Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

The relationship of family characteristics to the nutritional status of pre-school children

Delfina B. Aguillon,* Ma. Minda Caedo,** Jesse C. Arnold,*** and R.W. Engel**

* National Nutrition Council, Metro Manila, Philippines
** United States Agency for International Development, Manila, Philippines
*** Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA

Several investigators have studied the effect of income source and economic status, population pressure, nutrient intake, employment, traditions, habits, and practices on nutritional status (1-12). However, only a few of these studies have indicated the variables most significantly related to preschool child malnutrition. In his study of three population classes in rural India, Levinson (7) reported that the most significant variables for malnutrition in young children are caloric intake and diarrhoeal infection. Caloric intake is, in turn, determined by the mother's knowledge of proper nutrition and by income, the latter becoming increasingly more significant as the child grows older. The result of the study made by Kreysler (6) in Botswana indicated that the economic factors of a non-supporting father and lack of cropland or livestock are the most important in predicting malnutrition in children.

The present study investigates the influence of such factors as education of parents, household size, family income, home ownership, birth order, and type of infant feeding on the nutritional status of the pre-school child, using percentage of standard weight for age as the dependent variable to characterize families, and applying the information obtained in predicting nutritional status of families for nutrition programme targeting.

METHODOLOGY

The data for the present study were collected between May and July 1978. Sixty five municipalities in six provinces in Luzon and one province in the Visayas were selected for data collection. The selection of municipalities was based on the availability and willingness of the area Home Management Technicians of the Bureau of Agricultural Extension to be trained and to serve as data collectors. Within these 65 municipalities, 729 families with one or more pre-school children were selected and classified into four nutritional groups (normal and first-, second-, and third degree malnutrition), using the pre-school child with the lowest nutritional status in the family as the basis for classification. The Gomez classification was adapted to Philippine child-growth standards (13) with percentage of standard weight for age as the basis for categorizing, as follows: normal, 91 per cent or more; first-degree malnutrition, 76-90 per cent; second-degree, 61-75 per cent; and third-degree, 60 per cent or less. The children were weighed to the nearest 0.1 kg on paediatric scales.

An effort was made to include approximately equal numbers of families from rural, urban, and coastal areas (table 1). The sample of families is selective, with a higher sampling ratio of families with malnourished pre-school children because the study was designed to compare families having normal pre-school children with those having malnourished ones. The results presented here are therefore not representative of the total population. Where appropriate, however, reference will be made to differences noted between rural, urban, and coastal areas.

TABLE 1. Distribution of Families within Nutritional Categories by Type of Area

*Based on the nutritional status of the most poorly nourished pre-school child in the family.

TABLE 2. Percentages of Families within Nutritional Categories Related to Selected Socio-economic Variables

Figures in parentheses indicate the total number of families in each category considered in the analysis for each variable- e.g., of 141 families in the "normal" category, 40.4 per cent 157 families) owned their own home. These totals differ slightly in some cases from those shown in table 1 because of incomplete data for some of the variables.

An interview method was used in the study. For this purpose, a three-part questionnaire directed to the homemaker was developed, pretested, and refined. Part I collected information on some household characteristics related to the family unit, such as education of parents, income, home ownership, and size of family. Part II asked for information on infant-feeding practices. Because of omission of responses for a variable under study in some interview schedules, small differences in the number of families in each nutrition category will be observed in the tables presented. The data were analysed using the chisquare test.

Relationships between nutritional status {in terms of percentage of standard weight for age) and different family characteristic variables were also evaluated using the Pearson moment correlations. Stepwise regression models were computed to help determine which variables should yield the best predictions of nutritional status. The ability of several regression models to classify pre-schoolers as normal or malnourished was studied, and a model with a high probability of correct classification is reported here.

FINDINGS
Part I

Table 2 shows the percentage of families fitting selected measurements of socio-economic status within each of the four nutritional categories.

Home and lot ownership. About 30 per cent of the families surveyed owned the house and lot where they lived. The rest paid rent for either or both house and lot, or lived rent free by virtue of affinity to the owner or by assuming caretaker responsibilities. Eighty-four, or about 12 per cent, of the families neither indicated home or lot ownership nor said they were paying rent; hence, they were assumed to be squatters.

A progressive increase in home and lot ownership from the severely underweight family category [third degree) to the normal category was noted. These differences in proportion of families in each nutritional category proved to be highly significant. Further analysis of the data revealed that, in rural areas, more than twice as many families in the normal category (55.8 per cent) as third-degree ones 122.4 per cent) owned their homes and lots.

