Insurers using more physician profiling, delegates told
■ Payers push to measure physician efficiency, but experts say economic profiling can be inaccurate.
By Kevin B. O’Reilly — Posted July 10, 2006
Chicago -- Insurers increasingly are profiling doctors on how cost-effectively they provide care despite major methodological limitations, a panel told delegates at the American Medical Association Annual Meeting in June.
Physician profiling by private health plans is happening right now, Nancy Nielsen, MD, speaker of the AMA House of Delegates, who moderated the panel of experts, told physicians who attended a session, "Efficiency Measures: Necessary, Necessary Evil, or Just Evil? The Centers for Medicare & Medicaid Services also is pilot-testing privately developed physician efficiency measures.
Just because physicians might be unaware of the profiling "doesn't mean it's not being used to calculate your efficiency," Dr. Nielsen said.
Reasons for the increased use of efficiency measures are pretty straightforward, the experts said. The doubling of health care costs since 1999 and Medicare's impending insolvency in 2018 have created new professional duties for physicians to keep an eye on costs, said Troyen Brennan, MD, MPH, Aetna's chief medical officer.
"The more we spend per capita on health care, the more people lose insurance," Dr. Brennan said. "The more we provide for each individual patient, the more we actually deprive other patients who won't be able to access care at all. It's a problem for government, it's a problem for individual employers and it's a problem for physicians, because there's a whole host of patients who have no access to care."
Yet the efficiency metrics payers have employed to give patients financial incentives to steer clear of high-cost physicians have major limitations, the experts said.
For example, payers use episode-grouping software to batch together claims relating to the same diagnosis, said J. William Thomas, PhD, a health economist at the University of Southern Maine. From there, an expected cost of providing care for the diagnosis is generated by looking at the average cost for other physicians to provide the same type of care. The physician's actual cost is divided by the expected cost to reach a ratio. Scores under 1 mean that physicians provided care more efficiently than average, and scores over 1 signify less- efficient care.
"Can we screw up the measurement process?" Dr. Thomas asked. "Absolutely."
While some payers pair the efficiency measures with quality data, others don't, or payers give higher priority to cost data, Dr. Thomas said. And because payers have access only to claims data and not the complete medical record, he said, the efficiency metrics do not adjust for illness severity or comorbidities that might make treatment for an individual more costly.
But "the key issue to misclassifying physicians as efficient or inefficient" is sample size, Dr. Thomas said. It takes a sample size of 50 episodes of care to get accurate data, but that's often impossible given that each payer represents only a small slice of a physician's patient population.
"Because of differing sizes of databases from payer to payer, doctors could be rated as cost-efficient by one plan and average or cost-inefficient by another plan," Dr. Thomas said.
Groups such as the Ambulatory Care Quality Alliance and the National Committee for Quality Assurance are working on ways to aggregate physician performance data across public and private payers to get more representative, and thus more accurate, sample sizes.
The AMA doesn't have policy specifically addressing efficiency measures, but the Association has called for physician-specific health care data analysis to be transparent to physicians, done with physician input and released only if useful to physicians and patients.
Separately, the AMA has partnered with NCQA and Mathematica to develop 140 measures of quality -- not cost efficiency -- for CMS' Physician Voluntary Reporting Program by the end of this year. At press time, 98 had been completed.