When More is Less

Last week, Dartmouth’s Center for the Evaluative Clinical Sciences released the latest version of its “atlas” of medical care in the U.S. Since 1993, the Dartmouth Atlas Project has produced a fascinating series of studies examining the geographic variability in health care resources and their utilization. A consistent theme throughout the life of the project has been that the availability of resources drives their use: surgical rates—even for elective surgery—are higher in communities with more surgeons; hospitalization rates are higher in areas with more hospital beds. The newest report, called “The Care of Patients with Severe Chronic Illness: A Report on the Medicare Program by the Dartmouth Atlas Project” draws the same conclusions about the care for people with chronic disease (available in entirety online here)

Chronic disease—conditions such as diabetes, cancer, and heart disease—is the major health problem among Americans today, afflicting some 90 million people and accounting for 7 out of 10 deaths. The care of people with chronic illness accounts for over 75% of all U.S. health care expenditures. And most of the people with chronic conditions are elderly.

The principal finding of the Dartmouth study is that Medicare spends much more per enrollee in some states than in others. Spending for patients with severe chronic illness during their last 2 years of life ranges from almost $40,000/person in New Jersey, Washington, D.C., California, New York and Maryland, to under $25,000 in states including Idaho, Iowa, West Virginia and North Dakota. The variability is not due to higher rates of sickness in some regions. In fact, differences in illness levels are “virtually unrelated to differences in spending.” The other disturbing finding of the study is that the extra spending does not buy longer life or better quality of life. On the contrary, those with chronic illness in high spending regions have shorter life expectancies and less satisfaction with their care.

Spending on chronic illness varies by state, with those states that have greater reliance on primary care than on specialists spending less money and depending less on intensive care units. In Florida, for instance, Medicare patients who died spend an average of 4.7 days in the ICU during the last 6 months of life (a marker of the aggressiveness of care), whereas patients in North Dakota spend only 1.5 days. Academic medical centers also vary in the way they manage chronic illness: during the last 6 months of life, for example, patients who use New York University Hospital have an average of 76 physician visits, compared to 24 visits for patients who use the Mayo Clinic. In general, acute hospitals are dramatically over-used (in Hawaii, patients spend an average of 16.4 days in the hospital during their last 6 months of life, compared to 7.3 days in Utah) and hospice care is under-utilized (while in Arizona, 44.7% of dying patients are enrolled in hospice, in Alaska it is only 6.7% and the national average is 27.2%).

The authors conclude that what we need is “a population-based, community wide integrated system for managing severe chronic illness.” I agree. Specialists have no incentive to refer patients to hospice care; they do have incentives to order diagnostic tests and to use ICU care. Hospitals likewise, unless they are part of a network of care, have no incentive to enable patients to die at home; they benefit if their beds are filled. If we had a comprehensive system—and the closest that any segment of the U.S. comes to such an approach is the Veterans Administration system—Medicare would reduce its costs on patients with chronic disease by 30%, while simultaneously improving care.