By Christopher Guadagnino, Ph.D.
Ronald A. Paulus, M.D., is executive vice president and chief technology and innovation officer for Geisinger Health System.
PND: What was the genesis of Geisinger’s Care Model Redesign project, which was profiled in the September/October 2008 issue of Health Affairs?
RAP: Our board of directors asked our management team and senior clinician leaders to do something novel and valuable for our patients, and change the paradigm from some of the current, perverse payment incentives to something that was much more aligned across different stakeholder constituents for delivering the best care at the best possible price – basically increase the value that we offer to our community. In a way, it was a response to various pay-for-performance programs, in which each payor was coming up with their own plan of what “good care” was and choosing metrics that were often simply process-oriented and not outcome-oriented. The question was, Could we as an integrated health system and as a provider-led initiative come up with a better way of doing pay-for-performance?
PND: What are the core principles of the initiative?
RAP: Several-fold. First, we want to apply the most recent evidence about what should be done for patients to give them the best possible outcomes, translating the medical literature’s evidence and/or industry consensus into process steps that can be used to get patients all the things that they need, and hopefully none of the things that they don’t need during a care encounter. Number two is alignment and rationalization of incentives so that doing the right thing is actually rewarded, rather than punished. The third core principle is a multidisciplinary team-based approach to care: physicians, nurses, pharmacists and other care team members are interacting in a coordinated fashion. Fourth is that the right thing to do is “hard-wired” into the process, so that by using our electronic health record and other electronic infrastructure, we don’t rely on individual heroism or goodwill or memory to see that these right things get done. Lastly, the patient and his or her family is actively engaged in the care process and their preferences are taken into account, communicating in a peer-like way with the caregiving team. All of these principles are applied within a broader framework of accountable parties who are leading these initiatives, metrics of performance that can be tracked and trended over time, and a feedback loop that can enhance performance.
PND: What was the process used to develop these projects?
RAP: We started with a clinical business case that lays out the benefits we expected to accrue from an outcome standpoint, a patient satisfaction standpoint, and an efficiency standpoint. The second aspect of our process is that it’s collaborative, and we start with a multidisciplinary team that includes physicians, nurses, nurse practitioners, physician assistants. The third, and maybe most important aspect of all is that we started our innovation initiatives in what we call the “sweet spot” at Geisinger: focused where we can collaboratively work with Geisinger Health Plan and our provider group and focus first – not exclusively, but first – on the patients for whom we provide the majority of care and for whom we pay the majority of care. That sweet spot enables us to tweak incentives, look for alignments, and redesign our clinical processes around what adds the most value to the patient and to the system. We rolled out the clinical part of the programs to all patients, regardless of payor, and we can systematically choose whether to roll out the economic parameters to the other payors depending on our contract and on their level of interest. We try to incorporate very specific targets for our redesign initiatives so that we know what constitutes success or failure. We take pieces of care process improvement methodologies, like Six Sigma or lean reengineering, and try to learn as we go, reusing things we’ve deployed in other projects in the overall innovation architecture.
PND: Can you describe the specific projects involved?
RAP: Our Personal Health Navigator medical home demonstration project, which involves accountable primary care physician practices, is designed to help coordinate and facilitate care for a patient and their family – whether it is provided directly by that practice site, referred specialty care, end-of-life planning – over an entire lifetime. For each patient, there is a primary care physician “team captain” and a group of other caregivers and facilitators: nurses who work in that practice, mid-level practitioners like nurse practitioners and physician assistants, front office staff. They work to: (1) make sure that patient’s preventive health care, acute care and health maintenance are up to evidence-based standards; (2) make it comfortable and desirable, so that that patient or family can feel comfortable raising issues of concern, sharing personal information and the like in a way that achieves their outcome goals; and (3) deliver care in an efficient, close-to-home manner. Our doctors, our care management nurses and our health plan care managers all use the same information available electronically through our electronic health record.
