Hygiene and health-seeking behaviours of households as predictors of nutritional insecurity among preschool children in urban slums in Ethiopia - the case of Addis Ababa
Gugsa Abate, Wambui Kogi-Makau, Nelson M Muroki
Applied Nutrition Programme, University of Nairobi, Kenya

S A J Clin Nutr 2001 May Vol 14 No 2 pp 56-61

Abstract
The objective of the study was to establish hygiene and health-seeking practices most likely to be predictors of nutritional insecurity among children living in slums. A cross-sectional study was conducted from March to May 1997 comparing 192 households with and 192 without malnourished children. All the households with children in the 3 - 36-month age group were identified. Using underweight (weight-for-age) as an indicator of nutritional insecurity, the households were classified into two groups, namely nutritionally secure and insecure households. Subsequently, sampling frames for each set of households were established and used to select the study households randomly. Four slums in Addis Ababa, Ethiopia, constituted the study sites. The results indicated that there was not a significant difference between secure and insecure households with regard to prevalence of immunisation and dietary (food withholding) habits during episodes of diarrhoea. After adjusting (by means of logistical regression) for covariates, six household behaviours were established as having the power to predict exposure to childhood nutritional insecurity in urban slums of Ethiopia. The presence of children's faeces inside the house, failure to have diarrhoea treated at a health facility, prolonged storage of cooked foods (beyond 24 hours), feeding children with unwashed hands, and poor handling of drinking water and foods are risk factors that can predict nutritional insecurity. Advice with a view to achieving sustainable behaviour change in households, namely good personal and household hygiene practices and increased utilisation of health facilities is recommended as being essential in addressing challenges to nutritional insecurity and in optimising the success of public health programmes.

The majority (79%) of the residents of Addis Ababa live in low-grade and congested slums.1 Characterised by high population density, poor housing conditions, lack of proper drainage and sanitation inter alia, urban slums are unique in that they constitute an environment unfit for human habitation.2 Each of these environmental conditions may, in some way, bear negatively on the general wellbeing of children^3-5 as such an environment entails risk of infection and parasitic infestation,1 conditions known to contribute to increased nutritional risk.^4-8

Studies addressing nutritional problems have established that malnutrition is one of the major health-related problems in urban slums.^9-11 For example, a study conducted in a Nairobi slum12 found that 86.2% of the preschool children were stunted. Hofvander and Eksmyer,¹³ in a study conducted in rural and urban slums in Ethiopia, found that 3% of the children had severe protein-energy malnutrition (PEM).

It is fully recognised that adequate household hygiene behaviours can play a critical role in minimising the frequency of infectious diseases, and can possibly reduce nutritional risk in children.^4,14,15 The importance of examining the health-seeking behaviours of households, which include dietary skills and utilisation of health services for the treatment of illnesses, has recently gained increased attention; the implication being that understanding the role of such behaviours can facilitate translation of food and health care resources into improved child nutrition.^14,15 Hence, investigation of these two aspects of family behaviour (hygiene and health-seeking behaviour) in the context of a child's nutritional status and in an unhealthy urban setting in Ethiopia is paramount as little information exists on this subject.

This study, therefore, aimed to fill gaps in knowledge by identifying the hygiene and health-seeking practices likely to aggravate the prevalence of childhood nutritional insecurity in the urban slums in Ethiopia. It is envisaged that ultimately the findings will be used to formulate strategies that aim to improve and protect nutritional security by enhancing child care practices.

Materials and methods
Seven slum villages (kebeles) that had previously been identified by the World Bank as typical slums and the most congested parts of Addis Ababa¹⁶ constituted the pool of slums from which this study selected. As there was an ongoing World Bank-supported programme in three of the slums, the study opted to work in the four non-programme slums. The three kebeles were excluded from the study on the basis that the health-seeking and hygiene behaviours of households in these kebeles may have been modified by the existing programme, thereby biasing the results of this study.

