In his 30-plus years of studying healthcare utilization and costs across the United States, John E. Wennberg has found that just about the only constant in the nation's healthcare system is inconsistency.
Having compared hospitals and clinics from one region of the country to another, as well as those within the same region and the same healthcare system, Wennberg -- director of the Center for the Evaluative Clinical Sciences at Dartmouth Medical School; founding editor of the Dartmouth Atlas of Health Care; and featured speaker for the 2005 Eisenberg Legacy Lecture, hosted by CHP/PCOR on Nov. 3 -- has documented large unwarranted variations in the amount, intensity and cost of healthcare services provided to specific patient populations.
The variations, Wennberg explained at the Eisenberg lecture, are unwarranted because they can't be explained by illness, patient preferences or the dictates of evidence-based medicine. As a consequence, he pointed out, there is no consensus in the American medical community as to what is the "right" rate of particular medical interventions -- nor is there much discussion about the problem of variation, much less what to do about it.
"This never gets discussed at medical staff meetings. It is subliminal behavior," Wennberg said at the Eisenberg Legacy Lecture, held this year at Stanford in the Bechtel Conference Center. The annual lecture, which honors the memory of John Eisenberg -- a renowned health services researcher and former director of the Agency for Healthcare Research and Quality -- is funded by the California HealthCare Foundation and is co-sponsored by CHP/PCOR, the Institute for Health Policy Studies at UC San Francisco and the Center for Health Research/School of Public Health at UC Berkeley.
In his 90-minute talk on "Understanding Practice Variations," Wennberg presented a wealth of data illustrating the phenomenon, and he urged the healthcare community -- particularly academic medical centers -- to examine and address the variations. To remedy the nation's overdependence on costly, high-intensity interventions such as surgery and ICU stays, he called for improved physician communication and greater patient involvement in medical decision making. He also urged reform of the nation's healthcare financing system, with incentives to encourage healthcare providers to deliver more efficient, more effective care.
"Our current system of healthcare financing rewards the quantity of procedures, not the quality of decision making," said Wennberg, a nationally known pioneer in the study of healthcare utilization and quality. His talk drew some 80 attendees from the three university sponsors and from organizations including healthcare systems, health insurers, biotech companies and healthcare investment firms.
Wennberg distinguished among three types of medical services: "effective care" (interventions widely established as medically necessary), "preference-sensitive care" (treatments such as discretionary surgery, for which there are viable alternatives involving tradeoffs), and "supply-sensitive care" (services involved in managing disease, such as specialist referrals).
He then discussed his most recent research showing striking variations in all three types of care: a comparative study of the performance of individual California hospitals, published Nov. 16 as a Health Affairs Web exclusive, with CHP/PCOR fellow Laurence C. Baker as a co-author. The study found that some California hospitals spend as much as four times more than others to care for patients with similar chronic illnesses, with no gain in quality or patient satisfaction for those that spend more.
In the "effective care" category, Wennberg presented data from a Medicare reporting system that tracks hospitals' compliance with best-practice guidelines for treating health conditions including heart attack and pneumonia. Patients at the highest-ranking hospitals received the recommended care 85 percent of the time, while patients at the lowest-scoring hospitals received such care less than half the time. Among California's academic medical centers, the quality score for treating pneumonia ranged from 64.7 at Stanford Hospital to 52.3 at UCLA Medical Center.
Wennberg lamented that many clinically proven treatments, ranging from beta blockers for heart disease to antibiotics following surgery, are consistently underused. "The irony is that most of these proven interventions don't cost that much," he said.
In the realm of "preference-sensitive care," Wennberg presented data showing that rates of knee and back surgery among the population served by Stanford Hospital are more than double the rates of those surgeries in Los Angeles and San Francisco. "This is prima facie evidence that there's a problem with our clinical decision making," he said.
Wennberg's analyses reveal that such variations can't be explained by patient characteristics such as age, sex, ethnicity or patient preferences. "What matters most in determining the variation in use rates is not the condition you have or the treatment you may want, but the physician whom you ask for advice," he said. Wennberg has also found no clear association between the supply of certain types of surgeons and the rates of surgeries they commonly perform.
To reduce overuse of discretionary surgery, Wennberg advocated "shared decision making," in which patients take an active role in making decisions about their care. "When the choice of treatment involves tradeoffs, patient preference should be the driver of medical decisions, but too often the decision is made by the doctor," he said. A good example is whether a breast cancer patient should undergo a mastectomy or a lumpectomy: Both procedures are medically viable options, but have very different lifestyle implications for the patient.
Wennberg cited clinical studies showing that patients who have used decision aids are better informed about their treatment options and feel more comfortable with their chosen treatment. Patients using shared decision making also choose surgery significantly less often than other patients.
Regarding "supply-sensitive care," Wennberg's latest research has found wide variation among California hospitals in their number of inpatient days per patient, for Medicare enrollees in the last six months of life. Such patients spent an average of 19.2 days in the hospital at UCLA Medical Center, for example, compared with just 11.6 days at UC Davis.
Wennberg said the American medical community and the public must challenge the assumption that more care is better. In fact, the opposite appears to be true. He cited a recent Dartmouth study that examined whether patients with hip fractures, colon cancer and heart attacks who lived in U.S. regions with a higher intensity of medical interventions had better outcomes than those living in regions with less care. The study found that patients in the higher-intensity regions had higher mortality rates and worse functional measures.
Improving healthcare delivery in the United States will require major changes in healthcare financing as well as culture, Wennberg said. "This isn't just a matter of cajoling doctors to talk to their patients more -- it's a matter of changing the financial incentives in the system."
As a promising development, he cited the Medicare Health Care Quality Demonstration Programs that were created under the Medicare Modernization Act and are set to begin in early 2006. Under the programs, participating health systems will test major changes to improve healthcare quality and efficiency, including alternative payment systems and modifications to Medicare's traditional benefit package.
In addition to delivering the Eisenberg lecture, Wennberg led two CHP/PCOR-hosted discussion meetings with groups of faculty and postdoctoral trainees. "This event was a wonderful opportunity for us to bring together the health policy community and discuss with a national expert these important questions surrounding practice variation," said CHP/PCOR executive director Kathryn M. McDonald.