Medicare rewards for process, patient satisfaction not always linked to better outcomes
The movement to improve hospital quality has long assumed that close adherence to best medical practices leads to better outcomes. But a Modern Healthcare analysis of Medicare's first stabs at rewarding hospitals that do well on process-of-care measures and a patient-satisfaction survey while penalizing hospitals with high readmission rates shows only a minor correlation between the two.
The analysis, which uses Medicare's initial results for the value-based purchasing and readmissions reduction programs, also shows that hundreds of hospitals that did well in one program were penalized in the other.
Dozens of hospitals that got the highest marks for avoiding readmissions were heavily penalized for their performance on process and patient satisfaction, while dozens more that received the maximum penalty for high 30-day readmission rates got high marks on key processes and attention to patients.
The faint correlation between the two programs underscores the limits of the current quality measures, which have been central to the intensifying federal effort to reduce healthcare waste and errors and slow U.S. spending on medical care, experts say.
“The science isn't quite there yet” to identify which organizational processes will deliver desired outcomes such as lower readmission rates, said Dr. Rachel Werner, an associate professor of medicine at the University of Pennsylvania.
While researchers are working to collect data on readmission-reduction strategies, she said, measures have not yet been developed for programs that better coordinate medications, doctors' visits and medical care in settings outside the hospital. The CMS, meanwhile, has finalized plans to include more outcomes-based measures, such as risk-adjusted mortality, in future iterations of its pay-for-performance program.
The CMS announced its first round of rewards and penalties for value-based purchasing, or VBP, in December. It scored hospitals based on their performance on clinical processes such as how quickly a heart attack patient receives an angioplasty or whether a heart failure patient receives discharge instructions. The final VBP score, which raised or reduced Medicare reimbursements by as much as 1%, also included a composite measure of patient experience.
The readmissions reduction program, whose first round of penalties of up to 1% were announced last fall, penalizes hospitals with higher-than-expected numbers of quickly returning patients. Patients who return to the hospital within 30 days of a discharge are widely considered a flag for lapses in care that may have caused their return. Federal penalties are set to steadily increase in coming years for excess readmissions, which are anathema to patients, hospitals and policymakers.
The Modern Healthcare analysis, which looked at the nearly 3,000 hospitals that were graded by both Medicare quality programs, found only a weak likelihood that hospitals would do well in both.
The weak correlation can be illustrated by the fact that hospitals that would be expected to earn penalties in both programs if the measures were closely linked did not. More than 40% of hospitals that faced the maximum 1% penalty for readmissions earned bonuses for value-based purchasing scores on patient-satisfaction and process-of-care measures, a Modern Healthcare analysis of CMS data shows.
Meanwhile, a similar percentage of hospitals that did not face penalties for excess readmissions nonetheless paid penalties for performance on value-based purchasing measures. And among hospitals that earned some penalties on readmissions, the results were more evenly split for value-based purchasing bonuses and penalties.
Data used in the Modern Healthcare analysis does have some limits. The population varies somewhat between Medicare's two incentive programs. Only the performance measures are weighted for hospital size. And the two incentive programs draw data from different time periods to calculate performance.
One likely reason for the weak correlation between the two programs is that the 12 clinical process measures used by Medicare's VBP program do not capture everything hospitals do to promote high quality and prevent repeated hospital visits, some experts say.
“There is limited evidence about the effective treatment and strategies for most acute-care conditions,” said Dr. Mihaela Stefan, an assistant professor of medicine at Tufts University.
She recently led a study team that analyzed hospital scores on more than two dozen process-of-care quality measures and readmission rates for patients older than 65. The results, published in the Journal of General Internal Medicine in October, found hospitals with better marks for process-of-care measures did not see lower rates of readmissions than hospitals with worse marks. Still, Stefan said reporting process measures should continue despite the weak correlation with readmissions, since process-of-care measures are among a limited number of quality measures with some evidence to show they do benefit patients.
Not everyone agrees.