The physician gatekeeper care model of the 1990s was basically designed to prevent the patient from getting referrals, or to keep costs under control with capitated payments. Instead of capitated reimbursement, we pay primary care practices fee-for-service, which encourages the doctor to bring the patient in to be seen, and to be more involved in their care. We didn’t give them a financial incentive not to refer, but we created a gainsharing pool of dollars, based upon how efficiently that patient was managed over the course of the year on a total expenditure basis – including primary care, specialty care and hospitalization. The pool was created based upon savings by using a more efficient, value-based referral network of low-cost specialists, imaging facilities and other ancillary providers. However, our doctors and practices can’t earn a dime from that pool unless they hit a variety of agreed-upon quality metrics. Did they properly screen the patient? Did they get them on the right preventive medicines? Did they control their blood pressure? Did they control their lipids? Did they control their blood sugar, if they’re diabetic? What was the hospital readmission rate for their patients? We are about to start tracking formal patient satisfaction survey information. So, the pool was created based upon value, but access to the pool is based upon quality.
We also provided a 24/7/365 personal navigator – almost like a concierge – that they can call, who is always going to get the patient to the right person. We also funded, through the health plan in advance, a dedicated care management nurse embedded within each practice site, as well as payments directly to the doctor and the practice for all this extra work that we’re asking them to do. Those are sizable amounts: about $1,800 per doctor per month, and about $5,000 per practice, per month, per thousand Medicare members – on top of the existing fee-for-service reimbursement. In return, we’re asking these practices to totally change the way they’re caring for people, such as staying open extended hours and on Saturdays, so that things that might have gotten shunted off to the emergency room are now being cared for by a patient’s regular doctor. We provide feedback data to all our practice sites that compare each site to all the other sites, to the best practices within Geisinger, and to where they’re trending. There is some peer competition there, but it’s also how can they learn from what’s working at the other sites.
Our thesis was that patients with multiple chronic diseases, who are using most of the dollars in the health care system, probably need to get more care, not less, and if they got more care up front, they would utilize fewer hospital admissions and high-cost services that really swamp the payment model. What we found is that, yes, the number of their office visits goes up, and their pharmacy costs go up. But that is more than compensated for by a decrease in hospital, nursing home, and other related costs. In preliminary data with two practice sites during the first year, we saw a 20 percent reduction in all-cause hospital admissions – readmissions as well as new admissions – and a seven percent reduction in total medical spending. We’ve since expanded the program from two to 30 different physician practice sites – including a non-Geisinger multispecialty group – and we’re seeing similar results in the expanded sites: a marked reduction in all-cause admissions, an increase in pharmacy spending, and an increase in office-based spending, netting out to reduced cost, overall.
PND: How essential is an integrated delivery system to the success of this program?
RAP: From a health plan perspective, we have the incentive to try to expand this wherever the health plan insures patients, regardless of whether Geisinger is providing the care or not. We believe this is translatable beyond an integrated delivery system because, in our scenario, we’re not acting as a delivery system, we’re acting as an insurance company that is working with other aspects of the delivery system which are non-Geisinger to create a similar medical home model. And we’re applying a very similar strategy, which is an embedded care management nurse, stipends related to the extra work we’re expecting practices to do, and so on. Our hypothesis is that you don’t need an integrated delivery system to do this, you just need a health plan and a physician practice that are willing to work collaboratively together, with the health plan realizing that they’re going to have to actually pay for these extra services, and the physician practice realizing that they’re going to have to work in a different way and be accountable for delivering quality and value together. We don’t have the same benefit of a common electronic health record platform there, so we’re looking at workarounds for how to do that – other ways of sharing claims data from the health plan with the doctors, and sharing some of the clinical information from the doctors with the health plan. We firmly believe that other insurance companies and other practices that don’t have anything to do with Geisinger could recreate a model like this, but those health plans have to be willing to fund these kinds of transformational changes because the practices don’t have the money to do it on their own and, even if they had the intellectual incentive – knowing it’s the right thing to do – they can’t do it without that help.
PND: What was the second care innovation project?
RAP: The chronic disease care optimization program was the second of three. We looked at a variety of diseases, including diabetes, chronic kidney disease, coronary artery disease, and most recently disease prevention, where we apply the same principles: what does the evidence say about what should be done for these patients and how can we ensure that all of those things are done 100 percent of the time. There are about eight to 10 performance metrics per disease state, except for the prevention bundle, which has significantly more. Here again is an intersection between our ability to mine our database to identify where the gaps in care are; to hardwire that into our electronic health record, including standard order sets; having our nurses check where the gaps in care are while they’re rooming the patient and prepopulate an order set that can go to our clinicians; tracking and setting up automated reminders. We made dramatic changes in the percentage of patients that are achieving a variety of process and outcome metrics. Then we align financial incentives, giving physicians the opportunity to earn a bonus by meeting targets for improvement. Our physicians are salaried, but they have on average about 20 percent of their compensation at risk, at least half of which is based upon quality measures – their ability to achieve substantial improvements in population-based health measures, and how they’re doing as a team.
