Public health nutrition

Effects of iron-zinc supplementation on the iron, zinc, and vitamin A status of anaemic pre-school children in Indonesia
Street food: An important source of energy for the urban worker
Positive and negative deviance in growth of urban slum children in Bombay
Home gardening for combating vitamin A deficiency in rural India

Effects of iron-zinc supplementation on the iron, zinc, and vitamin A status of anaemic pre-school children in Indonesia

Abstract
Introduction
Materials and methods
Results
Discussion
References

Werner Schultink, M. Merzenich, Rainer Gross, Roger Shrimpton, and Drupadi Dillon

Werner Schultink and Rainer Gross are affiliated with the Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ), Eschborn, Germany, and the Regional SEAMEO Center for Community Nutrition in Jakarta, Indonesia. M. Merzenich is affiliated with Christian Albrecht University in Kiel, Germany, and the Regional SEAMEO Center for Community Nutrition in Jakarta. Roger Shrimpton is affiliated with UNICEF in Jakarta. Drupadi Dillon is affiliated with the Regional SEAMEO Center for Community Nutrition in Jakarta.

Mention of the names of firms and commercial products does not imply endorsement by the United Nations University.

Abstract

The effects of iron-zinc versus iron supplementation on iron, zinc, and vitamin A status were investigated in a double-blind trial. The subjects were 67 pre-school children from among 85 initially enrolled; all had haemoglobin <110 g/L and a height-for-age Z score below - 1.5. They were allocated at random to two groups: 33 children received supplementary iron-zinc (30 mg iron and 15 mg zinc per day), and 34 children received supplementary iron (30 mg iron per day) for eight weeks. All subjects initially were dewormed with mebendazole and pyrantel pamoate. The chronically malnourished anaemic children had a low vitamin A status, although 46 had received vitamin A capsules through the government programme three to nine months before the study. At the start of the study, 81.8% of the children had serum retinol values below 0.70 µmol/L. Only 3.2% of the children had serum zinc values below 10.7 µmol/L, and the zinc status appeared to be sufficient. The iron status improved more in the iron-supplemented group than in the iron-zinc-supplemented group (p <.001). A combined iron-zinc supplement was not effective in improving iron status. The change in serum zinc in the iron-zinc-supplemented group was 1.8 µmol/L higher (p = .037) than in the iron-supplemented group. Serum retinol increased significantly in both groups by about 0.16 µmol/L (p < .001). The changes in retinol status were probably due to the deworming treatment.

Introduction

Iron and vitamin A deficiencies are among the most important nutritional problems in developing countries [1]. Several studies have demonstrated that vitamin A and iron are metabolically related. Supplementation with vitamin A improved iron status in pregnant women [2] and children [3, 4] who had a marginal iron and vitamin A status. It has been hypothesized that vitamin A allows iron that is trapped in the liver to be released when transferrin synthesis is resumed [5].

Whereas the epidemiological aspects of vitamin A and iron deficiency are well known, much less information is available on zinc deficiency. Zinc deficiency is thought to be highly prevalent in developing countries, since foods rich in zinc, such as red meats and whole grains, are not consumed much. However, little is known about the actual prevalence of zinc deficiency in developing countries. A recent survey in Thailand reported that 70% of schoolchildren had serum zinc concentrations below 10.7 µmol/L [6]. A relation between zinc and vitamin A metabolism has been described [7]. Zinc supplementation of children with vitamin A deficiency has been reported to result in a significant increase in plasma vitamin A [8]. Zinc influences vitamin A metabolism because it is needed to mobilize vitamin A from the liver [7, 9]. Interaction between iron and zinc occurs during the absorption phase. When given as a supplement, iron inhibits zinc absorption if the ratio of iron to zinc is greater than 2 to 1 [10,11].

Iron, zinc, and vitamin A are therefore metabolically interlinked. In view of the public health problems of iron and vitamin A deficiency in developing countries and the possible high prevalence of zinc deficiency, it is important to gain more understanding of the relationships among these three micro-nutrients.

The objectives were to study the interactions among iron, zinc, and vitamin A in malnourished children, to obtain information on whether zinc deficiency may be a problem in urban children in Jakarta, to investigate whether zinc supplementation improves vitamin A status, and to determine whether a combined iron-zinc supplement improves zinc and iron status.

Materials and methods

The study was carried out from August to September 1993, in Tambora district of urban Jakarta. This area can be characterized as having a population of low socio-economic status living under poor hygienic conditions. The estimated population density is around 50,000/km². Pre-school children were selected for the study who were both stunted, as indicated by a height-for-age Z score below - 2, and anaemic, as indicated by a haemoglobin concentration below 110 g/L. The names of 380 children between 24 and 60 months of age were obtained from lists at the local health centre, and the height and weight of all these children were measured. Of these, 160 children (42%) had height-for-age Z scores below - 2 and could therefore be considered as possible subjects.

The haemoglobin level of these stunted children was determined in fingerprick blood samples, and 75 children (47% of the stunted children) were found to have haemoglobin levels below 110 g/L. Sample size calculations indicated that with at least 40 children per group, a between-group difference of 6 g/L haemoglobin and 1.6 µmol/L serum zinc could be detected with a power of 0.80 and a significance level of .05. Therefore, to ensure a total sample size of at least 80 children, several children were included in the study sample with height-for-age Z scores between - 1.5 and - 2.0 and haemoglobin levels below 110 g/L. Finally, the parents of 85 children gave written consent to participate in the study.

The study design was approved by the ethical committee of the Regional SEAMEO Center for Community Nutrition at the University of Indonesia, Jakarta.

The children were randomly assigned to two groups. One group (n = 43) received a commercially available iron-zinc supplement (as ferrous and zinc phosphate) (Budenheim, Budenheim, Germany). The other group (n = 42) received only an iron supplement (as ferrous sulphate) (Kimia Farma, Jakarta, Indonesia) which is normally used in Indonesia to treat iron deficiency in children. The supplements were provided in the form of a syrup and the subjects were instructed to take a daily dose of 5 ml for a period of eight weeks. The daily dose of the iron-zinc supplement was equivalent to 30 mg of elemental iron and 15 mg of elemental zinc. The daily dose of the iron supplement was equivalent to 30 mg of elemental iron. Both syrups were similar in appearance and taste, and supplementation was double-blinded.

During the supplementation period, intake of the syrup was supervised on a daily basis by one of the researchers (M.M.) and by health-care staff. A questionnaire was used to obtain information about previous intake of vitamin A supplements provided by the Indonesian Ministry of Health. Dietary vitamin A intake was assessed at the start of the study by calculating the usual pattern of food consumption using a food-frequency questionnaire, as described by the International Vitamin A Consultative Group [12].

A previous study in a similar environment in urban Jakarta showed that 50% to 70% of children under five years of age suffered from parasites (ascariasis, trichuriasis, or both) [13]. Therefore all children received a deworming treatment (100 mg of pyrantel pamoate and 150 mg of mebendazole) before the start of the supplementation to remove helminthiasis as a possible confounder.

At the start and finish of the eight-week supplementation period, a venous blood sample was taken from each subject. Serum retinol was analysed by high-performance liquid chromatography [14, 15]. Serum zinc was determined by atomic absorption spectroscopy [16, 17]. Haemoglobin was analysed by the cyanomethaemoglobin method [18]. Serum ferritin was analysed by the enzyme immunoassay procedure [18] using a commercial kit (Ramco Laboratories, Houston, Tex., USA). Erythrocyte protoporphyrin was assessed by measuring the zinc protoporphyrin/haem ratio with a haematofluorometer [18] (Protofluor Z, Helena Laboratories, Beaumont, Tex., USA).

Because the values of serum ferritin and the zinc protoporphyrin/haem ratio were not normally distributed, a log transformation was applied so that tests for normally distributed values could be used. Between-group differences in treatment effect were tested by using the interaction term between treatment effect (before-after) and treatment type (Fe versus Fe/Zn) of analysis of variance for repeated measures with SPSS/PC+ 4.0 [19].

Results

Complete data were obtained for 67 children. Selected characteristics of the subjects are presented in table 1.

At the start of the study, 93.1% of all children had ferritin concentrations below 12.0 µg/L, indicating the absence of iron stores, whereas 77.6% had protoporphyrin values higher than 70 µmol of zinc protoporphyrin per mole of haem, indicating an inadequate iron supply for haem production. Only 3.2% (n = 2) of the children had a deficient zinc status, as indicated by a zinc concentration below 10.7 µmol/L. Vitamin A deficiency occurred in 15.2% (n = 10) of all children, as indicated by serum retinol values below 0.35 µmol/L, while 65.1% of all children had a marginal status with serum retinol values between 0.35 and 0.70 µmol/L. The distribution of the serum retinol values is shown in figure 1.

