The longitudinal intervention study: Design and implementation

The study was planned as an intervention in which one group of mother-infant dyads received food supplementation while a second group remained untreated. Whereas the treated group would tell us if an adequate nutritional status could be attained even under adverse socioeconomic conditions, the untreated group would provide important information about the natural course and consequences of chronic malnutrition (Madrigal and Avila, 1990).

This project was conceived as a study of cases and not as an epidemiological survey. At that time there were data suggesting that the problem of marginal malnutrition was related to breast-feeding (Martinez and Chávez, 1967). For this reason it was decided to devote a substantial amount of effort to measuring the quantity of breast milk produced and ingested. A great deal of effort also went into making behavioral observations of the children. This component was included because the research team believed that suboptimal child behavior was one of the main consequences of malnutrition. Because behavioral studies can only be done by direct observation in the households, the sample size could not be large. Based on a statistical procedure for taking into account the possibility of dropouts from the study, it was estimated that 20 dyads would be enough to test the behavioral hypotheses.

The supplemented and unsupplemented groups did not enter the study at the same time, because the people in the villages could have questioned the provision of food to some but not to other children. For this reason, during the first year of the study, only the nonsupplemented women and their newborns were recruited into the study. During the second and third year of the study, all the women who became pregnant received food supplementation. At the time of birth of the children, dyads with similar physical and socioeconomic characteristics to the unsupplemented group were recruited for continued supplementation.

The United States National Institutes of Health (NIH) initially funded the study for four years and later extended this period to seven years. The Mexican Council for Science and Technology (CONACYT) funded the project for another seven years. The project has continued to be funded by smaller grants, one of which was provided by the United Nations University (UNU).

The study was initiated in February 1968, and the first three months were devoted to a general study of the community and to establishing a close relationship with the families. Following this period, all the pregnant women in the village were studied. By the end of the first year, in June 1969, a group of 41 mother-infant dyads had been recruited. Twenty of these 41 women were selected for the measurements of milk production and intake and behavioral observations. The selection was based on socioeconomic status and maternal health, age, and anthropometry. The growth of the children of the remaining 21 dyads was followed longitudinally into the adolescent period.

The women and their children born in 1968 and 1969 were not given supplemental food and did not receive any type of intervention except in emergency situations. These children grew up with the support of their families under the usual conditions of the village. They were breast-fed for a prolonged period of time, and weaning foods were usually introduced with hesitation and at a very late age in an insanitary environment that constantly led to constant infections.

These children were born with low birth weight. They grew well in the first three months but then their growth velocity declined and therefore they began to suffer malnutrition. Of the 20 children who were selected for the full study, two were treated and dropped from the study because they developed severe malnutrition, one with edema and the other with marasmus. One of them was replaced by another child. Another child died of an infection under very difficult circumstances and one child emigrated. Therefore this group had a final sample size for the full study of 17 children. The total number of newborns of that year was later further reduced from 41 to 36 due to two additional migrations. This group has been analyzed and included in several reports dealing with preschoolers and teenagers.

The following year a second group of pregnant women was recruited and supplemented twice per day with a nutritious drink immediately following the first report of amenorrhea. The drink was made by mixing milk with fruits and was designed to provide 400 kcal per day and appropriate amounts of iron, niacin, riboflavin, and vitamins A and C. The intake of the food supplement was monitored, and it was shown that it provided 325 kcal per day. Supplemented women had similar socioeconomic and physical characteristics as the women in the nonsupplemented group that was recruited the previous year.

Supplemented subjects were matched at birth with their counterparts in the unsupplemented group, according to the physical and social characteristics of the mother-infant dyed. In addition to the 20 supplemented women who were included in the full study, another 20 women were also given food supplements and studied in some aspects of their development. The fact that the experimental and the control groups entered the study at different times was necessary, not only for interactions with the community but for logistic reasons.

The children in the experimental group began to receive supplementary food as soon as they showed the first signs of growth faltering, at about 12 weeks of age. First, the children were offered a bottle with milk during the night. When the children started to request to be breast-fed more often, even though they were being offered the bottle with milk, they were also given fruits and vegetables. Afterwards the infants were fed ad libitum with milk and a variety of strained foods. The research team always advised the women to continue breast-feeding during this time.

When the children were four years old, they were supplemented twice per day with a sandwich and a glass of milk. When the children began attending school, they sometimes missed one of the daily episodes of supplementation because they preferred to remain playing at school during recess instead of going out to receive their supplement. However, the children always received the supplementation after school hours. The supplementation intervention ended when the children were 10 years of age

Throughout the study, special care was taken to ensure that the only between-group difference was the nutritional supplementation. Measures were taken to balance the amount of contact with research workers and any other procedure that could have been considered a nonnutritional intervention.

Throughout the 24 years of existence of the Centro Rural de Tezonteopan, a variety of nutrition and child development parameters have been studied. The unsupplemented children are, as this report is being written, 22 years old and therefore have become young adults. The children in the supplemented group are now between 17 and 20 years of age. The range of ages in the latter group is explained by the fact that the large amount of effort needed for planning and implementing the project caused a slowdown in the rate of recruitment of subjects during the last years of the intervention phase.

The different types of studies are presented in the Results. The special methodologies that were employed are presented and discussed below.

a. All the women were included in longitudinal follow-ups of anthropometry and in a special study on fertility and reproduction.

b. The studies on food intake during early infancy included measurements of milk volume using a 72-hour test-weighing procedure at 2, 8, 16, 24, 36, 56, and 78 weeks of age.

c. The behavioral follow-up looking at mother-child interactions was the most significant component of the study. These observations were made between breast-feeding episodes while the observer was seated in a corner of the house pretending to read a book. Every 40 seconds the investigator looked over the book to make a "visual photograph" of the mother-child interaction (e.g., holding, feeding, kissing, verbalizing, degree of physical activity, etc.). Seventeen parameters were captured and written down every time a "visual photograph" was created. This procedure was carried out for 1.5 hours during the morning after the child woke up and for another 1.5 hours in the afternoon. This methodology is derived from that used for ethological observations of primates and is the one that captured the most important differences between the supplemented and unsupplemented children.

d. A similar methodology was used during the school period. The researchers made a hole in the wall of the classroom so that they could make behavioral observations of the children (e.g., standing, sleeping, attention span) while attending class. Each observation period lasted 1.5 hours and yielded information that discriminated between supplemented and unsupplemented children. Several national and international knowledge and problem-solving tests were also administered during the school period.

e. The physical activity of the infants was evaluated by observing the number of contacts that the heel made with the bed. Afterwards physical activity was measured as the number of steps taken in a specified period of time (10 minutes per hour for 10 consecutive hours).

f. The longitudinal assessments involving neurological, psychological, and cognitive measurements were done following traditional methods (A Chávez and Martínez, 1982).

g. Morbidity was recorded daily and the study also included a microbiological assessment of fecal contamination of household objects and members.

The experimental design was selected to test the hypothesis that nutrition during early childhood is an environmental factor that has a strong negative impact on long-term human development and function. The design of the study also contributed to a better understanding of the development of children in a deprived environment. It was also possible to study the life cycle in the families, since it included the follow-up of subjects from the time that they were in their mother's womb until they became pregnant.

The final objective of the project was to identify the critical point at which interventions that will achieve an optimal development in socioeconomically disadvantaged populations are most cost-effective.