Location of residence. The families live in barangays (villages) one to seven kilometers from the poblacion (town centre), about two-thirds of them within a distance of three kilometers. Analysis of the distance of the barangays from the poblacion, however, did not show any significant differences among family categories.

Income source. Of the 719 families from whom we obtained complete data on income, 214 (29.8 per cent) earned their living from farming, while 505 (70.2 per cent) depended on other sources for their livelihood. Farm families cultivated from one to ten hectares of land. Farm size, however, was not found useful in characterizing family categories.

Since about 60 per cent of the subject families earned less than 2,000 pesos (US$1 = P8.00) per year, the relationship of income to nutritional status was investigated by determining the proportion of families in this income bracket within each nutritional category. Table 2 shows that, among families in the third-degree category, 75 per cent of the non-farm families and about 84 per cent of the farm families had annual incomes of P2,000 or less. It should be noted, however, that among families in the normal category the proportion of farm families with incomes of P2,000 or less (71.4 per cent) is more than twice that of non farm families (32.6 per cent).

Income expenditures. The respondents were asked to rank six common expenditure items (food, clothing, housing, education, medicine, recreation) according to the share of their income spent on each. For all categories of families, food ranked highest, with clothing ranking second and recreation last. Among the normal and first-degree categories of families, housing ranked third. On the other hand, among the second- and third-degree families, medicine had a proportionally higher claim to family earnings than housing. Children in these family categories are more prone to sickness and therefore may demand a higher amount of income for medical attention. Education ranked fifth in all categories except among the families in the normal category, where it ranked fourth.

Education of parents. While mothers in 79.5 per cent of third-degree families terminated their education in the elementary grades (0-6 years of school), mothers in only 54 per cent of families in the normal category did so (table 2). The table further shows, at the other extreme, that the mothers in 20.1 per cent of the families in the normal category had achieved more than ten years of education compared to only 6.2 per cent for third-degree families. These significant trends were evident in all areas but were most pronounced in the urban areas. The educational attainments of the fathers followed essentially the same pattern. For both mothers and fathers, education was extended somewhat longer in the urban areas and was shortest in the rural areas.

Age of parents. No significant relationships were observed between the ages of either fathers or mothers and family categories.

Household size. Table 3 shows the relationship between family categories and number of household members. Families in the normal category had an average number of 5.6 members, while those in the third-degree category were characterized by an average family size of 6.8. Those in first and second degree categories were intermediate, with 5.8 and 6.7 members respectively. This pattern prevailed and was statistically significant in all areas- rural, urban, and coastal- whether only the total household size or the number of pre-schoolers (6 years old or under) in the family was considered in the analysis. The relationship had still greater significance when children 15 years old and under were grouped together.

Birth order, sex, and age. The usefulness of birth order, sex, and age of the suject children in characterizing families was investigated. Data showing the relationship of these variables to family categories are presented in table 4. In half of the families characterized as normal and first-degree, children of birth order two or less predominate, whereas in those in the second- and third-degree categories, children of the same birth order constitute only about 30 per cent. The reverse was true for high birth order. Nearly twice as many families in the second- and third-degree categories had children of birth order four or more compared with the normal and first-degree families.

TABLE 3. Average Numbers of Household Members within Nutritional Categories

TABLE 4. Percentage Distribution of Families within Nutritional Categories by the Birth Order and by the Sex and Age of the Subject Child

Females tended to concentrate in families in the second-and third-degree categories, while males were more frequent in those in the normal and first-degree classifications. Overall, in 59.5 per cent of the families in the normal category the subject child was male in contrast with only 32.4 per cent male subjects in third-degree category families. This was not unexpected. Earlier studies in the Philippines (14, 15) revealed a higher prevalence of malnutrition among female pre-school children than among males. Female subjects in the present study constituted 67.6 per cent of those in the third-degree category and 64.1 per cent of all subjects in both second- and third-degree families.

Because the slowing of growth associated with the weaning process generally becomes important in the latter part of the first year of life, it was expected that subjects under one year of age would be concentrated in families classified as normal or first-degree. Data of the present study, however, supported this expectation only with regard to male subjects. Table 4 shows that in all age groups, males appear more frequently in the normal and first-degree family categories than in those in second- and third-degree. The proportion of females in the two groups of combined nutritional categories is similar in all age groups except in the 12-23-month-old group. In this segment of the subject population, more than twice the number of children are in the second- and third-degree categories than in the normal and first-degree classification.

