Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Early feeding practices and the nutrition status of preschool children in rural Hubei province, China

Jing Chen and Douglas Taren


A nutrition surveillance survey was conducted in June 1990 in four poor rural counties of Hubei Province, China, that included 25 villages and 3,564 children, to determine early feeding practices and the nutrition status of preschool children. Data were collected through interviews, anthropometric measurements, analysis of haemoglobin values, and physical examinations. More than 90% of the infants under 6 months of age were being breast-fed, as were approximately 75% of those 12 months old and 26% of those 24 months old. A trend to introducing solid foods into children's diets at younger ages appeared to have occurred between 1984 and 1988; only 15% of the children ate solid food at 6 months of age and 58% at 12 months in 1984, compared with 23% and 78% respectively in 1988. The children's anthropometric measurements were comparable to international standards for the first months of life, and then weight-forage and height-for-age Z scores deviated negatively. Physical examinations suggested that several micronutrient deficiencies might be present The data also indicated that during a time of sweeping economic changes and expanded health care availability, breast-feeding was maintained, food was introduced into the diets of children at younger ages, and acute malnutrition was not common.


The economic status of the People's Republic of China has improved significantly over the past 30 years, with a change in the growth rate of the gross national product (GNP) from 4.1% for 1965-1980 to 8.2% for 1980-1989 [1]. National indicators of nutrition status indicate significant improvements in infant mortality, mortality under the age of five years, and life expectancy (tables 1 and 2) [2]. At the same time, the country has increased the breadth and number of health and nutrition programmes that benefit children, such as greater access to safe drinking water, improved sewage disposal, increased immunization coverage, and mass campaigns against parasitic diseases [3]. Growth studies from urban areas reported a significant increase in the mean weight of older children between the 1950s and the 1980s but little change in preschool children [4].

Information about the health and nutrition status of children living in the rural areas of China is limited. One method to obtain such information is through surveillance programmes that target populations at the greatest risk for nutritional problems [5]. These programmes have the advantage of being able to determine how nutrition status changes over time and can be used to evaluate various interventions [6]. Such a survey was conducted in rural Hubei province as part of the national nutrition surveillance programme to determine if the nutritional patterns of rural children had changed during the 1980s.



In June 1990 the Ministry of Health conducted a cross-sectional survey in four counties of Hubei province, a central location for heavy industry, transportation, and communication, located on the Yangtze River. The population of the province in the years preceding the survey was estimated to be about 35 million.

TABLE 1. Health and nutrition indicators, People's Republic of China



Under-5 mortality (per 1000)  




Infant mortality (per 1,000)  




Low birthweight, 1980-88 (%)


Immunization coverage, 1989-90 (%)  


DPT 97  




Breast-feeding, 1980-91 (%)  
at 3 months


at 6 months


Underweight, 0-4 yrs, 1980-91 (%)  
moderate and severe




Data from ref. 2.

TABLE 2. Indicators of socio-economic status, People's Republic of China



Population, 1990 (1.000s)  

1.139, 100

under 16 years


under 6 years


Number of births (1,000s)


GNP per capita, 1989 (US$) 350  
Urbanized (%)


Urbanized growth rate, 1980-90 (%)


Fertility rate


Life expectancy (years)  




Adult literacy, 1990 (%)  




Access to safe water, 1988-90 (%)  






Access to health services, 1985-90 (%)  






Data from ref. 2.

The counties selected for the study have a geographic distribution of hilly and plain areas and were chosen because they were considered economically disadvantaged and therefore at greater risk for poor nutrition status. They had an annual per capita income between US$40 and US$100 in 1989. More than 90% of the households had electricity, and the main sources of drinking water were wells and rivers. Ninety-five per cent of the children in the area went to school at five years of age, and more than 90% of those under six years were up to date with their immunizations. The major diseases in the preschool children were respiratory and gastrointestinal infections; infant mortality was less than 50 per 1,000.

Health officials from each county selected two townships to participate in the survey on the basis of the criteria of accessibility, willingness to participate, and having at least 400 children under six years of age. From these eight townships the largest 25 villages were selected as sample villages, and from these villages 3,564 children (at least 95% of those who met the criterion of being under six years of age, based on regional birth and health records) were recruited and participated in the survey. Another 17] also met the age criterion but were not included in the analysis because missing data prevented the computation of ages or standardized anthropometric values.

