Agenda, Scope, and Goals

MDRC recently conducted an evaluation of coordinated care programs in New York and  completed a similar evaluation in Colorado.

The evaluations. In both states, the core of the evaluation was an impact study that used state administrative records to examine the effects of care management on health care choices and health care costs. An implementation study explored how care management approaches differ from place to place, to help illuminate what made such programs effective.

These evaluations were among the first large-scale, rigorous studies of coordinated care for high-needs Medicaid recipients. They provided crucial information to state, local, and national policymakers about whether such interventions were effective at increasing the quality of care and reducing health care costs. At the national level, some policymakers see successful cost containment as a necessary aspect of any serious effort to enact national health insurance. At the state level, cost containment for high-cost cases has become a leading priority for state Medicaid programs.

Design, Sites, and Data Sources

MDRC recently conducted an evaluation of coordinated care programs in New York and completed a similar evaluation in Colorado.

Colorado. The evaluation in Colorado took place in six counties: Adams, Arapahoe, Boulder, Weld, Denver, and Jefferson. In these counties, two versions of coordinated care were tested, with services provided by Colorado Access or Kaiser Permanente Colorado.

Both organizations developed coordinated care programs that were similar to ones they had previously used with high-needs Medicare patients. Each program had several goals. One was to reduce the use of emergency departments for routine care by establishing a primary care provider for each patient and helping the patient make and keep appointments. A second was to help patients comply with treatment recommendations — such as taking medications, making follow-up appointments, and eating well — to keep chronic conditions from worsening to the point where hospitalizations were needed. Care managers began by assessing each individual’s health care and social service needs in order to develop individual care plans. A patient whose care needs seemed to be well managed received limited care management, while one who was struggling with multiple chronic conditions received more intensive care management. For example, a diabetic patient with pulmonary disease and a substance abuse problem might have been contacted by a care manager several times a week to assess the person’s medical needs, help him or her make appointments to see appropriate health care providers, and monitor his or her compliance with recommended actions, such as taking medications, following an appropriate diet, and checking blood sugar levels. The care manager would also have tried to work with the patient’s primary care physician to make sure someone was coordinating care among the variety of specialists that the patient was likely to see. Although protocols would provide care managers with suggestions for what was appropriate care, the care managers had substantial discretion to do what was necessary to help patients receive high-quality, appropriate care.

The evaluation of this program was built on a state plan to phase in coordinated care. In 2008, a randomly chosen portion of eligible Medicaid patients in Adams, Arapahoe, Boulder, and Weld counties who were in the fee-for-service system were sent a letter saying they had been enrolled in the Colorado Access managed care program and telling them how they could opt out of the program. The remainder stayed in fee-for-service care for at least two years from the date of random assignment. Because the group that was passively enrolled in the Colorado Access program was chosen at random, any differences that emerged in health care and health outcomes between the two groups over the following two years can be reliably attributed to the Colorado Access program. In 2009, a similar process began for the Kaiser Permanente Colorado pilot. Final reports on the two pilot programs have been published by MDRC.

New York. In New York, the state legislature approved funding for the Chronic Illness Demonstration Project (CIDP), which provided coordinated care for high-needs Medicaid beneficiaries with multiple chronic conditions in New York’s fee-for-service Medicaid system. In January 2009, the state announced that six programs had been funded, three in New York City, and one each in Nassau, Westchester, and Erie (Buffalo) Counties. Programs began operations in summer 2009.

Individuals were eligible for CIDP services if they were currently receiving fee-for-service Medicaid and were likely to be hospitalized in the coming year according to a predictive model developed by John Billings at New York University. This was a high-needs group. On average, its members were hospitalized about three times a year and used nearly $60,000 per year in Medicaid benefits. About three-fourths had been diagnosed with a mental illness and four-fifths had been diagnosed with a substance abuse problem. They also suffered from a range of other chronic conditions. Increasing primary care and reducing emergency department and hospital use for this group may result in substantial savings to the health care system. A final report on the study’s findings was published by MDRC in 2014.