The Charlotte, N.C.-based Premier healthcare alliance has been one of the leading organizations pushing the envelope on quality-based purchasing in the healthcare field for several years. Indeed, Premier healthcare alliance’s core program in this area, the CMS/Premier Hospital Quality Incentive Demonstration (HQID) program, which was launched collaboratively by Premier and the federal Centers for Medicare and Medicaid Services (CMS) in October 2003, has seen more than six years of impressive results, including an overall improvement in measured patient care quality of an average of 17.2 percent over the first four years of the HQID demonstration project, and an estimated 4,700 heart attack patients’ lives saved in the program’s first four years, according to Premier estimates. Meanwhile, a sister initiative, Premier’s QUEST: High Performing Hospitals initiative, has already racked up estimates of 8,043 lives and $577 million saved, as of last October, through that program’s focus on the elimination of avoidable hospital mortalities and on cost savings, across 157 participant organizations.

One of the questions that has been lingering in the minds of some in the healthcare industry has been whether pay-for-performance programs such as those sponsored by Premier can be useful across all types of hospitals, particularly those that care for impoverished and disadvantaged populations, and which typically are resource-poor themselves as institutions. Interestingly, a team of medical researchers led by Ashish K. Jha, M.D., published an article in the Sep. 7, 2010 edition of the Annals of Internal Medicine, entitled “The Effect of Financial Incentives on Hospitals That Serve Poor Patients,” which addressed that very issue. Jha and his co-authors found that, with regard to Premier’s HQID demonstration project specifically, “Among both pay-for-performance hospitals, and those in the national sample, hospitals with more poor patients had lower baseline performance than those with fewer poor patients…[But a]fter three years, hospitals that had more poor patients and received financial incentives caught up for all three conditions, whereas those with more poor patients among the national sample continued to lag.” The authors’ conclusion? “No evidence indicated that financial incentives widened the gap in performance between hospitals that serve poor patients and other hospitals. Pay-for-performance programs may be a promising quality improvement strategy for hospitals that serve poor patients.”

Richard Bankowitz, M.D., enterprise-wide chief medical officer at Premier Informatics, a division of Premier, spoke recently with HCI Editor-in-Chief Mark Hagland regarding the results detailed in the Annals of Internal Medicine study, which was not sponsored by Premier, and the implications of those results for pay-for-performance (P4P) programs going forward in healthcare more generally.

Healthcare Informatics: Were you involved at all in the development of the AIM article?

Richard Bankowitz, M.D.: No, not at all, and in fact, I didn’t even know it was being developed. But we had been looking at our own data in this area. We didn’t go and gather external data as these authors did, but we did an internal study, and found similar results. And we were delighted that those authors came to those conclusions.

HCI: One might assume that one would need a huge amount of resources to participate in and succeed in such programs as HQID, but that’s not the case?

Bankowitz: That’s right, the studies seem to suggest that resource-poor hospitals can improve. Part of this is that there’s a recognition that hospitals really need to address systems of care. A lot of these measures, which are based on evidence-based practices, tend to respond very well to putting in place systems that eliminate the variation that occurs. And the focus that something like the measurement provides, is really the key. And I do think that the financial rewards play some role. We’ve found in our own internal studies, when we’ve looked at those receiving a financial reward, by the third year, there really was no difference between the disproportionate-share hospitals [DSH], and the non-DSH. In the first year of HQID, there were relatively few DSH hospitals participating, but by the third year, their level of participation matched their representation in the industry.

We did find that the level of recognition was different. Our guess was that in some cases, hospitals might be motivated by getting the financial reward, but less so by the chance to get public recognition. Also, being involved in the program itself meant that these hospitals were able to access information on best practices, and we were able to share with them what the best hospitals were doing, so it didn’t require reinventing the wheel. So I think if there’s a combination of sharing with best practices, they can catch up pretty quickly.

HCI: Is there any cultural element to this that has to be addressed in these hospitals, in terms of serving largely uneducated and impoverished populations?

