Contents - Previous - Next


The main results were presented in four papers that can be summarized as follows:

Mortality (IV: Ascoli et al., 1967)

a. In the treatment village, deaths among children aged one to four years showed a 50% greater decline than expected from rates oft he preceding nine years. The decline for the control village was also greater than expected, exceeding that of the treatment village. The best result was in the feeding village, with a 300% improvement over the expected rate.

b. Infant mortality in the feeding village improved 19% during the study after it had remained the same during the preceding nine years. Infant mortality in the control village was unchanged. Although the decrease was greatest in the treatment village (36%), it followed the trend line of the previous nine years.

c. Deaths among males and females of preschool age were essentially equal.

d. There were three epidemics of measles in the feeding village during the five years, probably because it was so close to a market town. The relatively high case fatality rates from measles, 6.8% in one outbreak, remained so in the surrounding villages over the five years of the intervention. However, the only death from measles in the feeding village was in a child who did not participate in the study.

Total Infectious Disease Morbidity (V: Scrimshaw, Guzmán et al., 1968)

a. In all three villages, infectious illnesses were high in infancy and peaked in the second year of life, dropping rapidly thereafter, as shown in Figure 1. At all aces. they were highest in SMC, the treatment village.

b. The total illnesses observed in each village are given in Table 3. During the first three years, the amount of disease in the feeding village was markedly less than in the treatment village and only slightly less than that recorded in the control village.

c. Children participating in the program 75% or more of the time had the lowest total number of days ill and the shortest average duration of illness episodes. These figures were significantly less than those for children participating less than 25% of the time. Children who participated at least 25% of the time had fewer illnesses, on the average, than those who participated less often.

Illnesses Per Child Per Year, by Age in Years, in Three Guateleman Villages, May 1959 to April 1964 (Scrimshaw et al., 1968)

TABLE 3 Illnesses of Children, Ages 0 through 59 Months, in Three Guatemalan Villagesa

Year Ending April

No. of Childrenb

Annual Days of Illness per Child

Average Days Duration

Illnesses per Child per Year

Santa María Canqué (Treatment Village)


























Santa Catarina Barahona (Feeding Village)


























Santa Cruz Balanyá (Control Village)


























a By years, May 1959-April 1964.
b Based on actual count of children with disease instead of mid-year census figures used in previous mortality rate calculations.

Diarrheal Disease Mortality (VI: Gordon et al., 1968)

a. As shown in Table 4, the drop in diarrhea! disease in the first three years in the feeding village compared with the other two was very pronounced, and diarrhea prevalence was highest in the treatment village. In the fourth year, a new field director reassigned the two experienced and much liked persons responsible for distributing the food supplement in the village to another program of his own. The result was a drop in participation in the feeding program. A sharp increase in diarrhea morbidity followed in the feeding village to a level similar to that of the control village, although it was still less than that in the treatment village. By the fifth year, the feeding village morbidity from diarrhea equaled that of the treatment village and exceeded that of the control village.

b. In the feeding village, 34% of diarrhea! attacks lasted less than four days, whereas the corresponding percentages were 25% for the treatment village and 21% for the control village.

c. Overall, about 11% of children in a given quarter had an episode of moderate diarrhea, about 8% had mild diarrhea, and only 1% had severe diarrhea. In the second year of life, about 73% were without diarrhea in a given quarter, and in the fourth year this percentage increased to 86%.

d. The incidence of diarrhea was lowest in the dry season months of December to February and highest in the rainy season from May to August.

TABLE 4 Incidence of Acute Diarrheal Disease by Year for Children 0-4 Years Old

Santa María Cauqué (Treatment Village)

Santa Catarina Barahona (Feeding Village)

Santa Cruz Balanyá (Control Village)

Year ending April

No. of cases

Cases/100 children per year

No. of cases

Cases/100 children per year

No. of cases

Cases/100 children per year




































Respiratory Disease Morbidity (VI: Gordon et al., 1968)

a. Respiratory disease was consistently and significantly less prevalent in the feeding village than in the treatment village. However, it was lowest in the control village presumably due to its greater isolation.

b. Respiratory diseases decreased with increasing participation in the supplementary feeding program. Those who received food less than 25% of the time had 107 respiratory episodes per year, compared with 65 for those participating 25% to 74% of the time and 54 for those whose participation was 75% or greater.

