Contents - Previous - Next

Experimental design (I: Scrimshaw et al., 1967a)

The design called for the selection of three predominantly Mayan villages in the highlands of Guatemala, as similar as possible in all major characteristics. In one village, good medical care was provided for all residents. The second village received supplementary food for all preschool children and their mothers during pregnancy. The third was a control village receiving only the scant health attention provided by the government program. Mortality, disease-specific morbidity, anthropometric measurements, dietary intake of preschool children, clinical evaluation, and examination of fecal samples for parasites were done routinely for five years on all children under five years of age, as described below.

Search for Suitable Study Villages (II: Scrimshaw et al., 1967b)

The first step in any field study is securing authorization and understanding from government authorities at all levels, and the acquiescence of the local leaders and the communities. Therefore, the project was presensed to health officials at the national and departmental levels and to civil authorities at the departmental and municipal levels for each locality. It was also presented to schoolteachers and religious leaders for approval. Since Guatemala was one of the member countries of INCAP, approval of the national and departmental civil and health authorities was easily obtained, but careful and patient effort was required to assure cooperation of the local leaders. Only then was the project presented to the village, with a special effort to ensure the participation and support of both prestige leaders (informal) and clique leaders (heads of religious societies or other local groups). Conflict with other INCAP studies was avoided, and there were no other organizations working in the selected villages.

Eight years of prior field work by INCAP had found the communities in this area of the Guatemalan highlands to be cooperative and willing participants in intervention studies. Moreover, permanent migration was rare and dropout rates were low for preschool children. Although there was some short-term seasonal migration of adult males to the lowlands for work on sugar, cotton, and coffee plantations, their families usually remained behind in the home communities.

Reasonable proximity to a base of operations was essential for efficient logistics, but relative isolation was also required. In Guatemala, the highland rural population resides in small (200-2,000 persons), compact, and semi-isolated villages, many of them within 50 km of INCAP. The area is 2,000 meters above sea level and experiences a rainy season from about May to September and a dry season from about October to April. Roads are more or less passable even in the rainy season. Annual rainfall averages 120 cm and the yearly mean temperature is 18°C, with a low of 10°C in December and a high of 21 °C in May. The incidence of diarrhea varies with the season. It is lowest during the dry season from December to February and highest during the rainy season from May to August.

The prevalence of disease was Consideration of paramount importance in determining the suitability of villages in the area selected, the necessary duration of the study, and the size of the study populations. Previous INCAP work had shown incidences of diarrhea ranging from 180 to 400 cases/100 children under age five years, and a death rate of 20.5/1,000 children (Scrimshaw et al., 1962). Similar specific information was not available for upper respiratory infections, measles, and whooping cough, but death records indicated that their frequency was also high, despite government efforts to provide diphtheria-pertussis-tetanus (DPT) and measles vaccines to every young child. Seven of the 10 leading causes of death were listed as infectious diseases, including those caused by parasites (Béhar et al., 1958).

The high prevalence of nutritional deficits in the area had been well documented by INCAP. Mean weight-for-age for children under five years in this population fell below the 16th percentile, and kwashiorkor was clearly identified as a common and frequently fatal disease in the area (Scrimshaw and (Béhar 1965). Although some villages were visited occasionally by physicians of the National Health Service, practitioners of folk medicine generally attended to the health needs of the highland people, and obstetrical services were commonly provided by untrained, usually illiterate, midwives. For clinic or hospital service, travel to provincial capitals or to Guatemala City was the norm.

The time required for the preliminary survey of the area and its people (approximately one month) was unusually short because of extensive previous work by INCAP and experienced field personnel. Usually, this indispensable component in the planning of field studies requires a much longer period of time.

Basis for Village Selection (II: Scrimshaw et al., 1967b)

From available estimates of disease incidence and mortality for the selected area, it was determined that villages with a population of about 1,000, assuming that 16% of the total population were children under five years of age, should yield reliable data for evaluating changes in nutritional status and incidence of infections resulting from interventions. Twenty predominantly Mayan agricultural communities, with minimal contact with each other, were visited. Some of these villages were removed from consideration promptly, whereas others required repeated visits before final selection could be made.

The mayor was interviewed in each promising village to obtain data on total population, racial structure, and predominant economic activity. Civil registries were examined for deaths and their recorded causes. Village cemeteries were visited to verify random sets of death records. Inspection of houses for assignment into appearance categories provided a basis for gross comparisons of living conditions and sanitation among villages. Housewives were interviewed for information on eating patterns, with special reference to infants and young children. The three most comparable villages were selected, taking into account similar disease frequency and accessibility to INCAP.