TABLE 1. Selected characteristics of subjects^a

Characteristic	Iron-supplemented group	Iron-zinc-supplemented group
No.	34	33
Girls/boys	14/20	21/12
Age (mo)	37 ± 8	40+10
Height (cm)	85.3 ± 5.7	86.5 ± 6.6
Weight (kg)	10.8 ± 1.4	11.1 ± 1.8
Height-for-age Z score	-2.5 ± 0.7	-2.6 ± 0.9
Weight-for-age Z score	-2.4 ± 0.6	-2.4 ± 0.8
Weight-for-height Z score	-1.2 ± 0.6	-1.1 ± 0.8

a. Plus-minus values are means ± SD.

FIG. 1. Distribution of retinol concentrations before (open circles) and after (closed circles) intervention (n - 66)

The Indonesian government aims to supplement children with vitamin A capsules twice a year. The retinol level was not related to whether or not children had received vitamin A capsules during the months prior to the study. Children reported to have received a capsule three months before the start of the study had a retinol concentration of 0.52 + 0.17 µmol/L (n = 31), whereas children who had received a capsule nine months before the start of the study had a level of 0.46 ± 0.18 µmol/L (n = 15), and children who had not received any capsule at all had a level of 0.59 ± 0.21 µmol/L (n = 14) (ANOVA, p = .177). No information about the ingestion of vitamin A capsules was available for seven children.

TABLE 2. Blood values^a at the start and finish of the supplementation period

Value	Start	Finish	Difference
Iron-supplemented group (n = 34)
Haemoglobin (g/L)	98 ± 14	116 ± 7	18 ± 14bc
Protoporphyrin (µmol zinc protoporphyrin (µmol haem)	150 ± 91 (127)	80 ± 30 (75)	-70 ± 70bc
Serum ferritin (µg/L)	5.0 ± 2.4 (4.7)	14.7 ± 14.7 (9.7)	9.5 ± 14.4bc
Zinc (µmol/L)	13.5 ± 2.0	12.6 ± 2.4	-0.8 ± 2.8d
Retinol (µmol/L)	0.51 ± 0.18	0.68 ± 0.18	0.17 ± 0.25b
Iron-zinc-supplemented group (n = 33)e Haemoglobin (g/L)	104 ± 9	112 ± 10	8±8b
Protoporphyrin (µmol zinc protoporphyrin/mol haem)	120 ± 94 (98)	96 ± 70 (80)	-25 ± 28f
Serum ferritin (µg/L)	7.5 ± 4.0 (6.6)	9.0 ± 12.0 (6.5)	1.5 ± 10.6
Zinc (µmol/L)	13.2 ± 1.7	14.0 ± 3.4	0.89 ± 3.5
Retinol (µmol/L)	0.54 ± 0.18	0.69 ± 0.23	0.15 ± 0.21b

a. Mean ± SD (geometric mean).
b. Significant within-group change (p < .001).
c. Significant between-group difference in treatment effect (p < .01).
d. Significant between-group difference in treatment effect (p = .04).
e. For retinol n = 32.
f. Significant within-group change (p < .01).

The subjects who received the iron supplement showed significant increases in haemoglobin (p < .001) and serum ferritin (p < .001), and a significant decrease in the zinc protoporphyrin/haem ratio (p < .001) (table 2). The subjects who received the iron-zinc supplement had a much smaller but still significant increase in haemoglobin (p < .001) and a significant decrease in the zinc protoporphyrin/ haem ratio (p < .01), while the increase in serum ferritin was not significant (p = .52). The changes in haemoglobin, serum ferritin, and zinc protoporphyrin/ haem ratio were significantly larger in the iron-supplemented group than in the iron-zinc-supplemented group (p < .01 for all values). The significant between-group difference in treatment effect remained significant after correction for differences in initial haemoglobin levels (p < .01).

Both groups showed a significant increase in retinol level. After the two-month supplementation period, 9.1% of the children had serum retinol values below 0.35 µmol/L, while 40% had values between 0.35 and 0.70 µmol/L. Figure 1 shows that the distribution of serum retinol values shifted to the right at the end of the supplementation period. There were no significant differences between the groups for the changes in serum retinol.

In the iron-supplemented group the zinc level decreased non-significantly by 0.82 ± 2.85 µmol/L (p = .12), while in the iron-zinc-supplemented group there was a non-significant increase of 0.89 ± 3.49 µmol/L (p = .17). There was a significant between-group difference for the changes in zinc concentration (p = .037).

The changes in serum retinol levels were not correlated with the changes in zinc levels (r = .038), but they were negatively correlated with the retinol levels at the start of the study (r = -.384, p = .001).

Discussion

The subjects were chronically malnourished children with low height-for-age Z scores and haemoglobin concentrations below 110 g/L. Many children had vitamin A deficiency (15.2%) or marginal vitamin A status (65.1%), even though about 69% of the children were reported to have received vitamin A capsules (200,000 IU) between three and six months before the start of the survey. Similar observations on the limited effectiveness of vitamin A supplements were reported for children from Sudan [20] and Bangladesh [21].

Muhilal et al. [22] estimated that about 50% of Indonesian pre-school children had serum retinol values below 0.70 µmol/L. The results of the current study support this high estimated prevalence.

The mean serum zinc concentration was lower than the value of about 17 µmol/L in healthy Canadian children 4 and 5 years of age [23], similar to the 12.6 µmol/L in growth-retarded French children 2 to 5 years of age [24], but higher than the 10.2 µmol/L in Thai children 7 to 13 years of age [6]. Although 93.1% of the children in the current study had depleted iron stores at the start of the study, only 3.2% of the children had low serum zinc levels (<10.7 µmol/L). No information was collected on the dietary intake of zinc. However, on the basis of the serum levels, it seems that the dietary intake, possibly in combination with homeostatic responses of the body [25], was sufficient to prevent zinc deficiency. This is partially confirmed by the observation that a daily zinc supplementation of 15 mg for eight weeks did not lead to a significant increase in the average serum zinc level.

Previous studies of the effect of zinc supplementation on serum or plasma zinc concentration yielded different results. A more prolonged supplementation led to a significant improvement in zinc status among Thai children [26]; the mean zinc levels increased from 13.2 to 19.0 µmol/L after daily supplementation with 25 mg zinc for six months. Daily supplementation of Bangladeshi children recovering from clinical malnutrition with 4 mg of iron and 10 mg of zinc per kilogram of body weight for one month resulted in a significant increase in plasma zinc concentrations from 8.2 to 18.5 µmol/L [27]. Daily supplementation of Guatemalan schoolchildren with 10 mg zinc for 25 weeks resulted in a more modest, but still significant, increase from 14.2 to 16.2 µmol/L [28]. On the other hand, daily supplementation of Iranian children with 20 mg iron and 20 mg zinc for eight months did not result in any increase in plasma zinc levels, despite a low initial average zinc level of 10 µmol/L [29]. Twice-weekly supplementation with 70 mg zinc for 1.25 years did not increase the plasma zinc concentrations of Gambian pre-school children whose mean initial concentration was 15.2 µmol/L [30], and daily supplementation of US children 8 to 27 months of age with 5.7 mg zinc for six months failed to increase the initial level of 10.7 µmol/L [31]. The wide variation in the effect of zinc supplementation on serum and plasma zinc concentrations may also indicate that, although it is widely used as an indicator of zinc status, serum zinc concentration has limitations for the diagnosis of the actual zinc status [32].

In the iron-only group, significant improvements in haemoglobin, protoporphyrin, and ferritin concentrations occurred. Since there was no placebo group for iron supplementation, it cannot be concluded with complete certainty that the improvement in iron status was due to the iron supplements. However, in a previous study [13] of similar Indonesian subjects using a comparable treatment and including a placebo group, the effect of iron supplementation was similar to that observed in the present study, suggesting that the improvement in iron status was due to the iron supplement.

With respect to the effect of treatment on iron status, the changes were more than twice as large in the group supplemented with iron only as in the group supplemented with iron and zinc. In theory, this can have two explanations. First, the iron in the iron-zinc supplement was provided in the form of iron phosphate, which may be less readily absorbable than the iron sulphate given to the group given iron only [33]. The rate of absorption of ferrous pyrophosphate was reported to be about 50% to 70% of that of ferrous sulphate [34]. This difference in absorbability could partially explain the difference in the change in iron status between the two groups. Second, zinc could have had a negative effect on the absorption of iron. However, no information is available from the literature to support this hypothetical influence of zinc on iron status at the iron-to-zinc ratio that we used.

Zinc is reported to have a positive effect on vitamin A status [7, 8]. In this study the serum retinol of both groups of children showed the same increase, and therefore the zinc supplementation did not have a beneficial effect on serum retinol concentration. This lack of effect may be because the Indonesian children did not appear to be zinc deficient. A similar absence of improvement in serum retinol due to zinc supplementation was observed in Thai children [26].

