“For a small bump in cost, you get 53 depression-free days in year 1, and in the second year, you actually save money for the 54 days [gained],” Dr. Katon said.
To ascertain total costs, the researchers considered the cost of usual care as $0 for reference and calculated intervention-specific costs as a mean $591 per patient over the 2 years. Comparing other mean costs for intervention group vs. usual care, antidepressant medication was $416 higher for intervention patients; other medication costs were $126 lower (a net savings); outpatient specialty mental health care was $86 lower; and other outpatient costs were $501 lower for intervention patients.
Intervention-specific costs included psychiatrist and primary care supervision time, nurse time, overhead costs, and educational materials. Other ambulatory medical costs included primary care and specialty visits, emergency department use, urgent care visits, and laboratory and imaging charges. Researchers excluded costs of inpatient care and patient time. The cost of patient time is “difficult to do in the elderly, because most are not working,” he said.
Some figures were estimated. For example, 17%–24% of health care data were not available, Dr. Katon said. In addition, some organizations did not have pharmacy data. In cases where data were missing, imputation—which estimates costs by considering demographics, prior health care use, and other factors—was used to estimate costs.
The researchers estimated the incremental cost per quality-adjusted life year (QALY) for the intervention group. The range was $2,521 to $5,000. “It is widely accepted that anything that is under $10,000 per QALY for health care should be implemented immediately,” Dr. Katon said.
New interventions typically cost more with increased effectiveness, Dr. Katon said. “The holy grail is that an intervention that costs less with increased effectiveness should be implemented immediately.” For three of the eight organizations, the intervention saved money over the 2 years, with greater benefit, he added.
Reimbursement for collaborative care remains an issue. Psychiatrist supervision with the primary care physicians and depression care manager was not reimbursable, nor were the depression care manager's consultations with other providers (nonpatient treatment time). Follow-up telephone calls, likewise, were not reimbursed.
Despite the reimbursement issues, interest in the IMPACT model has been strong. “We're getting called all the time from health care organizations all over the United States with questions about how to implement this,” Dr. Katon said.