Random selection from a list of eligible villages within 7 5 km of the base of operations could result in technical and logistic complications. Small size of populations, possible diversity of ethnic stock, and localization in diametrically opposite directions from INCAP are examples of some of these complications. On the other hand, a properly qualified intentional choice of villages may avoid such complications.

The chosen villages were overwhelmingly Mayan, were located in the highland area surrounding Guatemala City, were accessible by all-weather roads, and had agriculture as the only source of livelihood. The villages were close enough to each other and to INCAP for supervisory visits to all in a single day, yet not close enough to allow significant interchange among them. The villages judged to best satisfy the above criteria were Santa María Cauqué (SMC), Santa Catarina Barahona (SCB), and Santa Cruz Balanya (SCZ).

Characteristics of the Villages Selected (II: Scrimshaw et al., 19676)

Previous INCAP research in Santa María Cauqué (SMC), a village with 923 inhabitants, provided ample information on its health and dietary practices and documented little or no change over the previous 10 years. The general environment of SMC was also well within the frame of reference for the area. Support for the previous activities of INCAP in this village, the health authorities of Guatemala had provided a simple clinic building, well suited for delivering rudimentary medical care and services to the community. Accordingly, SMC was selected for the "treatment" intervention.

With a population of 753, Santa Catarina Barahona (SCB) adjoined a larger community for which INCAP data showed environmental and health conditions similar to those of SMC. Furthermore, INCAP had evaluated physical growth in school children of both SCB and SMC, finding that these two villages had similar growth patterns and nutritional status. Because of the smaller area of SCB, administration of a feeding program would be less difficult. Accordingly, SCB was chosen as the "feeding" village. The community willingly provided two rooms for the feeding program.

The selected control village, Santa Cruz Balanya (SCZ), had the largest population (1,363 inhabitants), but this was considered desirable. SCZ had not been studied previously and required a longer period of reconnaissance. However, it proved to have characteristics and conditions broadly comparable to those of SMC and SCB. The population was cooperative and accepted the nature and aims of the program, and it voluntarily provided a small room as headquarters for the study. Although it was not to receive medical care or food, INCAP agreed to help the village in other ways, including providing a roof for their assembly hall, arranging for the improvement of their entry road, and contributing to various community activities.

The population in all three villages was predominantly Mayan Indian, as judged by dress and language used in the household. The proportion who did not identify themselves as Indian was 2% in SCB, 4% in SMC, and 6% in SCZ. An average of 44% were under age 15, and 17% were under five years of age.

The overall nutritional status of preschool children aged one to four years was poor (weight deficit of 10% or greater for age) in all three villages (84% in SMC, 88% in SCB, and 82% in SCZ). SCB had a smaller proportion of preschool children with second or third degree malnutrition (weight-for-age deficits of 25% or more) than either SCZ or SMC. In all three villages, calorie and protein intakes of the children were relatively poor, although children in SMC had a better diet, contradicting findings for weight-for-age deficits. Overt clinical signs of nutritional disease were few or nonexistent.

Intestinal parasites were found in 60% of the children, and over half of them had more than one parasite. Ascaris was found almost universally, while Entamoeba coli, Giardia lamblia, and Trichuris trichiura were present in 10% to 20% of the children. Deficiencies in environmental sanitation were universally evident.

Dwelling characteristics were similar in the three villages. Those classified as poor (straw roof, cornstalk walls, and dirt floors) were predominant. Without exception, chickens and dogs shared dwellings with household members.

Water for these villages, supplied from common sources, was distributed through public spouts dispersed throughout the villages, and individual households had to carry water by hand for varying distances. There was no prior information on morbidity, but a survey of deaths over the previous 10 years provided estimates of crude death rates of 24, 2 5, and 31/ 1,000 in SMC, SCB, and SCZ, respectively. Infant mortality was high: 136, 182, and 186/1,000 live births in SMC, SCB, and SCZ, respectively.

All children in the three villages were breast-fed, and the weaning process very rarely started before the third month of life. The commonest time for a systematic addition of supplements to the diet of an infant was eight to nine months after birth. Indeed, the majority of children continued to receive some breast milk for a prolonged period of time, often beyond age 24 months. Weaning practices were similar in the three villages, and pooled data for complete weaning in 267 children during the five years of study (Table 1) document the prolonged breast-feeding (V: Scrimshaw, Guzmán, et al., 1968).

A summation of all of the evidence indicated that despite some minor differences and discrepancies, the three villages were as nearly equivalent and comparable as could be expected when dealing with natural human populations (II: Scrimshaw et al., 1967b).