The increases in serum retinol are unlikely to have been caused by an increased consumption of vitamin A-rich food, since the food consumption pattern was monitored, and the usual intake pattern was reported during the survey. Other factors could possibly explain the significant improvement in retinol concentration. Part of the improvement could be caused by regression to the mean. The deworming treatment of the children at the start of the study could also have an effect on retinol concentration. The serum retinol of Brazilian children increased by 0.15 µmol/L after a deworming treatment only [35].

The average initial serum retinol level of these Brazilian children was comparable to that of our Indonesian children. It has been shown that supplementation of children suffering from intestinal parasites with vitamin A alone is ineffective [35, 36]. Infection may be another reason that the retinol concentrations of the Indonesian children were not related to whether or not the children had received vitamin A supplements [21, 35, 36].

The following conclusions can be drawn. Although they are not a perfect indicator, the serum zinc levels suggest that zinc deficiency is not as much of a problem as anticipated, even in stunted anaemic children. Furthermore, the combined iron-zinc supplement was not effective in improving either zinc or iron status, although the iron-zinc supplement prevented the zinc concentrations from decreasing, as seemed to happen in the iron-supplemented group. The vitamin A status was far from satisfactory, although many children had received vitamin A capsules. The improvement in vitamin A status could not be attributed to zinc treatment, as both groups showed an increase in serum retinol levels. This rise in serum retinol might be accounted for, in part, by the initial deworming provided to both groups. This improvement in vitamin A status without an evident increase in vitamin A intake confirms the need for further investigation of the interaction between parasites and micronutrient status, especially with respect to intervention programmes.

References

1. United Nations Administrative Committee on Coordination/Subcommittee on Nutrition. Second report on the world nutrition situation. Geneva: ACC/SCN, 1992.

2. Suharno D, West CE, Muhilal, Karyadi D, Hautvast JGAJ. Supplementation with vitamin A and iron for nutritional anaemia in pregnant women in West-Java, Indonesia. Lancet 1993;342:1325-8.

3. Mejia MJ, Chew F. Hematological effect of supplementing anemic children with vitamin A alone and in combination with iron. Am J Clin Nutr 1988:48:595-600.

4. Bloem MW, Wedel M, Egger RJ, Speek AJ, Schrijver J, Saowakontha S, Schreurs WHP. Iron metabolism and vitamin A deficiency in children in Northeast Thailand. Am J Clin Nutr 1989;50:332-8.

5. Turnham DI. Vitamin A, iron, and haemopoiesis. Lancet 1993;342:1312.

6. Udomkesmalee E, Dhanamitta S, Yhoung-Aree J, Rojroongwasinkul N, Smith JC Jr. Biochemical evidence suggestive of suboptimal zinc and vitamin A status in school children in Northeast Thailand. Am J Clin Nutr 1990:52:564-7.

7. Smith JC Jr, Brown ED, McDaniel EG, Chan W. Alterations in vitamin A metabolism during zinc deficiency and food and growth restriction. J Nutr 1976; 106:569-74.

8. Shingwekar AG, Mohanram M, Reddy V. Effect of zinc supplementation on plasma levels of vitamin A and retinol-binding protein in malnourished children. Clin Chim Acta 1979; 93:97-100.

9. Smith JC Jr, McDaniel EG, Fan FF, Halsted JA. Zinc: a trace element essential in vitamin A metabolism. Science 1973;181:954-5.

10. Solomons N, Pineda O, Viteri F, Sandstead HH. Studies on the bioavailability of zinc in humans: mechanism of the intestinal interaction of nonheme iron and zinc. J Nutr 1981;113:337-49.

11. Solomons N. Competitive interactions of iron and zinc in the diet; consequences for human nutrition. J Nutr 1986;116:927-35.

12. International Vitamin A Consultative Group. Guidelines for the development of a simplified dietary assessment to identify groups at risk for inadequate intake of vitamin A. Washington, DC: IVACG, 1989.

13. Angeles IT, Schultink JW, Matulessi P, Gross R, Sastroamidjojo S. Decreased rate of stunting among anemic Indonesian preschool children through iron supplementation. Am J Clin Nutr 1993;58:339-42.

14. Schindler R, Klopp A, Gorny C, Feldheim W. Comparison between three fluorometric micromethods for determination of vitamin A in serum. Int J Vit Nutr Res 1985;55:25-34.

15. Olson JA. Serum levels of vitamin A and carotenoids as reflectors of nutritional status. J Natl Cancer Inst 1984;73:1439-44.

16. Smith JC Jr, Butrimovitz GP, Purdy WC. Direct measurement of zinc in plasma by atomic absorption spectroscopy. Clin Chem 1979;25:1487-91.

17. Shaw JC, Bury AJ, Barber A, Mann L, Taylor A. A micro method for the analysis of zinc in plasma or serum by atomic absorption spectrophotometry using graphite furnace. Clin Chim Acta 1982;118:229-39.

18. International Nutritional Anemia Consultative Group. Measurements of iron status. Washington, DC: INACG, 1985

19. Norusis MJ. SPSS/PC+ advanced statistics 4.0. Chicago, 111, USA: SPSS, 1990.

20. Fawzi WW, Herrera GM, Willett WC, El Amin A, Nestel P, Lipsitz S, Spiegelman D, Mohamed KA. Vitamin A supplementation and dietary intake in relation to the risk of xerophthalmia. Am J Clin Nutr 1993; 58:385-91.

21. Rahman MM, Mahalanabis D, Alvarez JO, Waned MA, Islam MA, Habte D, Khaled MA. Acute respiratory infections prevent improvement of vitamin A status in young infants supplemented with vitamin A. J Nutr 1996;126:628-33.

22. Muhilal, Tarwotjo I, Kodyat B, Herman S, Permaesih D, Karyadi D, Wilbur S, Tielsch JM. Changing prevalence of xerophthalmia in Indonesia, 1977-1992. Eur J Clin Nutr 1994;48:708-14.

23. Smit Vanderkooy PD, Gibson RS. Food consumption patterns of Canadian preschool children in relation to zinc and growth status. Am J Clin Nutr 1987;45: 609-16.

24. Chakar A, Mokni R, Walravens PA, Chappuis P, Bleiberg-Daniel F, Mahu JL, Lemonnier D. Plasma zinc and copper in Paris area preschool children with growth impairment. Biol Trace Elem Res 1993;38:97-106

25. Johnson PE, Hunt CD, Milne DB, Mullen LK. Homeostatic control of zinc metabolism in men: zinc excretion and balance in men fed diets low in zinc. Am J Clin Nutr 1993;57:557-65.

26. Udomkesmalee E, Dhanamitta S, Sirisinha S, Charoen-kiatkul S, Tuntipopipat S, Banjong O, Rojroong-wasinkul N, Kramer TR, Smith JC. Effect of vitamin A and zinc supplementation on the nutriture of children in Northeast Thailand. Am J Clin Nutr 1992; 56:50-7.

27. Khanum S, Alam AN, Anwar I, Akbar Ali M, Mujibur Rahaman M. Effect of zinc supplementation on the dietary intake and weight gain of Bangladeshi children recovering from protein-energy malnutrition. Am J Clin Nutr 1988;42:709-14.

28. Cavan KR, Gibson RS, Grazioso CF, Isalgue AM, Ruz M, Solomons NW. Growth and body composition of periurban Guatemalan children in relation to zinc status: a longitudinal zinc intervention trial. Am J Clin Nutr 1993;57:344-52.

29. Mahloudji M, Reinhold JG, Haghsenass M, Ronaghy HA, Fox MRS, Halsted JA. Combined zinc and iron compared with iron supplementation of diets of 6- to 12-year old village schoolchildren in southern Iran. Am J Clin Nutr 1975;28:721-5.

30. Bates CJ, Evans PH, Dardenne M, Prentice A, Lunn PG, Northrop-Clewes CA, Hoare S, Cole TJ, Horan SJ, Longman SC, Stirling D, Aggett PJ. A trial of zinc supplementation in young rural Gambian children. Br J Nutr 1993;69:243-55.

31. Walravens PA, Hambridge KM, Koepfer DM. Zinc supplementation in infants with a nutritional pattern of failure to thrive: a double blind controlled study. Pediatrics 1989;83:532-8.

32. Gibson RS. Zinc nutrition in developing countries. Nutr Res Rev 1994;7:151-73.

33. International Nutritional Anemia Consultative Group. Iron deficiency in women. Washington, DC: INACG, 1984.

34. Brise H, Hallberg L. Absorbability of different iron compounds. Acta Med Scand 1962;171:23-31

35. Marinho HA, Shrimpton R, Giugliano R, Burini RC. Influence of enteral parasites on the blood vitamin A levels in preschool children orally supplemented with retinol and/or zinc. Eur J Clin Nutr 1991;45:539-44.

