Stephanie Fisher (corresponding author)
Centre for Global Health, 3–4 Foster Place, Trinity College, Dublin 2, Ireland
Babar Kabir
BRAC Centre (8th floor), 75 Mahakhali, Dhaka 1212, Bangladesh
Edward Lahiff
Centre for Global Health, 3–4 Foster Place, Trinity College, Dublin 2, Ireland
Malcolm MacLachlan
Centre for Global Health, 3–4 Foster Place, Trinity College, Dublin 2, Ireland
Attitudes, Diarrhoea, Hygiene, Knowledge, Practice
Mymensingh District
Knowledge Management
Water management
The study area The study was conducted in three rural villages, selected using purposive sampling, in the Mymensingh District of the Dhaka Division of Bangladesh. The villages, Kuchundhora, Boyalmara and Makiar Knda, which were within 7 km of each other, were chosen because of their geographical proximity, and their similarity in terms of accessibility, socioeconomic status, employment, education, and use of tubewell water. They are however split across two sub-districts; Kuchundhora and Boyalmara are situated within the Haluaghat upazila (sub-district) and Makiar Kndais situated within the Dhobaura upazila. The Haluaghat upazila is one of the 150 upazilas in which the BRAC WASH programme operates, but the Dhobaura upazila is not. Accordingly, the villages of Kuchundhora and Boyalmara are subject to the BRAC WASH intervention, and have been for precisely the same length of time, while the village of Makiar Knda has not received WASH equipment or education from BRAC or indeed any other NGO. Data collection and analysis By going door-to-door within the three study villages, every caregiver of a child under five years old (who was in most cases the mother, and in every case a woman) was invited to answer a pre-piloted 35-question survey conducted in the Bangla language by a BRAC fieldworker from the area. With the exception of one mother from Kuchundhora who was sick and therefore unable to partake, and one mother from Makiar Knda who was deemed insane and therefore excluded from the study, every caregiver of an under-five took part; a total of 107 women, 80 from the BRAC WASH villages and 27 from Makiar Knda. The questionnaire explored the caregiver’s knowledge of the causes of diarrhoea, her knowledge, attitude and practices of household water management, her water, sanitation and hygiene practices, and her child’s burden of diarrhoeal disease during the month prior to data collection. Wherever possible, the answers given concerning her water source and latrine type and cleanliness were verified by observation. Diarrhoea was defined according to the WHO definition, as three or more loose or watery stools within a 24-hour period. The impact of the BRAC WASH programme on the variables covered in the survey was determined using Fisher’s exact test (for qualitative answers), the Student’s t-test (for quantitative parametric answers) and the Mann–Whitney test (for quantitative non-parametric answers) to explore which answers were associated with the presence or absence of the intervention. A comparison between the two BRAC WASH villages was then performed in order to verify that any differences found to be significant by the first analysis were more likely to be due to the presence/absence of the intervention rather than some intrinsic difference between the three villages being studied. The same tests that were used to examine the impact of the intervention were then used to examine associations between the survey answers and the respondent’s child having suffered from diarrhoea within the previous month. Having found the factors associated with incidence of diarrhoeal disease among the under-fives, backward stepwise multiple logistic regression was used to eliminate those variables which did not play a significant role in predicting whether or not the child had suffered from diarrhoea, leaving only those which had a significant impact on the incidence of diarrhoea. All statistical calculations were performed using the computer programme SPSS Statistics 17.0. In addition, in each of the three villages, a focus group discussion was conducted among eight of the questionnaire respondents, selected using convenience sampling by the BRAC WASH fieldworkers. The focus group discussions were carried out in order to explore their answers, beliefs and attitudes in a greater depth than enabled by questionnaire. The discussion, facilitated by a BRAC fieldworker or interpreter with a note-taker present, was recorded using a digital voice recorder. Following the focus group, the recording was transcribed and translated by the interpreter present at the discussion. The three translated transcripts were analysed by identifying emerging themes and patterns in order to support, expand upon or dispute the results already obtained from the questionnaire. Free and informed consent of the participants was obtained for all aspects of the methodology, and the study protocol was approved by the Faculty of Health Sciences Ethics Committee, Trinity College, Dublin, on 2 April 2009.
Journal of Water and Health 9 9 9 909.19 9 9 92011
Journal