c. About one-third of all respiratory illnesses in the three villages had a duration of less than one week. Approximately another half ended in the second week, and a further sixth within the third week.

d. Respiratory illnesses of a month or more were most frequent in, the treatment village, approximating 6% of the total; in the other two villages the proportion was about 2%.

e. Both the availability of treatment and the feeding saved lives from respiratory disease. Case fatalities due to respiratory disease were more than 50% less in the treatment (1.2%) end feeding villages (1.7%) than in the control village (4%). This difference also held for mortality, which was 15.5/1,000 preschool children at risk in the treatment village and 12.8 in the feeding village, in contrast to 23 in the control.

Parasitic Infections (VI: Gordon et al., 1968)

a. By six months of age, 21% of children harbored at least one intestinal parasite. The rate increased progressively with age to reach 89% in children between four and five years of age. Ascaris and Trichuris were the most common, but they only rarely produced identifiable symptoms.

b. Despite the sanitary measures attempted in the treatment village, there was no evidence of success either in controlling general disease incidence or in improving parasite rates in children. The parasite rate was unchanged after three years in all three villages.

Physical Growth (VII: Guzmán et al., 1968)

a. No differences between boys and girls in mean height and weight were observed. Boys and girls in the feeding village tended to grow faster in both height and weight than those in either the treatment or the control village. There were no differences between the treatment and control villages in these measurements. The differences in growth between the feeding village and the other two villages resulted in a net difference of approximately 3 cm in height and 1 kg in weight by the end of the fifth year of life.

b. Although boys and girls had equal head circumferences at the start of the study, the head circumference of boys increased faster than that of girls in the three villages. In both sexes, the head circumference at birth was about 70% of the measurement at five years of age, and at 12 months it was about 90%. Preschool children of both sexes in the feeding village had greater head circumferences than their counterparts from either the treatment or the control village. For all three villages, these measurements were significantly smaller than those of US children of comparable ages.

c. Wrist ossification T scores were low in all three villages compared with reference standards for well-nourished children. The feeding village had a slightly higher score than the other two at the start, and the difference increased significantly during the study. It showed no improvement in either the treatment or the control village.

Collateral studies

The series of nine papers summarized here covered the specifically defined objectives of the study. They are rich in details and insights that cannot be presented in this relatively brief summary. In addition, even more reports by study investigators arose from collateral findings or were stimulated by the initial results. These included additional observations on diarrhea! disease (Gordon, 1964, 1965; Gordon et al., 1964a,b,c, 1965b,c; Gordon and Scrimshaw, 1965a,b; Mata, 1964, 1967; Mata et al., 1965, 1966; Catalan et al., 1965; Ascoli and Mata, 1965), bone maturation (Garn et al., 1964a,b, 1967; Rohmann et al., 1964; (Béhar et al., 1964), chicken pox (Salomon et al., 1966) and measles (Gordon et al., 1965a; Scrimshaw et al., 1966), and child growth and development (Béhar 1968; Guzmán, 1968; Guzmán, et al., 1966; Viteri et al., 1966). The populations were also used by other investigators for "piggyback" studies of oral conditions (Sweeney and Guzmán, 1966), and breast-feeding and weaning practices (de González, 1963; de González and (Béhar 1966). SMC subsequently became the site for an eight-year study of children from birth to five years of age summarized by Leonardo Mata in The Children of Santa María Canqué (1978) and summarized in Chapter 2.

Discussion of results

The results of this study have been available for more than 20 years and they have been commented upon and interpreted in a variety of ways. The penultimate paper (VIII: (Béhar et al., 1968) was perhaps too frank in pointing out confounding factors, including the inevitable variations among apparently well-matched villages. The result was that too often reviews of this study emphasized the confounding factors, instead of stressing the important policy significance of the results.

In our summary of results in this paper, we have described the significant beneficial effects of the nutrition supplement, especially during the first three years when the project was the responsibility of the original principal investigator. This occurred even though participation was far short of program goals. What was equally striking, and much less anticipated, was the almost total lack of demonstrable benefits for preschool children from the costly and high-quality medical care offered in the treatment village. The only significant benefit associated with the medical care was that the lives of several severely ill children were saved, but not any more than were saved by the food supplement.

