ProvenCare

by Glenn D Steele

  In the most recent phase of the Geisinger evolution, the system entered into a frenzy of merger and acquisition activity. This has led to significant changes in governance structure, with the foundation board delegating some primary responsibility for operating performance, regional strategic growth, and administrative leadership to local market boards.

  LESSONS LEARNED

  • Effective governance is integral to organizational risk-taking and success.

  • Alignment between management aspiration and board aspiration is critical.

  • The most important relationship in any nonprofit organization is that between the board chairman and the CEO.

  • Vertical integration between insurance payer and care provider is best accomplished with both components in the same fiduciary.

  • Simple, straightforward governance is best.

  • As systems get larger and more complex, governance structures most often become larger and more complex.

  • Scaling innovation into larger and more complex systems will become a more difficult management and governance challenge.

  5

  Getting Started

  Glenn D. Steele Sr., the coauthor’s father, began to experience debilitating angina when Dr. Steele Jr. was in medical school. In the 1960s, coronary artery bypass grafting (CABG) had just begun, and there were only a few medical centers that had reasonable volume. Dr. Steele Sr. and his family chose to go to the center with the most experience and best outcomes.

  Convincing the 55-year-old physician that he needed to give up a two-pack-a-day cigarette habit became almost impossible after he met the world-famous cardiologist who performed the diagnostic coronary catheterization. The cardiologist’s pack of Lucky Strikes, rolled up in the sleeve of his undershirt, was clearly visible under the operating gown and radiation shield.

  The post-catheter discussion was simple: don’t leave the hospital until the heart surgery. So the family didn’t. Not even for dinner, since at the time there was a razor wire fence surrounding the all-concrete campus. When the next day his wife insisted that the family talk to the proposed surgeon, the floor nurse observed that it probably was best not to do this, since the surgeon chosen (not by the patient) to do the CABG didn’t speak English very well. However, he was undoubtedly the “best pair of hands” among the cardiac group.

  After waiting in the hospital for about 10 days to have this semi-emergency heart surgery, Dr. Steele Sr. had his procedure on a day when 52 coronary bypasses were performed. He was number 48. At the end of the day, the families of all the heart surgery patients were assembled in a conference room adjacent to the cardiac intensive care unit. A nurse at a podium called out the patient numbers and family names. The family members raised their hands in response. She announced that patient 48 had been returned to the operating room due to continued bleeding, but was doing OK now. At least the family felt a modicum of empathy from the strangers surrounding them as the nurse moderator moved on quickly to number 49. Despite selecting this superb institution based on its high volume and best outcomes of CABG surgery, and despite Dr. Steele Jr.’s presumed insensitivity as an aspiring surgeon in training, he knew the overall experience was not optimal.

  This was during the days of full indemnity medical coverage, and when the family received the bill, the insurance company had been charged for two operations: the first one planned and the second one performed to repair problems related to the first. As a budding physician, Dr. Steele Jr. began to think beyond the physiology and anatomy of disease, surgical brilliance, and acute-care responsiveness. How could the system be so screwed up, and what could the quality, cost, and patient experience be if the entire episode were reengineered? Some 35 years later, what better place for this reengineering to occur than Geisinger?

  Initiating change is never easy. Coming off the dysfunctional Geisinger/Hershey merger made the operational turnaround at Geisinger easier. But now the institution was getting into the really tough stuff: changing how doctors practiced and had been taught to practice, and altering or at least redirecting some of the fundamental personality traits that got them selected for medical school in the first place: independence, creativity, ability to work incredibly hard, need for autonomy, and belief in the sacrosanct doctor/patient relationship that often celebrated individual variation to a fault. We were questioning how Geisinger providers had done things for most of their careers with good results, most of the time. We were about to suggest in our new strategy that the number one goal was fundamental innovation and continuous improvement in how we provided care. Manipulating the strategic conversation to get apparent buy-in would be easy compared to getting the work reengineered.

  WE START WITH HEART

  Our choices were intentional regarding how, where, and with whom to begin ProvenCare Acute. We wanted to create a flywheel effect within the organization. If patients were better served through reengineered cardiology care, we felt certain that clinical leaders in orthopedics, gastroenterology, rheumatology, endocrinology, and others would move quickly to join the reengineering effort.

  What were the ground rules for our innovation beta test? We wanted to start with something that already was performing well. This seemed counterintuitive to some colleagues who felt we should pick a clinical service obviously in trouble or producing suboptimal outcomes. Our starting assumption, however, was that clinical leadership would likely be the most important lever in getting key professionals to change how they gave care. And we assumed that the best clinical leaders were most likely in the service lines or departments already producing the best results. Credible physician champions would be integral in socializing the new institutional commitment to default best practice every time for every patient. Our intent was to move a service with pretty good outcomes to as close to perfect outcomes as possible.