Part II

The study also attempted to explore the extent to which the type of infant feeding practiced would be useful in characterizing families. Information on this phase of the study is presented in table 5. Mothers in families in the normal group practiced breast-feeding much more widely (69.8 per cent) than those in families categorized as third-degree (28.0 per cent). More than half the mothers in the latter group bottle-fed their babies, while only about 12 per cent of those in families in the normal category did so. In all areas, most of the families that bottle-fed their infants exclusively were in the third-degree category. In the rural areas, bottle-feeding was unusually high, being practiced by two of every three families in the third-degree category. Mixed feeding was practiced by 17.7 per cent of ail families, with slight deviations from this value among the four family categories.

Table 5 also shows categorization of families by age at weaning. Weaning is defined in this study as the time when breast or bottle milk is no longer the main food for the infant. Distinct differences between family categories and weaning practices may be noted, with early weaning (first six months) being practiced by 53.9 per cent of families in the third-degree category and by only 16.7 per cent among those in the normal category. Over 70 per cent of the mothers in families of the normal category continued breast-feeding for ten months or longer, compared with less than 30 per cent of third-degree families. Early weaning was particularly common in the urban areas, where 36.8 per cent of mothers had weaned their infants at six months or before, compared to only 20.4 per cent in the coastal families with similar weaning practices.

TABLE 5. Percentage Distribution of Families within Nutritional Categories by Feeding Practices

The age at which supplementary foods were introduced also varied. Early introduction of supplementary food (by three months of age) was practiced by 21.7 per cent of the normal category families, but by only 10.3 per cent of third-degree families. Whereas 22.9 per cent of third-degree families delayed supplemental feeding until seven months, only 14.4 per cent of families in the normal category put off giving additional food to infants for this long a period.

Other informational variables of note include the finding that mothers in families of the normal category had better knowledge of nutrition and infant-feeding practices than those in families categorized as third-degree. Also, while in the third-degree category families 17.4 per cent of the children were taken care of by a family member other than the mother, this was true for only 9.9 per cent of families in the normal category.

DISCUSSION

Table 6 presents correlation coefficients between the percentage of standard weight for age of the subject children and the different variables found useful in categorizing families. The table shows that the largest correlation was found between the percentage standard weight and type of infant feeding. Several studies (12, 16, 17) have indicated that breast-fed children are less likely to be malnourished than are bottle-fed children. However, a study by Ritchie and Naismith (18) showed that, although no differences in rate of increase in weight or length were observed between breast-fed and artificially fed infants within the first six weeks of life, the average velocities of growth between the two groups from the second to the sixth month were significantly different. The results suggest that the substitution of the bottle for the breast, under certain circumstances, can be less important than such factors as sanitation, income of parents, education of the mother, and access to health services.

Next in importance in characterizing families was income, which was found to have a highly significant correlation with nutritional status among non-farm families. This was followed by age at weaning, or length of breast-feeding.

TABLE 6. Correlation Coefficient between the Most Important Variables and Standard Weight for Age of the Subject Children

All coefficients were highly significant with the exception of farm income (0.05 <= p<= 0.10).

In a study of 92 families in Tanzania, Gupta and Mwambe (4) observed that, although more than half of the families had a fairly stable source of income, their children were malnourished. Other studies (5, 9, 19) obtained similar results, further suggesting the influence of negative nutritional behaviour of the mother (even with higher income) on the nutritional status of her child because of ignorance and certain cultural practices.

Results of the present study with regard to infant feeding and weaning revealed that those practices generally considered desirable were characteristic of the normal and first-degree categories of families, while those considered undesirable (bottle-feeding, early weaning, and delayed introduction of supplementary food) were observed more often in the second- and third-degree categories.

Several investigators have reported that introduction of supplementary food before the age of four months does not have any significant benefit (see 20, 21).In a World Health Organization nine-country study on breast-feeding, it was observed that weight gain was unusually low in infants who did not receive any food supplements until after six months (22). Available evidence therefore suggests that four to six months is perhaps the most appropriate time to introduce supplementary foods, even when the mother's milk supply is ample. Special consideration should be given to those who cannot be breast-fed or those who receive insufficient milk.

Aliling and Elequin (23) noted that as the level of education of mothers increased, a corresponding decrease in the incidence of second- and third-degree malnutrition among their pre-school children could be expected. The significant correlation between the educational attainment of the mothers and the percentage of standard weight of the subject children in the present study tends to confirm this observation.