A comparative cross-sectional study was carried out from March to May 1997 in the four non-programme slum kebeles. All four study kebeles are located in woreda 3 (district), and are named kebele 31, kebele 34, kebele 44 and kebele 45. According to the national 1994 census,¹⁷ the population of these kebeles is estimated to be 7 529, 4 879, 9 011 and 3 293 for kebeles 31, 34, 44 and 45 respectively. The population composition of the study kebeles is heterogeneous, comprising various ethnic groups with different languages and cultural backgrounds.

Each household in each kebele was initially screened by means of a house-to- house preliminary survey to determine whether children between the ages of 6 and 36 months were residing in the home at the time. Six hundred and thirty-two eligible households (270 from kebele 31, 103 from kebele 34, 184 from kebele 44, and 75 from kebele 45) were identified, and a total of 758 children from the above age bracket were registered. Thereafter, age and sex as well as weight and height measurements of children were recorded to determine their nutritional status. The anthropometric data were collected based on the methods described by the United Nations.18 Two weight measurements were obtained using a Salter weighing scale and the average was calculated and recorded to the nearest 0.1 kg. The children were weighed with minimal clothing and without shoes. Supine length for children under the age of 24 months and height for those above that age were obtained using length/height wooden measuring boards, and the average of the two measurements was recorded to the nearest 0.1 cm as the child's height.

The anthropometric data were translated into nutritional indices using the Epinut module of the Epi-Info computer package. Z-scores were calculated in terms of weight-for-age (WFA), height-for-age and weight-for-height using the National Centre for Health Statistics (NCHS) reference figures.¹⁹ The WFA index was used to classify the children into the categories malnourished and well nourished. Households where the child's WFA index was below -2 Z-score on the NCHS reference were put into the category of nutritionally insecure households, while those whose index was above a -2 Z-score on the same reference were classified as nutritionally secure households.

A register of each set of households was developed and used as the sampling frame. The required samples were drawn from among the sampling frame of each group using a random sampling method. A total of 384 households (192 for each group) was selected. The statistical formula recommended for comparative studies²⁰ was used to calculate the desired sample size for each group of households. If more than one child resided either in a malnourished or well- nourished household, then the youngest child was selected for the study. Information pertaining to immunisation, presence of diarrhoea in the index child, environmental sanitation, household hygiene practices and type of treatment and feeding practices during diarrhoea were collected from sampled households. This was done by interviewing mothers or the substitute child minders using a pre-planned structured questionnaire as the data collection tool.

Data were collected with the assistance of two fieldworkers who had completed grade twelve level of education. The two fieldworkers had also participated in other surveys and as such were familiar with interviewing and anthropometric measuring techniques. Nevertheless, further training in anthropometry and interviewing techniques was done. Pre-testing of the questionnaire was undertaken and some modifications were made before it was applied.

Data validity and reliability were achieved through close supervision of enumerators by the principal investigator. Vaccination records were used to ascertain ages of the children. However, mothers in the survey areas were easily able to recall the ages of their children below 2 years of age. When reliable documentary evidence was not available and when there was a problem with age recall, interviewers used a local event calendar to determine the month and year of a child's birth. At the end of each day, the completed questionnaires were checked to ascertain that all questions had been answered correctly and consistently. A 2 kg iron bar was used regularly to check scale accuracy and ensure that measurements were correct. Finally, at regular intervals the investigator and field assistant compared the measurement values from the same child to ascertain that measuring techniques were similar and to reduce inter- observer error.

Comparison of the various parameters, namely sanitation, hygiene and health- seeking practices in the two groups of households (malnourished and well nourished), was done by administering chi-square at P-values of less than 0.05 level of significance. Ranking of risk factors of malnutrition was done using multiple logistical regression (step-wise multivariate analysis).

Results
Of the 758 eligible children for whom anthropometric data were obtained, 51.2% were male and 48.8% female. Slightly over one-third (34.7%) were underweight, nearly half (48.4%) were stunted, and a small percentage (3.4%) were wasted.

