Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Merrill S. Read and Jean-Pierre Habicht
In the mid-1960s a longitudinal, multidisciplinary nutrition intervention study was undertaken in rural Guatemala by the Institute of Nutrition of Central America and Panama in conjunction with the US National Institute of Child Health and Human Development. The goal was to elucidate the relationship between undernutrition in pregnancy and early childhood, health, and subsequent behavioural development in infants and young children. Extensive detailed planning coupled with three years of pilot studies in the field preceded the initiation of the longitudinal study in 1969. This article outlines the problems encountered in planning and implementing the study, and their resolution. Many of these experiences will be helpful to others considering community-based intervention studies.
The INCAP longitudinal study on the effects of early nutrition supplementation in child growth and development (1969-1977) arose from the confluence of two factors: the interest of the Institute of Nutrition of Central America and Panama (INCAP) in the functional and behavioural consequences of early malnutrition, and the establishment of the US National Institute of Child Health and Human Development (NICHD), mandated to explore interdisciplinary factors in human development, especially mental retardation. Within the United States enthusiasm ran high for the Headstart Program and other components of the War on Poverty, including an attack on hunger. Internationally, this was a time of optimism concerning the Green Revolution and the promise it held for feeding millions; with this came the desire to document the health and performance benefits of better nutrition and improved agriculture.
In the early 1960s protein deficiency defined hunger and was seen as the major nutritional problem of poverty. Did hunger and malnutrition in poor children effect their learning ability? These concerns strongly influenced the longitudinal study design.
The story begins in Guatemala. Dr. Moisés Behar, director of INCAP from 1961 to 1974, was actively interested in the impact of malnutrition on growth and physical development and was increasingly concerned about the possibility of a simultaneous impact on learning. Dr. Joachín Cravioto, a well-known Mexican paediatrician, was supported by the Pan American Health Organization (PAHO) to come to Guatemala to explore possibilities for a community study of growth, nutrition, and neurointegrative development. He arrived in early 1962 and soon began a study in the small village of Magdalena Milpas Altas, about 40 km from Guatemala City. This study was ecological in design, meaning that many variables (physical growth, family social environment, food availability, school attendance, etc.) were assessed without an intervention; measures of inter-sensory organization and development were completed for children 6-11 years old. Analysis of these diverse data suggested that neuromotor development was related to physical growth, at least in poor rural children, and growth in turn was related to food intake or availability. However, other factors contributed to the outcome as well. Cravioto suggested that unravelling the causal pathways would require multidisciplinary, prospective studies [1].
After Cravioto returned to Mexico in early 1964, Behar continued to provide leadership in exploring models for additional studies focused on the impact of early malnutrition on learning and behaviour to be carried out under INCAP's direction. Dr. Nevin S. Scrimshaw, then chairman of the Department of Nutrition and Food Science at the Massachusetts Institute of Technology, who had founded INCAP and followed its developments closely, also provided strong impetus through his interest in the subject. This interest led to a conference on Malnutrition, Learning, and Behavior in 1967, with participants from 38 countries. The proceedings from this conference [2] greatly influenced the development of thinking on the subject.
Dr. John E. Gordon, who had founded and served as chairman of the Department of Epidemiology at Harvard School of Public Health, worked for many summers at INCAP through 1969. He, too, had a strong influence on the development of this study. He and the INCAP scientists discussed possibilities and eventually settled on a three-village intervention model: one village to receive a complete supplement, one to receive a placebo as a social control, and the third to be observed without intervention.
It was above all thanks to Gordon that the lessons learned in the previous large nutrition field study undertaken by INCAP [3, 4] were brought to bear in the design and initial phases of this study. Thus the importance of replicate villages within the three different treatments was recognized for the first time in nutrition intervention trials that used villages as the unit of treatment.
Particularly important was the need for a full-blown pilot study that would test intervention, data-collection methods, and logistics. The pilot project revealed deficiencies in all of these areas, many of which would have been difficult to remedy after the study itself had begun. As noted below, so many changes still had to be made during the first years of the intervention study that the final design would have been severely compromised if it had not been for the pilot study. Factors of unexpected importance in the final study were the strong esprit de corps and the principle that the study personnel should live in the field, which were developed during the pilot study under the leadership of Dr. Cipriano Canosa, a Spanish paediatrician with field experience in Colombia who arrived in September 1964.
