Breaking News

Medicare’s pay-for-performance

Joseph Cacchione, M.D.
Joseph Cacchione, M.D.

By Miriam Reisman.


Twelve hospitals in Pennsylvania are participating in one of the largest pay-for-performance (P4P) demonstrations in the country. The Centers for Medicare and Medicaid Services (CMS) has entered the P4P arena with several initiatives aimed at improving the quality of patient care, including the Premier, Inc. Hospital Quality Incentive Demonstration (HQID), a three-year project that began in 2003 and involves 266 hospitals across 38 states, including 12 from Pa. Experience from this project is expected to help shape the nationwide value-based purchasing program authorized under the 2006 Deficit Reduction Act and scheduled to be launched in fiscal year 2009.

Recently released second-year results of HQID show that the program has made significant progress. According to CMS, patients treated at hospitals participating in the program are living longer and receiving recommended treatments more frequently. Citing these results as solid evidence that pay-for-performance works, CMS, working in partnership with Premier, Inc., a nationwide organization of not-for-profit hospitals, has extended the program for three more years.

Researchers have begun to analyze the impact of the program, raising critical questions about the value of pay-for-performance efforts, the return on those efforts for hospitals and providers, whether the benefits outweigh the costs, and whether financial incentives are an effective way to stimulate quality improvement.

Participating Pa. hospitals present the HQID project as a promising strategy to improve the care of their patients, describing significant changes in care processes that have led to improved outcomes and, in some cases, potential cost savings. Their experiences also reveal important challenges that accompany the pay-for-performance movement, such as engaging physicians in quality improvement activities linked to P4P and sustaining improvement with adequate information technology and other infrastructure.

Demo Details

Under the HQID program, participating hospitals report process and outcome measures in five clinical areas – acute myocardial infarction, heart failure, coronary artery bypass graft, pneumonia, and hip and knee replacement – and are scored on a set of 33 standardized and industry-recognized quality indicators related to each condition measured. According to Premier, the clinical areas included in the program were each selected based on the fulfillment of two main criteria: national consensus on evidence-based measures of care and the availability of a clinical population at both high volume and high cost under the Medicare program.

Hospitals are ranked and rewarded using a composite scoring method. First, they are individually scored on the quality measures related to each clinical condition measured in the project. The score for each disease category is determined by dividing the total number of achieved interventions (numerator) by the total number of opportunities for the same targeted interventions (denominator). An opportunity for targeted intervention exists whenever a patient meets the criteria to be included in the target population for a particular measure. These individual scores are then “rolled up” into an annual composite quality score, which is used to establish baseline performance and identify the hospitals’ decile ranking.

Under its current payment system, CMS rewards only those in the top 10th and 20th percentile (1st and 2nd decile) of high-scoring hospitals in each of the five clinical areas. Hospitals scoring in the top 10 percent for a given set of quality measures in Year One receive a two percent bonus of the diagnosis-related group (DRG)-based prospective payment for the patients with the condition among all Medicare fee-for-service (FFS) beneficiaries, and hospitals in the second decile receive one percent added to their Medicare payment for that condition. Additionally, hospitals in the top 50 percent of each clinical area receive public recognition on the CMS website.

In the third year of the demonstration, those hospitals that do not meet a predetermined baseline performance on quality measures are subject to financial penalties. The baselines are clinical thresholds set at the Year One cut-off scores for the lower 9th and 10th decile hospitals. Hospitals receive one percent lower DRG payment for clinical conditions that score below the 9th decile baseline level and two percent less if they score below the 10th decile baseline level.

CMS recently acknowledged that its current bonus structure does not provide enough incentives to those average hospitals that make efforts to improve, and the agency plans to implement a new three-level financing structure that will take effect in the coming years. The top 20 percent will still be rewarded, but under the new structure, hospitals that exceed the median level of performance established two years ago will receive incentives. Rewards will also go to lower-performing hospitals that make considerable quality improvements.