Socio-economic and medical information for each child was obtained from a parent through a series of 36 questions, including the parent's education and occupation, the child's birth order, birthweight, history of acute or chronic diseases, how long the child was breast-fed, and the timing and types of solid foods that were introduced. A physical examination was performed for each child to identify signs and symptoms associated with protein-energy malnutrition and deficiencies of thiamine, riboflavin, ascorbic acid, vitamin A, vitamin D, and iron.

The percentage of children who were currently being breast-fed was determined. For those who had stopped breast-feeding, the age of weaning was recorded retrospectively to determine whether there had been a change in the length of breast-feeding. The proportions of children who were breast-fed for 12, 24, and 36 months were compared across each age cohort using chi-square analysis.

The parents were asked at what age rice, egg, vegetables, fruit, soya beans, meat, and other foods were introduced into the children's diets. From these data, the age at which the first of these foods was introduced was determined for each child. The average age at which the first solid food was introduced was compared across each age cohort by analysis of variance, and the proportions of children consuming solid food at 6, 12, 18, and 24 months of age were compared by chi-square analysis. The pattern for introducing each food into the diet was also determined, and the proportions of children eating the various foods at 6, 12, 18, and 24 months were compared by chi-square analysis.

Anthropometric measurements were taken by four teams of investigators who were trained as recommended by the United Nations [7]. The children wore sleeveless shirts, shorts, and no shoes. Children were weighed with platform-beam scales, calibrated daily, and the weights were recorded to the nearest 25 g; infants were weighed either on an infant scale or in a container that had a known weight. Length and height measurements were recorded to the nearest 0.1 cm. Children over three years of age were measured with a stadiometer; those under three were measured in a recumbent position using an infant measuring board with a fixed headboard and movable footboard. Blood was obtained by finger prick, and haemoglobin concentrations were determined by the cyanomethaemoglobin method [8].

Data were entered using dBASE III software (Ashton-Tate, Culver City, Calif.). The weight and height information was converted to Z-score values for weight for age (WAZ), height for age (HAZ), and weight for height (WHZ) based on the NCHS/CDC/ WHO standards using the EPINUT version 1.00 software (US Centers for Disease Control, Atlanta, Ga.). The data were then transferred to a Hewlett Packard minicomputer, and the statistical analyses were conducted using the SAS software programme [9]

The mean growth patterns of the children were compared with the NCHS/CDC/WHO standards [10]. The percentages of children with scores more than two standard deviations below the mean were determined by six-month age groups up to two years of age, and then by one-year age groups up to six.

FIG. 1. Study population by age and sex (N= 3,564; *226 children 60 months old or older)



Sociodemographic status

A total of 1,977 boys and 1,587 girls were involved in the study. Their age and sex distributions are shown in figure 1. Children under one year old represented 18.3% of the sample, and one-year-olds 22.8%. Children two, three, four, and five years old constituted 19.4%, 18.0%, 15.1%, and 6.4% of the sample respectively. There was a prominent yearly periodicity in the occurrence of ages, with the number of births increasing toward the month of January and decreasing afterward for each year. In the samples, 1.7% were twins, 45.1% were first born, and 39.8% were second born. Of the 3,564 children, 5.5% had a low birthweight. A majority of the parents had graduated either from middle school (45.7% of fathers, 39.8% of mothers) or from high school (27.6%, 13.2%).


At the time of the study, more than 90% of the children under six months old were being breast-fed (fig. 2), as were 75% of the one-year-olds and 26% of the two-year-olds. Breast-feeding continued until three years of age for 3% of children. Retrospective data about when children had stopped breast-feeding indicated no significant difference for those two to five years old.

FIG. 2. Percentages of children who were currently breast-feeding at the time of the survey, by age

FIG. 3. Ages at which solid food was first introduced into children's diets, according to mothers' reports, by the child's year of birth


Introduction of solid food

Figure 3 shows the ages at which solid food was introduced. A significant temporal change occurred between 1984 and 1988 with respect to earlier introduction of solid foods. For example, 25% of the children born in 1987-1988 were consuming solid food at six months of age, compared with 15% of those born in 1984-1985 (X2 = 14.0, p < .001). Similarly, 80% of those born in 1987-1988 were consuming solid food at 12 months, compared with 58% of those born in 1984-1985 (X2 = 58.3, p < .001).