Health seeking practices of the sample households
Table I shows the distribution of households by prevalence of immunisation, types of diarrhoea treatment and foods withheld during diarrhoea episodes. As shown in the Table, the proportion of children in the malnourished group who had been fully immunised for age (80.2%) was not significantly different (P < 0.05) from that of the well-nourished group of children (77.6%).

Significantly more children in malnourished households (13.5%) had suffered diarrhoea than in well-nourished households (4.2%) (P < 0.01). The proportion of households that took children to hospitals/clinics during the bout of diarrhoea was significantly higher in the well-nourished households (76.0%) than in the households of malnourished children (58.9%) (P < 0.05). On the other hand, there was no significant difference in the prevalence of home treatment of diarrhoea between the two types of households.

The proportion of mothers in the malnourished households who reported that they withheld food when their children had diarrhoea (38.5%) was not significantly different from that observed in well-nourished households (40.1%). The data show that fruit and/or vegetables, cow's milk and injera were the three foods most commonly withheld. With the exception of porridge and/or potato, more mothers in the well-nourished households withheld the foods listed in Table I than mothers in malnourished households. However, porridge and/or potato were withheld in relatively more malnourished than well-nourished households.
Environmental sanitation, personal and household hygiene practices of the sample households (Table II)
A similar proportion and high number of households in both study groups (91.4% in malnourished and 89.2% in well-nourished households) had no drainage for dirty water. A significantly higher (P < 0.05) proportion of malnourished households (79%) had stagnant and dirty water in their compounds compared with well-nourished households (68.2%). Human faeces were observed within about 10 m of most households in both groups (90.1% in malnourished and 85.4% in well- nourished households). In the interior of the houses, human faeces were observed in a significantly higher proportion of malnourished households (78.6%) than well-nourished households (28.6%) (P < 0.0001). There were also animal (dog) faeces within about 10 m of almost three-quarters (73.4%) of the malnourished households, and a lower percentage (64.6%) of well-nourished households.

A significantly higher proportion of well-nourished households (84.3%) than malnourished households (59.9%) stored children's food for 0 - 24 hours after cooking (P < 0.0001). However, the proportion of households that stored food for more than 24 hours was significantly higher in the malnourished group (22.9%) than in the well-nourished group (8.3%) (P < 0.001).

A significantly higher proportion of malnourished households (22.3%) did not cover their food compared with well-nourished households (5.2%) (P < 0.0001). A significant proportion of well-nourished households (46.9%) practised hand washing with soap before feeding children compared with malnourished households (29.7%) (P < 0.001).

Multivariate logistical regression analysis was performed to examine the effect of each variable further while controlling for the confounding effects of others, and adjusted risk odds ratios were estimated. Health facility-based management of children with diarrhoea, presence of stagnant water in the compound and child waste inside the house, storing children's food for more than 24 hours, washing hands with soap, and food and water handling were the variables selected for multivariate analysis. These were selected due to their apparent statistical significance in the bivariate analysis.

The variables health facility-based management, presence of children's waste, prolonged storage of food, hand washing, and food and water handling practices remained significant, as illustrated in Table III. Stagnant water in the compound was not significantly related to the outcome variable, although in the earlier cross tabulations there was a suggestion of a significant effect. The risk ratio indicated that the children who did not receive health facility-based treatment for diarrhoea were about two times more likely to be exposed to malnutrition than children who did (P < 0.05). Similarly, the likelihood of malnutrition among children of households where there was child waste inside the house was about seven and a half times greater than among those whose home environment was free from human faeces (P < 0.0001). The risk of childhood malnutrition in households that had stored children's food for more than 24 hours was about three times greater than the risk among households that did not follow this practice (P < 0.05). Likewise, the practice of feeding children with unwashed hands was two and a half times more likely to expose children to malnutrition than feeding with washed hands (P < 0.01). The risk of malnutrition was three times as great among children in households that stored drinking water in uncovered receptacles than it was in households using covered water containers (P < 0.01). Finally, the odds of malnutrition among children from households where food was not covered during storage was approximately three and a half times greater than for children from households that covered food (P < 0.01).

