Trimming Health-Care Costs Without Reforming the System
In the heated debate over health-care reform, one inconvenient fact is often ignored. There's little evidence to support the use of many of today's routine treatments and procedures. By some estimates, the portion of medicine that has been proven truly effective is still in the range of 25%. That means billions of dollars are being spent on care that may not be effective. And there's even a raging debate over whether the country should do "comparative effectiveness" research to try to figure out what does really work.
That's the bad news. The good news is that these woes mean there is enormous room for improvement, if doctors can understand how to care for people more effectively. The latest proof of the gains that are possible comes from a study published Oct. 1 in The American Journal of Managed Care by researchers from Kaiser Permanente.
In the study, patients with diabetes or heart disease were given a simple, low-cost regimen of aspirin, a generic cholesterol-lowering drug, and a blood-pressure drug. Compared with similar patients not taking the combination, these patients had 60%-80% fewer heart attacks and strokes in a two-year period. Plus, the approach saved hundreds of dollars per patient. "This is an example of an opportunity that has been sitting there for more than a decade," says Dr. David Eddy, founder and chief medical officer emeritus of Archimedes Inc., a private health-care research firm whose work paved the way for the study. "It shows how we can be smarter at determining the right treatments and find clever, simple ways of delivering those treatments."
At the same time, the study also offers a cautionary tale of how hard it is—and how long it can take—to prove that changes in treatment really are effective. This particular story starts way back in the early 1990s, when Eddy began developing a computer simulation he dubbed Archimedes. The idea was to create a SimCity-like world in silicon, where virtual doctors conduct virtual clinical trials on virtual patients.
Modeling a Cholesterol Drug's Benefits
Eddy showed that the predictions of the model almost exactly matched the results from clinical trials. Then it was time to tackle a real-world problem.
At the time, Kaiser was prescribing to its patients what was then a relatively new cholesterol-lowering drug, Mevacor, from Merck (MRK). "We were treating everyone who walked in the door," recalls Dr. James Dudl, diabetes expert at the Kaiser Permanente Care Management Institute. "We thought the drug would do spectacular things."
But maybe not. When Kaiser plugged the data into the Archimedes model, the computer simulation predicted that the net benefit was tiny. "We were treating the wrong population with an expensive drug," says Dudl.
Was there a better approach? Dudl began to think of ways to target high-risk patients, such as those with diabetes and other conditions like hypertension and heart disease. The conventional wisdom was that the best treatment for diabetes was keeping blood-sugar levels consistently low, which would help ward off complications like heart disease. But Dudl wondered what would happen if he flipped that around, aiming treatment at the downstream problems instead of blood sugar. His idea: give patients a trio of generic medicines—aspirin, a cholesterol-lowering statin, and a blood-pressure-lowering ACE inhibitor.
Using Archimedes and thousands of virtual patients, Eddy compared the drug combination to the traditional approach. The model took about a half-hour to simulate a 30-year trial, and the results were startling. Controlling blood sugar accomplished little, but the simple three-drug combination would cut heart attacks and strokes by 71%.
$8 Billion in Potential Cost Savings
At a pivotal meeting of the board of the Care Management Institute in 2003, Eddy presented the results and made an impassioned plea to implement the findings. "I told them, 'This is as good as it gets to improve care and lower costs, which doesn't happen often in medicine,' " Eddy recalls. "'If you don't implement this,' I said, 'you might as well close up shop.' "
Kaiser listened. As reported in the new study, the company prescribed the drug combination to 68,560 Kaiser Permanente members in California with diabetes or heart disease. Researchers followed the fates of those patients for two years and compared them to 101,464 similar patients who didn't take the combination.
The results mirrored Archimedes' prediction almost exactly. Patients who took the drugs some of the time had a 60% reduction in heart attacks and strokes. For those who adhered more closely to the drug regimen, the benefit was an 80% reduction. "We're extremely happy with the results," says Dudl.
Dudl and his fellow researchers didn't do a cost analysis of the program. "Kaiser's motivation was to improve the quality of care and let the cost chips fall where they may," says Eddy. But Eddy wasn't so reticent. By reducing heart attacks and strokes, he calculated, the program saved about $350 per person treated. Multiply that by the number of diabetics in the country (23 million) and the potential for cost savings is huge: $8 billion.
"The general point here is that quantitative thinking is just beginning to enter health care," says Eddy. Imagine, he says, trying to optimize the operations of an airline without having any data on the numbers of passengers, the cost of fuel, and other basic information. The new study shows it's possible to get that key information in health care, and put it to use to both improve quality and cut costs. "This is equivalent to the Wright brothers flight at Kitty Hawk," says Eddy. "It's saying that something can be done."