Working with a nongovernmental organization (NGO) that promotes the development of private health insurance programs for low-income groups in the third world, Milliman consultants undertook a study of a program that our client was sponsoring in Nigeria. When the program began in 2007, the organization anticipated substantial premium subsidies. However, by early 2011 the organization was also paying significantly more to providers than its estimates used to set premium levels. In addition, many providers were claiming that the fee schedule was too low.

The Milliman team visited six hospitals that had contracts with the program, three in Lagos and three in an outlying state. Working with data collected at each hospital and the program administrator, we found that the main driver of the excess provider payouts was significantly higher payments for program members diagnosed with hypertension. The program was also paying higher than expected amounts for diabetes and tuberculosis. A review of utilization rates for service provided to these members suggested that the current payment methodology was encouraging providers to provide unnecessary services.

While all the hospitals were profitable overall, with current practice patterns most were likely losing money on our client’s patients. Self-paying patients were paying significantly more than what our client was paying, but our client’s fee schedule was similar to other local health plans and significantly higher than what the government paid for members of the civil service health plan. In addition, the hospitals were not operating efficiently, and so we developed a model to estimate a hospital’s costs when it improved efficiency. The model illustrated that if the hospitals achieved moderate improvements in efficiency, they would be able to make reasonable profits at the current fees paid by our client’s program.

Based on our analyses and discussions with stakeholders, we concluded that there was little reason for our client to significantly increase the fees it paid its providers. However, to prevent possible disruption to its provider network, we recommended that our client consider a small increase in fees and changes to its payment methodology that would include better incentives for providers to improve efficiency.

The Nigeria project reinforced a number of the difficulties that we encounter when developing effective microinsurance programs, for example:

• Data is limited and of poor quality. Although it was supposedly based on ICD-9 codes, the hospitals’ data often deviated from ICD standards, and it was difficult to audit the accuracy of most of the data we were given.
• Designing an affordable plan, even with limited benefits, is challenging for low-income populations. Premiums for the Nigeria program are significantly subsidized, and it is difficult to foresee a time in the near future when subsidies will not be required.

At the same time, it is interesting to reflect that this program was also dealing with many of the challenges faced by insurance plans in the developed world—for example, the pros and cons of capitation versus fee for service, determining the provider incentives to encourage efficiency, and achieving a balance between provider income expectations and affordable premium rates.

Besides the author of this report, Sudhanshu Bansal, Stefan Becker, and Rob Parke participated in this project.