Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
The data were analysed using the Statistical Package for Social Sciences. Although our target children were those from 12 to 66 months of age, responses from mothers of children 0 to 66 months of age were included in this portion of the analysis; mothers of infants were considered part of our audience because the infants would soon be within the target age range.
Interviewee responses were analysed and compared across conditions for the following five key questions regarding child-feeding practices and nutrition services. (The terms in parentheses are names by which the questions are referred to in the following discussion and the related graphs.)
First, it was necessary to see whether the answers to question 5 reflected the assumed differences in the degree of intervention between VTRL and VTRS exposures. The data showed a significant difference in intensity of treatment (X2 = 72.8; p < .001). Fifty per cent of the mothers in the VTRS villages reported they had seen no showings, in contrast to fewer than 20 per cent of the mothers in the VTRL villages.
Answers to the other four questions were rated "acceptable" if they met the following criteria:
1. Complete meals. Respondent mentioned at least one of each of the following foods:
2. Nutri-Pak for snack. Respondent named Nutri-Pak.
3. Specific BNS activities. Respondent named specific correct activities, such as
4. Specific Nutri-Pak information. Respondent gave correct and specific
FIG.2. Percentages of Mothers Giving "Acceptable" Responses to Questions, by Level of Intervention-Rural Leyte, 1979 (mothers of children 0 - 66 months old; VTRL N = 483, VTRS N = 373, BNS N = 200, comparison villages N = 269; chi-square significant at p < .001 for each question)
FIG.3. Mean Nutritional Status of Preschool Children by Level of Intervention-Rural Leyte,1979 (ages 12 - 66 months; Harvard standards; VTRL N = 703, VTRS N = 520, BNS N = 295, comparison villages N = 397; one-way anova significant at p < .001)
The differences in the answers to these four questions according to the levels of intervention were significant, with mothers in the VTRL villages consistently having the highest percentage of acceptable answers, those in the VTRS group the second highest, and those in the BNS and comparison groups the lowest (fig. 2). Comparing the proportion of mothers giving acceptable answers to the complete meals question with the responses to the VTR frequency question, we again find a significant difference, with those who had seen no VTR shows having the lowest score (33.9 per cent acceptable answers) and those who had seen four shows having the highest score (62.5 per cent acceptable answers).
1979 Results-Nutritional Status
Nutritional status was measured by percentage of standard weight for age using the Harvard standards, with separate standards for male and female. Among the three intervention groups (VTRL, VTRS, and BNS), VTRL had the highest nutritional status (mean of 78.4), VTRS had the second highest (mean of 77.0), and BNS the lowest (mean of 75.5); the comparison group was included for completeness, but represented a higher socio-economic group (fig. 3).
After the data were divided according to age groups, a two way analysis of variance showed significant difference among groups exposed to different conditions but no significant difference among age groups and no significant interaction effects.
To determine whether the higher means indicated less severe malnutrition, we also looked at the data using the Gomez classification scheme (fig. 4). The results show significant differences among conditions with respect to the Gomez classification. Children from the VTRL villages showed the most first-degree and the least second- and third-degree malnutrition, and the children from the BNS villages showed the least first-degree and the most second and third-degree malnutrition. Those in the VTRS villages were in the middle.
1981 Sample Selection and Data Collection
By February of 1981, all Waray-Waray-speaking BNS villages in Leyte accessible by road had Nutri-bus operations for at least 18 months. Matching control villages without intervention were not available in Leyte. An effort was made to find groups based on degree of intervention. All Nutri-bus villages meeting the same criteria used in 1979 were assigned a score based on daily field reports from the Nutri-bus staff regarding services rendered. Scores were based on number of VTR showings, number and type of supervisory calls to the BNS, and amount of Nutri-Pak supplied, with VTR shows most heavily weighted. Those villages with a score of 60 or above were assigned to the "high-service" group and those with a score of 30 or below were assigned to the "low-service" group..
As in 1979, the objective was to collect data on all the preschoolers in each sample village. The same methods described earlier were used to estimate the preschool population, and the percentage of the estimated number of preschoolers actually measured ranged, by condition, from an average of 88.1 to 107.6 per cent. Any village where fewer than 80 per cent of the estimated number of preschoolers were measured was excluded from the analysis.
Data were collected using the same techniques, types of teams, and equipment described for the 1979 data collection.
