This was a cross-sectional study conducted in Klang, Malaysia involving 682 mother-infant pairs with infants up to six months attending the Maternal and Child Health (MCH) section of the government health clinics over four months in 2006. An analysis of bed-sharing practices has already been published . In Klang, there are eight government health clinics with their subsidiary community clinics providing health needs for the community. Two of the health clinics including their subsidiary community clinics were chosen for the study by random selection where one health clinic (Bukit Kuda Health Clinic) serves an urban population while the other (Kapar Health Clinic) serves a rural population.
The sampling method used was universal sampling. All eligible mother-infant pairs who attended the clinics during the study period were included in the study. Data collection was carried out according to a schedule that was prepared. Data was collected from the two identified areas in the study at alternate weeks with designated days for each clinic. The specific days selected for data collection for each clinic coincided with the days when immunization was scheduled for the infants.
The inclusion criteria in the study included all mother-infant pairs visiting Bukit Kuda and Kapar Health Clinic including the subsidiary community clinics between 19 June and 19 October 2006. Mother-infant pairs were excluded from the study if mothers were less than 18 years old, have a child with congenital malformation, infants more than six months old, infant taken care of by caregivers during night time and infant not accompanied by mother at the clinic.
After consent from eligible mothers, a face-to-face interview using a pre-tested pre-coded structured questionnaire was conducted by the author in the clinics at respective days during the study period. The questionnaire was pre-tested among 35 mother-infant pairs from another health clinic in the same district. Data regarding maternal socio-demographic variables (area of residence, ethnicity, age, body mass index, marital status, education level, occupation, monthly household income and smoking status), paternal socio-demographic variables (ethnicity, age, education level, occupation, smoking status), biomedical variables (parity, number of antenatal visits, mode of delivery, infant gender, infant gestational age, infant birth weight, breastfeeding class attendance), household variables (number of people staying in the house, number of rooms in the house, number of children under 12 years staying in the house and occupancy rate) and psychosocial variables (husbands' support toward breastfeeding, bed-sharing practice) were collected. The interview was conducted in 'Bahasa Malaysia' as it was the national language and widely used by the respondents. For mothers of Chinese ethnicity not well versed in 'Bahasa Malaysia', the interview was conducted in Mandarin language. In order to ensure no eligible mother-infant pairs were missed from the interview, the author attended the clinics at least fifteen minutes before the immunization session started. Liaison with the nurses at the clinics was also made to ensure that all eligible mother-infant pairs were interviewed. Socio-demographic data were also collected from all eligible mother-infant pairs who refused to participate in the study. A yellow sticker was placed on the right upper corner of the infants' immunization card after the interview to prevent interviewing the same mother-infant pair during the entire four months of data collection.
In this study, exclusive breastfeeding was defined as the infant having received only breast milk from the mother (either directly from the breast or expressed) and no other liquids or solids with the exception of drops or syrups consisting of vitamins, mineral supplements, or medicines over the one month period prior to the interview. A recall over a one month period was used in this study because this duration coincided with the immunization days where infants were called in every month for the first six months of life.
Bed-sharing was defined as an infant sharing a bed with mother within arm's reach of the mother. A bed was defined as either a sleeping mattress placed on a bed frame or a sleeping mattress placed on the floor. Mothers were defined as a 'bed-sharer' if they shared a bed with their infant for three or more times a week for all or part of the night.
Occupation was defined as the working status of the mother. Working encompasses any form of employment with contribution to income tax.
All data in the questionnaire were coded and entered into Statistical Package for the Social Sciences (SPSS) version 11.0. The dependent variable was exclusive breastfeeding. Univariate associations between various factors and exclusive breastfeeding adjusting for infant age were identified and multiple logistic regression was used to construct the model to examine the independent association of various factors to exclusive breastfeeding while simultaneously controlling for potential confounders. The statistical test used was binary logistic regression and the level of significance was set at p < 0.05. Odds ratio (OR) and 95% Confidence Interval (CI) were calculated as a measurement of association between an independent variable and the outcome. All variables associated with exclusive breastfeeding in the univariate analyses were included in the initial multivariate model. Variables were excluded from the final logistic model if they were not associated with exclusive breastfeeding and their removal from the model did not materially affect the association of other variables in the model based on the algorithm proposed by Hosmer and Lemeshow .
The project received ethical approval from the University Malaya Medical Faculty Ethics Committee and Medical Research Committee Ministry of Health Malaysia. Signed informed consent was obtained from all the participants. Confidentiality of the data and the privacy of mothers were respected at all times.