Promising National Results

In its second-year-end report, CMS reported that the pilot program has “steadily improved the quality of patient care” according to the latest results of the program. The average improvement in the quality performance scores of the program’s second year was 6.7 percentage points, and the total improvement over the first two years was 11.8 percentage points. Additionally, according to CMS, the gap between lower-ranking hospitals and top-ranking hospitals is closing as hospitals with lower quality scores continue to improve.

Comparison data for the two-year study period show that participating hospitals overall experienced nearly 1,300 fewer deaths in treating heart attack patients, and they generally have scored higher on quality measures than other U.S. hospitals. In addition, patients received approximately 150,000 additional recommended evidence-based clinical quality measures, such as smoking cessation and discharge instructions. Thus far, CMS reports, it has awarded incentive payments of $8.7 million to 115 top-performing hospitals, representing the top 20 percent of hospitals in each of the project’s five clinical areas.

Challenges and Improvements

Thomas G. McCarter, M.D.

Coronary artery bypass graft (CABG) presented Main Line Health with another opportunity for improvement, particularly around the measure associated with appropriate administration of prophylactic antibiotics prior to surgery to prevent surgical wound infections. Three of the health system’s hospitals – Bryn Mawr, Lankenau, and Paoli – have reached 100 percent of the P4P goal and maintained this level of performance. Paoli also reached 100 percent in two additional measures: prescribing aspirin at discharge and discontinuing antibiotics within 24 hours after surgery. While the cardiothoracic surgeons were already heavily invested in improving perioperative antibiotic use, says McCarter, the project provided motivation and allowed them to focus more on the details of the process, from ordering the antibiotic to administering it within the recommended time to providing the proper documentation.

The challenge of physician buy-in is a universal one, and because the HQID program rewards hospitals and not physicians, participants have had to develop their own incentives to motivate staff. For instance, report cards that evaluate the performance of physicians have played a significant role in promoting pay-for-performance activities at Meadville Medical Center, according to David E. McNamara, M.D., vice president of medical affairs. In response to the HQID program, as well as several other pay-performance projects the hospital has undertaken, Meadville, a community hospital in Crawford County, Pa., implemented a highly detailed report card system that evaluates on a “real-time” basis the physician’s quality of care processes for each individual patient. In addition to showing the physician’s compliance percentage for each relevant clinical measure, the report card explains each measure, looks at how the physician scores against both internal goals and state and national benchmarks, and makes suggestions for a path to improvement.

According to McNamara, the report cards were initially met with some resistance by physicians, who questioned the validity of the data. The hospital responded by providing physicians with more transparent, timely and detailed information about their scores. Building the infrastructure for this system involved manual data collection, a tedious and labor-intensive task, but one that was well worth the effort, says McNamara. As the system became more sophisticated and the accuracy of data improved, he says, the report cards were better received by physicians and helped improve their performance. The hospital also provides free gas cards valued at $25 to physicians achieving 100 percent compliance in any of the clinical areas for one fiscal quarter. In addition to these small financial incentives, says McNamara, peer pressure also plays a factor in influencing physician behavior. “When they get their report card, they know where they stand against their peers, against the hospital, against our goal, and against our benchmark.”

Jeffrey Jahre, M.D.

Like a report card, the profile report identifies for each physician, among other information, which P4P quality measures are not being met. “We can profile right down to the individual physician practitioner who was responsible for the care provided,” notes Jahre. For instance, in the treatment of hip and knee replacement patients, a profile will identify each case where a physician may not have properly administered an antibiotic within one hour prior to surgery. The reports are sent to the chief of anesthesia, the physicians and the administrator of the hospital. The chief of anesthesia reviews these cases one-on-one with the physicians in order to help them effectively assess and improve their performance on these particular measures.

