Quality improvement, pay for performance, and "outcomes measurement": what makes sense?

The author argues against the use of pay-for-performance programs based on broad-based measures of patient outcomes in behavioral health care. He describes various problems with such programs. The purpose for collecting data is often not clear. Generic instruments do not measure improvement in specific mental disorders. Risk adjustment systems for behavioral health populations are not adequate. Mandated use of different instruments by payers is burdensome and precludes meaningful comparisons. The methodology for using outcomes-based approaches needs further development, and therefore the data collected will have little utility. The author discusses alternative approaches.