Field Procedures and Methods (III: Scrimshaw et al., 1967c)

The Feeding Intervention

The feeding program implemented in SCB was designed to improve substantially the nutritional status of preschool children without intentionally altering the sanitary conditions of this village. The program was planned and supervised by an experienced nutritionist. A high-quality protein food consisting of 18.5 g Incaparina3, 30 g dried skim milk, and 17.5 g sugar was offered six days a week as a mid-morning snack to preschool children and pregnant and nursing women. Cooked as a gruel, an 8-oz (227 g) serving supplied 15 g good-quality protein, 225 calories, and micronutrients. A banana was included with all servings, providing 125 additional calories.

Since breast-feeding usually satisfies nutrient requirements up to four to six months of age, the supplement was not encouraged for infants under this age range without evidence of growth retardation. The food was prepared and distributed at one of the rooms provided for the program.

TABLE 1 Completed Weaning in 267 Children in the Three Study Villages.

Month Weaning Completed

Total Children





6 - 11



12 - 17



18 - 23



24 - 35



36 - 47



Occasionally, however, in case of incapacitating illness, the supplement was taken to the home of the recipient. Individual records were kept of both attendance and approximate amount of supplement consumed (III: Scrimshaw et al., 1967c).

The food supplement was designed to increase the calorie intake of preschool children by one-third and to double their intake of high-quality protein. In practice, however, this was not the case. As shown in Table 2, not only did the amount of supplement received by a child vary greatly, but also the proportion of children who received 75% or more of the prescribed amount of supplement declined progressively from 51% to 18% during the five-year study. Conversely, the proportion of children who did not take part in the feeding program increased from 11% in the first year to 37% in the second year, and settled around 20% for the remainder of the time.

In an attempt further to improve the nutritional status of the preschool children in this village, nutrition education activities were carried out during the first 18 months of the study by a resident home economist who talked with the mothers collectively and individually. Visual aids and demonstration techniques were used extensively to ensure correct delivery and reception of intended messages. As follow-up, the leaders organized community meetings to inform all housewives in the village. New couples received special attention by home visits of a health educator. In all instances, the focus was on proper use and handling of local food resources for maximum nutritional benefits to the family and, more specifically, for children under five years of age. Unfortunately, the sequential dietary surveys provided no evidence that this program was effective.

The Medical Intervention

The full-time services of the same physician and nurse, both with graduate public health training, were provided five days per week with all necessary medicines to dispense without charge. The clinic had offices, a waiting room, a simple laboratory, and living quarters for the nurse and the field-workers who recorded morbidity by household visits. The more sophisticated laboratories of INCAP were available for backup, including diagnostic microbiology. A part-time sanitarian promoted latrine construction and personal hygiene for the first three years. A safe water supply was ensured at all times, although households still had to carry water from a central fountain. Immunizations were left to the routine of the public health service, and this proved to be a serious deficiency in the program. The services were provided to the entire community, with special attention to children under five years of age. Detailed individual charts were kept within standardized family files.

TABLE 2 Food Supplement Received by Preschool Children of Santa Catarina Barahonaa

Percentage of Participation






Year Ending April

Total Children







































































aBy percent of prescribed amounts and years of the study, May 1959-April 1964.

The Control Community

As a control community, SCZ was to remain without change in medical, sanitary, or nutritional practices. A physician visited this village at weekly intervals to supervise field operations but did not offer medical care. A medical and laboratory team visited SCZ at quarterly intervals to carry out the routine surveys required by the study. These personnel were seldom involved in medical problems of the village, and resident field-works refrained from giving medical or nutritional advice. A variety of social and educational activities-evening movies, puppet shows, athletic events-were provided for this village. Schoolrooms were improved and the access road to the village was rebuilt and graded as a token exchange of benefits to ensure continued cooperation (III: Scrimshaw et al., 1967c).

In both SCB and SCZ, there were rare occasions when it was humanely and ethically imperative to provide emergency medical services to seriously ill persons. Most often this consisted only in providing transportation to a hospital, and such incidents seldom involved preschool children.

Activities Common to All Three Villages

In addition to village-specific activities and the keeping of a detailed diary of significant events and progress in the continuing program, the study plan required basic data collection by common procedures in the three study villages, as described below (III: Scrimshaw et al., 1967c).

Population censuses taken at the start of the study and repeated at yearly intervals thereafter recorded changes in family structure and available facilities. This information was used for stratification of families by social and economic conditions. The censuses also documented changes in village characteristics for evaluating possible association with study interventions.

