Linda Buchwald, MD, has been caring for patients with some of the most complex multiple sclerosis cases in Massachusetts for over 30 years. A member of the Harvard, Tufts, and Boston University faculties, she is chief of neurology at the Harvard-affiliated Mount Auburn Hospital, and has served on the medical advisory board of the National Multiple Sclerosis Society.
So it came as a shock when insurance companies began notifying her that their new ranking programs, which place physicians in one of three tiers based on a combination of cost and quality measures, relegated her near the bottom, in either Tier 2 or Tier 3. Patients treated by physicians in the lower tiers are either charged higher co-payment fees, or must change physicians.
Dr. Buchwald began calling health plans and found it virtually impossible to get through by phone or by e-mail. “Finally, I got an e-mail back from Unicare with a list of quality measures. Of about 90 quality measures, the only one relevant to neurology was this: does the physician order vitamin B12 on dementia patients? I don't see dementia patients much anymore, so there were really no quality measures at all relevant to my practice,” she said. “And when I finally reached the Tufts plan, they said they didn't have any neurology measures yet.”
LAWSUITS CHALLENGING POLICY
Physician tiering is the latest cost-containment wave to sweep health insurance plans — and it's fraught with problems like those faced by Dr. Buchwald. She's one of five lead plaintiffs filing suit with the Massachusetts Medical Society to stop the state's General Insurance Commission from implementing its tiering program — the Clinical Performance Initiative — until and unless the program adheres to specific standards: transparency; fair notice; formal feedback and correction processes; meaningful physician involvement in the development of the standards and measures; demonstration of the program's accuracy, validity, and reliability; and submission of the program to an independent oversight authority.
Massachusetts is just one of several states taking action against tiering programs. In November 2007, New York Attorney General Andrew Cuomo settled a complaint against CIGNA and other national health insurers, under which they will take several steps to improve their ranking programs, including ensuring transparency, fairness, due process, and independent oversight. In Connecticut, a suit by the Fairfield County Medical Association, which seeks to halt flawed ranking programs developed by United Healthcare and CIGNA, is still pending. And in Washington state, Regence Blue Shield agreed to include physicians in developing quality measures after the state medical association filed suit over its physician-ranking program in 2006.
THE AANPA POSITION
In July, the AANPA published a policy paper on physician ranking programs, seeking “active involvement in the development, implementation, and evaluation of any physician profiling initiatives that would cover neurologists.”
The paper, available online at aan.com, praised Attorney General Cuomo's actions, and stated the Academy's support for the American Medical Association's stance, urging that “any entities choosing to design and implement physician profiling programs adhere to several principles to help ensure clarity and fairness.” Those principles include physician involvement in developing standards and measures, transparency of the process, advance notification to physicians before the release of rankings and the opportunity to appeal disputed rankings, the performance of pilot studies, and the need for demonstrated validity and accuracy of the measures used.
“The statement says that Academy members want to participate in this process,” said William Henderson, the administrator for Upstate Neurology Consultants in Albany, NY, who served on the workgroup arranged by the Payment Policy Subcommittee of the AANPA Medical Economics and Management Committee, which developed the statement. “From what I can see, the insurers have been doing their own thing. New programs are just announced, with no involvement of physician groups beforehand.”
“As neurology groups begin using electronic health records, they now have clinical data which are far more accurate than the claims data an insurer might have,” said Henderson. “We have a lot of information to bring to the table.”
The AANPA statement also highlights another key principle: “Ensure that ranking for doctors is not based solely on cost of care and that these rankings clearly identify the degree to which a ranking is based on cost alone. Further, cost ratings alone should not be used to select physicians.”
This is a key concern, said Gloria Galloway, MD, professor of neurology and pediatrics at Nationwide Children's Hospital at Ohio State University, and a member of the payment policy subcommittee. “Rankings for doctors in general, and certainly neurologists in particular, should not be based solely on cost, and if they are based in part on cost, that needs to be transparent. Right now, they are not, and the degree to which the cost portion goes into the total ranking is unclear. Neurologists are particularly vulnerable, given the complexity and number of disorders their patients often face.”
This is the problem for Dr. Buchwald, who found literally no quality measures relevant to her practice used by the insurance companies that ranked her in Tiers 2 or 3. “My ranking is negatively affected by my commitment to treating neurologically complex patients, often with advanced disease,” she said. “Even patients with more benign or early disease are often expensive because they are treated with immune modulators that cost $26,000 a year).”
LIMITATIONS OF SOFTWARE-BASED DATA
The two primary software programs used by health insurance companies to help develop physician rankings are supposed to adjust for patient complexity. “They look at all the claims and diagnoses for a particular person in a given year, and come up with a risk score,” said William Thomas, PhD, professor of health policy and management at the University of Southern Maine in Portland and an expert on measuring health care providers' performance in terms of efficiency and quality of care. “So a person who has a many chronic conditions will have a higher risk score, and a fundamentally healthy person would have a lower risk score. But the reality is these systems don't work very well at all.”
The software doesn't differentiate between comorbidities that might have an effect on a particular condition, he said, adding that revised software is supposed to address this issue, but it's unclear how well it works.
Ranking programs are also fraught with other errors, the Massachusetts Medical Society suit charges. One physician had his entire ranking lowered when a computer miscategorized a patient's radiotherapy treatment as an “office visit.” Another dug through reams of data and files to match his own patient records to the insurers' claim records, and found that 68 percent of the patients attributed to him weren't his — he had merely read their EKGs or exercise tolerance tests.
But many doctors do not have the time or capability to run such checks. “I work 95 hours a week,” Dr. Buchwald said. “ I don't have the time to go through their claims data and try to hitch it up with patients. It's a Herculean task, which would also depend on the cooperation of the health plans.”
Dr. Thomas said the problem is “dirty data.” “There are errors in claims databases — like diagnostic coding errors — that can influence physician scores,” he said.
Another big issue, he adds, is sample size — how many episodes of care are available to calculate a physician's score. The National Committee for Quality Assurance recommends at least 30 episodes per physician; Dr. Thomas said his research indicates that the required minimum should be at least 50 episodes. “The consequence of that for health plans, especially those with a small market share, is that they can accurately profile only a small part of their provider network,” Dr. Thomas said. “If they're tiering all their physicians, where does that put you?”
HOW RELIABLE ARE THE RANKINGS?
Dr. Thomas is now developing a paper that assesses the reliability of physician rankings by clinical specialty. “What we find is that there are huge differences in reliability,” he said. “On a scale of 0, meaning completely unreliable, up to 1, meaning completely reliable, the range was 0 up to 0.93. For some clinical specialties, there was absolutely no information on performance. You can go through the calculation and come up with numbers, but they don't mean anything at all,” he said. “The most reliable specialty was dermatology.”
And neurology? Its measures were squarely in the middle, with a median reliability score of 0.58. (Although it should be noted that 58 out of 100 points is a failing grade in most scoring systems.)
That's a strong argument for more involvement of neurologists in the process of developing these measures, said Marianna Spanaki-Varelas, MD, PhD, senior staff neurologist at Detroit's Henry Ford Health System and also a member of the payment policy subcommittee. “We want to define neurology-specific quality measures, and also to be part of the process in collecting and validating the data,” she said. “We want to make certain that we have the right sample size so that the data are statistically significant.”
In the meantime, said Dr. Buchwald, the current tiering systems are demoralizing for physicians and patients. “It's humiliating — they're saying you're a bad practitioner,” she said. “We really think we're doing our best to do it right: looking at cost, risk, benefit, and outcomes on everything we order. I understand they're trying to contain costs, but it is dangerous thinking to have these flawed systems mandated.” •