36. Tanumihardjo SA, Parmaesih D, Muherdiyantiningsih, Rustan E, Rusmil K, Fatah AC, Wilbur S, Muhilal, Karyadi D, Olson JA. Vitamin A status of Indonesian children infected with Ascaris lumbricoides and dosing with vitamin A supplements and albendazole. J Nutr 1996;126:451-7.

Street food: An important source of energy for the urban worker

Abstract
Introduction
Materials and methods
Results
Discussion
Acknowledgements
References

T. Sujatha, Veena Shatrugna, G. V. Narasimha Rao, G. Chenna Krishna Reddy, K. S. Padmavathi, and P. Vidyasagar

The authors are affiliated with the National Institute of Nutrition, Indian Council of Medical Research in Hyderabad, India.

Abstract

Urbanization and migration have changed the patterns of living and eating. The easy availability of foods in the cities and small towns has helped workers cope with long periods of absence from home. Under these circumstances, existing methods for the assessment of dietary intakes of individuals and larger populations have not been entirely satisfactory. An urban study was therefore planned to investigate the caloric intakes of men from the low socio-economic group. The problems of using the 24-hour-recall method were investigated in 51 randomly selected households. Men from these households were interviewed individually for a recall of street foods eaten by them. It was significant that the energy derived from street foods helped men from the low-income group meet their recommended dietary allowance for energy.

Introduction

It has been acknowledged that the existing methods for the assessment of dietary intakes of individuals and larger populations have not been entirely satisfactory [1, 2]. Certain built-in limitations may vitiate the findings and thus influence the interpretation of the results [3]. The critical need to improve dietary assessment methods has been stated repeatedly [2, 4, 5].

One of the widely used methods for assessing dietary intakes of individuals is the 24-hour recall, in which adult women of the household are contacted for information on the food intake of all the individual family members for the preceding day. Although food consumed outside the house is supposed to be assessed by this method, it assumes that most of the meals are cooked and consumed at home, and that women are able accurately to recall the quantities eaten by each member of the household.

It has now been acknowledged that with rapid urbanization, more than 40% to 50% of the population will live and work in urban areas by the turn of the century [6]. Apart from certain advantages, migration would impose a large number of constraints on the population, especially in relation to their dietary habits. With long hours of continuous work in the factories at great distances from home and the unavailability of a rapid transport system, it is logical that the majority of the urban workers will be unlikely to depend on food prepared at home. Access to regular cash income increases the possibility of the intake of energy-rich snacks, often fried in fats and oils, or meals from the innumerable food stalls, restaurants, and canteens scattered all over the city, thus altering the meal intake patterns of the urban workers.

At this juncture street foods have assumed importance in international research bodies. The Food and Agricultural Organization has initiated a number of investigations to highlight the various concerns regarding street foods. The studies include nutrient contribution, hygiene, sanitation, socio-economic and regulatory aspects, and licensing systems regarding street foods [6].

Studies from a few countries in the South-East Asian region suggest that 25% and 30% of household expenditures are spent on street foods in Kuala Lumpur and Lloilo, respectively [7]. A report from Bangkok highlights the fact that 90% of the population eats out most of the time [8]. Studies from India [9], Indonesia [10], Nigeria [II], and Peru [12] have looked into the major nutrients derived from freshly prepared street foods. A study on total dietary intake using diary records among students from Bogor University indicated that street foods constituted the largest part (78%) of their total energy intake [13]. Investigation into the licensing requirements and legislation revealed a lack of uniformity in this region. In the Kuala Lumpur Federal Territory of Malaysia, there are more than 25,000 vendors, of whom over 10,000 remain unlicensed [14]. A study in Pune City, India, showed an appreciable number of unlicensed street food vendors [9].

With the growing importance of food vendors in Indian cities, it was decided to investigate the methodological problems with the 24-hour-recall method by directly interviewing urban men. In addition, the energy contribution of food obtained from wayside stalls by men from a large urban slum was assessed.

Materials and methods

A total of 51 households belonging to the low socio-economic group were randomly selected from an urban slum situated in the busiest part of Hyderabad. The men of these households were largely employed as mobile vendors, auto drivers, rickshaw pullers, hamalies (head load workers), daily wage workers in small factories, carpenters, and so forth. The women were either housewives, servant maids, or home-based workers, such as bead makers, tailors, and so forth.

Women were contacted on three consecutive days for an assessment of the dietary intakes of the men, using the 24-hour-recall method. The differences between the days fell just short of statistical significance F_2,150 = 0.06). The energy value of these diets was calculated using the information published in the Nutritive Value of Indian Foods [15]. During the course of the interview, it was observed that women were not aware of the foods consumed by their spouses outside the home.

It was therefore decided to interview the men on the same days that the women were contacted for the 24-hour dietary-recall survey. Here also there were no differences between days (F_2,150 = 0.75, not significant). Male investigators interviewed the men individually to obtain the details of the food and beverages consumed by them outside their homes (outside calories). The foods consumed at home were also verified with the men. The caloric value of these foods and beverages were calculated separately using the information on cooked foods generated at the National Institute of Nutrition [16].

The data were analysed by analysis of variance (ANOVA).

Results

The mean age of the men was 34.2 ± 6.7 (SD) years, ranging from 23 to 52 years. All the men were married and lived in nuclear families. They had at least one pre-school child and had been employed continuously over the past two years. According to the 24-hour-recall method, the men’s mean energy intake from foods consumed at home was 1,960 ± 75 (SE) kcal. However, it was observed that all the men in the study had significant outside energy intakes from food and beverages obtained from street vendors. These included beverages, such as sweetened tea with milk (one cup provided 80 kcal), or toddy, a cheap, locally brewed liquor (consumption ranged from 100 to 1,300 ml per day). Many ate snacks (usually prepared from gram flour and fried), whereas a few ate meals that consisted of rice and pulse preparations sold at cheap hotels. The mean caloric intake from outside foods was 472 + 57 kcal. The total energy intake, calculated by adding the intake derived from 24-hour recall by the wives and the intake outside the home reported by the men, was significantly larger than the intake derived from 24-hour recall by the wives (2,432 vs. 1,960 kcal, p <.05) (table 1).

An attempt was made to investigate the effect of occupation on the men’s intake of street foods. It was hypothesized that certain occupations, such as vending, rickshaw pulling, auto driving, and so forth, which required mobility, increased men’s access to street foods and thus influenced their outside energy intakes. However, low- and high-mobility workers did not have different amounts of energy intake from street foods. Moreover, the absence of women from the home because of their own work did not contribute to the increase in the men’s intake of street foods.

As shown in table 1, the energy contribution of the various street foods averaged 472 ± 57 kcal (range, 40-1,541 kcal) in the 51 male workers studied. Table 2 shows the contribution of each of four categories of street foods to the total energy consumed outside the home, for the subgroups of men who reported their consumption. All 51 subjects reported consuming tea on the day of the 24-hour recall. Twenty-three men consumed toddy on the reported day, for a mean energy contribution of 140 ± 28 kcal (range, 40-494 kcal), which represented 28% of the total of 498 ± 85 kcal consumed outside the home by that subgroup. The 146 ± 10 kcal (range, 40-320 kcal) contributed by sweetened tea drinks represented 31% of the total energy from street food. A subgroup of 34 men reported consuming “snacks,” with a derived energy content of 196 ± 22 kcal (range, 55-534 kcal), which was 33% of the daily intake outside the home. Finally, slightly more than one-fifth of the men (11 individuals) consumed full meals from street food sources. Their total energy from outside sources was 1,021 + 136 kcal; these meals contributed an average of 615 ± 93 kcal (range, 280-954 kcal), which was 60% of the total non-household contribution.

TABLE 1. Daily energy intakes^a of men inside and outside the home in relation to mobility of occupation

Mobility	n	Home intakeb	Intake outside homec	% calories from outside	Total intake
Lowd	31	1,842 ± 82e (784-2,606)	413 ± 67 (40-1,541)	21	2,255 ± 68f (1,088-2,735)
Highg	20	2,144 ± 135h (1,147-3,270)	563 ± 102 (52-1,491)	18	2,707 ± 125i (1,812-3,606)
Total	51	1,960 ± 75e (784-3,270)	472 ± 57 (40-1,541)	19	2,432 ± 71f (1,088-3,606)

a. kcal ± SE (range).
b. According to recall by wife.
c. According to interview with man.
d. Plumber, electrician, salesman, etc.
e. and f.Significantly different at p < .05.
g. Rickshaw puller, carpenter, coolie, etc.
h. and i. Significantly different at p < .001.