FIG. 4. Gomez Classification of Preschool Children by Level of Intervention-Rural Leyte, 1979 (age 12-66 months; Harvard standards; total N = 1,518; chi-square significant at p < .001)
FIG. 5. Percentages of Mothers Giving "Acceptable" Responses to Questions, by Year-Villages in Rural Leyte Common to 1979 and 1981 Surveys (mothers of children 0-66 months old; 1979 N = 428, 1981 N = 464: chi-square significant at p < .001 for "complete meals" question only)
The responses of those interviewed were analysed and compared across conditions for the same questions used in 1979 to determine mothers' reports of number of VTR shows attended. To support the assumption that the villages were grouped by degree of intervention, a significant difference would be expected for reported frequency of VTR shows between high-service and low-service villages. That assumption was not supported by the data. There were no significant differences in recorded VTR frequency between high- and low-service villages (X² = 7 4, p > .05).
For the rest of the analysis, the high- and low-service groups were abandoned on the assumption that the field reports were not sufficiently accurate to be used. Instead, the data in 1981 were compared with those in 1979.
1979 vs. 1981 Results-Knowledge/Attitudes
For the rural villages common to both the 1979 and 1981 samples, there is a significant improvement in the proportion of mothers describing complete meals to give to their two-year-olds (fig. 5). Both the "specific Nutri-Pak information" question and the "specific BNS activities" question showed higher scores, but these differences are not significant. In both cases, however, the proportion of mothers giving acceptable responses-76.3 and 87.1 per cent respectively in 1981 -may be approaching a ceiling effect that cannot be much improved, considering that the project does not reach all mothers. In any event, these scores show that the 1979 levels of knowledge have been well maintained (fig 5).
The "Nutri-Pak for snack" question showed a decrease in proportion of mothers with acceptable answers, but the difference is not significant (fig. 5). This is not surprising, since the availability of Nutri-Pak had greatly decreased and Nutri-Pak was not available in most villages in 1981.
1979 vs. 1981 Results-Nutritional Status
Among the common rural villages, the mean nutritional status of the preschool children, as measured by percentage of standard weight for age, using the Harvard standards, improved significantly from 77.65 in 1979 to 79.21 in 1981 (T = 3.0, p < .005).
Looking at the means of the common rural villages individually, it can be seen that all but one improved. The village that did not improve decreased by only one-tenth of one percentage point (fig. 6). Disaggregating the average mean nutritional status of the common rural village children by age groups shows that all ages have higher means, but the greatest improvements were in the 24 to 35-month and 36- to 47-month age groups. Applying the Gomez classification to the common rural villages also shows significant improvement, with children moving from third- and second-degree to first-degree categories of malnutrition (fig. 7).
FIG. 6. Change in Mean Nutritional Status of Preschool Children in Nine Villages ("A"-"l") in Rural Leyte from 1979 to 1981 (ages 12-66 months; Harvard standards; sign test significant at p < .001)
FIG. 7. Gomez Classification of Preschool Children, by Year-Villages in Rural Leyte Common to 1979 and 1981 Surveys (ages 12-66 months; Harvard standards; 1979 N = 637, 1981 N = 696; chi-square significant at p < .05)
The evidence does not support a null hypothesis of no effect from the Nutri-bus intervention. Both the knowledge and attitude data and the nutritional status data within 1979 and between 1979 and 1981 support the assumption of positive impact by the Nutri-bus project. However, since the data lacked well-matched control groups and statistically random assignment of villages to the different approaches, the question remains whether or not there is an acceptable alternative explanation for the results found.
Regarding the 1979 findings, the sample selection criteria endeavoured to control for differences in access to health and nutrition facilities, health and nutrition programmes, past health and nutrition status, socio-economic status, distance from the municipal capital, and sources of income. In the 1979 - 1981 comparison of rural villages, the same villages were analysed in both years, and only those that continued to meet the criteria were included.
One possible alternative explanation for the 1979 differences could be the political priorities that were used in assigning the buses to villages. However, the government personnel who made the assignments had no reason (nor means} that we know of to systematically assign the best villages to the bus first, then the second-best, and save the worst for last. Furthermore, that would not explain the differences in knowledge, since the comparison villages, known to be the best off overall, had the lowest knowledge scores.