  The most credible non-Geisinger quality outcome metrics at the time were from the Pennsylvania Health Care Cost Containment Council (PHC4), an independent state agency focused on addressing the increasing cost of healthcare. At least theoretically, PHC4 was meant to stimulate competition in the healthcare market by providing individual consumers and group purchasers of health services comparable information about the most efficient and effective healthcare providers and by giving data to providers for identifying opportunities to contain costs and improve care quality.1 We routinely used these metrics to determine our quality and cost competitive position benchmarked against all other providers in Pennsylvania.

  We wanted to pick a high-volume, high-cost, high-visibility hospital-associated episode of care at the top of the PHC4 reports. We would then attempt to encompass in the redesign effort everything from diagnosis through post-acute-care rehabilitation.

  Another ground rule was to choose a care pathway in which the main disciplines had established unambiguous indications for the intervention; had evaluated data or actual scientific evidence to achieve consensus on what should and should not be done throughout the entire episode; and, finally, had agreed on short- and ideally long-term outcome metrics.

  When we started the ProvenCare Acute process in 2004–2005, there was little best practice or evidence-based consensus process available for off-the-shelf use. No one had attempted to bake it into an entire reengineered care pathway to evaluate whether it could be done or would actually affect patient care quality and/or cost outcomes.

  The key postulate to unlocking value would be eliminating unjustified or ambiguous indications for treatment in the first place. We would then apply already-identified, evidence-based or consensus-based best practice recommendations (ideally available from discipline-based study groups) to begin identifying and minimizing unjustified variation for each component of the care. Surprisingly, few professional disciplines had developed any kind of formalized best practice consensus process at the time we started our reengineering, and almost none of the disciplines had established outcome registries. As a result, our choice of clinical service reengineering targets became qu
ite easy. In fact, outside the relatively limited universe of randomized clinical trials (not applicable to most of our patients due to age, comorbid disease processes, and simple lack of geographic access), only cardiology and cardiothoracic surgery had any immediately available consensus processes and readily available outcome metrics. Our choice of heart disease, specifically elective CABG and interventional cardiology for stent placement, took advantage of a combination of Geisinger’s strengths in those services plus the readily available best practice recommendations we could bake into attempts at reengineering the treatment episodes.

  Once our best-probability reengineering wins were under way, the buzz created by the early results affected our internal Geisinger momentum (the flywheel effect) and had influence on other disciplines’ best practice consensus processes and registries, first and most prominently, orthopedics through the Hospital for Special Surgery in New York City and the American Association of Hip and Knee Surgeons.

  Our choice of the Geisinger Community Practice Service Line (CPSL) for our first reengineering target in ambulatory care was again quite intentional. CPSL was our first multidisciplinary service line, it was large (composed of approximately 250 primary care physicians in 2001), and it was geographically dispersed, with 55 sites providing care in 47 of Pennsylvania’s 67 counties. It primarily was responsible for the care of nearly 30,000 type 2 diabetes patients. And most important, similar to our heart care clinical services, CPSL was incredibly well run and Geisinger’s most innovative care delivery group.

  SUCCESS FACTORS IN OPERATIONALIZING AND SCALING

  Innovation was our top strategic target, so we began to socialize our definition of fundamental innovation in caregiving using a repetitive top-down narrative. But it was the men and women providing care in the service lines or in the discipline-based units who were expected to pick what they most wanted as their innovation targets. The job of leadership was to prioritize a limited number of beta tests believed most likely to succeed in providing real benefits to patients. It probably was a combination of wisdom and luck that led us to choose heart care for the initial ProvenCare Acute reengineering and a reengineered care pathway for type 2 diabetes for the ProvenCare Chronic bundled best practice redesign commitment. Getting early success was our goal. Even with the inefficiencies of any beta test, ProvenCare Heart began to show higher quality and lower cost for the CABG patients within several years. ProvenCare Chronic demonstrated immediate benefit in achieving improved intermediate and process metrics and surprisingly showed dramatic effects in lowering diabetes-related disease effects within three years. As expected, this early success led to incredible internal and much more than expected external affirmation.

  Some of this was deserved, such as our American Surgical Association presentation and paper2 and the Bloomsburg (Pennsylvania) Press Enterprise editorial on the effect of reengineering on teacher compensation.3 Some was over the top and focused more on the sexy packaging of our single-priced so-called warranty, as it was referred to in the New York Times,4 instead of the substantive default best practice commitment that was the core of our reengineering from the beginning to end of a particular treatment. But no matter—both the internal flywheel effect and the external affirmation meant scaling within Geisinger became assured.

  Of course, learning from our failures also helped us move forward to scale innovation. Our first attempt to reengineer how we cared for autism patients and their families is a good example. Our researchers looked at why there was on average a year-and-a-half wait for an initial appointment. We already had recruited a group of excellent autism psychiatrists and psychologists, but demand was staggering, and the referral lineup and waiting times simply were unacceptable. Our researchers developed an extraordinarily interesting automatic writing device that could transmit a significant amount of information and allow families to provide input to the doctor’s office prior to seeing one of our autism specialists. That sounded great, but there was one big problem. The researchers didn’t include caregivers in the instrument design and its application into the redesigned clinical pathway. No matter how cool a new device or technology might be, if respected clinical leaders don’t buy into the care redesign, it isn’t going to happen.

