The researchers collected health status data using 26 metrics for patients admitted to a Connecticut hospital, from the moment patients arrived until they left. Researchers measured improvement during the course of the hospital stay, much as teacher performance may be measured by students' test scores during an academic year.
The novel approach may have broader applications with fewer flaws than existing measures of physicians' quality and efficiency, researchers say. Healthcare performance and quality is often measured by whether providers follow the correct procedure to care for patients. Measures of results—quicker recovery and better health—are less common.
The development of performance measures is a growing priority for employers burdened with ever-rising health benefit costs; public benefit programs with budget constraints; and patients, who increasingly have health plans that encourage comparison shopping. But options are few, and those that exist are limited.
“There's not really good measurements or mechanisms to measure value,” said Ann Boynton, deputy executive officer for benefit programs, policy and planning, for the California Public Employees' Retirement System, which spent more than $7.5 billion last year on health benefits for more than 1 million people. CalPERS has entered into multiple ventures to test incentives for providers to deliver high-quality, lower-cost care.
Boynton said such ventures have shown value for specific procedures, such as births, but lack the measures that can identify which doctors, medical groups or hospitals deliver the greatest overall value. “There's really not a lot out there.”
Hospitals and medical groups, too, are in the market for ratings as public and private insurers move to link a growing share of providers' payments to delivery of high-quality, low-cost care under accountable care and other incentive-based contracts. Medicare introduced accountable care contracts in 2012 and continues to expand their use. Major health systems report a growing number of incentive contracts. Dignity Health, San Francisco, has 74 such contracts with more on the way. Banner Health, Phoenix, Ariz., ended last year with $458 million in revenue from incentive-based contracts.
Ratings may be used to target incentive payments to high-value providers within networks. Some medical groups, hospitals and accountable care organizations already share performance results with providers and award individual performance incentives.
“No surprise, provider groups are aligning their compensation model to mimic how they themselves are getting paid in the marketplace,” said Dr. Marcus Zachary, senior medical director for the San Francisco-based Brown & Toland Medical Group. “That just makes good sense.”
Metrics of performance increasingly take into account results, such as what percentage of doctors' diabetic patients have well-controlled blood sugar, he said. Brown & Toland is shifting its quality improvement focus this year to put more emphasis on results and less focus on process, he said. “That's where the market is going.”
For policymakers, that market shift is considered critical to maintaining a recent slowdown in the nation's historically fast growth in health expenditures. “It's incredibly obvious, even with the slowdown in spending, that healthcare costs are very important,” said Dr. Leora Horwitz, director of the Center for Healthcare Innovation and Delivery Science at New York University's School of Medicine, and one of three researchers that developed the value measure.
But practically, consumers care not just about cost but the results they get, she said. “Quality and outcomes and value to patients is equally, if not more, important.” That's why she and her colleagues Jason Fletcher, a healthcare economist at the University of Wisconsin-Madison, and Elizabeth Bradley, a public health professor at Yale University, sought to develop a measure of results and cost.
Early quality-improvement efforts targeted processes but the research on the effectiveness of those processes has significant limits, Horwitz said. Multiple processes contribute to results and it's often unclear which processes are most critical to good outcomes. “There is a relatively limited evidence base for most things,” she said. “That's not a radical statement, although it's disturbing. First of all, we're not even sure what processes we ought to be measuring.”
Measuring results, however, is a challenge for several reasons. Outcome measures are limited. “Report cards measure mortality or readmission rates,” said Fletcher. And an accurate statistical measure of performance can be undermined when metrics are so discrete and potentially rare among the small pool of patients of any one doctor, said Fletcher.
In the experiment with the new outcome-based measure carried out in Connecticut, researchers compared average improvement for each doctors' patients through a six-month window.(Researchers also looked at average improvement for the doctors during a second six-month period and found doctors' results were largely the same despite a completely different pool of patients.)
Researchers acknowledged that limits of the work include the potential contribution of other hospital caregivers. But the method may still be useful.
“The method allows physicians' added value to be estimated on the basis of every patients' health status rather than on the basis of relatively rate events such as readmission or mortality, and is therefore less sensitive to sampling variation from a small number of patients per physician,” the wrote.