Meanwhile, the NICHD had been established by the US Congress in October 1962 under Public Law 87-838, which also created the National Institute of General Medical Sciences. These two institutes were the first within the US National Institutes of Health that were established for the study of basic health sciences rather than specific disease states, and each embraced both biological and behavioural concerns. To a considerable degree, the NlCHD's mandate grew out of President Kennedy's concern for mental retardation and broader issues of maternal and child health. The specific mandate was to be "an institute for the conduct and support of research and training related to maternal health, child health, and human development . . . and the basic sciences relating to the processes of human growth and development, including prenatal development."
The first director of the NICHD, Dr. Robert Aldrich, was appointed in February 1963, and the institute became functional in June of that year. From the outset, Aldrich was concerned with the need for an integrated, multidisciplinary approach to health status and needs, especially in relation to questions of brain damage, mental retardation, reproductive wastage, and the concept of life-span outcomes. Early planning also recognized the importance of international research and joint studies abroad. International interest was implemented by the appointment of 12 international fellows in paediatric health to be stationed overseas; it was hoped that these assignments might lead to a network of collaborating research centres. One of the first of these fellows was Dr. Delbert Dayton, a paediatric neurologist, who was sent to INCAP in February 1965 to develop skills in nutrition and neurointegrative development. Dayton joined Canosa in the pilot studies, with special interest in neurological assessment, morbidity, and bone growth. Dr. João Salomón, a Brazilian epidemiologist, also joined the team.
Canosa had two major tasks: to refine and flesh out the plan for the three-village study, and to obtain funding for such a project. He felt strongly that the prenatal nutrition and health status of the mother were potentially significant factors in infant and child development. Thus they became important components of the plan as it was refined. The working hypothesis focused on whether nutritional supplementation of mildly to moderately malnourished pregnant women and their children would influence the children's mental development. The basic three-village comparison strategy remained intact.
Pilot studies were initiated with INCAP support in early 1965 in Acatenango, a relatively isolated but accessible community of about 2,000 inhabitants. The government of Guatemala had already built a health centre there, which it turned over to INCAP for the duration of the study. Acatenango had another advantage: the nearby small village of Los Planes, a sort of suburb of Acatenango, had many of the characteristics thought to be necessary for the larger study, and it was incorporated into the pilot project.
A full home census identified 320 preschool children in Acatenango and 99 in Los Planes. Regularly scheduled home visits every two weeks were established to document morbidity and pregnancy status. The first home dietary studies were undertaken in both villages as well. The reward for participation was a photograph of the family or of the child being studied; cameras were a rarity in rural Guatemala at that time, and the photographs were much appreciated.
Meanwhile, having become aware of the interests of the NICHD, Canosa went to Washington in January 1965, where he discussed his plans with Dr. Dwain Walcher, leader of the Growth and Development Program in NICHD. Walcher became interested and served as a major facilitator in initiating NICHD support. The first contract was awarded to PAHO in June 1965 for a study at INCAP entitled "A Nutritional Survey of Indian Communities in Highland Guatemala," a title it carried officially throughout the 14-year contract phase of the project. The first-year award was for US$133,910, for the purpose of surveying suitable communities for a longitudinal study, and to begin the development of appropriate test instruments. Dr. David Kallen, a social psychologists, was the first NICHD project officer for the INCAP study; in 1966 this responsibility was transferred to Dr. M.S. Read.
An early incident in the INCAP-NICHD relationship, although not central to the study, illustrates the problem-solving mood of the times. To facilitate the work in Los Planes there clearly had to be a centre for health care and behavioural test administration. The NICHD was not permitted to construct buildings under research funds, particularly overseas; it could, however, support essential remodelling. The team in Guatemala, together with the people of Los Planes, built a wooden frame with a thatch roof at the appropriate location. With this in place, it was possible to use NICHD funds to remodel the structure into a health centre. The centre remained for the village when the project moved on to other locations.
