. There is no indication as to whether these biases are growing or not. A better understanding of the phenomenon will require more geographically disaggregated nutrition surveys in the large countries, as well as more surveys in countries not heretofore investigated.
As for pre-school children, gender biases in malnutrition are not infrequent among them, but in this case it is boys who more often are worse off. However, for this age group, in many of the countries studied the biases seem to be disappearing, especially those once affecting girls.
1 Senior Officer, FAO Population Programme Service. The guidance and support of the FAO Nutrition Planning, Assessment and Evaluation Service are gratefully acknowledged.
2 Given the intention of outlining a broad picture of the situation, this note leaves aside a host of local surveys and concentrates on: national surveys; surveys at State or Province level in large countries such as Brazil, China, India or Indonesia; national rural surveys; and a few surveys of capital cities. In the tables and text, surveys are national ones unless otherwise indicated: in the other cases, the area covered is indicated following the date of survey. Statistical significance is taken at the 95% level.
3 BMI = weight / [height] 2.
4 www.fao.org/es/ESN/ncp/list.htm
5 Each survey of a sub-national entity for which separate results are presented within a national operation is counted as a survey. In these cases, individual results for such entities are indicated in italics in the tables and text.
6 The following surveys reveal no significant gender bias in chronic energy deficiency among adults: Chile 1992 Metropolitan area, China 1992 (Urban), Fiji 1993 (Indians), India 1994 (Arunachal Pradesh, Chandigarh, Delhi, Haryana, Himachal Pradesh, Manipur, Nagaland, Punjab, Rajasthan, Tripura), Panama 1995, Sri Lanka 1995, Turkey 1999 Ankara, Venezuela 1981-7, Vietnam 1995 Ha Tay Province.
7 This total includes the State of Jharkhand, created inside the former limits of Bihar in 2000.
8 According to the same source, the 10 other States or Territories covered in the 1994 survey (listed in endnote 6) total 131 million inhabitants.
9 Like with adults, the nutritional situation of older children and adolescents (˜ ages 6-18) is assessed by means of the BMI. However, very few surveys address this issue. Among those reviewed for this note, none revealed a gender bias in BMI deficit.
10 www.who.int/nutgrowthdb/. As a rule, the surveys reviewed are based on the classical age bracket for such assessments, i.e. 0 to 5 years. Occasionally, however, the lower limit of the age bracket was age 3 or 6 months, or the upper limit was age 3, 4, or 6 years.
11 These are: Albania 1996-8. Algeria 1987, 1995. Angola 1996. Argentina 1995-6. Azerbaidjan 1996, 2000. Bahrain 1989. Bangladesh 1985-6, 1989-90, 1992, 1995-6, 1996-7. Bhutan 1986-8, 1989. Bolivia 1989, 1998. Botswana 1996. Brazil 1986 Nordeste, 1987 Ceara, 1989, 1989 Rio Grande do Norte, 1990 Ceara, 1991 Maranhao, 1991 Pernambuco, 1991 Piaui, 1996. Burkina Faso 1992-3, 1998-9. Burundi 1987. Cameroon 1991. Central African Republic 1994-5, 1995. China 1987 (Inner Mongolia, Heilongjiang, Zhejiang, Shandong, Sichuan), 1992, 1995, 1998, 2000. Colombia 1986, 1995. Comoros 1995-6. Congo 1987. Côte d'Ivoire 1986. Dem. Rep. of Congo 1995. Djibouti 1996. Egypt 1988, 1992-3, 1994-5, 1995-6, 1996, 1997, 1997-8, 1998. El Salvador 1993, 1994, 1998. Eritrea 1995-6. Georgia 1999. Ghana 1987-8, 1988, 1993-4, 1998-9. Guatemala 1987, 1995, 1998-9. Guinea 1999. Guyana 1997. Haiti 1994-5. Honduras 1987, 1991-2, 1993-4. India 1988-90 (Andhra Pradesh, Gujarat, Karnataka, Kerala, Madhya Pradesh, Maharashtra, Orissa, Tamil Nadu), 1991-2 (Andhra Pradesh, Gujarat, Karnataka, Kerala, Maharashtra, Orissa, Tamil Nadu, West Bengal), 1994 (Andhra Pradesh, Arunachal Pradesh, Assam, Dadra & Nagar Haveli, Daman & Diu, Goa, Gujarat, Haryana, Himachal Pradesh, Karnataka, Kerala, Madhya Pradesh, Maharashtra, Manipur, Meghalaya, Mizoram, Nagaland, Orissa, Punjab, Sikkim, Tamil Nadu, Tripura), 1996-7 (Andhra Pradesh, Gujarat, Karnataka, Kerala, Maharashtra, Tamil Nadu), 1998-9, 1998-9 Andhra Pradesh. Indonesia 1999 Central Java and 