[Maternal factors associated with the duration of breast-feeding in peripheral areas of Guadalajara, Mexico].

The aim of the present study is to determine the duration of breast-feeding by mothers covered by the Mexican Social Security Institute and living in Tlaquepaque and Tonalá, outlying areas of the city of Guadalajara, Mexico; to identify possible maternal factors associated with early weaning; and to consider the social importance of this phenomenon. A sample of live-born infants and their mothers were selected from the study area. The children had reached at least one month of age between May 1990 and April 1991. Through multistage probabilistic sampling, random selection was made of primary health care units within the study areas, of family physicians' offices within those units, and finally, of all children meeting the study criteria in each of those offices. After visits to 166 dwellings, 141 mothers were interviewed, or 91% of the projected sample size (155). A 33-item questionnaire was prepared for the interviews. The interviewers were social workers trained to do this work, which they carried out between June and July 1991. Logistic regression models were used to estimate the relative risk (RR) associated with each variable and the probability that the infant was weaned early because of the presence or absence of certain variables. To ensure valid results, several regression models were developed with a view to selecting the one that best fit the data. The population attributable risk (PAR) was also calculated. The findings indicate that few mothers breast-feed their children in these areas, since 34.8% of the children were breast-fed for less than one month. Three maternal risk factors were statistically associated with early weaning (p < 0.05): maternal age under 20 years (RR = 3.75; CI 95%: 1.53-9.19); single marital status (RR = 2.88; CI 95%: 1.08-7.69); and social class, i.e., the mother's belonging to a "non-worker" social group (RR = 2.72; CI 95%: 1.17-6.28). The probability of an infant being breast-fed for less than one month was 0.84 when the three maternal risk factors were present, and 0.15 when they were absent. The high estimated PAR reflects the high proportion of mothers who were under the age of 20 or did not belong to the working class. The results underscore the importance of reducing the effect of these risk factors so that breast-feeding can become more widely practiced in the population studied.