One possible alternative explanation of the improvement between 1979 and 1981 could be based on economic conditions, if there had been an improvement during that period. In fact, the 1979 - 1981 period was the beginning of the world-wide recession. Table 2 compares key consumer price indexes with primary and secondary sources of income for the sample families in the rural area. Conventional wisdom would tell us that, if the economy from 1979 - 1981 had any effect on the nutritional status of the sample children, it would have been a negative one.
TABLE 2. Changes in Cost of Living, Measured by Consumer Price Indexes, Compared with Prices Received by Farmers for Principal Income Crops
|Jan. 1979||Jan. 1981||Change (%)|
|Consumer price indexesa|
|fuel, light, water||244.8||418.0||+70.7|
|Prices received by farmers (pesos)b|
a. Year Book, Philippine Census and Statistics, NEDA, 1980 and
b. Bureau of Agriculture Extension
Putting the Results in Perspective
To understand the results better, it might be useful to see how these findings compare with changes in nutritional status elsewhere in the Philippines. Two surveys meant to reflect the national condition, each of one year's duration, were conducted during approximately the same period as the Nutri-bus study. While these surveys dealt with populations more representative of the nation as a whole, the Nutri-bus areas studied were rural only and in depressed areas. Also, these surveys used Philippine standards for determining percentage of standard weight for age, whereas the Nutri-bus project used the more stringent Harvard conditions. However, using percentage of change in Gomez classification should help somewhat to control for those differences. Although the comparison is not perfect, the percentages of reduction of children with third- and second degree malnutrition does seem to indicate more improvement in the Nutri-bus areas than in the nation as a whole (table 3).
It is also interesting to compare the changes in the Nutri-bus areas with changes found in feeding programmes with children of approximately the same ages and in programmes of approximately the same duration. The two feeding projects shown in table 4 provide an idea of the magnitude of improvement that can be expected when all the children in a sample are given free food. The percentage increase achieved in the Nutri-bus areas seems quite respectable in comparison with the feeding programmes, especially considering that the Nutri-bus areas received no free food and that all children in these areas were measured to determine the average improvement, even those whose mothers had not been reached by the intervention.
TABLE 3. Comparison of Nutri-bus Evaluation with Nationwide Findings in Similar Period
Change in Gomez Classification (%)
|Nationwide studiesa||+ 9.8||+ 0.8||- 0.7||- 30.7|
|Nutri-bus evaluationb||+ 7.5||+ 12.9||- 13.0||- 51.0|
a. Combined findings of two nationwide studies of percentage
of change in malnutrition in preschool children (two years,
1976-1977 and 1979-1980i, OPT data, Philippine standards,
geographic and SES mix: Seventeen Province Survey (1979 N =
61,755; 1980 N = 91,574); index Municipalities Survey (1976 N =
644,087; 1977 N = 470,507).
b. Study of preschool children in rural Leyte (two years, 1979-1981), data gathered by project team with clinical scales, Harvard standards, rural and lower SES 11979 N = 637; 1981 N = 696).
TABLE 4. Comparison of Nutritional Status of Preschoolers in Nutri-bus Area with Findings in Two Studies of Similar Duration and with Similar Age Groups but with More Intensive Intervention
|Type of Programme||Country||Authority||Age of Children (months)||Percentage of Daily Calorie Requirement in Rationa||Duration of Programme (months)||Average Increase in Percentage Points of Standard Weight for Age|
|On-site feedingb||India||Narangwal Health Research Center, 1974||0-36||66||24||3.0|
|Take-home feedingc||Philippines||Asia Research Organization 1976||24-60||25||13-18||1.0|
a. FAO requirement for 12- to 47-month-old children is 1,360
b. Underweight children identified through periodic surveys were served food supplements by project staff twice daily at village feeding centre (or, in some cases, at home). Only children receiving the food supplements were measured; same children at pre- and post-measurements; average age of children increased at post-measurement. Source: Study I, Supplementary Feeding, prepared by the Harvard Institute for International Development for the Office of Nutrition US/AID, 1981.
c. Free food given. Only children receiving the food were measured; same children at pre- and post-measurements; average age of children increased at post-measurement. Source: Ibid
d. No free food given. All children in village measured, whether or not reached by the programme; not all the same children at pre- and post measurements; average age of children unchanged at post-measurement.
The following conclusions may be drawn, using a sequence of questions designed for the evaluation of programmes during implementation (5).