There was a significant difference in the age at which various foods were introduced (X2 = 455, p < .001), but the order of foods did not vary among the age groups. Figure 4 shows the timing of the introduction of various foods into the diets of children who were two years old at the time of the survey. Rice was almost always introduced first; 60% of the children were consuming rice by the age of 12 months. The next two foods were usually eggs and vegetables, followed by soya beans and fruit; 50%, 35%, 27%, and 20% of the children respectively were consuming these foods by 12 months. Meat tended to be introduced last, with 20% consuming it by 12 months.

FIG. 4. Ages when specific foods were introduced into the diets of children who were between 24 and 35 months old at the time of the survey

FIG. 5. Mean weight and length/height measurements of children by sex and age, compared with NCHS reference standards


Anthropometric measures

Low birthweight ( < 2,500 g) was reported in 5.5% of the children. The mean WAZ and WHZ values for both boys and girls under six months were greater than zero, and the mean HAZ value was near zero. The scores did not indicate any consistent difference between the growth of boys and girls for any of the age groups studied compared with the reference standards.

After six months of age, the mean WAZ fluctuated around-1.00 SD for both boys and girls (fig. 5). The percentage of children with a WAZ below -2.00 SD was lowest during the first year of life (6.1%) and greatest at 12-17 months (32.6%) (fig. 6).

The mean HAZ was comparable to the NCHS median within the first few months of life. At 12 months it was about -1.00, and by 36 months it was about -2.00. The percentage of children below-2.00 SD for HAZ increased from 9% for children under one year old to 44% for those 18-23 months old. However, two-year-olds showed a decline in the percentage with low HAZ (30.5%), but this increased in each age group until five years of age (43%).

The mean WHZ values started above the NCHS standard mean until about 65 cm in length (fig. 7). They remained between the NCHS mean and -1.00 SD for children between two and five years of age. Those 12-17 months had the greatest percentage with WHZ values below -2.00 (9.2%), compared with less than 2% for those three and four years old.

FIG. 6. Percentages of children with weight-for-age (WAZ), height-for-age (HAZ), and weight-for-height (WHZ) Z score values less than-2.00 in different age groups



The proportion of children with haemoglobin concentrations less than 11 g/dl was greatest in infants 611 months old (42%). After one year of age the proportion of children who had low haemoglobin values decreased (fig. 8). Haemoglobin levels were positively correlated with the time foods were introduced into the diet. The correlation coefficients were significant with probability below .001 for rice (r=.17), eggs (r=.17), fruit (r=.20), vegetables (r=.20), soya beans (r=.22), protein (r=.20), and other foods (r=.20). Children who had consumed milk had slightly higher haemoglobins than those who had not (11.7 1.3 vs. 11.1 1.4; p < .05). Haemoglobin was not correlated with WAZ, HAZ, or WHZ values nor with the presence of a recent infection. It was not associated with the level of education for mothers or fathers.

Signs and symptoms of nutritional deficiency

Clinical impressions of rickets were recorded in 6.4% of the children. Children suspected of having rickets represented 4.0% of the total sample, moderate disease 2.2%, and severe 0.2%. Riboflavin deficiency was suspected in 4.1 %, vitamin C deficiency in 0.7%, and vitamin A deficiency in 0.3%. Other suspected deficiencies were thiamine (0.2%) and iodine (0 1 % ).


Our findings indicate that there were, indeed, changes in the feeding patterns of preschool children in the area surveyed during a time in which sweeping changes in economic policies and substantial economic growth were occurring in China. During that same time, public health interventions were increased, with expanded primary care coverage and extended nutrition education.

Breast-feeding was the predominant infant-feeding method until about six months of age, with only 25% of mothers reporting that solid food was introduced by this age. However, between 1984 and 1989 rice and other foods were being given to children at younger ages. Nonetheless, breast-feeding continued to supply nutrients to nearly 75% of the children at 12 months of age, and breast-feeding was common in one- and two-year-olds. Extending breast-feeding beyond the first year of life has been associated with greater anthropometric measurements even after controlling for the introduction of other foods, infections, and various socio-economic factors [11]. The breast-feeding rates were greater than have been reported for most Asian settings (in rural or urban areas), where indications are that between 60% and 80% of infants are breast-fed for six months [12].

FIG. 7. Mean weight for length/height of children, by sex, compared with the NCHS reference standards

FIG. 8. Percentages of children with haemoglobin concentrations less than 11.0 g/dl, by age group

It appears that the initiation and duration of breast-feeding were greater in rural Hubei than in other areas of China. However, no national data are available on the proportion of infants who are breast-fed for 12 months. Our findings are consistent with a series of studies suggesting that breast-feeding rates in China are higher in rural areas and decrease with greater urbanization [13-15]. For example, in Shanghai 32.3% of infants were reported to be breast-feeding for six months, compared with 73.7% in the suburbs [15]. According to one national study, 60.3% of infants in rural areas breast-feed for six months, compared with 34.4% in urban areas [13]. However, the surveyed areas of Hubei province had the greatest rate, with 94.7% of the children 12-47 months old being breast-fed for at least six months, and 61.7% for at least 12 months.