Food item	No. of subjects consuming	Daily caloric equivalent	% of all calories from outside	Total daily calories from outside
Toddy	23	140 ± 28	28	498 ± 85
Tea	51	146 ± 10	31	472 ± 57
Snacks	34	196 ± 22	33	593 ± 74
Meal	11	615 ± 93	60	1,021 ± 136

a. kcal ± SE.

Discussion

Over the past 50 years or more, numerous studies have attempted to establish the reliability and validity of various methods used for dietary surveys, but it has become recognized that the methods and the criteria used to judge them are variable [2].

Dietary assessment by the 24-hour-recall method has been found to be advantageous for obvious reasons. It requires a minimum of time because it relies on a cost-effective tool, the memory of the respondent over a short period of time. However, Beaton et al. [19] demonstrated the inter- and intraindividual variation in this method. The accuracy of the intake data is affected by inaccurate recall of amounts and poor memory of past dietary practices [3]. Mertz and co-workers demonstrated an underreporting (18%) of caloric intake by both men and women, when reported intakes were compared with the requirements for maintaining their body weight [20].

An unexplained difference of 1,195 kcal/day between expenditure and intake was reported in a study on Gambian women [21]. Energy expenditures of women were measured during peak agricultural activity by the doubly labelled water method, and a figure of 2,490 kcal was obtained. The estimated energy intakes were only 1,152 kcal, with a deficit of 1,453 kcal, of which 203 kcal was accounted for by weight changes due to fat oxidation. The authors inferred that the estimates of energy intake might be substantially in error. Carefully conducted studies from our institute have also suggested that the 24 hour-recall method may underestimate energy intake in urban men [22].

TABLE 3. Daily home and total energy intakes^a of men with occupations of low and high mobility as a percentage of RDA

Mobility	n	Home intakeb	% RDA	Total intakec	% RDA	RDA
LOWd	11	1,693 ± 114 (784-2.284)	75	2,136 ± 84 (1,798-2,726)	94	2,272 ± 94 (2,046-2,923)
Highe	33	2,035 ± 94 (1,147-3,270)	76	2,543 ± 85 (1,652-3,606)	95	2,664 ± 37 (2,046-3,035)
Total	44	1,950 ± 79 (784-3,270)	76	2,441 ± 72 (1,652-3,606)	95	2,566 ± 42 (2,046-3,035)

a. kcal ± SE (range).
b. According to recall by wife.
c. Home intake according to recall by wife + intake outside home according to interview with man.
d. Plumber, electrician, salesman, etc.
e. Rickshaw puller, carpenter, coolie, etc.

A study on street food consumption in Haiti showed that 14% of the RDA of men aged 22 to 23 years was met by food intake outside the home [23]. In the United States, a higher proportion of men (69%) reported obtaining and eating food away from home [24]. In the present sample drawn from the urban low socio-economic group, about 20% of calories were derived from street foods. In fact, when the outside calories were added to the home calories, the energy gap in this group of men narrowed and closely approached the RDA. The correlation between intake and requirement is absent when only the home calorie intake is considered. In the study group, 70% of the men required 200 to 500 kcal from street foods to bridge the calorie gap. It is possible that the lack of association reported by some studies between the intake and the nutritional requirements of populations in India also may be attributed to the absence of information about the energy derived from street foods.

There have been numerous attempts to develop a method that would be easy and quick to administer but that would achieve a more representative result when compared with the 24-hour-recall method [24]. This study suggests that it is imperative to extend the 24-hour-recall method to include individual interviews with the subjects to arrive at the total energy intake of urban men. In addition, the study points to the significant contribution of street foods to filling the energy gap in adult men. What is disturbing, however, is the quality of the food supplement. It appears that fried snacks, sweetened tea, and so forth contributed to an increased intake of fat and empty calories, with little vitamin and mineral content. However, the micronutrient quality of these foods needs further study and analysis.

Acknowledgements

We are grateful to the director of the National Institute of Nutrition, Hyderabad, for encouragement during the study. We also thank Dr. Leela Raman, Dr. Ghafoorunissa, and Ms. B. V. S. Thimmayamma for their valuable suggestions, and K. Usha Rani, Md. Nazeema Begum, A. Padma Siromani, and B. P. Prema Kumari for their help.

References

1. Bingham SA. The dietary assessment of individuals: methods, accuracy, new techniques and recommendations. Nutr Abstr Rev 1987;57:705-41.

2. Medlin C, Skinner JD. Individual dietary intake methodology: a 50 year review of progress. J Am Diet Assoc 1988;88:1250-7.

3. National Research Council, Food and Nutrition Board. Implications for reducing chronic disease risk. Methodological considerations in evaluating the evidence. 4th ed. Committee report on diet and health. Washington, DC: National Academy Press, 1992:23-38.

4. Dwyer JT, Krall EA, Coleman KA. The problem of memory in nutritional epidemiology research. J Am Diet Assoc 1987;87:1509-12.

5. Woteki C. Improving estimates of food and nutrient intakes. Applications to individuals and groups. J Am Diet Assoc 1985;85:295-9.

6. Dawson RJ. International activities in street foods. In: Proceedings of the first Asian conference on food safety, The challenges of the 90s. Kuala Lumpur, Malaysia. Rome: Food and Agriculture Organization, 1990:129-34.

7. Food and Agriculture Organization and Food Technology Development Center. Report of the FAO regional workshop on street foods in Asia. Yogyakarta, Indonesia: Bogor Agricultural University and FAO, 1986.

8. Food and Agriculture Organization. Street foods. FAO food and nutrition paper No. 46. Rome: FAO, 1989.

9. Food and Agriculture Organization and State Public Health Laboratory. Report of the study on street food in Pune. Pune, India: Government of Maharashtra and FAO, 1986.

10. Food and Agriculture Organization and Food Technology Development Centre. Study on street foods in Bogor. Bogor, Indonesia: Bogor Agricultural University and FAO, 1984.

11. Food and Agriculture Organization and Department of Human Nutrition. Report of the study on street foods in Ibadan. Characteristics of food vendors and consumers: implications for quality and safety. Ibadan, Nigeria: University of Ibadan and FAO, 1987.

12. Food and Agriculture Organization/Pan American Health Organization. Final report and working papers of the workshop on street foods in Latin America. Lima, Peru: FAO, 1985.

13. Netherlands Organization for Applied Scientific Research (TNO). Nutrition and Food Research: Cooperation with Third World Countries. Annual Report. Zeist, Netherlands: TNO Nutrition and Food Research, 1989:28-30.

14. Food and Agriculture Organization. Assessment of economic impact of street foods in Penang, Malaysia. Rome: FAO, 1990.

15. Gopalan C, Ramastri BV, Balasubramanian SC. Revised by Narasinga Rao B, Deosthale, Pant KC. Nutritive value of Indian foods. Special publication. Hyderabad: National Institute of Nutrition, Indian Council of Medical Research, 1989.

16. Pasricha S. Count what you eat. Hyderabad: National Institute of Nutrition, Indian Council of Medical Research, 1995.

17. Food and Agriculture Organization/World Health Organization/United Nations University. Energy and protein requirements. Technical Report Series No. 724. Geneva: WHO, 1985.

18. Indian Council of Medical Research. Nutrient requirements and recommended dietary allowances for Indians. A report of the expert group of the Indian Council of Medical Research. Hyderabad: National Institute of Nutrition, 1995.

19. Beaton GH, Milner J, Corey P, McGuire V, Cousins M, Stewart E, de Ramos M, Hewitt D, Grambsch PV, Kassim N, Little JA. Sources of variance in 24 hour dietary recall data: implications for nutrition study design and interpretation. Am J Clin Nutr 1979:32:2546-59.

20. Mertz W, Tsui JC, Judd JT, Reiser S, Hallfrisch J, Morris ER, Steele P, Lashley E. What are people really eating: the relation between energy intake derived from estimated diet records and intake determined to maintain body weight. Am J Clin Nutr 1991,54:291-3.

21. Singh J, Prentice AM, Diaz E, Coward WA, Ashford J, Sawyer M, Whitehead RG. Energy expenditure of Gambian women during peak agricultural activity measured by the doubly labelled water method. Br J Nutr 1989;62:315-29.

22. Ghafoorunissa, Reddy V, Sesikaran B. Palmolein and groundnut oil have comparable effects on blood lipids and platelet aggregation in healthy Indian subjects. Lipids 1995;30:1163-9.

23. Webb RE, Hyatt SA. Haitian street foods and their nutritional contribution to dietary intake. Ecol Food Nutr 1987;21:199-209.

24. United States Department of Agriculture. Nationwide food consumption survey. Continuing survey of food intakes of individuals. Men 19-50 yr. 1 day. Report No. 85-3. Hyattsville, Md, USA: Nutrition Monitoring Division, Human Nutrition Information Service, 1985.

