Want better prevention performance? First, you must have a system

Physician variation means gaps in care persist across nation

Despite a national focus on cancer screenings, immunizations, and other preventive care, an analysis of patterns of care in the United States shows these trends:

  • Preventive services are highly variable across the country and even within a region.

  • Whether a patient receives preventive care isn’t related to access to physicians or even continuity of care.

  • A physician who is thorough in ordering mammograms for older women may fail to provide regular foot exams for diabetics.

    Those findings from the Dartmouth Atlas of Health Care — 1999 underscore the need for physicians to rethink their processes of care, says David Wennberg, MD, MPH, director of the Center for Outcomes Research and Evaluation at Maine Medical Center in Portland, and editor of the Atlas chapter on ambulatory care.

    The Labanon, NH-based Dartmouth Atlas is supported by a grant from the Robert Wood Johnson Foundation and published by the American Hospital Association Health Forum. It was first published in 1996 with a second edition in 1998. (See editor’s note for ordering information, p. 63.)

    "I think there needs to be a more systematic approach to the way we take care of patients for preventive services as well as chronic diseases," Wennberg continues. "It can be as simple as a flow sheet or as advanced as an electronic medical record, which posts reminders for the front office staff or physicians about what services are needed."

    Without a system to remind them which patients need preventive care, physicians fall far short of national goals for cancer screening, immunization, and diabetic care, the Atlas researchers found.

    "We have to assume that the cause of the variation is the practice styles of individual physicians or groups of physicians about the usefulness of some of this [preventive care]," says Atlas editor Megan McAndrew Cooper, MBA, MS. "In some places they put a lot of value on some kind of screening but not on others."

    With its vast database of Medicare claims, the Atlas provides a detailed picture of the patterns of care in fee-for-service medicine. The country was divided into 306 "hospital referral regions," based on patterns of care. Each edition presents a new focus, but many of the measures have been followed over time. "The patterns are extremely persistent from year to year," says Cooper.

    In some regards, the Atlas findings are discouraging. For example, the data show that areas with a greater supply of hospital beds have higher rates of hospitalization. Yet a greater supply of physicians in a community does not equate to better screening and other preventive care.

    "If you have a lot of beds available and a person presents with certain symptoms, that influences whether they’re admitted," says Cooper. "Clinicians have a very hard time with this because they see themselves as making decisions based on a set of clinical rules." Those rules, however, are influenced by a variety of factors, including available bed space.

    Meanwhile, managed care seems to have a limited impact on the overall provision of preventive services. The National Committee for Quality Assurance in Washington, DC, began developing quality measures in the late 1980s with an emphasis on preventive health. Yet areas of the country with higher managed care penetration didn’t show consistently high performance on preventive care.

    "We looked at profiles of hospital referral regions in northern and southern California," says Cooper. "They have a lot of managed care and a lot of doctors. They are just as idiosyncratic in complying with guidelines."

    For example, a region could have high rates of providing eye exams for diabetics but low rates of screening for lipids in the same population. "We don’t see any correlation between screening for colorectal cancer and screening for breast cancer," notes Cooper.

    "What I see in those maps is opportunity," says Gerald O’Connor, PhD, DSc, associate director of the Center for the Evaluative Clinical Sciences at Dartmouth-Hitchcock Medical Center in Lebanon, NH. O’Connor was a founder of the Northern New England Cardiovascular Disease Study Group, a consortium in Maine, New Hampshire, Vermont, and Massachusetts that shares information on the treatment of cardiovascular disease. After data feedback, training in continuous quality improvement, and site visits to other facilities, the 23 cardiac surgeons in the project recorded a 24% reduction in hospital mortality rate.1

    That reduced mortality rate has persisted, and northern New England now has one of the lowest mortality rates for coronary artery bypass graft surgery in the country. (For more information on the northern New England project, see related story, p. 64.)

    Yet during the first three years of measurement, the surgeons simply observed and studied their rates. They learned that the variation in their mortality rates was not due to case mix, as they had originally believed. In other words, the surgeons or hospitals with the highest mortality rates did not have sicker or older patients.

    O’Connor likens the recent Atlas findings as the first stage of the study group. Physicians are just learning about the differences in rates for various kinds of care, he notes.

    "The question is, will the professional societies and groups of physicians step up to the plate? Now that we know this, how can we put in systems of care?" he asks. "How can we remember the last time someone had a mammogram?"

    Solving that problem shouldn’t be as complex as trying to determine why patients die after bypass surgery. I think they’re quite tractable problems," says O’Connor, who is also professor of medicine and community and family medicine at Dartmouth Medical School in Hanover, NH.

    O’Connor also advises physicians not to rely too heavily on anecdotal information, but to gain more power by aggregating data. "There’s nothing about the person who died this month that tells you what to do differently in the operating room," he says. "But you can see in aggregate data what’s impossible to see as an individual."

    Ultimately, physicians enjoy the academic aspect of uncovering the cause of variation and measuring how interventions improve outcomes.

    "It’s been very empowering for people to be clinical scientists, to learn from the their patients and do a better job tomorrow than they did yesterday," he says.

    [Editor’s note: To order a copy of the Dartmouth Atlas contact the AHA, P.O. Box 92683, Chicago, IL 60675-2683. Telephone: (800) 242-2626. Fax: (312) 422-4505. Web site: www.aha.org/shoppingcart. Book version, #044401: $295 for members, $350 for nonmembers; CD-ROM version, #044452: $1,095 for members, $1,295 for nonmembers. Shipping and handling: $12.95.]


    1. O’Connor GT, Plume SK, Olmstead EM, et al. A regional intervention to improve the hospital mortality associated with coronary artery bypass graft surgery. JAMA 1996; 275:841-846.