Bankowitz: It’s an interesting question, and I don’t think we have any hard and fast data. But when you start looking beyond data around mortality, and towards data around questions like readmissions, I think that certainly must be a factor. It’s well known that the level of medical literacy is going to vary greatly; and those individuals who aren’t as medically literate may not understand as readily the importance of taking one’s medications, for example. And if you look at some populations and the numbers of prescriptions that don’t get filled, sometimes it’s as high as 40 percent; and if you look at areas in which that is the case, such as asthma, where ED visits or admissions are higher, certainly, whatever socioeconomic factors are at play, they’re going to show up in the ED with greater frequency. In this case, it’s a little bit different, though, because most of the processes under study are under the direct control of the caregiver, in terms of ordering aspirin after a heart attack and so forth. But obviously, in terms of issues like outcomes around complications and readmissions, those factors come into play.

HCI: What does this Annals of Internal Medicine study say about the overall robustness of P4P programs?

Bankowitz: Well, a couple of things. First of all, incentives do work. And whether it involves getting recognized as a high-performing hospital or obtaining the funding, incentives do matter. At the very beginning of the HQID program, adherence to best practices in many areas was abysmal. Look at pneumonia care: improvement in that area [since then] has been tremendous. And overall, we’ve seen dramatic and sustained improvement after several years of HQID. And even though the incentives have been relatively small, we’ve asked some of the hospitals about the resources expended, and they’ll say, we did get the money, but we’ve probably spent more resources accomplishing the goals involved. But these incentives get the chief medical officer and the chief financial officer aligned, for the first time. And so now we can have a conversation for the business case for this; and that’s an important conversation.

HCI: Value-based purchasing was incorporated into the federal healthcare reform legislation passed in March. Did the provisions around value-based purchasing in that legislation reflect the HQID program’s modeling work in that area?

Bankowitz: I do think that the value-based purchasing put forth in the healthcare reform draws heavily on the evidence from the HQID program, because it’s a very broad program, and it’s taken place over six years. I do think it’s important that the healthcare reform legislation speaks of pay for improvement, not just pay for performance. If you compare absolute attainment, it’s true that the non-DSH hospitals did somewhat better. But it’s important not just to look at the attainment level, but also at the levels of improvement as well. And CMS had published a working draft on value-based purchasing before the healthcare reform legislation was drafted, and had already incorporated that concept. The other thing that we’ve felt strongly about at Premier and have seen it make its way into the legislation, is that we wanted it to be basically revenue-neutral, in the sense that, if hospitals don’t attain their objectives, they’re going to lose some money, and rather than having that [forfeited] money disappear into the general coffers, it could instead potentially be used to help hospitals improve; and that is in the legislative language.

HCI: What are the lessons for CIOs in all this work to date?

Bankowitz: There are a couple of lessons. The first lesson would be that, because we know that hospitals were able to perform across a wide variety of settings—some had robust health information systems in place, and some didn’t—I think it’s important that HIT systems follow and not lead, in the sense that you need to address your processes of care first, design your workflow, and then design your HIT system to facilitate that improvement, rather than digitizing obsolete processes. The second lesson, I would say, is that, while it is true that you want to hardwire in important processes and make the right thing easy to do, it’s also worth remembering that whatever information systems hospitals put in are going to need to be fairly flexible and fairly robust. And no one can determine with certainty what the evidence-based practices will be in a few years. So it will be impossible to try to simply design towards specific measurements. You need in all cases to design to eliminate variation and provide for efficient processes, and then no matter what measures are established, you’ll do well.

HCI: Will the energies coming out of the federal American Reinvestment and Recovery Act/Health Information Technology for Economic and Clinical Health (ARRA-HITECH) Act and out of healthcare reform-based value-based purchasing dovetail?

Bankowitz: Well, we certainly don’t want to require measures that will require a lot of manual labor; and whatever measures we develop need to be supported by information systems. And as we go forward and install these systems across the country, it would be a real shame if we couldn’t derive the data we needed to support these programs from automated systems, and had to go backwards into manual collection of data.

HCI: Do you remain optimistic as we go forward in this area?

Bankowitz: I do, for a number of reasons. Premier is in the process of supporting several improvement programs, including the QUEST program, and we’re seeing many hospitals improving in the areas of using evidence-based care, improving quality, and reducing cost. And I’m very optimistic that, given the right kinds of use of tools and the right kinds of collaboration, hospitals can really improve. I also look forward to alignments around the automation of the collection of data. And I think everybody’s on the same page, so I’m optimistic about that as well.