25. Block G. A review of validations of dietary assessment methods. Am J Epidemiol 1982;115:492-505.

Positive and negative deviance in growth of urban slum children in Bombay

Abstract
Introduction
Materials and methods
Results
Discussion
Acknowledgements
References

Shameera S. Merchant and Shobha A. Udipi

The authors are affiliated with the Food Science and Nutrition Department in S. N. D. T. Women’s University, Mumbai, India.

Abstract

This study was undertaken on 500 children to examine the psychological, behavioural, socio-economic, and physiological correlates that distinguish positive deviants from median growers and negative deviants in an urban slum of Bombay. Children were monitored for six months and then placed into one of these three categories. Mothers of 50 children from each category were interviewed. Positive deviance was associated with better maternal child-care wisdom and technology in terms of motivation to weigh, groom, interact with, and actively feed the child. Mothers of positive deviants were prompt in utilizing health services and had better nutrition knowledge scores. Negative deviance was associated with short gestational age, illiteracy, early age at marriage, poor maternal decision-making power, less concern about child welfare, and inadequate social support. Although median growers came from families with more floor space, the interpregnancy interval was short, suggesting that mothers were not able to devote sufficient time to their children. The study brings out the importance of addressing maternal behavioural and psychosocial correlates in programmes aimed at improving child health and nutrition.

Introduction

Undernutrition has been found to persist in households and communities where income and reserves are both stable and sufficient to avoid such conditions. There are also well-nourished individuals among the undernourished in communities where both health care and food supply are chronically short [1]. There are women who can cope and manage to rear healthy and active children and who yet belong to the same community as others who cannot [2]. Thus, maternal technology is a distinct determinant of the malnutrition complex [3]. Positive deviants are children who demonstrate better growth than other children in low-income families living in impoverished environments where a majority of the children suffer from nutritional growth retardation and frequent ill health.

Positive deviance has been viewed as “a social, behavioural, and physiological adaptation to nutritional stress and limited food availability” [4]. Modification and operationalization of the concept of positive deviance and the addition of the terms negative deviance and median growth have been suggested [5]. Research on positive and negative deviance can lead to programme applications such as nutrition and health education, community development pilot projects, and so forth. It is also important to identify aspects of maternal technology that favour positive deviance in the growth of children, with specific reference to behavioural practices of the mothers who rear their children in a particular manner in various cultural settings.

One study of deviance was conducted in rural Tamil Nadu [5], but none have been done in urban areas of India. Landholding may be a significant variable in a rural setting but not in an urban slum, where the aetiological factors and mechanisms may be quite different. Therefore the present study was undertaken in urban Bombay (1) to examine the psychosocial, behavioural, socio-economic, and physiological correlates that may distinguish positive deviants from median growers and negative deviants; and (2) to identify those correlates that need to be addressed in programmes promoting child health and nutrition.

FIG. 1. Flow chart of the study design

Materials and methods

Parental and household characteristics
Mothers’ child-care “wisdom” knowledge, attitudes, and behaviour
Child characteristics

All children (n = 500) 6 to 36 months of age from four areas of an urban slum in Bombay were monitored for 6 months from June to December 1992. Weights were measured every month, and heights were measured in the first and sixth months using standard techniques. Figure 1 summarizes the study design. Rank according to growth velocity and growth over a six-month period were used as criteria for classifying the children.

Children were classified as positive deviants, median growers, or negative deviants, according to whether they consistently tracked at the high, medium, or low end, respectively, of the growth spectrum, which takes into account both growth velocity and attained weight. The mothers of 50 children from each deviant category and the mothers of 50 median growers were interviewed using a semi-structured interview schedule to obtain relevant information. The conceptual framework used to study the variables associated with the deviant pattern of growth is shown in figure 2. More than 100 variables were investigated to determine their influence on children who were placed in a deviant category. The conceptual framework represents a broad statement of the causal relationships at work in the environment. The variables investigated are summarized below.

Parental and household characteristics

Maternal characteristics: health and nutritional status

» Mother’s age at marriage and reproductive history (number of live births, number of stillbirths, number of abortions and miscarriages, number of living children, causes of child death(s), child spacing);

» Food intake during pregnancy (whether it increased, decreased, or remained unchanged as compared with the pre-pregnant state; which items increased by how much and why; whether any special traditional foods were consumed during pregnancy).

» Illiterate;
» Can read and write but had no schooling;

FIG. 2. Conceptual framework: effect of mothers’ child-care “wisdom,” and child’s characteristics on child’s nutrition and health status [adapted from ref. 4]

» Studied up to the 4th standard;
» Studied up to the 10th standard;
» Studied above the 10th standard but did not graduate;
» Graduated;
» Did postgraduate study.

Proxied on the basis of whether the mother was abandoned or divorced by the father of the child or whether the father had married again.

Household structure and socio-economic status

Income or wealth

» Cash income;

» Possession of luxury items (TV, cable connection, etc.);

» Frequency of purchase and consumption of food items such as milk and milk products, non-vegetarian items, vegetables, cereals, and pulses; whether daily, twice a week, once a week, once in 15 days, once a month, rarely or never;

» Type of house

- Kuccha: two or more walls made of material other than bricks: tin, asbestos, cloth, plastic, tarpaulin, etc.;

- Semi-pucca: one of the four walls not of bricks;

- Pucca: all four walls made of bricks.

» Floor space (approximate square feet);

» Maternal wealth (proxied on the basis of gold jewelry possessed by the mother); the jewelry worn by the mother was recorded and it was verified whether it was made of gold. The amount and kind of jewelry worn (bangles, rings, earrings, chain, nose ring) were assigned scores: none = 0, light = 1, heavy = 2. Thus the maximum possible score was 10.

» Number of persons and their relationship with each other; type of family (joint, nuclear, or extended); number of children under 6 years of age; number of mothers in family with children under 6 years of age; number of adults over 18 years of age.

Mothers’ child-care “wisdom” knowledge, attitudes, and behaviour

Nutrition and health

The mothers’ nutrition knowledge and awareness were quantified by assigning scores to the responses to the following:

» How often should the child be fed? (frequently/5-6 times a day = 2; 2-3 times a day = 1; “whenever time permits” = 0);

» What should the management be during diarrhoea? (gives oral rehydration solution [ORS] and knows correct amount of ingredients to be used/gives soft diet/visits doctor = 5; gives ORS but does not know correct concentration of ingredients = 1; restricts food/no treatment/does not know what to do =0);

» How do you judge whether your child is growing normally? (height/weight = 5; physical milestones/ comparison with other children = 4; frequency of falling ill = 3; appearance, “by looking”/physical activity of child/appetite = 2; “speaks well, understands well” = 1; does not know/free from nazar [evil eye] = 0);

» Why do you weigh the child? (to assess growth pattern of child = 5; “community worker called to weigh”/does not know = 0);

» Whether mother knows weight of child (yes = 5; no = 0);

» Whether mother knows extent of change in weight from previous weighing to current weighing (yes = 5; no = 0);

» If your child loses weight, what do you attribute it to? (does not eat well/illness/does not take care of child/”vitamin” deficiency = 5; does not know/ nazar = 0);

» What do you do about it? (dietary modifications/ visit doctor = 5; give tonics = 3; does not know/ does not perceive the need to take action/takes child to hakim [traditional healer] = 0);

» Do you feel breastfeeding is important? (yes = 5; no = 0);

» Why is breastfeeding important? (“it is full of vitamins”/”good for child’s health, has strength”/ “child gets whatever mother eats”/child does not get diarrhoea, free from infections/easy to digest = 5; custom, belief/economical = 2; does not know = 0);

» Was colostrum given to the child? (yes = 5; no = 0);

» Why were supplementary foods given to the child? (good for child’s health/insufficient milk = 5; told by friend, doctor, nutrition educator = 4; “so that the child will stop breastfeeding and be habituated to oaf/child’s demand/mother’s pregnancy/ working mother/mother’s illness = 2);

» How do you portion child’s food? (according to child’s age = 5; as much as child wants = 0);

» Is the child immunized? (complete and age-appropriate = 5; incomplete = 2; no = 0);

» Scores for supplementary feedings (table 1).

The general health, hygiene, and sanitation of the household were assessed by observing the condition of the floor, cleanliness of vessels, and storage conditions of food and water. The classification was done qualitatively as follows:

» Condition of floor

- Good: clean, well swept, free from dirt;
- Average: an otherwise clean floor, but with food particles or dirt on it;
- Poor: not swept, dirt or leftover food on the floor, abundant flies on the floor.

» Cleanliness of vessels: a similar categorization in terms of good, average, or poor was made to distinguish the extent of cleanliness;

» Storage condition of food: whether the food containers were kept on the floor or at a higher level (platform, table, shelf, etc.); whether the food was covered or exposed;

» Storage condition of water: cleanliness of handas (water containers); whether kept covered or uncovered.

