The results, published today (Tuesday, April 23) in two papers in the journal Health Services Research, are particularly encouraging since the patients studied were both poor and severely mentally ill, factors many feared would leave them vulnerable to lower quality care in cost reduction settings.

"Conventional wisdom says these patients are the ones who fall through the cracks," said Joan Bloom, professor of health and policy management at UC Berkeley's School of Public Health and lead author of one paper. "The fear is that these patients are prime targets for cost reduction because they tend to incur higher treatment costs on average and because they are less likely to speak out about reduced quality of care. Our findings show that it doesn't have to be that way."

Bloom is the principal investigator of the National Institute of Mental Health (NIMH) grant that funded the study. She is the lead author of the paper that analyzed the cost of direct treatment services before and after the implementation of the managed care program. Brian Cuffel, former adjunct assistant professor at UC San Francisco's Department of Psychiatry, is the lead author of the two-year patient outcomes paper.

Bloom described the pair of analyses as "a silver-lining story" in the debate over managed care. "We were pleasantly surprised at the findings," she said. "The study results indicate that managed care can work well under certain circumstances."

Colorado passed legislation in 1992 to switch its Medicaid mental health services from the traditional fee-for-service system to one that is under managed care, also known as capitation. Part of the law mandated that the first two years of implementation be evaluated as a pilot program, which began in 1995.

At least 32 states - including New York, Utah, Massachusetts, Minnesota and California - are in some stage of switching their Medicaid mental health patients over to managed care programs.

Researchers were able to study two models of managed care delivery because Colorado contracted with both for-profit and non-profit agencies to deliver mental health services. "This is probably the first study that really looks at the nature of different models of capitation," said Teh-wei Hu, professor of health economics at UC Berkeley's School of Public Health and study co-investigator.

In western and southern Colorado, a for-profit managed behavioral health firm formed a joint venture with an existing provider agency to deliver services. The legislation called for profits to be capped at 5 percent, with any additional revenue from cost savings reinvested into the program to increase access to mental health services.

In other regions, existing non-profit agencies contract directly with the Colorado Department of Human Services to deliver care through community mental health centers. Mental health providers, including those in the fee-for-service program, were divided into 17 regions based on county borders. Organizations holding contracts to deliver Medicaid mental health services for the state are called Mental Health Assessment and Service Agencies.

Researchers studied adult patients who were diagnosed with schizophrenia or bipolar disorder, or who had at least one 24-hour psychiatric hospital stay in the prior year.

They used claims data and interviews regarding utilization of services by 522 severely mentally ill patients. For the analysis in both studies, patients were randomly selected from each region and then matched by gender and prior service costs.

The researchers found that non-profit managed care areas saw a 19 percent drop in costs per person while for-profit managed care areas saw cost reductions of 70 percent per person.

The cost reductions for consumers in the managed care regions reflect the fact that they were less likely to use services than their fee-for-service counterparts. Differences in costs between the managed care region were also due to reductions in the cost of outpatient care found only in the joint venture areas.

Along with the reduced costs, the study also found changes in access patterns in managed care plans. Patients in non-profit capitation programs reported an increased incidence of services being refused, discontinued or reduced. Patients in for-profit managed care plans also reported significantly longer waiting times.

But the reports of reduced access did not translate into measurably poorer outcomes for the patients, according to the second paper, led by Brian Cuffel. "Ultimately, the patient's well-being is what we care about," said Cuffel.

The 591 patients in the outcomes paper were evaluated on the severity of psychiatric symptoms, their ability to function, their quality of life and public welfare. Medicaid patients enrolled after capitation began were included in the analysis. Trained and experienced interviewers used psychiatric rating scales to assess patient issues such as alcohol or drug abuse, suicide, physical health, degree of social activity, occurrence of homelessness, feelings of personal safety and financial stability.

No significant differences were found for most outcome measures among Medicaid patients in the managed care and fee-for-service groups. However, after two years, the rate of homelessness was lower for those in non-profit managed care plans compared with fee-for-service regions. Patients in the for-profit joint venture capitation plans also reported relatively lower ratings of suicide and substance abuse problems, and higher levels of social contact and daily activities compared with fee-for-service patients.

"We weren't able to tease out certain aspects of care that led to the few outcome differences we did find," said Cuffel, who is now vice president of research at United Behavioral Health in San Francisco. "The fact that we didn't really see consistent findings - where patients in one program clearly fared better than another - makes us feel that the differences weren't directly caused by capitation. Perhaps the differences reflected the existing strengths of some community mental health centers over others."

Tom Barrett, director of Mental Health Services for the Colorado Department of Human Services, pointed out that the primary goal of the capitation plans was to increase mental health services in the community. He said the capitation rollout worked well because it involved an unprecedented level of community involvement. "We made a concerted effort to solicit input from health care providers, health care advocates, family members, consumers and legislators to improve the quality of services," said Barrett.

Hu, at UC Berkeley, said capitation programs can help create a more predictable, stable financial environment, which can then translate into expanded overall services. "I think Colorado is a good example for other states to follow," Hu said.

In addition to Joan Bloom, Brian Cuffel and Teh-wei Hu, the authors on both papers are Neal Wallace, assistant professor of health economics at Portland State University, and Jaclyn Hausman, research analyst at UC Berkeley's School of Public Health. Richard Scheffler, professor of health economics and public policy at UC Berkeley's School of Public Health, and Mei-Ling Sheu, assistant professor of health management at Taipei Medical University, were co-authors of the cost-effectiveness paper.