The differences in these results may also be associated with attitudinal differences in the populations. A greater percentage of women in the suburbs (94.2%) believed that breast-feeding is best for the child, compared with 87.8% of women living in Shanghai [15]. Other attitudinal differences associated with increased amounts of breast milk (a proxy for successful breast-feeding) included believing that nothing was better than breast milk and that extended breast-feeding was appropriate [15]. Not having had a difficult labour and not having a premature infant were also associated with greater production of breast milk [15].

The health education campaigns in Hubei province did not focus on breast-feeding, since the practice is ubiquitous and the supply of breast-milk substitutes is limited. These may be two reasons why no change in breast-feeding rates were seen despite the country's increased prosperity.

Exclusive breast-feeding stops when solid foods, soups, and liquids other than breast milk are given to infants. Preceding the current survey, the nutrition surveillance programme in Hubei put an emphasis on the earlier introduction of nutrient-rich foods that would increase the intake of protein, zinc, and vitamin C. The programme may have contributed to the significant change that was demonstrated in this study. However, the earlier introduction of solid foods could occur for many other reasons that were not determined in the present analysis, such as increased financial resources and improved crop production [1].

An analysis of Malaysian data reported that digit preferences are present when women retrospectively report the duration that they breast-fed their children [16]. Similarly, preferences for 6,12, 18, 24, 30, and 36 months were observed in the current study. Although the retrospective nature of the current data could be considered a limitation, the data from Malaysia suggest that the time frame used for retrospective data had less effect on the reliability of measurements than socioeconomic status or ethnicity, both of which were fairly equal in the present study.

We believe that solid food is being introduced earlier to children. What has not been measured is the amount being introduced and consumed. It is also recognized that solid food is being given later than current international recommendations; in the cohort born between 1987 and 1988, more than 75% of the children were reported to have been first given solid foods after six months of age.

The positive association between haemoglobin values and the eating of solid food may be a marker for one of the benefits of earlier introduction of solid food. In a more global approach, this association illustrates how improved economic status improves nutrition status, since families with more resources are able to offer infants solid food earlier and may have access to more informal sources of health information. This would be a reason why haemoglobin values were greater in children who received milk other than breast milk, since these would be children whose families were more able to afford milk products. Without additional information, however, it is difficult to determine what factors are associated with the higher haemoglobins.

The infants in rural Hubei province were slightly heavier than the reference standards. The strongest evidence for this is the WAZ and HAZ values recorded for children under three months old; the mean Z scores of both groups were positive. Additional support is the small percentage of infants with birthweights under 2,500 g. Although the low-birth-weight rate reported in this study would include only survivors, the weights of infants one to three months old support the view that these Chinese children were larger than the reference standards. It was only after the age at which solid foods are recommended to be introduced that the children in our study showed a negative deviation from the reference growth standards. This deviation could occur either because of the later introduction of solid food or because the growth of breast-fed infants may actually differ from that of bottle-fed infants in terms of pattern, velocity, and tissue composition [17-20].

National reference growth data for China, collected in 1985 from a representative sample of 175,290 middle-income children from four northern provinces and six southern provinces, suggest that the weights of infants in the first month of life are also greater than international references [21, 22]. The data for the children from the four poor counties in Hubei province were comparable with the national growth data for the first six months of life, but thereafter they had lower weight and height values for their ages. Their decreased growth rate after 12 months of age could be the result of such things as being fed solid foods at an older age, less overall intake, or an increased frequency of infectious diseases, including those due to parasites.

The results of two separate analyses indicate that growth rates in China are highest in children from urban areas, lower in children from the suburbs, and even lower in children from rural areas [21, 22]. Compared with children in other rural areas with a mid-level socioeconomic status, the boys in Hubei province had lower mean weight and height values after one year of age, and the girls after six months.