Closing the feedback loop is the most effective way to ensure physician engagement, says David Nash, M.D., chair of Jefferson Medical College’s Department of Health Policy. “If you give physicians non-punitive information regarding their individual performance, relative to a local peer group, and it’s reliable and thoughtful data, they will participate to make the appropriate changes in their day-to-day work.” In addition, he says, hospitals should provide physicians with appropriate education and training to help them better understand pay-for-performance, as well as give them the tools to make their own improvements. There are multiple ways for hospitals to align incentives between the hospital and the physician, says Nash, such as investing in an electronic medical record system and hiring additional staff such as physician extenders and hospitalists.

Patient outcomes are determined not just by what physicians do, but also by nurses and care coordinators. St. Luke’s, along with other hospitals, has been relying heavily on nursing resources to meet their P4P performance goals. “We’ve learned that interventions are not just physician-related interventions. Some of the interventions are nursing-sensitive,” says Jahre, citing smoking cessation advice and counseling, and discharge instructions for heart failure patients. In order to achieve top decile performance and improve index scores, notes Jahre, hospitals must be able to perform across every measure within a clinical condition. In order to accomplish this, St. Luke’s sets performance goals for the nurse managers of their patient care units much the same way as they do for physicians. The nurses receive their performance feedback data on a monthly basis, and their supervisors are able to utilize this objective data in their annual review.

One of the key lessons that Jahre has learned from the pay-for-performance experience is that, unless a unless a patient care intervention is specifically designated as the responsibility of either the physician or nurse, no one takes ownership and the intervention is not performed. “If you make something everybody’s responsibility, then it’s nobody’s responsibility.” To remedy this problem, the hospital assigns certain interventions to physicians and others to nurses. For instance, smoking cessation advice and counseling for heart failure patients at St. Luke’s is now a part of the nurse’s discharge instructions. In certain instances, says Jahre, a team of nurses might be assigned to perform concurrent evaluations of each patient’s chart to assure that all of the various measures for the heart failure are being met. While this process has proved to be both costly and time-consuming, Jahre says it has been effective in heart failure patients and may eventually be implemented in the other clinical areas as well.

Nurses are also playing a critical role in heart failure treatment at Meadville, where the CHF program includes a patient care coordinator who is responsible for conducting all patient follow-up and documenting patient status in the database. The patient coordinator, a registered nurse, works with the physicians and caregivers while the patients are in the hospital, to ensure that each of the core measures is met and also provide follow-up after discharge. A patient’s weight two days after discharge is a primary determinant of readmission in CHF, explains McNamara, so follow-up is critical in the patient’s overall treatment. “If the weight was going up, or there was a red flag, we intervened really quickly.” With such interventions, McNamara reports a decline in CHF readmission from 11 percent to four percent, along with dramatic cost savings.

As the P4P participants have learned, data plays a huge role not only in positively influencing physician performance, but also in helping to determine hospital ranking for incentive payments. “This whole program is dependent upon good or accurate physician documentation,” says Joseph Cacchione, M.D., executive vice president of Saint Vincent Health System in Erie, Pa.

According to Cacchione, one of the potential pitfalls in pay-for-performance reporting is identifying the patients who should be included in the denominator, or the number of all eligible patients who should have received the prescribed treatment indicated by a particular measure, as well as the patients for whom the treatment is not indicated due to adverse events or other reasons. For example, if a hospital reports 100 patients with congestive heart failure (denominator), and only 80 of those are on an ACE inhibitor (numerator), the rate for that performance measure is 80 percent. However, if it is documented in those 20 patients that there was a contraindication for that ACE inhibitor, then the denominator becomes 80 and the performance rate is 100 percent.

In an effort to assure accurate physician documentation, Saint Vincent utilizes specially trained outcomes care managers to remind physicians to document patients who are not candidates for guideline-based care so that they are not counted in the denominator. “That is a huge part of any pay-for-performance program,” says Cacchione. “You want to make sure that you are being graded on what you should be doing and not on erroneous data.”