It was hoped that the extensive data base established in Acatenango and Los Planes might permit them to be used as test sites for the longitudinal study. This idea was soon abandoned when it was realized the extent of social contamination that had occurred as a result of the presence of the research team and their frequent visits to homes. This experience emphasized the importance of equalizing social interventions across all study villages if one wished to attempt to tease out the impact of nutrition and health care alone. This became a major concern throughout the pilot study and in the project as implemented.
With NICHD support assured, Canosa and his colleagues began to search for communities that might be suitable for intervention studies. This was an arduous and painstaking exercise involving the examination of census data and hundreds of village visits and evaluations.
Meanwhile, methodological issues were under study in Acatenango and Los Planes. Dietary, clinical, and anthropometric assessment techniques were developed. It was the consensus of the team members that drawing blood samples for clinical chemistry assays was not acceptable to the population and would have to be omitted from the study protocol. However, medical care would be continued as prescribed by the Guatemalan Ministry of Health, but with project physicians providing the service.
Initial psychological testing was based on the Gesell test, translated from English into the Mayan language spoken in Acatenango. An American psychologist, Dr. Robert Klein, was enlisted in the fall of 1966 to develop the behavioural test techniques further. This area was considered to be the most difficult to quantify. It quickly became apparent that in the full study it would not be possible to administer behavioural tests across the different Mayan languages that sample size requirements would demand. Therefore it was decided to develop the tests in Spanish and to concentrate on villages where Spanish was commonly spoken. Concern was widely felt about the appropriateness of translating standard United States tests into Spanish for children in very different cultural settings. Klein set out to resolve these issues and to develop an appropriate test battery, which was tested initially in Guatemala City in early 1967 [5].
Considerable discussion centred around assuring adequate nutrition to study participants. Various methods for supplement delivery were first tested in Los Planes in early 1967. It was concluded that providing a supplement to all citizens at a central facility near the clinic would work best. The volume consumed by pregnant or lactating mothers or their children could be monitored on each visit to the centre.
The design that was ultimately agreed upon was to include three sets of three matched villages, one village in each set to receive protein supplementation, one to receive a calorie drink, and one to serve as a no-intervention supercontrol. Ethical concerns and fiscal constraints led to abandonment of the last concept, as discussed below. It was concluded that the resultant three pairs of villages receiving intervention would offer adequate control for unexpected social, economic, or health (epidemic) changes, while also providing a statistically desirable population size. Unfortunately, budget limitations ultimately restricted the intervention to only two sets of villages. However, the existence of the extra set of well-studied villages which received no intervention made it possible to conceptualize a follow-up cross-sectional study conducted in 1988-1989 [6].
By 1969 the goals were enunciated as follows:
The goal of this long-term prospective study is an integrative analysis of the nutritional and sociocultural factors affecting mental and physical development, i.e., to determine the effects of malnutrition, per se, as they interact with other variables that affect early development. To achieve this the project is concerned with the study of psychological and social characteristics of all preschool children and their families in four villages while providing two of the villages with nutritional supplements and all of the villages with medical care and social stimulation. [7]
The major motivation for the original longitudinal study was in fact less all-encompassing. It was to elucidate the long-term expression of intra-uterine and preschool malnutrition on behaviour. This explains the experimental design, which began by collecting data in pregnancy and following the child to 7 years of age. Research in the 1960s had indicated that behavioural test scores at 3-4 years correlated more highly with tests at ages up to 18 years than did scores at earlier ages [8]. However, the impact of malnutrition on behavioural and motor abilities at an earlier age were unknown. In addition, it was considered desirable to continue the intervention well beyond 3 years of age, an age thought both then and now to be the end of risk of significant malnutrition in these children. Nutritional supplementation was to produce an unambiguous contrast between well-nourished and malnourished children.
The longitudinal study was finally begun in January-March 1969 in four villages: Conacaste, Espíritu Santo, Santo Domingo, and San Juan. Conacaste and Santo Domingo were paired as large villages, and Espíritu Santo and San Juan as small ones. The interventions in each pair of study villages included a high-protein, high-calorie supplement (atole) or a non-protein, low-calorie supplement (fresco). The composition of the supplements was modified in 1970, after the project was begun, in response to new information about dietary deficiencies in iron, riboflavin, and vitamin A. The protein and calorie supplements were provided freely to all residents of the village; special efforts (e.g., home visits) were made to draw pregnant and lactating women and children under 7 years of age to the centre. The number of cups of supplement consumed was recorded.