There has been a significant temporal trend of increased growth rates among Chinese children during the past two decades. The rates for older children are now comparable to NCHS/WHO standards, especially for stature [4]. Data from several countries indicate that changing environmental conditions have significantly increased the growth of children. Between 1963 and 1967 the heights of southern Chinese children in Hong Kong were comparable to international standards, and deviations were strongly associated with socio-economic status [23]. The average heights of Thai children also increased between ]962 and 1975 [24]. Furthermore, a secular trend of increased growth was reported for Japanese children who immigrated to the United States, which may be due to increased intake of protein and fat [25].

In conclusion, the dietary and growth patterns of children in rural Hubei may be changing. Although much of this change may be due to economic and health care policies, the persistence of breast-feeding and the introduction of solid foods into the diets of infants at younger ages may be having equally important effects.



We extend our appreciation to Goubing Zhang and Gouching Yi for their assistance with the research design and data-collection procedures. We are grateful for the help of Xiao Fang, Xiao Yeng Liu, Ann - Xing Wang, and Liping He in data collection. We also acknowledge Kathleen Jaegers for her careful editing and typing of the final manuscript.



1. World Bank. World development report 1991. Washington, DC: World Bank, 1991.

2. UNICEF. The state of the world's children 1992. New York: UNICEF, 1992.

3. World Bank. World development report 1990. Poverty. Washington, DC: World Bank, 1990.

4. United Nations Administrative Committee on Coordination, Subcommittee on Nutrition. Update on the nutrition situation: recent trends in nutrition in 33 countries. Geneva: WHO, 1989:112-15.

5. Habicht J-P, Mason JB. Nutritional surveillance: principles and practice. In: McLaren DS, ed. Nutrition in the community. London: John Wiley & Sons. 1983:21744.

6. Mason JB, Habicht J-P, Tabatabai H. Valverde V. Nutritional surveillance. Geneva: WHO, 1984.

7. United Nations, National Household Survey Capability Programme, Co-operation for Development and Statistical Office. How to weigh and measure children: assessing the nutritional status of young children in household surveys. New York: United Nations, 1986.

8. Drabkin DL, Austin JH. Spectrophotometric studies: spectrophotometric constants for common hemoglobin derivatives in human, dog, and rabbit blood. J Biol Chem 1932;98:719-33.

9. SAS Institute. SAS/STAT user's guide, release 6.03 edition. Cary, NC, USA: SAS Institute Inc., 1988.

10. Hamill PVD, Drizd TA, Johnson CL. NCHS growth curves for children birth-18 years. Vital and Health Statistics, series 11, no. 1651, DHEW publication PHS 78-165. Washington, DC: US Government Printing Office, 1977.

11. Taren DL, Chen J. A positive association between extended breast-feeding and growth in rural Hubei province, PRC. Am J Clin Nutr 1993;58:862-7.

12. World Health Organization. The prevalence and duration of breast-feeding: a critical review of available in formation. World Health Stat Q 1982;35:92-116.

13. Yun YP, Kang ZS, Ling LJ, Xin QC. Breast feeding of infants between 0-6 months old in 20 provinces, municipalities, and autonomous regions in the People's Republic of China. J Trop Pediatr 1989;35:277-80.

14. Meehan KF. Breast feeding in an urban district in Shanghai, People's Republic of China: a descriptive study of feeding patterns and hospital practices as they relate to breast feeding. J Trop Pediatr 1990;36:75-9.

15. Su ZF. Breast-feeding of infants 0-6 months of age in the People's Republic of China. Chinese Med J 1981; 61 :5804

16. Haaga JG. Reliability of retrospective survey data on infant feeding. Demography 1988;25:307-14.

17. Dewey KG, Heinig MJ, Nommsen LA, Lonnerdal B. Infant growth and breast-feeding. Am J Clin Nutr 1989; 50:116-17.

18. Whitehead RG. Nutritional aspects of human lactation. Lancet 1983;1:167-9.

19. Dewey KG, Heinig, MJ, Nommsen LA, Lonnerdal B. Adequacy of energy intake among breast-fed infants in the DARLING study: relationships to growth velocity, morbidity and activity levels. J Pediatr 1991;119: 53847.

20. Whitehead RG, Paul AA. Growth charts and the assessment of infant feeding practices in the western world and in developing countries. Early Hum Dev 1984;9:187-207.

21. The Ten Province Children Growth and Development Study Group. The ten province city and rural survey: 1985 survey on the physical development of Chinese children. Chinese Med J 1987;67:429-32.

22. Nine City Children Growth and Development Study Group. The nine city and suburb survey: 1985 survey on the physical development of Chinese children. Chinese Med J 1987;67:423-8.

Contents - Previous - Next