According to Cacchione, care management teams have also been helpful in encouraging physicians’ adherence to guidelines and removing variability in care. In response to the CMS program, as well as various other P4P projects that it has undertaken, Saint Vincent implemented a “quality plan” in which outcomes care managers are trained to remind physicians to perform recommended interventions – a “check-double-check” system that Cacchione compares to an airline pilot’s checklist before takeoff and landing.

Based on the American College of Cardiology’s Guidelines Applied in Practice (GAP), the program has been successful in helping Saint Vincent reduce its mortality rate in acute myocardial infarction, one of the HQID quality measures. According to Cacchione, since implementing the GAP program, the hospital has experienced a reduction in its AMI mortality from 12 percent to approximately seven percent. “Not only are we giving patients the right medicines at the right time and under the right circumstances, but we’re improving the chance that they’ll survive, in this case, a heart attack.” The hospital has also implemented guideline-based care in congestive heart failure, and while guideline adherence has not improved the mortality rate in CHF, it has contributed to reduced readmission and length of stay for CHF patients.

Cost-Benefit Issue

While many of these pay-for-performance efforts are ultimately intended to reduce health care expenses, as most of the hospitals noted, the changes required to achieve this goal are quite costly. First, in order to participate in the program, hospitals were required to purchase Premier’s database software program, a cost estimated by one hospital at $300,000. In addition, costs for meeting pay-for-performance goals include developing new IT-based resources, producing patient education materials, increasing staff hours, and hiring additional personnel to perform tasks such as concurrent coding, abstracting data from medical records, and documentation of patient encounters. Many of the hospitals say they have not yet looked at the implications of these added costs.

“Do we save by doing these systems better? That’s the next phase,” says Geno Merli, M.D., chief medical officer of Thomas Jefferson University Hospital in Philadelphia. Merli says that, while the hospital has been devoting significant resources to accomplish the P4P targets, its main focus has been exploring how to set up the best systems to deliver care services.

Data generated from the program has been tremendously useful around these efforts, says Merli. “Focusing on data and producing data helps us to begin to look at how to do delivery of the care. It makes us think about care with a much more focused attention.” For instance, the hospital’s data on smoking cessation revealed that physicians’ compliance with this measure was less than optimal. As a result, the hospital enhanced its physician order entry system to include a pop-up screen to remind physicians about patients for whom this intervention should be performed. The physician can then choose to perform the intervention him or herself or refer the patient for consultation in the hospital’s smoking cessation treatment program.

According to Merli, the data has helped the hospital to establish better systems in several other measures, including time to antibiotics for patients with pneumonia, door-to-needle time for patients with acute coronary syndrome, ACE inhibitors in CHF patients, and perioperative antibiotics for hip and knee patients. Most of the systems are being developed so that they are sustainable, says Merli. “Perhaps once we improve in those areas, then we move onto other areas to achieve targets.”

It may be too early for hospitals to assess the actual cost savings due to the improvements in care, and some say the financial rewards for performance barely offset the costs of participating in the program. So, what is the return for hospitals in pay-for-performance? This is a question currently being addressed by researchers and analysts. In a study published earlier this year in the New England Journal of Medicine, researchers compared results over a two-year period from 207 hospitals taking part in the CMS project with 406 hospitals that were participating in public reporting of data but that were not offered financial incentives. Results showed P4P hospitals achieved only modestly greater improvements in quality than did hospitals engaged only in public reporting. The study authors say that additional research is required to determine whether the benefits of pay-for-performance programs outweigh their costs.

“While there is little published evidence about the specific economic return on investment for this program, process measures of quality clearly have improved across the board, and we are going to infer that clinical outcomes have improved as well,” says Nash. That CMS rewards hospitals based principally on process improvement rather than on clinical outcomes improvement has more to do with the quality of the tools than anything else, he says. “We don’t yet have granular enough outcome measures at the individual patient level to which we can readily attach incentives. However, we should not sacrifice the good on the altar of the perfect. And therefore, overall, the CMS program is a major step in the right direction.”

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.