Further comparisons made in this study showed that large family size was significantly negatively associated with good nutritional status. Studies done by Ballweg (1), Wray and Aguirre (12), Aliling and Elequin (23), and Guzmán (24) gave similar results, indicating that large families are more prone to have malnourished children. These findings could be attributed to the inability of mothers to provide adequate care to their young children, especially in cases where there is more than one pre-school child in the family. Very often they are tended only by an older brother or sister, or in some instances are left to fend for themselves. Added to this is the effect of poor intra-family food distribution, where older members of the family get the larger share of food in the household.

Use of Data for Model Building

Several regression models were developed using those family characteristic variables (table 6) associated with nutritional status. These were then investigated relative to their ability to classify families into the four categories. A sample model with reasonably good classification ability, and applicable to all the geographic areas, was as follows:

Y = 87.04 + 0.62 x₁ + 0.52 x₂ - 2.03x₃ + 3 x₄ +2.39x₅ -0.35x₆ +0.005x₆ - 11.16x₇ + 0.94 x₈ - 5.04 x₉

where Y = percentage of standard weight for age, x1 = number of years of formal education of mother, x2 = number of years of formal education of father, x3 = total number of household members, X4 = total family income in pesos per year divided by 1,000, x5 = a classification variable indicating whether the family is farm (1 ) or non-farm (0), x6 = age of subject child in months, X7 = a classification variable indicating whether the child has been weaned (1) or not weaned (0), x8 = age in months when the child was weaned, if it has been, and x9 = type of infant feeding practices (breast = 1, mixed = 2, bottle = 3).

The classification ability of this and other models was investigated and will be presented in a separate report. If normal and first-degree were grouped together in a normal category, and second. and third-degree in an underweight category, the above model accurately categorized about four out of five families (80 per cent classification probability).

With some caution, direct physical interpretations can be drawn from the model in terms of value to the child in percentage of standard body weight. Each year of a mother's education is worth about 0.6 percentage point to the child. For the father's education each year is worth about 0.5 percentage point. Each household member costs the child 2 percentage points, and each P 1,000 increase in income is worth 3 percentage points. With respect to infant-feeding practices, the model indicates that 11.6 percentage points should be subtracted if the child has been weaned, and 0.94 per cent should be added back for each month of age until the child is weaned. This implies that the child will have about a 1 percentage point gain from each month the weaning age is delayed beyond 12 months. The model further suggests that breast-feeding, as opposed to bottle-feeding, is worth about 10 percentage points. It is to be reiterated here that in this study infant-feeding practice was found to be the single best indicator of the child's nutritional status as measured by weight for age.

Table 7 illustrates how the above model might be used to predict percentage of standard weight for age as a basis for classifying families. The three pairs of columns under the heading "Family Characteristics" illustrate, for three hypothetical children in different families, the calculation of the contribution to the child's percentage of standard weight of each of the nine variables- the value assigned to that variable being given in parentheses, followed by the product obtained by multiplying that value by the variable coefficient. Each case considers a child of 14 months of age. The model predicts that the hypothetical child with poor inputs will be only 52A 1 per cent of standard weight, while for the children with average and good inputs, the predicted percentages of standard weight are 79.97 and 106.39, respectively.

CONCLUSIONS

Income, educational attainment of parents, and family size are significantly associated with the nutritional status of the pre-school child. Attention to them in multisectoral development planning can significantly influence nutritional status. Moreover, the study also identifies infant-feeding and weaning practices as highly important, and these can be improved if relatively short-term nutrition education is given to mothers.

The usefulness, particularly in field application, of the model derived from this study in planning and targeting intervention activities needs to be tested and possibly refined and adapted to other existing field situations.

TABLE 7. Example of Regression Model Classification

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the Home Management Technicians of the Bureau of Agricultural Extension for their conscientious efforts in the data collection. We are also indebted to Ms. Ma. Paz Geniza for her technical assistance, and to Ms. Sharon Myers for the statistical programming and computing. Finally, we wish to acknowledge the financial support of USAID through contracts AID 492-1536 and AID/ASIA-C1136.

REFERENCES

1. J.A. Ballweg, "Family Characteristics and Nutrition Problems of Preschool Children in Fond Parisien, Haiti, "J. Trop. Pediat. Environ. Child Hlth., 18: 229 (19721.

2. A. Chávez, C. Martinez, M. Muñoz,, P. Arroyo, and H. Bourges, "Ecological Factors in the Nutrition and Development of Children in Poor Rural Areas," in Proceedings of the Western Hemisphere Nutrition Congress III (Futura, Mt. Kisco, N.Y. USA, 1972).