Discussion
High vaccination coverage against childhood diseases has been reported to be a safeguard for better nutrition and health.^21,22 The results of this study, however, showed that immunisation status, which was found to be similar in the two groups of children, did not show an association with the children's nutritional status. This confirms the importance of other factors in the causation of malnutrition and suggests that while immunisation may be a necessary condition, on its own it cannot provide adequate protection against nutritional insecurity. This is not surprising considering the fact that the aetiology of malnutrition is multifaceted.^15,23

Withholding of food during illness is considered one of the factors that brings about malnutrition as it reduces food intake of children at a time when they have a higher nutrient requirement to cope with illness.^24-27 The results of this study, however, did not show an association between food withholding and malnutrition.

The six variables found to be significant risk factors for malnutrition in this study are: failure to have children with diarrhoea treated at health institutions, child waste inside the house, prolonged storage of cooked food, feeding with unwashed hands, and storage of food and water in uncovered receptacles. The observation that a high proportion of the well-nourished households took their children to hospitals or clinics for treatment of diarrhoea suggests that management of diarrhoea in health institutions is more effective in terms of maintaining adequate nutritional status than household management of diarrhoea. The present study, therefore, recognises the general consensus that the role of health facilities in determining children's nutrition merits special attention. This may be due to the fact that the type of diarrhoea management offered in health institutions may have a protective effect (probably, as reported by Tesfaye,28 the use of antibiotics offers protection against other opportunistic infections). Also, it could be that the benefit is derived as a result of health information provided in such institutions. Poor hygiene practices and unsanitary household conditions are associated with a high prevalence of infections,14,15 which in turn are associated with nutritional insecurity.^4-8 These conditions can result in contamination of food or fluids. It has been suggested that optimal child care practices must include protecting children's food from contamination in order to reduce the risk of infection.^29,30

As is evident from the results of this study, the practice of feeding children with unwashed hands is a risk factor that exposes children to malnutrition. This could be explained by the fact that the care giver's hands might be a major route of transmission of potential pathogens capable of causing infections such as diarrhoea, thus contributing to the synergism between nutritional insecurity and disease which has already been substantiated.^4-8

The finding that prolonged storage of cooked food is a risk factor in childhood malnutrition is not surprising since such a practice is known to impact negatively on child survival.^24,30 This is due to microbial development that occurs during prolonged storage of food at temperatures favourable to that process.^29,31-34 Usually, food is prepared once, in the morning, and then used for subsequent feeding for the whole day. The left-over food at each meal is kept to be re-fed to the child, usually without re-heating and without taking precautions to prevent contamination. Considering the unhygienic domestic environment of most households in the malnourished group, such food can be subject to microbial invasion^,29,31-34 thereby sustaining the synergistic cycle of infection and malnutrition.

The observation that malnutrition was more common in households that had stored food and drinking water in uncovered receptables is not unexpected, because such practice is highly likely to result in contamination that exacerbates the chances of negative nutritional consequences.^4-8

The results of this study also show the significance of child waste inside the house in determining children's nutritional status. The presence of faeces inside the house, as observed in a significantly higher proportion of the malnourished households in this study, leads to acceptance of this factor as a contributor to food contamination.²⁹ This finding is not unexpected since such a situation involved in the malnutrition-infection vicious cycle,^15,30 confirming the importance of poor domestic hygiene in the aetiology of malnutrition. It is also probable that the poor household hygiene practices observed in the malnourished households are partly responsible for the significantly higher prevalence of diarrhoea among the malnourished children. In view of the finding that the well-nourished households had better sanitation and personal hygiene practices, it seems that health education programmes aimed at encouraging good hygiene may be helpful in enhancing nutritional security in children.