» Antenatal care (whether the mother went to the doctor during pregnancy; number of visits; whether she was immunized);

» Child care (stage of illness of child when medical aid was sought; food intake of child during illness; whether it was restricted, increased, or unchanged; whether dietary modifications in consistency and form were made).

Breastfeeding, bottle-feeding, and weaning behaviour

» Breastfeeding and bottle-feeding behaviour (whether child was breastfed or bottle-fed; day of introduction of breastmilk; whether colostrum was given; reasons for giving colostrum; whether and why mothers considered breastfeeding important; duration of breastfeeding; reasons for stopping breastfeeding);

» Weaning behaviour (the ideal age for introduction of supplementary foods; reasons for introduction; mother’s perception of child’s appetite).

A scoring pattern for supplementary foods was developed (table 1). A score ranging from +5 to - 2 was assigned for each category. A score of +5 indicated the most appropriate age (in months) for introducing the food. As the introduction of the food was delayed progressively, lower scores were assigned. Similarly, if a particular food was introduced much earlier than appropriate, a score of - 2 was assigned.

Thus, if a mother of a 12-month-old child (who had not yet been introduced to meat and eggs) had introduced rice or roti by 8 to 10 months, mashed vegetables by 10 to 12 months, soups, juices, and fruits by 8 to 10 months, dal by 6 to 8 months, and bread and biscuits by 8 to 10 months, the ages of introduction of the various foods would have been delayed. Thus, it was anticipated that the mother would also introduce meat and eggs at a later age, and hence a lower score (3) would be assigned. Thus, the total score would be [4+3+4+4+4+4+4+4+3+3] =37.

TABLE 1. Scoring pattern for supplementary foods

Food	Age at introduction (mo)
	<4	4-6	6-8	8-10	10-12	12-18	18-24	>24
Rice, roti		5	5	4	4	1	0	0
Mashed leafy vegetables		5	5	4	3	2	1	0
Other vegetables		5	5	5	4	3	2	0
Soups, juices		5	5	4	3	2	1	0
Fruits		5	5	4	4	3	2	0
Eggs		5	5	4	4	3	2	0
Dais		5	4	3	2	1	0	0
Meat, fish, poultry			5	5	4	3	2	0
Biscuits	-2	5	5	4	3	2	1	0
Pav (bread)		5	5	4	2	1	0	0
Maximum possible score								50

The mother’s decision-making power

The decisions were classified into various categories:

» Household decisions (education, health care, in-trahousehold food and task allocation);

» Financial decisions (budgeting, spending family income as well as income earned by the mother herself);

» Decisions regarding self (food intake during pregnancy and lactation; breastfeeding and weaning of the child; whether she can take a job outside the house);

» Social decisions (outings; attendance at functions such as child naming, funerals, or religious ceremonies).

Mental and behavioural development

» Manner in which the child is fed: the mothers were asked how the child was fed, and five categories of feeding patterns were recorded:

- Child feeds self without supervision or encouragement from parents;
- Child eats from the same plate as siblings;
- Child feeds self but is supervised and encouraged by parents;
- Mother or grandmother feeds the child;
- Child is not served food but is left to the mercy of the neighbours or siblings to be fed.

» The frequency of consumption of eggs, non-vegetarian items, bread, roti, fruits, and vegetables by child (daily, sometimes, or rarely);

» The frequency of feeding the child.

An inspection of the child’s nails, skin, face, and clothes was done:

» Nails: overgrown and dirty; short but dirty; well-trimmed and clean;
» Skin: boils or rashes; dirty; clean;
» Face: runny nose; muddy; clean and free from dirt;
» Clothes: child not fully clothed; clothes torn, dirty, or neat and clean.

The motivation of the mother to weigh her child was assessed on the basis of her willingness and eagerness to weigh the child, her interest in knowing the weight of the child as well as the change in weight from the previous month, her enthusiasm to know what should be done in case of weight loss, and her attendance during growth monitoring. The mothers were classified into the following categories:

» Totally uninterested: mother extremely reluctant to have child weighed; needed persuasion by the community worker or researcher who would eventually bring the child to be weighed; mother seldom interested in knowing the weight of the child;

» Uninterested: mother brought the child to be weighed or sent someone to weigh the child when persuaded to do so, but was not interested in knowing the weight of the child or the change in weight from the previous month;

» Slightly interested: the mother herself (or sometimes a guardian) came to weigh the child without much persuasion and listened to the report about the child’s weight if it was given by the community worker or researcher, but never took the initiative to ask about the weight change;

» Highly interested: mother eagerly waited each month for her child to be weighed; inquired without being persuaded to weigh the child; showed interest in what was to be done if the child lost weight.

Variable	Positive deviants		Median growers		Negative deviants
	Mean + SD	%	Mean ± SD	%	Mean ± SD	%
Pre-term (<37 wk)		0		2		14
Weight gain or loss (kg)a	1.31 + 0.37		0.66 ± 0.23		-0.17 ± 0.37
Height gain (cm)b	5.64 ± 3.07		4.73 ± 2.59		3.97 ± 1.87

a. p < .0001.
b. p £.01.

Child characteristics

Age, sex, birth order, birthweight, gestational age (full-term >37 weeks, pre-term <37 weeks), number of siblings, and birth interval. Comparisons were made for the variables between the three groups using SPSS (Statistical Package for Social Sciences). The chi-square test and analysis of variance were used for tests of statistical significance.

Results

Child characteristics
Parental and household characteristics
Mothers’ child-care “wisdom” knowledge, attitudes, and behaviour

The results are discussed according to the conceptual framework presented in the methodology.

Child characteristics

There were no significant differences among the three groups of children in age (17.5 + 9.1 months), sex, birth order (2.9 ± 2), or birthweight (2.72 ± 0.6 kg). However, there were significant differences in gestational age and attained weight and height, with a larger number of negative deviants being pre-term (table 2). Positive deviants had a mean weight gain of 1.31 ± 0.37 kg over the six-month study period, which was twice the amount of weight gained by median growers. In contrast, negative deviants had an overall weight loss of 0.17 ± 0.37 kg (fig. 3). Similarly, positive deviants showed a mean height gain of 5.64 ± 3.07 cm, which was approximately 1 cm more than that of median growers (4.73 ± 2.59 cm). The latter was approximately 1 cm greater than that of negative deviants (3.97 ± 1.87 cm) (fig. 4).

Parental and household characteristics

Household structure and socio-economic status

Income and wealth. There were no significant differences among the three groups in total income (Rs 2,860 ± 1,842), per capita monthly income (Rs 387 ± 186), and economic dependency ratio (0.267 + 0.126). Thus, economic status, which is a critical aspect of positive deviance research, had no confounding effects in this study. The three groups did not differ significantly in possession of luxury items and maternal wealth, but mothers of negative-deviant children were more likely to possess gold jewelry. The channelization of income, however, differed, with households having positive-deviant children purchasing and consuming fruits, leafy vegetables, and biscuits more often (p < .05). Median growers came from homes with more floor space, whereas positive deviants were more likely to live in pucca houses (table 3).

Household size and structure. No significant differences were observed with reference to number of family members, children, and type of family.

Parents’ educational status

More mothers than fathers were illiterate. Illiteracy of both parents predisposed the child to negative deviance (p < .05) (fig. 5).

Social disruption and conflict with husband

Although the three groups did not differ significantly, the percentage of mothers who reported marital disharmony was highest among mothers of negative deviants (14%), followed by mothers of positive deviants (6%) and mothers of median growers (4%).

Maternal characteristics, health, and nutritional status

Mothers of negative-deviant children were married at a younger age and had their first pregnancy earlier (p < .05). A higher percentage of mothers of median growers were either pregnant or lactating at the time of the study; mothers of median growers also had more abortions or medical terminations of pregnancy (p < .05). Equal percentages (26%) of mothers of positive deviants and mothers of median growers increased their food intake during pregnancy, as compared with only 18% of mothers of negative-deviant children (p < .05) (table 3).

FIG. 3. Weight gain or loss (kg) of children over six months according to growth category

Mothers’ child-care “wisdom” knowledge, attitudes, and behaviour

Nutrition and health

Mothers of positive-deviant children achieved a higher mean score for nutrition knowledge (83.6 ± 16.3), followed by mothers of median growers (80.1 ± 13.0; not significant) and mothers of negative deviants (71.9 ± 20.4; p < .05) (table 4). Analysis of variance for nutrition knowledge score according to mother’s educational status, father’s educational status, and mother’s age in relation to deviant category showed no significant differences among the three groups. Positive deviants came from homes with good storage practices for food and water as compared with the other two groups (fig. 6).