1. Is the intervention performing as expected?
a. Are services being delivered to target groups? Yes, but improvements are needed. Approximately 66 per cent of the mothers in the sample villages in 1981 reported seeing VTR shows. However, this is down from 75 per cent in the VTRL treatment in common rural villages in 1979. This may indicate need for better field management and/or greater variety in shows. Reports of the use of Nutri-Pak as a snack dropped nearly 50 per cent from 1979 to 1981. Nutri-Pak was not available in most villages A better distribution system is needed.
b. Is gross outcome acceptable to management? Yes, but improvements are needed. The stated objective was to test the null hypothesis. Although the data do not meet the requirements of experimental design, the evidence favours the rejection of the null hypothesis. Since successful field projects to improve nutritional status are so few, it was decided before the evaluation that any statistically significant improvement found would be taken to indicate that a useful intervention was being pursued and efforts would be focused on increasing the impact of the intervention.
c. Where should improvements be sought? The three areas where improvements are being most vigorously sought are field management systems, videotape production efficiency, and the Nutri-Pak distribution system.
2. Is it worth continuing the project? Yes, especially if the recommended improvements can be implemented, When the findings are compared with national trends and more intensive feeding programmes, the improvement in the Nutri-bus areas seems good. While this evaluation does not prove that the Nutri-bus project was totally or even partially responsible for the improvements, the evidence favours that interpretation. Plausible alternative explanations were explored and rejected.
3. Should the programme be extended to other locations? Yes, if the recommended improvements can be implemented. The Nutri-bus project is still under development. Improvements are needed in the areas listed above, as well as in precision of evaluation.
4. Is there a causal link to improved nutrition? Although these data cannot definitely answer this question, the presumption is yes.
Improvements in nutritional knowledge and attitudes and nutritional status were found in the rural Nutri-bus areas surveyed. Feeding standards improved, as measured by the mothers' reports of what they fed children. In 1979 mothers in the VTRL villages were 55 per cent more likely to describe complete meals than those in the villages of the comparison group. Mothers in the VTRL villages were 473 per cent more likely to name Nutri-Pak as a good snack and 71 per cent more likely to give correct specific descriptions of Nutri-Pak than the BNS village mothers (fig. 2). Mothers in 1981 were 27 per cent more likely to describe complete meals (fig. 5) than those inter viewed in 1979.
As for nutritional status, in 1979 there was 25 per cent less third" and second-degree malnutrition among children of the VTRL villages than among those of the BNS villages and 29 per cent more first-degree malnutrition and normal nutritional status, using the Harvard standards for weight for age and the Gomez classification (fig. 4). Comparison of the nutritional status results of 1979 and 1981 shows a 17 per cent decrease in third- and second-degree malnutrition and a 12 per cent increase in first-degree malnutrition and normal status (fig. 7).
Evidence from the 1979 and 1981 evaluation of the Nutri-bus project favours the assumption of its effectiveness. It also reinforces the recognized need for improvements in the field management system, video-tape production efficiency, and Nutri-Pak distribution system. The Nutrition Center of the Philippines has committed itself to the continuing development and evaluation of the Nutri-bus project.
Dr. Richard M. Lockwood, Development Communications Consultants, Inc., made significant contributions to the data analysis and preparation of this report. Dr. Nevin S. Scrimshaw and Dr. William Rand, Massachusetts Institute of Technology; Dr. Marian Zeitlin, Tufts University; Dr. Joe D. Wray, Columbia University; and Dr. Barbara Underwood, National Eye Institute, National Institutes of Health, offered comments and suggestions on results and interpretation of the data. The Coca Cola Company contributed funding for this evaluation.
1. F.S. Solon, "Nutrition and Government Policy in the Philippines" (Paper presented at the Conference on Nutrition and Government Policy, Bellagio, Italy, 1975).
2. National Nutrition Council in coordination with the NEDA and other co-operating agencies, Philippine Nutrition Program, 1978-82 (National Nutrition Council, Makati, Philippines, 1977).
3. Philippine Nutrition Program: Implementing Guideline (National Nutrition Council, Makati, Philippines, 1981).
4. Nutrition Center of the Philippines, "Fact Sheet # 7'' (NCP, Manila, Philippines, 1976).
5. J.J.P Habicht, J. Mason, and H Tabatabai, "Basic Concepts for Design of Evaluation during Program Implementation," in D. Sahn, R. Lockwood, and N.S. Scrimshaw, eds., Evaluation of the Impact of Food and Nutrition Programs, (United Nations University, Tokyo, in press).
Contents - Previous - Next