In order to avoid accepting this unexpected conclusion, confounding factors have been overemphasized by some reviewers who have mistakenly cited the results as minimal or inconclusive. Some have sought to discount the results by suggesting that the medical treatment must have been inadequate. This was certainly not the case. The clinic building was attractive and well equipped and had laboratory backup provided by INCAP. No household was more than about five minutes from the clinic. Both the physician and the nurse had advanced training and, with occasional rare exceptions, were in the village five days a week throughout the study. Moreover, there was virtually no limit on free availability of medicines. No industrialized country, much less any developing country, could possibly afford to provide this level of care to a community of fewer than a thousand persons, except on an experimental basis. The only reasonable conclusion is that a program of medical care alone, no matter how sophisticated, cannot prevent the malnutrition and infectious disease morbidity that so greatly affect populations of this type in developing countries.

The main health problems among preschool children in the village were diarrhea, respiratory infections, and the communicable diseases of childhood. Immunization against the latter is a preventive, not a curative, measure. In this case, immunization was the responsibility of the government and was only erratically provided, despite repeated requests. Similarly, the availability of treatment did not prevent the spread of respiratory disease.

As for diarrhea! disease, nothing in the clinic program had any influence on the contact spread of diarrhea! infection from contaminated hands and a dirty physical environment. For an experimental period of three years, every child coming to the clinic with diarrhea received either an antibiotic or sulfonamide, but in comparison with prior diarrhea! prevalence or with rates in the other two villages, the high prevalence rates were not affected. The mere presence of the clinic did not lead to the improvement in personal hygiene needed to reduce contact spread of diarrheal infection.

If providing curative medicine of this quantity and quaky does so little for the health of young children, the sporadic and poor-quaky government curative services certainly did not contribute to bringing down the high morbidity and mortally among the preschool children in Guatemala. The only potentially effective preventive measure by the government health service at the time of the study was immunization, but the visits by mobile teams to the villages were erratic and poorly planned. There were also occasional breaks in the cold-chain that rendered the results unreliable. As a result, outbreaks of measles occurred shortly after a visit by an immunization team.

The policy significance of the results in the feeding village has often been discounted because their magnitude seemed small. However, despite the fact that nothing was introduced in the feeding village that would directly affect exposure to the infectious diseases that caused most of the ill health in Guatemalan villages, their prevalence, severity, and consequences were reduced, even when only one-half of the children received 50% or more of the prescribed supplement.

There are several different ways of looking at the feeding village results. The most common view has been that the differences were small, but so was the average degree of participation. The remarkable finding was the occurrence of any significant responses to this single intervention under such adverse circumstances. The results lend credence to the value of nutritional improvement. Equally, they confirm that nutritional improvement alone is not sufficient to overcome the lack of other health measures.

There was an attempt to include preventive services in the treatment village, and in fact, in the summary of the final paper (IX: Scrimshaw et al., 1969), reference is made to a "program of preventive medicine and medical care." However, the actions taken were either irrelevant or failures of execution.

Potable water was ensured, but the infections harming young children were not water-borne. Young children became infected by contact spread of diarrhea, but because the villagers still had to carry the water considerable distances to their homes, there was no increase in its use for personal hygiene. Additional pit latrines were built, but there was little evidence of either their use by young children or an association between diarrhea and availability of latrines. Promotion of breast-feeding would have been meaningless in these populations, in which prolonged breast-feeding is universal, and the health statistics would have been immeasurably worse if this had not been the case.

In retrospect, it is clear that for the program in the treatment village to have been a success, it would have needed to incorporate effective primary health care that stressed prevention. An effective program of immunization would have eliminated the effects of chicken pox, whooping cough, and measles that are periodic problems in all three villages. Improved availability of water and emphasis on its use for personal hygiene would have reduced the contact spread of diarrhea! disease. Periodic weighing of all children until they were eating the same diet as other family members could have detected and prevented most moderate to severe malnutrition if properly used to monitor nutritional status and to stimulate remedial efforts by the mother. Both preventive medicine and supplementary feeding together would probably have been required to effectively reduce both morbidly and mortally of preschool children to levels characteristic: of industrialized countries.

