Erik Muther is Executive Director of the Pennsylvania Health Care Quality Alliance.
PND: Your organization recently conducted a survey to learn how consumers use health care quality information. What were its key findings?
EM: With the help of the Center for Opinion Research at Franklin & Marshall College, we surveyed 537 Pennsylvanians online and 375 by telephone to measure their attitudes towards consumer interest in hospital quality data, its impact on the hospital selection process, and their perception about the content and usability of our website, phcqa.org. In response to questions about their most influential and trusted sources of health quality information, the answer is still overwhelmingly doctor recommendations, friends and family, and personal experiences – taken together, those encompassed 60 to 70 percent of the responses. But there is a growing number of people who do look at quality ratings and Internet-based health care information for making health care decisions: between 10 and 12 percent, relative to the other sources.
PND: What are obstacles to more widespread consumer use of online health care quality information?
EM: People use health care information in a variety of ways, but not to the extent that they use information when they purchase other services or products. Most people don’t directly pay for their health care – they have health insurance – so in most cases they’re not directly purchasing at the “retail price” the services that they get. So, they tend to be focused more on ensuring that they get the service that they want, and don’t spend as much time looking up information to ensure that the service is of good quality relative to other services of a similar type. Also, each government or private insurance entity provides a source of quality information that tends not to be very standardized. One health plan may have a contract with HealthGrades, for example, whereas another health plan may have WebMD Health Services to provide health care quality information. There’s not a clear, single source for people to go to and obtain quality information. When things tend to be more disparate and less intuitive, it becomes overwhelming and a little daunting. In our survey, there was a very large number of people who came to our website and were surprised that it was relatively easy to use, that it was not more confusing and complex. That either means they had a perception that these sorts of websites are very confusing, or they had not been to other websites and were surprised to see that it was actually relatively straightforward.
PND: What is the goal of your organization?
EM: The Pennsylvania Health Care Quality Alliance is a multi-stakeholder collaborative of organizations that came together to develop a common statewide approach to transparency of hospital quality measurement using measures that are evidence-based and actionable. The goal was to help providers evaluate and improve the quality of patient care, give insurers the opportunity to assess and evaluate the performance of their provider networks in a common way, and ultimately enable consumers to see how their provider – hospitals, in this case – was performing, and also benchmark that performance against other hospitals. The website was the output of consensus measures the alliance developed. The alliance’s member organizations include the four Blue Cross Blue Shield plans in Pa. – Highmark, Independence Blue Cross, Capital BlueCross, Blue Cross of Northeastern Pennsylvania, the Hospital & Healthsystem Association of Pennsylvania, the Delaware Valley Healthcare Council, the Hospital Council of Western Pennsylvania, the Pennsylvania Medical Society, representatives from the Governor’s Office of Health Care Reform, and from the U.S. Department of Health and Human Services.
PND: How do the interests of these various stakeholders coincide, and how do they diverge?
EM: One alignment is an agreement that an effort to develop a standardized quality approach should build upon existing state, federal, private and public data sources, for example, using the Pennsylvania Health Care Cost Containment Council’s measure for hospital-acquired infections. The alliance would look at that measure, examine how it is collected and reported, and identify whether it should be adjusted or left the same. The goal is to avoid duplication of existing measures and not unreasonably burden hospitals with additional reporting standards. Payors and hospitals would then use the agreed-upon measure set as the primary basis for future quality improvement efforts. For example, if Independence Blue Cross is going to partner with the University of Pennsylvania to promote health care quality, to the extent possible, they should try to leverage the standard approach that was developed by the alliance, rather than having to recreate effectiveness measures for their program. Similarly, for pay-for-performance programs, there should be some attention paid to using these quality measures that they’ve collectively agreed on. Another issue of alignment is that the alliance should promote ways for lesser performers to learn from better performers – create a forum to share best practices – and not be just another ranking and reporting type of activity.
Points of divergence were about where price and cost comes into play – whether the alliance focuses on quality and cost in parallel, or separately. There was agreement that quality should be the primary focus and that, once we developed a consensus on how to measure quality, then we could start to entertain questions about how cost factors in, such as whether getting higher quality requires more resources. Another divergent point was how to calculate and identify who the benefits of quality care will go to, other than the patients, obviously. The question of cost savings coming from that improved care was a divergent point. Some of the hospitals felt that there should be additional compensation provided to them if they significantly improve quality, because it would be the health insurance company that benefits from the resulting lower cost. The perspective of the health plans was that it is often difficult to measure what the lower cost is of that improved quality because quality is measured across a fairly extensive period of time. There are member turnover and other issues, where people roll in and out of insurance, and there isn’t always a clear benefit to the insurance company from that improved quality.
PND: What about metrics, themselves – have there been instances in which insurers wanted to include a metric that hospitals objected to?
EM: There are two that come to mind. There are a few National Quality Forum (NQF) efficiency measures, and the health plans were interested in trying to introduce an efficiency measure into this agreed-upon approach to health care quality. The hospitals were hesitant to do that, primarily because the efficiency measures haven’t been vetted over an extensive period of time, and there is a question about whether they really measure efficiency well. The other category was health care-associated infections. Certain individuals on our measures and methods workgroup, which is made up of physician leaders from both hospitals and health plans, disagreed on the utility and usefulness of some PHC4 infection measures, primarily because of methodological concerns related to data collection. For example, surgical site infections that are counted and tracked by PHC4 are only those that are actually reported in the hospital. But, if somebody two or three days after discharge notices they are having swelling or discoloration in the area they had surgery – if their doctor looks at it and gives them antibiotics for an infection, the hospital may never know about that. You would be missing a lot of infections and there is no way in the PHC4 model to be able to count that data.
