Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Socio-economic and demographic characteristics
Table 1 summarizes the socioeconomic and demographic characteristics of the four groups of sample families. The 1985 control and intervention samples and the 1986 intervention sample were similar in mean age of children, sex ratio, means of the fathers' and mothers' education, and socio-economic scores.
TABLE 1. Demographic and socio-economic variables of the sample households
(N = 96)
(N = 111)
(N = 78)
|Intervention 1986 (N = 90)|
|Mean (SD)||%||Mean (SD)||%||Mean (SD)||%||Mean (SD)||%|
|Child's age (months)||13.4 (3.2)b||13.0 (2.0)b||14.0 (2.5)b||13.7 (2.7)b|
|Father's education (years)*||2.6 (3.8)c||2.3 (3 7)c||3.8 (3.5)d||2 5 (3.4)c|
|Mother's education (years)**||1.1 (2.2)c||0.8 (2.0)e||1.8 (2.7)f||0.5 (1.4)e|
|Landholding (acres)§||0.3 (0.8)g||2.4 (4.0)h||1.4 (3.5)h||1.4 (2.5)h|
|Major source of income|
|fishing and day labour||30||21||28||10|
Pairwise comparisons: CI85 = 1985 control vs. 1985
intervention. CI86 = 1986 control vs. 1986 intervention. C85-86 =
1985 control vs. 1986 control. I85-86 = 1985 intervention vs.
No significant differences between values in the same horizontal row with the same superior lever.
+CI85, p = .001. Cl86, p = .000001.
*Cl86, p = .05.
**C85-86 p = .01. Cl86, p = 007
§Cl85 p =.0001. C85-86, p = .01. I85-86, p = .07.
The percentage of Hindus among the 1985 control families was twice as high (45%) as in the 1985 intervention group (23%) (chi-square test, p < .0001). In 1985 the intervention families owned a larger mean amount of land than the controls, but in 1986 the difference in land ownership was not significant.
Overall, 19% of the mothers and 45% of the fathers had attended school. However, the 1986 control parents had more education than the others: 18% of the fathers and 10% of the mothers had ten or more years of school, whereas the corresponding figures were 7%9% for the fathers and 0%-2% for the mothers in the other groups. The 1986 control group also had the highest socio-economic scores and radio ownership.
Hygiene and sanitation knowledge and conditions
The percentage of mothers who correctly identified dirty and disease-causing items was significantly higher in the 1986 intervention sample than in either the baseline samples or the 1986 control sample (see FIG. 1. Mothers' hygiene knowledge: percentages identifying substances as disease-causing).
Regression analysis (table 2) indicates that mothers' knowledge of germs and disease-causing substances was determined by the intervention, mothers' and fathers' education, and membership in the 1986 control group. The structured behavioural observations of sanitation and hygiene practices had been conducted among the members of the 1986 control sample.
TABLE 2. Standardized regression coefficients of socio-economic and other variables on mothers' knowledge of germs, sanitation practices, children's contact with faeces, and dryness of play areas
|Independent variable||Knowledge of germs||Sanitation practices||Faecal contact||Dryness of play area|
|Socio-economic status||0.004 (0.01)||0.07(0.11)*||-0.02(-0.06)||0.02(0.13)**|
|Amount of land held||NE||-0.0005 (-0.04)||-0 007 (-0.03)||0,005 (0.13)**|
|Mothers' knowledge of germs||NE||0.32(0.17)*****||-0.04 (-0.14)**||0.05 (0-09)|
|Intervention 1986||2.52(0.55)*****||5.21(5.57)****||-0 43(-0.22)****||0.84(0 30)***|
NE = not in equation.
*p=.03. **p=.01. ***p=.001. ****p =.0001. *****p=.00001 .
Behaviours and conditions
No significant differences were found between the baseline samples with respect to the hygiene and sanitation variables shown in table 3. The 1986 intervention sample was higher than the 1986 control and the baseline samples in the percentages of mothers who were most attentive to their babies, checking on them continuously, and who either threw away food that fell on the gound or washed it before giving it back to the baby, and was much higher in the percentages who cleaned their hands with ash after defecating and who removed the babies' faeces from the ground completely with a scraper. It was also somewhat higher than the 1986 control sample and much higher than the baseline samples in the percentage who cleaned the baby immediately when it defecated.
