Employers seek to reduce variation in medical care

Dartmouth Atlas projects focus on patient choice

Employer groups are targeting variation in physician practice as an area of quality improvement and possible cost savings.

For several years, the Dartmouth Atlas of Health Care has reported wide variation in medical care based on Medicare data, but that information alone had little impact on patterns of care. Then in 1997, the Chicago-based Midwest Business Group on Health sought to show that employers could be the catalyst for change.

The 11-state business coalition hasn’t proven yet that projects based on Atlas data can reduce variation. But they are heralding the tool as a way to find overuse and underuse of medical services and to focus improvement projects.

Three pilot projects in Muncie, IN; Wichita, KS; and Milwaukee are addressing percutaneous transluminal coronary angioplasty rates that are significantly higher than the national average. (See chart below) In about six months, Muncie is expected to be the first to report data.

"There are two tests that we were trying to pass with this project," says Jim Mortimer, president of the Midwest Business Group. "One was to find data that were understandable to employers, so they could find the most appealing opportunities for improving the system. The second test was that the data were acceptable and viewed as valid by clinicians."

While the project encountered some obstacles, including skepticism on the part of physicians, those goals have been met, he says.

Variation in U.S. care disturbing’

The variation presented in the Dartmouth Atlas makes U.S. medicine seem somewhat arbitrary. Even when adjusted for demographic differences and incidence of heart disease, the number of angioplasties performed varies from two per 1,000 to 14 per 1,000, with a U.S. rate of six per 1,000.

"This variation is reflecting something very disturbing about the way medicine is practiced," says Megan McAndrew Cooper, MBA, MS, editor of the atlas, which is produced by the Center for the Evaluative Clinical Sciences at Dartmouth Medical School in Hanover, NH.

"There is very little absolute, shared knowledge about what’s best under what circumstances. Physicians are making a lot of decisions on an ad hoc basis, and they don’t have an enormous amount of feedback about their own work, what others are doing, and outcomes," she explains.

"What happens to you is more a function of where you live and which doctor you go to for care than it is about some knowledge base about what works or an informed decision by the person affected," Cooper adds.

In fact, she contends that drastic measures to save Medicare from insolvency wouldn’t be necessary if the nation addressed differences in use of medical procedures and cost around the country. But those differences aren’t decreasing on their own, even with managed care.

"The patterns of variation are very stable over time," Cooper says. "Places with very high bed rates have stayed very high. Everybody’s been reduced some, but the rate in reduction in bed capacity is the same in high-rate areas as it in low-rate areas."

The Midwest Business Group on Health saw the same potential to hold back increases in health costs by ensuring their employees receive appropriate care. The Robert Wood Johnson Foundation in Princeton, NJ, funded a study by the coalition, the Hospital Research and Education Trust of the American Hospital Association and researchers at Dartmouth.

The first product of the collaborative was a publication, A Health Care Purchaser’s Guide to Using the Dartmouth Atlas.

A Muncie, IN, health system provides the business coalition’s best example of what can be done with data from the Dartmouth Atlas. General Motors Corp. worked with the Cardinal Health System to analyze 44 variables, comparing them with 14 other regions that have General Motors sites. They produced a matrix that helped identify problems areas. The team selected angioplasty as an area for study. But in working with area cardiologists, it first had to verify that the Medicare data reflected the pattern of care in the community at large.

When it calculated rates using a state database from the Indiana Hospital Association, the team found the same pattern of variation (although not necessarily the same frequencies). "That validated the findings from the Dartmouth Atlas," says Mortimer.

Cardiologists met to discuss and analyze the variation. "When you have physicians from different communities with different practice signatures discussing how they manage patients with a particular condition, they realize there are different ways of [doing things]," says David Wennberg, MD, MPH, senior research associate at the Maine Medical Assessment Foundation in Manchester, which uses statewide Medicare and Medicaid data to address variation. "The data is a way to get conversations rolling."

The project team, which included support from Ball State University in Muncie, IN, and the Indiana Hospital & Health Association, decided to focus on patient awareness as a way to reduce the angioplasty rate.

The Shared Decision-Making videos, produced by the Dartmouth-affiliated Foundation for Informed Medical Decision Making, explain the risks, benefits, and options of various treatments. Studies show that use of the shared decision-making program can reduce the use of invasive procedures to treat heart disease.¹

Docs aren’t always enthusiastic

The collaborative between physicians and health care purchasers hasn’t been easy to establish in all markets. In Milwaukee, which has an angioplasty rate of 9.1 per 1,000, physicians have been less willing to become involved in the project, says Mortimer.

The project team there is now trying to work with physician leaders, he says.

So far, employers have steered away from any talk about "selective contracting." But Mortimer says he doesn’t discount that possibility.

Although the Dartmouth Atlas information is purely population-based, physician- or medical group-specific information could be obtained from other databases.

"[Employers] may wind up rewarding the practitioners who are working with them by channeling people in that direction," he says. "They can be more confident that people will get better care. That’s the whole idea of quality improvement — better value, less waste, less overuse."

Reference

1. Morgan MW, Deber RB, Llewellyn-Thomas HA, et al. A randomized trial of the ischemic heart disease shared decision-making program: An evaluation of a decision aid. J Gen Intern Med 1997; 12(supp):12-62.