PND: Were any metrics excluded because they failed to meet the alliance’s “evidence-based and actionable” criteria?
EM: Two categories have been delayed, not necessarily discarded. One is preventive measures to address ventilator-associated pneumonia (VAP), many of which have conflicting clinical literature. There wasn’t a clear actionable list of ways that hospitals could specifically reduce their VAP rate, and we wanted to wait until the evidence had borne that out a bit more. The other delayed category is urinary tract infections (UTIs) because there wasn’t an adequate risk adjustment for that measure. Clearly, if a hospital has a larger number of patients who are in the ICU or are spending long periods in the hospital, they are more likely to develop a UTI related more to being on a device for a long period of time, rather than to direct quality of care issues. We felt it was worth waiting until there is a way to present that in a way that is clear and actionable for hospitals to address.
PND: What quality information is the alliance focusing on, and how is it different from what is already available from other entities?
EM: The measures cover heart attack, heart failure, pneumonia, infection prevention, and appropriate care measures – which indicate the percent of patients who received all of the evidence-based care for their respective conditions. Within each of those categories we have outcome and process measures. The process measures come primarily from the Centers for Medicare & Medicaid Services all-payer data (not just Medicare patients), and most of our outcome measures come from PHC4. We try to identify measures from all domains of quality: effectiveness of care, timeliness of care, safety issues, and patient experience data. We’re trying to get as broad a set as possible, and all of our measures are NQF-endorsed, except for readmission rate for heart failure for the reason of complication or infection.
The alliance convened for the first time in January 2007, so it is still in the early stages of its activities. But, as a first pass, there was a focus on avoiding recreating the wheel and duplicating areas that had been fairly well-covered by other reporting entities or sources of data. As we begin to look into areas where there has not been a long history of data collection or public reporting, we may have to look at other sources or collectively agree on how to measure something. At least for the moment, we’re more of a clearinghouse for measures that are otherwise available in other locations – offering data aggregation into a “one-stop shop” that adds value to users. The exception is the appropriate care measures, which are patient-level composite measures that our state quality improvement organization – Quality Insights of Pennsylvania – calculates for us because they have access to that patient-level database.
The decision to focus on hospital quality measures was primarily driven by the long history that Pa. has on hospital reporting to PHC4, and there was already groundwork laid by a variety of other initiatives like the Hospital Quality Alliance. So, it seemed like a logical starting point. There is an understanding that, at some point, our alliance will look at other settings of care, including outpatient and perhaps long-term care facilities.
We’re starting to go through the next set of potential measures to be included in a formally adopted consensus measure set of health care quality indicators for hospitals. We are looking at some that have recently gone through the NQF process and in some cases have already been adopted by CMS, for example, more granular mortality measures for heart attack and pneumonia. We may look in more detail at measures that are reported to a variety of physician specialty organizations, such as the Society for Thoracic Surgery database and the American College of Cardiology clinical registry.
PND: How is the alliance’s information disseminated?
EM: It’s disseminated through a website. According to our survey results, such a large number of health care decisions are made from doctor recommendations and from personal experience of friends and family, so it probably doesn’t make sense to spend a large amount of time or resources on direct-to-consumer marketing or communication. That’s sort of a double-edged sword because you also want the end consumer to become more aware of these sorts of resources. But, if they’re primarily making most of their decisions in consultation with their family and their physician, it makes more sense to explore ways to promote what the alliance is doing within the physician and professional realm.
Most of the national surveys that I’ve seen, including ones sponsored by the California Health Care Foundation and the RAND Corporation have shown that the number of consumers who look up health care quality information to help them make a health care decision has historically been in the mid- to low-single-digit percent range. We saw 10 to 12 percent on ours, and the Kaiser Family Foundation in a recent survey had seen their number break into the double digits for the first time. We’re still talking about a low number, but a very fast growth rate, which is fairly encouraging.
While our goal in creating the website was to communicate this information to the consumer, the real work of the alliance is collaboration to come up with a consensus measure set. Since that is the primary focus, and not necessarily to create a consumer resource, I think the alliance for the moment is comfortable with not spending a lot of time and resources in trying to raise awareness of the site’s availability.
PND: Is your primary focus the provider community, as a tool to improve their quality?
EM: I would say yes. It’s the providers, primarily, who are looking at this information and using it for benchmarking and other purposes. Right now our aggregated data can be looked at in a variety of ways: most recent, quarterly, and by individual hospital as far back as the data goes. What we are going to do, probably in the next year, is create something that allows for providers to do even more sophisticated analysis. One of the decisions the alliance is going to need to make fairly soon is how much more we want to invest in that capability. There has been some discussion about whether it would be worth collecting measures from hospitals for benchmarking purposes only, for example. There may be some measures that could be of tremendous value to hospitals or physicians to understand where their institution or practice is relative to others, both for the purposes of benchmarking and quality improvement, but maybe measures that you don’t necessarily want to get published until it’s understood how best to compare those measures across providers. Physicians and hospitals have a very good understanding of what happens within their own environment, but a very poor understanding of what happens outside of their environment. The only entities that can really help to fill that gap for them are the insurance companies, who follow people regardless of where they get care. Finding a way to align data and provide mutual data exchange between these organizations is going to benefit health insurance companies, hospitals and physicians.