Training traditional birth attendants in clean delivery does not prevent postpartum infection

Objective: To compare the maternal outcome, in terms of postpartum infectio n, of deliveries conducted by trained traditional birth attendants (TBAs) w ith those conducted by untrained birth attendants. Methods: The study took place in a rural area of Bangladesh where a local N GO (BRAC) had previously undertaken TEA training. Demographic surveillance in the study site allowed the systematic identification of pregnant women. Pregnant women were recruited continuously over a period of 18 months. Data on the delivery circumstances were collected shortly after delivery while data on postpartum morbidity were collected prospectively at 2 and 6 weeks. All women with complete records who had delivered at home with a nonformal birth attendant (800) were included in the analysis. The intervention inve stigated was TEA training in hygienic delivery comprising the 'three cleans ' (hand-washing with soap, clean cord care, clean surface). The key outcome measure was maternal postpartum genital tract infection diagnosed by a sym ptom complex of any two out of three symptoms: foul discharge, fever, lower abdominal pain. Results: Trained TBAs were significantly more likely to practice hygienic d elivery than untrained TBAs (45.0 vs. 19.3%, p < 0.0001). However, no signi ficant difference in levels of postpartum infection was found when deliveri es by trained TBAs and untrained TBAs were compared. The practice of hygien ic delivery itself also had no significant effect on postpartum infection. Logistic regression models confirmed that TEA training and hygienic deliver y had no independent effect on postpartum outcome. Other factors, such as p re-existing infection, long labour and insertion of hands into the vagina w ere found to be highly significant. Conclusions: Trained TBAs are more likely to practice hygienic delivery tha n those that are untrained. However, hygienic delivery practices do not pre vent postpartum infection in this community. Training TBAs to wash their ha nds is not an effective strategy to prevent maternal postpartum infection. More rigorous evaluation is needed, not only of TEA training programmes as a whole, but also of the effectiveness of the individual components of the training.