To achieve the health-related Millennium Development Goals, it will be necessary to improve the delivery of health services, particularly to poor people. One proposed approach to improving both coverage and quality of care is to contract with NGOs or other non-state entities to deliver health services. This paper reviews experiences with contracting in which some coherent form of evaluation was carried out, using at least before and after, or controlled evaluation designs. Ten examples of contracting with NGOs were found and from these studies, it appears that in developing countries, contracting with non-state providers to deliver primary health or nutrition services can be very effective and that improvements can be rapid. These results were found in a variety of services and settings. Six of the ten studies compared contractor performance to government provision of the same services and all six found that the contractors achieved better results. There are a number of concerns about contracting that have been raised, including scale, cost, sustainability and contract management. Scaling up appears not to be a problem. Four of the examples studied involved populations of many millions of beneficiaries. On the issue of cost, four of the studies found that NGOs performed better even when provided the same resources as the public sector. Of the nine studies with three or more years of elapsed experience, seven have been continued and expanded. Contract management was seen as a significant issue in at least three of the examples reviewed, however, it did not prevent contractors in those instances from being successful. Based on the success thus far, health services contracting appears to improve service delivery and may help achieve the MDGs. The approach should be carefully expanded in developing countries using large scale pilots initially. Future efforts at contracting should include rigorous evaluations.