TABLE 3. Sanitation and hygiene variables of households (percentages)
(N = 96)
(N = 111)
(N = 78)
(N = 90)
|Cleaning hands after defecation*|
|water and soap||1||2||14||0|
|water and ash||0||1||0||83|
|water and earth||99||97||86||17|
|Removal of faeces from groundt|
|scraping with hoe/dirt thrower||43||35||23||96|
|wiping with straw/leaves||79||87||76||6|
|covering with ash/sand||8||11||0||0|
|rubbing with foot||14||7||21||0|
|Cleaning of baby after defecation|
|as soon as mother sees it||50||43||91||97|
|sometimes not until job at hand is done||46||55||8||3|
|usually not until job is done||4||2||1||0|
|Treatment of food dropped on ground§|
|cleaned and given back to baby||21||24||31||51|
|given back to baby||67||64||55||12|
|Mother's checking on baby+|
|more than once per half hour||44||39||23||23|
|less than once per half hour||22||15||20||6|
Pairwise comparisons as for table 1.
* C85-86. p=.0007. I85-86, p=.000001. CI86, p=.000001.
t C85-86, p = .007. I85-86, p = .000001. CI86, p = .000001.
C85-86, p = .000001. I85-86, p = .000001.
§ I85-86,p= .000001. CI86, p= .000001.
+ C85-86, p = .00009. I85-86, p = .001.
The percentage of the babies' outdoor play areas that were clean was also highest in the 1986 intervention group (see FIG. 2. Households in which the child's play area was rated clean (chi-square test: difference between 1986 control and intervention samples significant at p < .001)). Similarly, a smaller percentage of that sample was reported to have been in contact with faeces than in the other groups, though the control group had a lower contact rate than the intervention group at baseline; there was a 35% change in the children's oral contact with faeces at the intervention site after about a year, compared with 7% at the control site (see FIG. 3. Children's reported oral contact with faeces during the two weeks preceding the survey (chi-square test: differences between control and intervention samples significant at p = .05 in 1985 and p = .0001 in 1986)).
The second regression in table 2 indicates that hygiene and sanitation practices were determined by the mothers' knowledge about disease-causing matter, the mothers' and fathers' level of education, socio-economic status, and the intervention.
The third regression indicates that the mothers' germ knowledge, the mothers' and fathers' education, and participation in either the 1986 control or the 1986 intervention group had significant effects on reducing babies' reported contact with faecal matter. In addition, the intervention, the family's socio economic status, and the amount of land had a significant association with the dryness of babies' play areas.
Children's morbidity and growth
The baseline surveys were administered at different timein July at the control site, which recorded a very high rate of diarrhoea (57% past two-week prevalence), and in September at the intervention site, with a rate of 20%. An attempt was made to arrive at a rough estimate of diarrhoeal prevalence in July for the intervention site, based on mothers' recall of whether the baby had diarrhoea two months previously. The recall yielded a rate of 56%, suggesting no significant difference between the control and the intervention samples in July (see FIG. 4. Prevalence of diarrhoea in children 9-18 months old. The data for the control site in 1985 and for both sites in 1986 represent mothers' reports for the two weeks preceding the survey; the 1985 survey at the intervention site was conducted about two months after that at the control site, towards the end of the diarrhoea season, and the value represents the mothers' recall for the time of the survey at the control site, two months previously.). The final survey (1986) was administered at the same time in both sites. The children's past two-week prevalence of diarrhoea decreased at the end of the intervention period in both sites, but the difference between sites was not significant.
Table 4 shows that dryness of the child's play area, and hygiene and sanitation practices were significantly associated with lower diarrhoeal morbidity. The intervention did not show a direct effect on diarrhoeal morbidity as measured.
TABLE 4. Standardized regression coefficients of variables on children's diarrhoeal morbidity
|Sanitation practices||-0.21 (-0.11)*|
|Dryness of play area||-0.90 (-0.17)**|
|Control 1986||-0.45 (-0.03)|
|Intervention 1986||-0.77 (-0.05)|
Adjusted R2=09. Significant F, p=.0001.
Figure 5 (see FIG. 5. Mean weight-for-age Z scores of children 9-18 months old, based on the WHO standard (1986 intervention sample significantly higher than all other samples at p<.05). Children below -3.5 SD: 1985 control 29%, intervention28%; 1986 control 21%, intervention, 18%) illustrates the mean WAZ scores of the children in the baseline and final surveys. In the 1986 samples, the intervention children had a significantly higher mean WAZ (p<.05) than the control children, while the difference between the 1985 samples was not significant. Multiple regression analysis indicates that diarrhoeal morbidity, socio-economic status, and the intervention significantly influenced the growth status of the sample children (table 5).
TABLE 5. Standardized regression coefficients of variables on children's growth status (weight-for-age Z scores)
|Age of child||-0.18 (-0.09)*|
|Socio-economic status||0.03 (0.15)**|
|Diarrhoea morbidity scale||-0.02 (-0.13)**|
|Respiratory infection||-0.002 (-0.05)|
|Control 1986||0.06 (0.04)|
|Intervention 1986||0.34 (0.11)*|
Adjusted R2 =.07 Significant F, p=.002.
*p=.03 **p =.001 .
Contents - Previous - Next