Institutionalizing and sustaining social change in health systems: the case of Uganda

The key to high impact health services is institutionalizing and sustaining programme evaluation. Uganda represents a success story in the use of a specific programme evaluation method: Lot Quality Assurance Sampling (LQAS). Institutionalization is defined by two C's: competent programme evaluators and control mechanisms that effectively use evaluation data to improve health services. Sustainability means continued training and funding for the evaluation approach. Social science literature that researches institutionalization has emphasized 'stability', whereas in global health, the issue is determining how to improve the impact of services by 'changing' programmes. In Uganda, we measured the extent of the institutionalization and sustainability of evaluating programmes that produce change in nine districts sampled to represent three largely rural regions and varying levels of effective health programmes. We used the proportion of mothers with children aged 0-11 months who delivered in a health facility as the principal indicator to measure programme effectiveness. Interviews and focus groups were conducted among directors, evaluation supervisors, data collectors in the district health offices, and informant interviews conducted individually at the central government level. Seven of the nine districts demonstrated a high level of institutionalization of evaluation. The two others had only conducted one round of programme evaluation. When we control for the availability of health facilities, we find that the degree of institutionalization is moderately related to the prevalence of the delivery of a baby in a health facility. Evaluation was institutionalized at the central government level. Sustainability existed at both levels. Several measures indicate that lessons from the nine district case studies may be relevant to the 74 districts that had at least two rounds of programme evaluation. We note that there is an association between the evaluation data being used to change health services, and the four separate indicators being used to measure women's health and child survival services. We conclude that the two C's (competent evaluators and control mechanisms) have been critical for sustaining programme evaluation in Uganda.