PND: Would that model transfer to non-Geisinger physicians?
RAP: Other health plans have their own pay-for-performance programs for non-Geisinger doctors, but for the typical fee-for-service doctor out there, there is no way that they’re going to have 20 percent of their compensation at risk, or half of that 20 percent based upon quality. That’s just not the way that the world works and I think it’s one of its deficiencies. That’s one of the things that policymakers need to grapple with. Physicians are important professionals in our society, and they should be rewarded for outcomes, rather than being paid for widgets of service. We also have this perverse incentive where physicians and hospitals are, frankly, often paid more if things don’t go well. Hospitals are paid for a readmission that could have been avoided. Doctors are maybe paid for follow-up care from a problem that potentially could have been avoided.
PND: What results have you seen from the program?
RAP: We’ve seen statistically significant increases in many measures across time and we have more than doubled, and often tripled, the percentage of patients getting all the relevant measures associated with their disease. We’ve compared how the Geisinger doctors compared to non-Geisinger doctors who are serving the health plan, and with statistical significance we’re better 60 to 80 percent of the time than the non-Geisinger doctor group. That improvement has been steadily increasing from quarter to quarter.
PND: What is your third innovation project?
RAP: It is our ProvenCare acute program for those patients who, despite great care coordination, the medical home, and chronic disease care optimization, still need some major intervention – specifically, elective coronary artery bypass surgery. How do you optimize that intervention? The principles here are again the same: a multidisciplinary team, defining outcome goals – quality and efficiency metrics, and hardwiring into the electronic health record, and incentive alignments. We devised a single price for all services across an entire 90-day episode bundle – the hospital fees, doctor fees, consultant fees, follow-up fees – as the incentive package. The episode was defined from the start of the office visit where the decision is made to have surgery through a 90-day post-discharge from the hospital. That rewards the buyer because we looked at the historic readmission rate and added only half of that cost to the price, so the buyer of care is 50 percent better off on readmission cost, on average, than they were before the program started. Their payment is also completely predictable, and they have no risk of outliers – those unusually high-cost cases. At the same time, we had a financial incentive on the hospital side to get readmission rate and other costs down, because 50 percent of that amount was still bundled into the price. We are taking risk, at the provider level, for all of the care, any complications and any related readmissions for 90 days, which the press labeled as a “warranty.”
PND: How is the single fee distributed among clinicians?
RAP: Our doctors are salaried, so reimbursement would not change for them. The big distribution was between the hospital and consulting physicians, who may not be Geisinger doctors. We set up an overall clinical enterprise incentive: if care could be improved, the incentive went to the clinical enterprise getting the lump sum payment – the cardiovascular service line – rather than to the doctors directly. We’re taking an active look at whether to have future incentives filter more directly to the individual caregivers.
PND: How does this program differ from traditional pay-for-performance programs for hospitals?
RAP: Traditionally, almost all pay-for-performance has been for ambulatory services and primary care. There’s very little, if any, specialist pay-for-performance programs and very little inpatient. Also, most pay-for-performance programs have been about, “if you do some process step a certain percentage of the time, you get some bonus.” But because any given payor typically represents a small percentage of an overall practice’s revenue, and because patients who meet the criteria are even a smaller percent, you quickly go down to where you’re at less than one percent, in terms of dollars, so the incentives don’t make a difference. This program is quite different: providers initiated it, not the payor; it is focused on specialty care; it is inpatient care; and we assume risk for this whole bundle – which had not been done before.
PND: What results have you seen from the program?
RAP: On an overall basis, pretty much all of the clinical performance measures which were tied to the Society of Thoracic Surgeons clinical outcomes database improved between 15 and 60 percent, length of stay fell by about half of a day, the hospital’s financial status improved significantly, and readmissions fell by 44 percent. The program has now been expanded to hip replacement surgery, cataract surgery, bariatric surgery and percutaneous coronary intervention. We’re in the process of applying it to our first non-surgical area – perinatal care, and also to spinal surgery.