All the villages received primary health care. In 1969 Dr. Jean-Pierre Habicht, newly arrived, and Dr. José Maria Reyna-Barrios, a Guatemalan physician, refocused and restructured the health care services to make them more effective and efficient. A spin-off from these studies was the development of an effective, relatively low-cost, clinic-centred medical care system that used subprofessional paramedical personnel. Training of community midwives in hygiene and improved delivery practices also was instituted.
It is pertinent to note that the research plan was built around the concept of intervention cohorts. All the children involved at the date the study began were to be followed as age-related cohorts to permit exploration of the impact of the child's age at the introduction of supplementation. The largest cohort was to be the group in which the mother received a supplement during pregnancy and lactation, and the child to the age of 7 years.
By this time Canosa was leading a large multidisciplinary team that worked full-time in the study villages. Klein was in charge of the psychological testing and socio-anthropological aspects of the study, including census information. Habicht, who had arrived in January 1969 to replace Salomón and Dayton, who had returned to their own countries, was in charge of setting dietary, morbidity, and anthropometric measurements. Dr. Charles Yarbrough was hired to develop and manage the data compilation and statistical computing.
In January 1970 Canosa left INCAP and returned to Spain; the direction of the project then was assumed by Klein. In that same year the fresco supplement was made equivalent to the atole in its concentration of vitamins and minerals thought to be limited in the food supply, and fluoride was eventually added to both supplements to combat rampant caries in young children.
Subsequent changes in the professional staff are shown in table 1. Of particular interest is the effort in the early years of the project to identify, train on the job, and send out for further training outstanding young Latin American professionals who then returned to the project. Among them were Reynaldo Martorell, until recently at Stanford University and now professor at Cornell University; Aaron Lechtig, now UNICEF nutrition representative for East and Southern Africa; and Hernán Delgado, now director of INCAP. Martorell completed his doctoral dissertation at INCAP, and Lechtig and Delgado attended the Harvard School of Public Health for postgraduate training. Two others who made significant contributions to the data processing activities were Pedro Arenales (1970-1977) and Bill Owens (1975-1977).
TABLE 1. Full-time doctoral staff of the Division of Human Development, 1969-1977
Dates |
Nationality |
|
| Directors | ||
| Cipriano Canosa, MD | 1969-70 | Spain |
| Robert E. Klein, PhD | 1970-77 | USA |
| Field directors | ||
| Juan Jacobo Erdmenger, MD | 1969-72 | Guatemala |
| Guillermo Guzman, MD | 1972-77 | Guatemala |
| Psycho-social section | ||
| Robert E. Klein, PhD (section head) | 1969-77 | USA |
| Otto Gilbert, PhD | 1969-71 | Guatemala |
| Victor Mejía Piveral, Lie | 1970-77 | Guatemala |
| Barry Lester, PhD | 1971-73 | USA |
| Frederick Morrison, PhD | 1971-73 | USA |
| Charles Teller, PhD | 1972-77 | USA |
| Robert Lasky, PhD | 1972-74 | USA |
| Patrice Engle, PhD | 1973-77 | USA |
| Mark Irwin, PhD | 1974-77 | USA |
| John Townsend, PhD | 1975-77 | USA |
| Biomedical and epidemiology section | ||
| Jean-Pierre Habicht, MD, PhD (section head) | 1969-74 | Switzerland |
| Aaron Lechtig, MD (section head, 1974-77) | 1970-77 | Peru |
| José Maria Reyna-Barrios, MD | 1969-70 | Guatemala |
| Oscar Naranjo, MD | 1969-70 | Colombia |
| Ricardo Blanco, MD | 1969 | Guatemala |
| Reynaldo Martorell, PhD | 1972-77 | Honduras |
| Hernán Delgado, MD | 1972-77 | Chile |
| Victor Valverde, PhD | 1972-77 | El Salvador |
| Statistical computing section | ||
| Charles Yarbrough, PhD (section head) | 1970-77 | USA |
| Peter Russell, PhD | 1969-77 | USA |
Visits to the site by multidisciplinary teams approximately twice a year for advice and evaluation were provided under the NICHD contract. The first was held in June 1966. Many illustrious scientists served as members of these groups: Drs. Nancy Bailey, Josef Brozek, B. David Coursin, Frank Faulkner, Howard Freeman, John Gordon, Samuel Greenhouse, Jerome Kagan, David Kallen, Hamish Munro, Ricardo Ramos-Galvan, Harriet Rheingold, Henry Ricciutti, Alex Roche, A. Kimbal Romney, Nevin Scrimshaw, Edward Sussman, and others. A number of these scientists also served on smaller specialized consultative groups.