'West Sumatra (rural), 1999-2000 East Java, Lombok and West Java (rural). Iran 1995, 1998. Jamaica 1990, 1991, 1992, 1994, 1995. Jordan 1990, 1997. Kazakhstan 1995. Kuwait 1994-5. Kyrgyzstan 1997. Lao 1993, 1994, 1997. Lebanon 1996. Libya 1995. Madagascar 1992, 1993-4, 1995. Malawi 1992. Malaysia 1988, 1992-95 rural. Maldives 1994, 1995. Mali 1995-6. Mauritania 1990-1. Mexico 1988, 1998-9. Mongolia 1992. Morocco 1987, 1992. Mozambique 1989, 1997. Namibia 1992. Nepal 1995, 1996, 1997-8. Nicaragua 1993, 1997-8. Niger 1992, 1998. Nigeria 1990, 1993, 1999. Oman 1991, 1995. Pakistan 1985-7, 1990-1, 1990-4. Panama 1997. Paraguay 1990. Peru 1991-2. Philippines 1987, 1989-90, 1993, 1998. Qatar 1995. Romania 1991. Russian Federation 1993, 1995. Rwanda 1991-2, 1992, 1992. Senegal 1986, 1992-3, 1996. Seychelles 1987-8. Solomon Islands 1989. Somalia 2000. Sri Lanka 1995. Sudan 1992-3. Tajikistan 1994, 1996. Tanzania 1991-2, 1996, 1999. Thailand 1987, 1995. Togo 1988, 1998. Tunisia 1988, 1996-7. Turkey 1993, 1998. Uganda 1995. Uruguay 1992-3. Uzbekistan 1996. Venezuela 1981-7. Vietnam 1985, 1990, 1995, 1995 Hanoi, 1997, 1998. West Bank & Gaza Strip 1996. Yemen 1996, 1997. Yugoslavia 1996. Zambia 1992. Zimbabwe 1988, 1994, 1999.
12 A.E. Sommerfelt and F. Arnold, 1998. Sex differentials in the nutritional status of young children. In: Too young to die: genes or gender?, New York, United Nations, pp. 133-153. Having examined 41 Demographic and Health Surveys, the authors concluded: "If anything, the data suggests that girls are less likely to be undernourished than boys".
13 L.C. Chen, E. Huq and S. D'Souza, 1981, Sex bias in the family allocation of food and health care in rural Bangladesh, Population and Development Review, vol. 7, No. 1, pp. 55-70; and K.H. Brown, R.E. Black, S. Becker, S. Nahar and J. Sawyer, 1982, Consumption of foods and nutrients by weanlings in rural Bangladesh, American Journal of Clinical Nutrition, vol. 36, No. 5, pp. 878-889.
14 C. Makinson, 1986, Sex differentials in infant and child mortality in Egypt. Princeton University, unpublished doctoral dissertation.
15 A.B. Roberts, P. Roberts, T. Tiva and K. Tulimanu, 1981, Malnutrition and anaemia in Gilbertese preschool children: a case finding and epidemiological survey, Journal of Tropical Pediatrics, vol. 30, No. 4, pp. 237-239.
16 Indeed, in each of the three countries just reviewed, at least one survey found an anti-male bias for one of the indicators (see Tables 2, 3, 4).
17 B.D. Miller, 1981, The endangered sex: neglect of female children in rural North India, Ithaca, Cornell University Press; and Monica Das Gupta, 1987, Selective discrimination against female children in rural Punjab, Population and Development Review, vol. 13, No. 1, pp. 77-100.
18 Another detrimental phenomenon for females, namely excess mortality - whose possible linkages with malnutrition are much discussed - could be evoked at this stage. First, could excess female mortality create a statistical artefact that would hide the real extent of female undernutrition? This cannot happen unless a mortality crisis removes by chance unusual numbers of undernourished girls (and not boys) from the population shortly before a nutrition survey is taken. As for unbalances between the numbers of boys and girls present in the population surveyed, they are irrelevant since the indices are calculated separately on the population of each sex.
The countries where under-five mortality is markedly higher for females than for males seem to be Bangladesh, Papua New Guinea, Nepal, Samoa, India, China and the Maldives (by growing importance of the relative gap between sex-specific rates). Smaller gaps are found in Guinea, Brunei Darussalam, Yemen, Barbados, Mongolia and Vietnam. There is excess male under-five mortality in the remaining 158 countries for which data is available (WHO, 2000, World Health Report 1999, Geneva). As for adult age groups (15-59), they present excess male mortality in all the countries of the world except the Maldives; the countries in the left column of Table 1 therefore are in this same situation, but for the reasons given above this does not affect the validity of the findings reflected in that table.
|
