|05. The Role of Traditional Birth Attendants in the Reduction of Maternal Mortality|
Evaluation of TBA training could potentially take place at several stages in the implementation process. Possible components for evaluation include: inputs (e.g. expenditure, technical assistance; programme management); outputs (e.g. numbers trained; behaviour change; % deliveries with trained TBAs; % and nature of cases referred); and outcomes (i.e. maternal morbidity and mortality or perinatal mortality). Our view is that one of the reasons for continuing debate over TBA training is the haphazard way the programmes have been evaluated. Despite high expenditure on the programmes, there are surprisingly few methodologically sound evaluations, even of programme outputs. Among the many studies documented in the literature problems with sample size, study design, control or comparison groups, and statistical analysis are extremely frequent. The impression overall is that many of these evaluations were not planned as an integral part of the programme process, but initiated as an afterthought. Even evaluation of the TBA training process has not been as frequent or as rigorous as might be expected. Studies that have been done present a mixed picture. Several studies report that TBAs practice what they have learnt during their subsequent work in the community (Lartson et al . 1987, Akpala 1994). However, adoption of improved practices is not universal (Bemara et al. 1990) and the extra confidence gained from the training experience may lead to a higher incidence of dangerous procedures and sometimes delays in referral (Eades et al . 1993). There is also evidence that training does not substantially alter the belief systems of TBAs and will therefore have little impact on practices that are rooted in these beliefs (Goodburn et al. 1995).
TBA training as a package of interventions has rarely been submitted to any kind of rigorous assessment in terms of outcomes (Smith 1996). Cost-effectiveness studies, which should take account of impact in relation to limited resources and competing priorities, are even rarer. Many authors clearly assume that the value of TBA training is self evident (Semali 1992, Amin & Khan 1989, Islam et al . 1984, Bayoumi 1976, Mathur et al . 1983).