These site-visit teams and consultants identified and helped resolve many of the difficult issues that surrounded a project of this type:
» Issues related to behavioural testing:
—the need for tests that were (a) suitable to the culture, (b) sufficiently sensitive to discriminate impacts in specific domains, and (c) likely to be able to predict behavioural change early in life;
—the need to control for the impact of frequent testing on test performance;
—the question of cognitive scores versus the social meaning of behavioural differences.
» Issues related to the social stimulation arising from
—Who attends and why?
—What means might be used to increase attendance and equalize the social impact (television? a lottery? prizes or gifts?)?
—Does children's attendance change their development pattern through play? Does it decrease their anxiety so that their test scores improve?
» The issue of the symbolic and cultural meaning of atole (a food) versus fresco (a refreshment) in terms of its effects on attendance and parental perception of the child's health and well-being.
» The need for standardized terminology and criteria for tracking morbidity. » Issues related to nutrition status:
—Are home dietary surveys needed frequently, or only periodically?
—Can recording the number of cups of supplement consumed be substituted for dietary assessment?
—Is physical growth a reasonable surrogate for nutrition status?
One of the most heavily discussed and controversial questions was the ethics of intervention research in under-served, high-risk populations in developing countries. It will be recalled that this was the period of increasing sensitivity to research ethics, generated in part by Dr. Sol Krugman's widely publicized hepatitis vaccine tests using mentally retarded children in New York City. The INCAP staff and the site visitors wrestled with ethical questions for many months, questions such as what benefits the village populations would gain in return for the intensive, prolonged testing and measurements. One answer was the provision of health care not otherwise available in the community. But could health care be provided without plans to include nutritional treatment of clinical malnutrition? In the primary villages, such children could be treated and tracked as a separate subgroup, but what about the supercontrol villages? Could status and change be measured at periodic intervals in these villages if no continuing health care or other benefit were to be given? Considerations such as these contributed to the elimination of the supercontrol part of the original plan, although it was hoped to return to these villages at the end of the study for a new survey to assess changes independent of interventions.
Another major concern centred on whether a nonnutritive placebo could be used in villages known to have extensive childhood malnutrition. If a placebo were to be used, how safe were saccharine or sugar substitutes, particularly in malnourished pregnant women and young children? Cyclamates were suspected to be carcinogenic. Subsequently, it was proved that saccharine is carcinogenic in high doses in rats. United States support for the INCAP study would have ceased had this placebo been used. Thus, the use of fresco, a popular drink containing calories, was agreed upon. This was fortunate because it permitted differentiation between the effect of protein and calories, which was particularly important because calories turned out to be more limiting than protein in the diet of pregnant women.
Finally, when the project ended, what were the residual ethical obligations to the communities? How would they be satisfied? The resultant protocols were reviewed at INCAP, PAHO, NICHD, and WHO, and it was judged that they more than satisfied current ethical standards.
Some have said that these frequent oversight and consultative visits took excessive time from research, especially in report writing and even premature publication. Others point out that new insights were provided, pitfalls avoided, and creative new approaches devised; preparation of annual reports also provided regular opportunities to assess progress across disciplinary lines. Undoubtedly both points of view were partly true. The resultant project and its data base is one of the most unique and extensively controlled and documented in the scientific literature. No other study has comparable perinatal data with follow-up into early childhood. The behavioural test battery adopted for the field study was broader in scope and more sophisticated than those used in high-risk populations in the United States. On the other hand, the sheer magnitude of the data files made analysis and timely publication difficult. In an attempt to satisfy the recommendations in all areas, the project grew in scope, complexity, and cost (table 2). Ultimately, it was the escalating cost that led to disaffection and then withdrawal of funding by the NICHD, particularly in the company of a changing philosophy of what types of research were important for understanding human health and achieving human potential. The NICHD and the world in which it prospered moved away from long-term studies of human development and toward shorter-term clinical types of research. Thus the NICHD became more like the other institutes within the National Institutes of Health.
