New Mexican Health-care Program Successful At Reducing Crippling Health Care Costs

The study was designed and led by Gary King, David Florence Professor of Government and director of the Institute for Quantitative Social Science at Harvard. The results are published in the current issue of The Lancet.

"The success of Seguro Popular in reducing catastrophic health expenditures is remarkable," says King, "not least because governmental money spent on the poor in many countries rarely reaches the intended recipients."

King's study of about 500,000 people is the largest-ever randomized health policy experiment. It features innovative research designs and statistical methods he and his colleagues developed that increase what we learn from an evaluation while simultaneously saving a great deal of money. The design includes several failsafe components that preserve the experimental randomization even if politics or other problems intervene, including those which have ruined most previous large scale public policy evaluations. The approach is now being implemented or considered for evaluations of many other public policy programs around the world.

Passed in 2003, Seguro Popular was developed to provide health care to 50 million Mexicans who otherwise lack coverage. Voluntary enrollment in the program, at no cost to the poor, provides access to health clinics, drugs, regular and preventative medical care, and the money to pay for it all. The program's primary goal is the reduction of catastrophic health expenses, those exceeding one-third of a household's yearly disposable income.

About a half a million people in 118,569 households were included in this study, which was conducted over 10 months. In the treatment clusters, 44 percent of households reported participating in the program, compared to 7.3 percent in control communities, which was approximately as expected. Among participating households, those suffering catastrophic health expenses were reduced by almost 60 percent, contributing to a 30 percent reduction in catastrophic health expenses across treatment communities.

The evaluation also highlighted areas in which the program was ineffective. Contrary to prior non-randomized studies, the researchers found no increase in utilization of health services, although longer-term research may show an increase. Health outcomes will also take longer to show an effect.

Before the program was instituted, 174 communities were paired up based on having similar background variables, such as the health of the community, size, and the number of schools. Then one community within each pair was randomly chosen to receive treatment: Families were encouraged to enroll in Seguro Popular, health facilities were built or upgraded, and medical personnel, drugs, and other supplies were provided. In the other community within each pair, no changes were made.

"One advantage of this design is that if one of the communities was to drop out of the study, due to interventions by politicians or for other reasons, the paired community would be removed as well, and the balance between the treated and control groups would not be affected. In contrast, classical randomized experiments are destroyed when even one community is lost. The matched pair design also decreased the margin of error to as little as one-sixth of what it would be with traditional experimental methods," says King. "That's the equivalent of collecting many more respondents, or randomizing throughout many more communities, for the same cost."

Seguro Popular in Mexico covers about the same number of people as are uninsured in America. King points out that there may be lessons for other countries to learn in the success of Seguro Popular.

King's co-authors were Emmanuela Gakidou of the University of Washington; Kosuke Imai of Princeton University; Jason Lakin, Clayton Nall, and Nirmala Ravishankar of Harvard; Ryan Moore of Washington University in St. Louis; Manett Vargas of the Ministry of Health in Mexico; Martha María Téllez-Rojo and Juan Eugenio Hernández Ávila of the National Institute of Public Health in Mexico; Mauricio Hernández Ávila of the Ministry of Health in Mexico and the National Institute of Public Health in Mexico; and Héctor Hernández Llamas of Conestadistica. Gakidou, Imai, Lakin, Nall, Ravishankar, Moore, and Vargas are all King's current or former students and affiliates at the Harvard Institute for Quantitative Social Sciences.

The research was funded by the Mexican Ministry of Health, the National Institute of Public Health in Mexico, and the Harvard Institute for Quantitative Social Science.