There remains the paradox that diarrhea! and respiratory diseases did not decrease in the treatment village, and actually increased in the treatment village and ultimately also in the feeding village. The only way that we were able to account for this was the increased number of persons leaving and entering the villages as a direct result of the program. Not only did the doctor, nurse, and community workers come each weekday, but visitors to the program were very frequent. Bringing mothers and children together, either for clinic visits or for the feeding program, facilitated the contact spread of both diarrhea! and respiratory disease. Credence to this interpretation is given by a sharp increase in morbidity in the feeding village during the last two years of the program, when many additional people entered the village on a daily basis for cognitive studies of the children.

By contrast, because of a lack of interesting activities to visit and a deliberate desire to keep the control village undisturbed, additional visitors to it were infrequent, as were activities that brought mothers and children into a confined area. As already indicated, the control village was more distant than either of the other two villages from Guatemala City, from other townships, and from a main highway. Moreover, it had its own market, whereas in the other two villages, families frequently traveled to larger markets in nearby towns.

Much was learned about the problems of conducting studies of this sort. It was noted that a field study as long as this one inevitably must experience errors as well as changes of personnel that will affect its operations. Moreover, administrative difficulties and unexpected events are likely to occur and disturb what was conceived as a sound plan. When multiple villages are involved, differences eventually emerge, no matter how carefully they were originally matched. Additionally, there is always the risk that a chance event in one village-flood, earthquake, or epidemic may weaken or completely invalidate the experimental design. We were aware of these hazards and attempted to minimize them.

In addition to the incomplete participation in the feeding program, mention must be made that those from 4 to 24 months of age, who needed the supplement most, were fewest in number, while those less in need, the three- and four-year-old children, were in the majority among participants in the feeding program. The latter could attend alone, whereas someone needed to take scarce time to bring in the younger ones. Moreover, only a few of the children attended regularly throughout the entire study. The number present on a given day was fairly uniform, but the composition of the group varied from day to day. When an effect shows through such strong confounding factors, it is either a strong effect or an artifact; it is incumbent on the investigators to try to distinguish between the two. In these studies, comparisons between the treatment and feeding villages are reliable and significant. However, comparisons involving the control village are less reliable and suspect because of the ways in which it differed from the other two.

Medical, social, and public health benefits of the study

In addition to the comparisons between medical care and feeding, discussed at length above, the study contributed to a better definition and understanding of the general health of young children in the area. It not only confirmed the high incidence of acute diarrhea! disease and upper respiratory infections, but also documented the prominence of the communicable diseases of childhood. The results showed that these diseases are of longer duration, more severe, and more likely to be fatal in this underprivileged population than in well-nourished populations.

At the time of planning this study, undernutrition and malnutrition were recognized as endemic among children of underprivileged populations, although they remained largely hidden until exacerbated by an infection. We postulated then that the differences in child morbidly and mortally observed between these populations and those more privileged arose from the synergism between malnutrition and infection, which produces an effect beyond the simple addition of their respective independent negative contributions to health. The passage of time and additional studies have strengthened our early conclusion.

This basic ecologic interaction between a host and its environment should be the focus of preventive efforts instead of the traditional disease specific approach still commonly used by narrow specialists. Two outstanding facts emerge, both sufficiently strongly supported by evidence from other less-privileged populations to characterize them as general principles. One is that in all regions, the highest death rates after infancy occur in the second year, being comparable in magnitude to the postneonatal infant mortality rate. It follows that preventive measures focused on this period are more likely to be effective than those that take the conventional approach of targeting children between one and four years of age. It is always easier to get compliance from the older preschool children at the expense of efforts to increase coverage of the children most in need.

Public health statistics tend to focus on mortality rates. Expressing the size of the problem by the number of persons who are ill, instead of by the number who die, was of considerable practical value. The demonstration of a multifactorial causality also brought recognition that a program for control must include measures against a variety of factors. Singling out individual elements, whether in succession or indiscriminately, will not be as effective as a common approach to diseases having an interlocking and interacting effect. Perhaps the most important reminder provided by this study is that an approach based on concerted action against the major factors, social as well as biological, can be expected to give better results than measures against any one singly or in succession, even those as important as malnutrition or infectious disease (IX: Scrimshaw et al., 1969).

Contents - Previous - Next