Mothers of positive deviants exhibited better health-seeking behaviour for their children as well as for themselves during pregnancy. They were more likely to take a sick child to the doctor immediately and to change the consistency and form of the diet, and less likely to restrict the child’s food intake during illness. Similarly, more mothers of positive-deviant children received antenatal care during pregnancy, and mothers of positive-deviant children had more antenatal-care visits (table 4).

Mothers of positive deviants initiated breastfeeding earlier and were more likely to breastfeed their infants. When these mothers were asked the ideal age to introduce solids (knowledge), they reported a mean age of 6.9 ± 3.0 months, which was earlier than the actual age (practice) that they introduced solids (7.7 ± 4.6 months), indicating that the mothers were not putting their knowledge into practice.

Maternal decision-making power

Mothers of negative-deviant children had less power to make decisions about seeking medical treatment for children during major illnesses and attending social functions such as puja and child naming (table 4).

Mental and behavioural development

Manner in which the child is fed. Negative-deviant children were less likely to receive parental supervision while feeding and were more likely to be left to be fed by siblings or neighbours (fig. 7). When mothers were asked how frequently their children should be fed (knowledge), no significant differences were seen among the three groups. In contrast, when mothers were asked how frequently they actually fed their children (practice), significant differences were seen, with positive deviants being fed more often, followed by median growers and then by negative deviants. Seventy-eight percent of positive deviants and 82% of negative deviants consumed roti (flat bread made from wheat flour) as the main staple. Negative deviants consumed fruits less often.

FIG. 4. Height gain (cm) of children over six months according to growth category

TABLE 3. Significant variables relating to parental and household characteristics

Variable	Positive deviants		Median growers		Negative deviants
	Mean + SD	%	Mean ± SD	%	Mean ± SD	%
Floor space (sq ft)a	157 ± 129		231 ± 272		113 ± 89
Gold jewelry (score)b	1.9 ± 1.8		1.6 ± 1.8		2.3 ± 3.2
Pucca housec		80		62		40
Illiterate fathera		22.4		18.4		48
Illiterate mothera		28		40		58
Mother’s age at marriage (yr)a	16.7 + 2.1		16.9 ± 2.1		15.6 ± 2.1
Abortions or medical terminations of pregnancy (no.)a	1.0 ±0		1.3 ± 0.5		1.0 ±0
Mother pregnant or lactatinga		22.4		47		28
Increased food intake during pregnancya		26		26		18

a. p £ .05.
b. Not significant.
c. p £ .0001.

FIG. 5. Percentage of children with illiterate parents according to growth category

FIG. 6. Cleanliness and storage of water and food in homes of children according to growth category

TABLE 4. Significant variables relating to mothers’ child-care “wisdom” knowledge, attitudes, and behaviour

Variable	Positive deviants		Median growers		Negative deviants
	Mean ± SD	%	Mean ± SD	%	Mean ± SD	%
Nutrition and health
Nutrition knowledge (scores)a	83.6 ± 16.3		80.1 ± 13.0		71.9 + 20.4
Good storage practices for food and watera		72		56		38.7
No antenatal care for mothera		2		2		16
Antenatal-care visits (no.)b	7.7 + 3.7		6.5 ± 3.5		5.7 ± 5.7
Child taken to doctor immediately when illb		57.1		40		28
Food restricted during child’s illnessc		69.4		64		88
Diet modified during child’s illnessd		44.9		20		24
Breastfeeding initiated within first 2 daysa		56		50		44
Reasons for introduction of supplementary foods (scores)b	3.6 ± 1.3		4.3 ±1.3		3.9 + 1.7
Full freedom for mother in treatment of child’s major illnessesb		38		16		12.5
No freedom for mother in social decisionsd		46		26		64.6
Mental and behavioural development
Child left to be fed by neighbours or siblingse		0		4		4
Child feedings per day (no.)e	4.9 ± 1.8		4.2 ± 1.4		3.5 ± 1.0
Roti consumed dailyb		82		78		64
Child’s nails cleanb		64		64		38
Boils or rash on child’s skin		4		12		24
Mother uninterested or highly uninterested in child’s weighta		8		20		38

a. p £.005.
b. p £ .05.
c. p £ .001.
d. p £ .01.
e. p £ .0001.

FIG. 7. Manner of feeding of children according to growth category

Motivation of the mother to weigh her child. Significant differences (p < .005) were seen among the three groups, with 38% of mothers of negative deviants being uninterested or highly uninterested, as compared with 20% of mothers of median growers and 8% of mothers of positive deviants (fig. 8). Overall, 76% to 80% of all the mothers were of the opinion that their children were growing normally, which meant that mothers of negative deviants did not perceive that their children were not growing well. They were also more likely to weigh their children because of the persuasion of the investigator or community worker.

Discussion

The results of this study indicate that given the same family access to economic resources proxied on total income, per capita income, and economic dependency ratio, and a similar environment, some children grow better than others. This can be attributed to a large extent to maternal, psychosocial, and behavioural correlates that may be termed “maternal child-care wisdom and technology.”

FIG. 8. Motivation of mothers to weigh child according to growth category

It has been proposed that negative and positive deviance be viewed as two different conditions rather than as two ends of a continuum, and that factors associated with negative deviance may not be inversely correlated with effects associated with positive devince [5]. The greater number of negative deviants among pre-term infants could be attributed to the fact that their mothers were married at a younger age and had fewer antenatal-care visits during pregnancy. A decrease in food intake during pregnancy may be due to a decreased appetite because of nausea and vomiting, but it may also be imposed by the mother to restrict the size of the baby and avoid a difficult birth [6]. Further, nausea may interfere with a woman’s ability to perform household tasks, and hence women may intentionally restrict food intake m.

The ability of the mother to use available resources wisely was the main determinant of whether the child was a positive or a negative deviant. The problem within households with negative-deviant children was not a lack of resources per se, but an inability of the mother to use them and channel them in the right direction. Mothers of negative deviants often used their resources to purchase luxury items such as television sets or gold jewelry, whereas mothers of positive deviants made more intelligent choices by spending their available income on nutritious foods such as green leafy vegetables and fruits and feeding these to their children, or by seeking immediate medical help for their children.

The failure of the mothers of negative deviants to perceive that their children were growing poorly, coupled with their inability to prioritize health problems, could be responsible for the poor growth of the children.

In addition to money, effective utilization of another important resource - time - was observed among mothers of positive deviants, who, despite household responsibilities, made an effort to feed, groom, and interact with their children, unlike mothers of negative deviants, who tended to be distracted by the entertainment value of the mass media and neglected child-care responsibilities. Mothers of positive deviants also came regularly to weigh their children.

The importance of parental education for the health and well-being of the child has been stressed by many. Parental education could function by lowering fatalistic attitudes to illness, increasing belief in the possibility of changing children’s health status and the acceptance of new ideas, generating greater confidence in dealing with health professionals, and producing more direct responsibility for child-rearing practices [8].

Nutrition knowledge scores did not depend on the educational status of the mother, indicating that female literacy, although highly desirable in itself, is not a precondition for improving nutrition knowledge, attitudes, and practices [9j.

Negative deviants were poorly groomed, lived in homes with poor sanitation and hygiene, more frequently defaecated in the open, were left to be fed by siblings or neighbours, and were not given foods like chappati, which require supervision from the mother.

In a Mexican squatter settlement, mothers distracted by many responsibilities would reach a point of impatience with breastfeeding and with any feeding that required their personal attention. In some cases, the mother would bond preferentially with one child and neglect the others. Under these circumstances, only those children who were able to feed themselves would get enough to eat, and those still depending on the breast would get token feed-ings [10].

In choosing among alternatives for feeding their infants, mothers may be influenced by a variety of factors: economic conditions, health characteristics and concerns (their own as well as their child’s), and requirements and desires related to allocation of their own time; the presence of alternative care-givers; beliefs and values related to the social acceptability of the choices; and advice from other people and media sources [11].

The decision-making power of the mothers also influenced whether the child became a deviant. Mothers of negative deviants had less freedom to take their children to a doctor in major illness. Lack of family support and poor perceptions on the part of the mother were responsible for this behaviour. In addition, the mothers of negative deviants could not attend social functions and thus were deprived of chances to meet other mothers and discuss their children’s health, share experiences, and learn from each other. Many young married women in India are subjected to social controls that impinge on their own as well as their children’s health. Decisions about health-seeking behaviour and health expenditure are often made by other family members [12].

Before developmental programmes are implemented, four parallel aspects of the status of Indian women, which are interdependent, must be understood and recognized [13]: the woman’s economic resource base; the public or political arena allowed to her by society; her family structure and the strength it provides as well as the limits it imposes; and the psychological and ideological sense about women in her society, a sense that shapes her own perceptions of self as well as the options she allows