Conclusion
This study concluded that low utilisation of health facilities, unhygienic household conditions, inadequate personal hygiene, and poor handling of drinking water and foods constitute a threat to children's nutritional security in urban slums in Ethiopia. Hence, as long as households fail to appreciate the link between these risk factors and malnutrition in children, the overall success of public health programmes may remain suboptimal.

The study, therefore, supports the need for a continued public education campaign among poor urban communities. Such campaigns should strongly emphasise the significance of hand washing before feeding young children, keeping drinking water/foods covered, keeping the house free from faecal material and seeking proper treatment for diarrhoea and other preventable diseases. There is a need to advise slum communities to avoid the practice of storing children's foods for more than 24 hours. There is also a need to establish how long foods can be stored safely. Further research should be undertaken in the slum settings of Ethiopia to investigate the level of contamination of complementary foods, as well as the incidence of infections and rate of malnutrition with increased duration of storage after cooking.

This study was conducted under the aegis of the University of Nairobi (UN), Kenya. The authors acknowledge both the United Nations University (UNU) and the UN for financing this research, the interviewers Ato Seid Endris and Woizero Brehane Gabissa for their excellent field work, and the mothers who participated in this study. The authors also express their sincere appreciation to those who provided assistance while analysing the data at the Unit of Applied Nutrition at the UN.