3. R.P. Devadas, "Social and Cultural Factors Influencing Malnutrition,"J. Home Econ., 62: 164 (1970).

4. B.M. Gupta and A. Mwambe, "Study of Malnourished Children in Tanga, Tanzania," pt.1, "Socio-enonomic Cultural Aspects," l Trop. Pediat Environ. Child Hlth.,., 22: 268 (1976).

5. S.S. Hijazi, Child Growth and Nutrition in Jordan:: A Study of Factors and Patterns (Royal Scientific Press, Amman, Jordan, 1 977).

6. J. Kreysler, "Some Notes on Social Reasons Leading to Malnutrition in Children Based on Information from Serowe Village," presented at a seminar on Aspects of Home, Health and Nutrition, Botswana (1978).

7. F.J. Levinson, Morinda: An Economic Analysis of Ma/nutrition among Young Children in Rural /India,, Cornell/MIT International Nutrition Policy Series (Cambridge, Mass., USA, 1974).

8. K. Rao, "Nutritional Status of Preschool Children and the Related Factors," Indian J. Nutr. Dietet., 15: 23311978).

9. I. G. Rawson and V. Valverde, "The Etiology of Malnutrition among Preschool Children in Rural Costa Rica," J. Trop. Pediat Environ. Child Hlth., 22: 12 (1976).

10. S. Schofield, Development and the Problems of Village Nutrition, Institute of Development Studies (Groom Helm Ltd., St. John's Road, London SW 11,1979).

11. L.S. Sims and P.M. Morris, "Nutritional Status of Preschoolers: An Ecologic Perspective," J. Amer. Dietet Assoc., 64: 492 (1974).

12. J.D. Wray and A. Aguirre, "Protein-Calorie Malnutrition in Candelaria, Colombia," pt.1, "Prevalence: Social and Demographic Causal Factors," J. Trop. Pediat., 15: 76 (1969).

13. J. Bulatao-Jayme, D. de la Paz, and C.C. Gervasio, "Recommended Height and Weight Standards for Filipinos," addendum, Phil. J. Nutr., 24: 203 (1971).

14. M. Caedo, V.B. Santiago, D. Diaz, and R.W. Engel, "Progress Report: Integrated Nutrition-Family Planning Program, Province of Bulacan, 1971-1972," Occasional Paper, Resource Library, USAID, US Embassy, Manila, Philippines.

15. V.B. Guzmán,, G.B. Roman, and I. Morelos, "Family Formation Patterns and Health: An International Collaborative Study in India, Iran, Lebanon, Philippines, and Turkey" (WHO, Geneva, 1976).

16. H. H. Kanawati and D.S. McLaren, "Failure to Thrive in Lebanon," pt. 2, "An Investigation of the Causes," Acta Paediat Scand., 62: 571 11973).

17. A.C.K Antrobus, "Child Growth and Related Factors in a Rural Community in St. Vincent," J. Trop. Pediat.. Environ. Child Hlth., 17: 188 (1971).

18. C.D. Ritchie and D.J. Naismith, "A Comparison of Growth in Wholly Breastfed Infants and in Artificially Fed Infants," Abstr. Comm. Prov. Nutr. Soc., 34: 118A (1975).

19. J.F. Mancebo and L.U. Onate, "The Nutritional Behavior of Some Mothers of Different Income Levels and the Nutritional Status of Their Preschool Children," Phil. J. Nutr., 32: 80 (1979).

20. P.S. Venkatachalam, T.P. Susheela, and P. Rau, "Effect of Nutritional Supplementation during Early Infancy on Growth of Infants," J. Trop. Pediat., 13: 70 (1976).

21. T. Becroft and K.V. Bailey, "Supplementary Feeding Trial in New Guinea Highland Infants," J. Trop Pediat., 11 :28 (1965).

22. World Health Organization (WHO) and United Nations International Children's Emergency Fund (UNICEF) Background Paper for the Meeting on Infant and Young Child Feeding, Geneva, 9-12 October 1979.

23. Y. Aliling and E.T. Elequin, "Food and Nutrition Attitudes of Rural Mothers in Eight Barrios of Cavite: Implications for a National Nutrition Education Program," Nutrisyon, 1:37 (1976).

24. V.B. Guzmán, "Child Health, Nutrition, and Family Size: A Comparative Study of Rural and Urban Children," Phil. J. Pediat., 22: 129 (1973)

Contents - Previous - Next