TABLE 2. Fiscal history of the INCAP longitudinal study
Funding (US$) |
||
Basic |
Supplementa |
|
| 1965 (June) | 133,910 |
|
| 1966 | 160,000 |
|
| 1967 | 269,051 |
|
| 1968 | 330,000 |
|
| 1969 | 365,000 |
|
| 1970 | 429,998 |
|
| 1971 | 580,684 |
|
| 1972 | 563,757 |
|
| 1973 | 543,685 |
|
| 1974 | 578,781 |
|
| 1975 | 727,512 |
35,661 |
| 1976 | 630,708 |
37,497 |
| 1977 | 325,039 |
76,842 |
| 1978 | 148,024 |
48,891 |
| Totals | 5,786,149 |
198,891 |
5,985,040 |
||
a. USAID supplement for continuation of studies, entitled "Influence of Maternal Nutrition on Infant Mortality and Morbidity"
The four-village study at INCAP sparked a great deal of interest, particularly as it became clear that many interactive factors confounded an understanding of the relationship between malnutrition and mental development. A series of studies involving nutrition and behavioural interventions were developed in Colombia and Mexico that complemented the work in Guatemala, many of which were also funded by NICHD and benefited from INCAP's experiences. These are described elsewhere [9]. The desirability of establishing a central data repository for these and other longitudinal studies of human health and development has been recognized [10], particularly since the opportunity and funding for such interdisciplinary studies will not come again soon; but this concept has not been realized, nor have procedures been agreed on to compile, document, and protect data files and to ensure access to them by future scientists. Only this Guatemala study is available to other investigators who were not involved in the field work.
For instance, interesting results have been reported from a secondary analysis of the data from the INCAP study [11]. Longitudinal analysis of individual child records was used to demonstrate a positive association between atole supplementation and growth in weight and height, offset by the negative impact of diarrhoea. Verbal development and school enrolment similarly were associated positively with the protein supplements, and total dietary intake also affected school achievement. The analytic approach used permitted the investigators to assess the effects of other non-nutritional factors on child growth as well. It is clear that secondary analysis by new techniques can yield innovative and provocative results. This instance, however, also reveals how desirable some intellectual contact with key members of the INCAP team would have been; it would have enabled the investigators to avoid the conclusion concerning the effect of protein, which is not now accepted by others.
The uniqueness of the population and the comprehensive data files from the INCAP study have made follow-up studies an intriguing and enticing goal as well. Studies using a subsample of 138 of the 6-8-year-old children living in three of the four villages [12, 13] led to the conclusion that "early nutritional stress compromises the young child's ability to respond to stimulation and to engage the environment in interaction. Over time, these deficits may inhibit the development of social competence and the ability to perform in routinely stressful situations." These studies were supported by the National Institute of Mental Health. The importance of such follow-ups provides the rationale for the 1988-1989 study reported in this issue.
Three other points may be worth noting: (1) The survival of long-term studies such as these at INCAP and elsewhere in Latin America depends on enthusiasm and flexibility. When enthusiasm by the sponsor wanes, innovative approaches to changing research needs are submerged under administrative control and creativity is stifled. (2) Different styles of leadership are required at different stages of design, implementation, and project maintenance. Charisma and conviction may be required in the early stages, confidence and perseverance in the implementation phase, and strong research management and administrative skills later to keep the project on course. These needs are each apparent in retrospect in the INCAP study, although they were not always acted on at the time. (3) The concept of a permanent repository for longitudinal data remains important as fashions in research rise and fall and as new questions are posed that require these data for analysis. Without access to past study files, pilot testing of new hypotheses that could provide guidance for relatively less expensive new short-term field studies on these populations cannot be carried out. The 19881989 follow-up study is an excellent example of this point.