References

1. Harpham T, Lusty T, Vaughan P. In the Shadow of the City: Community Health and the Urban Poor. Oxford: Oxford University Press, 1988: 24-31.
2. Manciaux M. Urban slum and squatter settlements: implications for child health. In: WHO/ISPCAN. Pre-congress Workshop on Child Abuse and Urban Slum Environments, 14 - 15 September 1984. Geneva: WHO, 1984: 20-23.
3. Getaneh T, Assefa A, Tadesse Z. Protein-energy malnutrition in urban children: Prevalence and determinants. Ethiop Med J 1998; 36: 153-166.
4. Gopaldas T, Patel P, Bakshi M. Selected socio-economic, environmental, maternal and child factors associated with the nutritional status of infants and toddlers. Food Nutr Bull 1988; 10(4): 29-34.
5. Melville B, Williams M, Francis V, Lawrence O, Collins L. Determinants of childhood malnutrition in Jamaica. Food Nutr Bull 1988; 10(1): 43-48.
6. Vella V, Tomkins A, Borghesi A, Migliori GB, Adriko BC, Crevatin E. Determinants of child nutrition and mortality in North West. Bull World Health Organ 1992; 70: 637-643.
7. Rowland MGM, Cole TJ, Whitehead RG. A quantitative study into the role of infection in determining nutritional status in Gambian village children. Br J Nutr 1977; 37: 441-450.
8. Rowland MGM, Rowland GJG, Cole TJ. Impact of infection on the growth of children from 0 - 2 years in an urban West African community. Am J Clin Nutr 1988; 47: 134- 138.
9. Khanjanasthiti P, Wray JD. Early protein calorie malnutrition in slum areas of Bangkok Municipality, 1970 - 1971. In: Harpham T, Lusty T, Vaughan P, eds. In the Shadow of the City: Community Health and the Urban Poor. Oxford: Oxford University Press, 1974: 50-67.
10. Jha SS. Urban health in underdeveloped countries with special reference to women and children. In: Harpham T, Lusty T, Vaughan P, eds. In the Shadow of the City: Community Health and the Urban Poor. Oxford: Oxford University Press, 1985: 44.
11. Basta SS. Nutrition and health in low income urban areas of the third world. In: Harpham T, Lusty T, Vaughan P, eds. In the Shadow of the City: Community Health and the Urban Poor. Oxford: Oxford University Press, 1977: 46- 47.
12. Waihenya EW, Kogi-Makau W, Muita JW. Maternal nutritional knowledge and the nutritional status of pre-school children in a Nairobi slum. East Afr Med J 1996; 73: 419-423.
13. Hofvander Y, Eksmyr R. Anthropometry of children in a typical rural district and an urban slum in Ethiopia: A cross-sectional survey of 1 093 children. Courrier 1970; xxi(1): 1-4.
14. Engle PL. Care and child nutrition. Theme paper for the International Conference on Nutrition (ICN). New York: UNICEF, 1992: 5-8, 13-47.
15. Latham MC. Human Nutrition in the Developing World. Rome: Food and Agricultural Organisation, 1997: 15-36.
16. Jembere T. A case study of the health component in kebele 41, ÔKefitegna' 3. In: Harpham T, Lusty T, Vaughan P, eds. In the Shadow of the City: Community Health and the Urban Poor. Oxford: Oxford University Press, 1985: 153-162.
17. Central Statistical Authority. The 1994 Population and Housing Census of Ethiopia: Results for Addis Ababa, Ethiopia. Addis Ababa: Central Statistical Authority, 1994; 1: 18-19.
18. United Nations. How to weigh and measure children Ñ Assessing the nutritional status of young children in household surveys. New York: United Nations, 1986: 8-11, 29-41, 60-66.
19. Hamil PV, Drizd TA, Johnson CL, Reed RB, Roche AP, Moore WM. Physical growth: national centre for health statistics percentiles. Am J Clin Nutr 1979; 32: 607-629.
20. Fisher AA, Laing JE, Townsend JW. Hand Book for Family Planning Operations Research Design. 2nd ed. New York: The Population Council, 1991: 40-46.
21. Devi PY, Geervani P. Determinants of nutrition status of rural pre-school children in Andhra Pradesh, India. Food Nutr Bull 1994; 15: 335-342.
22. Veteri F. Nutrition-related health consequences of urbanization. Food Nutr Bull 1987; 9(4): 33-49.
23. UNICEF. Strategies of Improving Nutrition of Children and Women in Developing Countries. New York: UNICEF, 1990; 1-12.
24. Cameron M, Hofvander Y. Manual on Feeding Infants and Young Children. Oxford: Oxford University Press, 1983: 33-37. 115-116.
25. Mitzner K, Scrimshaw N, Morgan R, eds. Improving the Nutritional Status of Children During the Weaning Period. Cambridge: International Food and Nutrition Program, 1984: 34, 81, 151.
26. Necla C, Buyukgebiz B, Buyukgebiz A, Namik C. Nutrition and diarrhoeal diseases. Food Nutr Bull 1988; 10(1): 52-54.
27. Martinez H, Tomkins AM. Nutritional management of diarrhoea. Food Nutr Bull 1995; 16: 349-355.
28. Tesfaye F. Management of childhood illnesses at seven health stations in Arsi zone. Ethiop J Health Dev 1998; 12(2): 81-86.
29. Wolde Tenssay Z, Mengistu A. Indigenous weaning foods: Hygiene and diarrhoeal diseases in rural Ethiopian setting, Jimma zone. Ethiop J Health Dev 1997; 11(2): 49-56.
30. Brown KH, Creed-Kanashiro H, Dewey KG. Optimal complementary feeding practices to prevent childhood malnutrition in developing countries. Food Nutr Bull 1995; 16: 320-336.
31. Erku WA, Ashenafi M. Microbiological load and microflora of weaning foods obtained from pediatric outpatients in Addis Ababa. Ethiop J Health Dev 1998; 12: 141-147.
32. Black RE, Lopez de Romana G, Brown KH, Grados BO, Kanashiro HC. Incidence and aetiology of infantile diarrhoea and major routes of transmission in Huascar, Peru. Am J Epidemiol 1989; 129: 785-799.
33. Black RE, Brown RH, Becker S, Abdul Alim ARM, Merson MH. Contamination of weaning foods and transmission of entrotoxigenic Escherichia coli diarrhoea in children in rural Bangladesh. Trans R Soc Trop Med Hyg 1982; 76: 259-264.
34. Ashenafi M. Growth potential of some food-borne pathogens in various traditional Ethiopian sauces. Ethiop J Health Dev 1996; 10: 41-45.

Last updated: 17-Feb-2004