Disease morbidity and information on injuries for children under five years of age were obtained through biweekly home visits. Occurrence, duration, treatment, and symptoms such as fever, coughing, respiratory distress, anorexia, nausea, vomiting, convulsions, skin rash, edema, and diarrhea were recorded. When diarrhea was present, the number of stools, their consistency, and the presence of blood were recorded. The nature and circumstances of injuries and of events in the life of a child with likely health significance were also noted. Programmed home visits were the responsibility of two full-time field-workers in residence at each village.

Deaths were documented by the field-workers, who learned of such occurrences soon after the event, as can be expected in a small community. The date and place of death, age, birth certificate number, and cause of death according to the civil registry were recorded for each event. In the course of weekly supervisory visits, the study epidemiologist asked the field-workers to provide details of circumstances for each death as needed to diagnose a probable cause and assign it to one of the broad classification groups used in this study.

Nutritional status was evaluated primarily by anthropometric measurements. Weight-for-age and tricipital skinfold thickness were considered measures of caloric intake adequacy for growth. Height-forage was used as an index of development and adequacy of protein intake. All measurements were made quarterly in blind duplicates by experienced INCAP personnel, using standardized procedures and equipment with routinely programmed checks of reliability. Periodic physical examinations were discarded in the first year because they were not useful. This was due to the scarcity of clinical signs, their lack of specificity, and poor reproducibility.

Roentgenograms of both wrists of preschool children were also made at the start of the study and at yearly intervals thereafter. Evaluation of ossification status in these roentgenograms provided an index of maturation.

Dietary adequacy was measured during home visits at the start of the study and at yearly intervals thereafter. All dietary surveys were conducted by experienced INCAP nutritionists. Data were obtained through a combination of observation, weighing, and interview during daily visits programmed over consecutively staggered three-day periods. All days of the week were represented equally. In each village, 10 to 14 families were randomly selected for these surveys. Seasonal food production and availability, with particular reference to the foods added or withdrawn from the diets of infants at the beginning or end of weaning, were also noted in the course of the dietary surveys.

Prevalence of enteric infectious agents was determined at quarterly intervals for comparison with the frequency of acute diarrhea! disease and the incidence of kwashiorkor. Rectal swabs were collected, promptly inoculated in the field into appropriate culture media, and incubated on the same day at the central INCAP laboratory. Fecal samples were examined for intestinal parasites and used in isolating enteroviruses. Routine procedures of INCAP were used for these investigations.

The epidemiology of acute diarrhea! disease was a special investigation conducted in the treatment village. Cases were evaluated epidemiologically and examined bacteriologically for enteric pathogens during the five years of study. These examinations were generally based on a single stool or rectal swab, but during a trial period of 18 months, fecal samples were collected on successive days until either a positive result or five successive negatives were obtained.

Data Management

Data management was the responsibility of the INCAP Division of Statistics and included the coding of data, production of interim reports, and eventual statistical analysis. All data collected on a daily basis were reviewed by the field supervisors. At this time, discrepancies and omissions were rectified, either by the field-workers or by personal inquiry by the supervisor. The data were submitted weekly to the project epidemiologist for review, reassurance of completeness, and classification.

In this study the infectious diseases category was predominant and included diseases of the digestive tract, respiratory diseases, common communicable diseases, parasitic diseases, and other infectious diseases. Other broad categories considered in this classification scheme were accidents, poisoning and violence, congenital malformations, neoplasms, nutritional diseases, metabolic and degenerative diseases, skin diseases, and other ill-defined diseases.

All diagnoses and classifications were made by a physician with reference only to clinical information collected by nonmedical field-works. Most of the methods required for collecting data as outlined above had been developed and used in earlier INCAP field studies in the region (III: Scrimshaw et al., 1967c).

Stages of the Study

Most of the technical and professional personnel needed for the project were already in residence at INCAP, had participated in previous studies, and were familiar with the proposed methodology. Accordingly, the tasks required for assembling staff, selecting methods, and defining field organization were completed in the unusually short period of two months, and a pilot study for testing personnel and procedures started in March 1959. Late in April 1959, the staff reviewed the protocol of standard operations and modified it in accord with the pilot experience. Data collection for the definitive study began May 1, 1959, and continued for five years, through April 1964.

In interpreting the data, a distinction must be made between the first three and the final two years of the project. During the first three years, the project was closely controlled by the original investigators. However, in the final two years, responsibility for the project was assigned to an individual whose priority was his own study of the cognitive performance of children who had been receiving the supplement in the feeding village. 'the consequences of this are considered both in describing and in discussing the results.

Contents - Previous - Next