This study compares the performance of 13 contracting (Special Operating Agency (SOA)) districts and 10 non-contracting (non-SOA) districts
This study compares the performance of 13 contracting (Special Operating Agency (SOA)) districts and 10 non-contracting (non-SOA) districts across four provinces of Cambodia in increasing coverage of basic health services. The 4 Minimum Package of Activity (MPA) indicators used are: full immunisation of children under one, antenatal care (two or more consultations), delivery by a trained professional, and delivery in a health facility.
Suspicions about the quality of the routine HMIS data mean that all conclusions about the relative performance of contracting and non-contracting districts have to be treated with some caution. This is particularly true of the data for immunisation and antenatal care.
Placing greater emphasis on the data on deliveries, by a trained professional and in a health facility, and in the two provinces included in our study where systematic comparison of SOA and non SOA districts is possible – Kampong Cham and Prey Veng – leads to the conclusion that there is some, but very weak, evidence that SOA districts perform better than non SOA districts over the period 2009-12.
However, clear selection bias and the existence of countervailing factors such as the greater resources available to SOA districts and the various parallel initiatives, particularly the nationwide midwifery scheme, pose considerable problems of attribution of the observed improvements of the four health indicators. Overall, there is little reason to believe that the better performance of SOA districts is due to the contracting mechanisms in these districts as such.
This research is supported by the Department for International Development’s ReBUILD Programme which is led by Liverpool School of Tropical Medicine
Vong, S.; Raven, J.; Newlands, D. Understanding Contracting in Cambodia. The performance of contracting and non-contracting districts in extending primary health coverage: analysis of secondary data. (2015) 40 pp. [ReBUILD RPC Research Report 13]