Dr. Ashish Jha, an associate professor of health policy and management at the Harvard School of Public Health, said early efforts to measure quality correctly reported on processes of care. But the more valid measure now is actual outcomes, which is of most concern to patients. “It's time to move on,” Jha said.
Others pointed to the possible disconnects between what goes on within the hospital and subsequent care. The rate at which patients return to the hospital within 30 days measures a system of healthcare that extends beyond the hospital, said Arnold Epstein, a professor and chairman of the Harvard School of Public Health's department of health policy and management. Many factors other than hospital care can prompt a readmission.
For hospitals that have fared well under one program yet floundered under the other, the results are confounding.
MidMichigan Medical Center-Midland, a 250-bed hospital, was deemed a top performer by the CMS in curbing readmissions and received no penalty under that program. Hospital officials credit the organization's success to its participation in several voluntary quality improvement initiatives, including ones led by the state's hospital association and the Institute for Healthcare Improvement.
But MidMichigan Medical Center-Midland was also among 1,423 hospitals that will lose up to 1% of their Medicare payments for 2013 under the CMS' value-based purchasing program. The hospital will have its payments docked by 0.71%, according to federal data.
“You would expect to see a correlation between the two, but in our case, you just don't,” said Kay Wagner, director of quality for four-hospital MidMichigan Health, the parent health system, also in Midland. “Our readmission rates are great but our process-of-care metrics for heart failure, for example, are not.”
Wagner argued that the hospital's different experiences in the two programs illustrate that quality measures—particularly those in use in federal programs—are still in their infancy. The dozen process-of-care measures that account for 70% of each hospital's score under value-based purchasing may not be the best predictors of patient outcomes, she said.
The remaining 30% of a hospital's VBP score is based on a composite measure of the Hospital Consumer Assessment of Healthcare Providers and Systems, or HCAHPS. The survey, an eight-item instrument, collects patient-reported data on experience-related variables such as hospital cleanliness, pain management and communication with physicians and nurses.
MidMichigan Medical Center-Midland was in the midst of a large-scale renovation during the baseline period for 2013's VBP program, which ran from July 1, 2009, until March 31, 2010. Those renovations—and the dust and noise that accompanied them—likely affected the hospital's HCAHPS scores, Wagner said.
Still, Wagner backed the programs and said the federal government's addition of outcomes measures, such as risk-adjusted mortality, in later versions of value-based purchasing will likely make a big improvement. “They're going to learn by trial and error,” she said.
Ukiah (Calif.) Valley Medical Center also ended up as a high scorer in the readmissions program, but fell into the penalty range for the CMS' pay-for-performance initiative. The 62-bed hospital received no readmission penalty but will lose 0.65% of its Medicare reimbursement in 2013 because of its score on value-based purchasing, according to CMS data.
Despite that inconsistency, hospital officials said the readmissions and VBP programs are good initial gauges of hospital quality. “They might not be directly correlated with each other right now, but it's a good beginning and I know they'll continue to improve,” said Heather Van Housen, Ukiah Valley Medical Center's vice president for patient care. “I do think the correlation will become tighter later on.”
Michael Motte, CEO of 189-bed St. Alexius Hospital, St. Louis, said he wasn't surprised by the difference between the hospital's performance in the value-based purchasing and readmissions programs.
Like MidMichigan Medical Center-Midland and Ukiah Valley, St. Alexius faces no readmissions penalty, a success Motte credited to close collaboration with local nursing homes.
But St. Alexius Hospital will see a 0.66% cut in its Medicare reimbursement this year, a penalty that will cost the hospital approximately $81,000, Motte says.
“They're two different programs focusing on different things,” he said of the hospital's contrasting performance. “They're going to have different results.”
Value-based purchasing is based on older data, Motte said, and the hospital has since enacted a number of changes that should boost its future score. They include leadership changes, implementation of health information technology systems and new initiatives aimed at increasing patient satisfaction.
“In my opinion, these are good metrics,” he said, “and I think we're moving in the right direction.”
TAKEAWAY: The weak link between hospitals' performance on two federal quality initiatives suggests there's a long way to go toward improving the metrics.