OBJECTIVES: In response to low utilization of primary health services in rural areas, the Government of Punjab contracted with a local non-governmental organization (NGO) to manage the basic health units in one district. METHODS: To evaluate the performance of the contractor, health facility surveys, household surveys, and routinely collected information were used to compare the experimental district (Rahim Yar Khan, RYK) with a contiguous and equally poor district (Bahawalpur, BWP). RESULTS: The evaluation found that contracting led to more than a 50% increase in out-patient visits in RYK compared to BWP. There was also increased satisfaction of the community with health services. Technical quality of care was equally poor in both districts and contracting also had little effect on the coverage of preventive services. The latter was likely the result of the NGO not being given managerial responsibility over vaccinators and other community health workers. CONCLUSIONS: Despite methodological limitations, this study found that contracting in management achieved important goals at the same cost to the Government, implying a large increase in efficiency. Contracting in management worked reasonably well in this context and has now been significantly expanded. The approach provides a plausible means for large-scale improvements of poorly performing primary health care systems.