  A second example of learning from failure was our redesign of cataract surgery. When Geisinger ophthalmologists presented their reengineering results, it was obvious that patient outcomes before and after redesign were close to perfect. The redesign efforts were worthy in getting the group to move together, but much less so in terms of actually improving the value proposition.

  Surprisingly, no one has ever asked us to present our failures and what we learned from them. Outsiders must assume we would either be too embarrassed to expose our failures, or they might suspect we would cloak them as hidden successes. Failures are as important to learning as the successes that receive much more focus.

  Discussion leading up to passage of the Patient Protection and Affordable Care Act (ACA) did not directly influence our overall innovation approach and its internal organization successes. But the debate did help, particularly since early on many of our clinical, insurance, and administrative leaders were involved. This was more a function of the Geisinger payer/provider structure than it was healthcare reform altruism. We simply were trying to take advantage of our relatively unique overlap between caring for and insuring the same patients. Our business model and professional pride of purpose were in remarkable confluence. If the patients we cared for stayed healthy, our insurance company made money. If our providers decreased unnecessary or bad care, our patients did better and our insurance company did well financially. Proving this alignment of professional aspiration and business success, first in Medicare managed care, then in commercial managed care, and most recently in Medicaid managed care, has made Geisinger the model for many virtual and real payer/provider vertically integrated models presently being constructed.

  At the heart of our reengineering innovation was an intent to use our structure, culture, and unusual payer/provider overlap to obtain benefit for those we served in both the insurance company and clinical enterprise sides of Geisinger. This was something we knew that other providers could not achieve easily.

  Changing the compensation plan also was important. Money flowed from the insurance company to the providers as a result of significantly lower total cost of care obtained through fewer complications, decreased test duplications, shorter lengths of stay, and for the patients in ProvenCare Chronic ambulatory reengineering, significant declines in cost of care from decreased hospital admission and readmission rates. If our 30,000 type 2 diabetes patients had significantly fewer strokes, heart attacks, and eye and kidney disease complications because of better care, they were certainly better off, their need for inpatient hospital care diminished, and the total cost of care decreased to the benefit of our insurance company. We could then shift a portion of that unlocked value through internal transfer pricing to our people who had actively changed how they provided care to patients in order to continue improving outcomes.

  The compensation plan was redesigned to set aside 20 percent of total compensation for clinicians based upon their achievement of specific patient care benefits. These were linked annually to the organization’s strategic aims, primarily innovation, and not to relative value units, patient panel size, or other volume metrics. Although compensation incentives were important, clinicians responded more directly to seeing patients do better. For example, our community practice leadership and physicians committed themselves to Geisinger’s diabetes care bundle, and the results in their patients (decreased number of heart attacks and strokes, better management of glucose levels, lower risk of developing diabetes complications over their lifetimes, etc.) furthered professional pride.

  If we hadn’t seen patient benefit from our redesigns, the compensation plan changes alone would have been ineffective in sustaining long-term buy-in to the continuous innovation strategy. Both Geisinger and non-Geisinger academ
ic and clinical colleagues quickly became interested in knowing how default best practice could be socialized among a group of highly productive and rambunctious specialists (even orthopedists) without our being accused of dictating cookbook medicine or managing to a static process nonresponsive to the rapid advances regularly being claimed and sometimes validated in diagnostics and therapeutics.

  Another key to our initial and ongoing ProvenCare success was finessing false polarities and showing them to be nonstoppers in our commitment to reengineering. Two prominent examples were “You’re making us practice cookbook medicine,” and “New discovery is changing medical care too fast to commit to any best practices for any period of time.” The way false polarities were finessed was extraordinarily important. We quite simply allowed people to make exceptions to whatever was in the best practice algorithm as long as they justified those exceptions in real time to their professional colleagues. In addition, we allowed revision, updating, or modification to the best practice default algorithm at any time, and the discussion to socialize change could be called by any of the people involved in the episode of caregiving. The changes had to be socialized in a consensus process similar to the socialization of the original best practice algorithm.

  But getting better outcomes at lower cost through care reengineering was the heart of the matter and the target of our internal and external professional interests. Most healthcare economists and public policy mavens were focusing on the problem of high price per unit in healthcare delivery. Geisinger was focused on removing unnecessary or hurtful units of work. As our early Geisinger reengineering road tests were expanded and sustained, we began to hear the next logical level of questioning. If the process worked at Geisinger, and if the results benefiting patients as well as the business model were valid and sustainable at Geisinger, could ProvenCare be applied elsewhere?