Background:  The Centers for Medicare and Medicaid Services’ designation of 32 accountable care organizations (ACOs) across the U.S. to enroll Medicare fee-for-service patients beginning in 2012 makes ACOs an important feature of the national healthcare landscape–at least for publicly insured patients.However in 2007, Blue Shield of California, along with provider and employer partner organizations, began exploring development of one of the first ACO-like programs in the country to serve commercial patients. Two years in the planning, the program was successfully launched in 2010 and has been delivering savings to consumers.

Located in the competitive Sacramento market of northern California, the program has been receiving national attention as an example of an innovative shared savings model involving a large insurer—Blue Shield of California; a purchaser—the California Public Employees Retirement System (CalPERS); a physician group—Hill Physicians Medical Group; and a hospital system—Dignity Health (formerly Catholic Healthcare West). The population served by the virtually integrated health organization includes approximately 42,000 CalPERS employees and their families covered by Blue Shield.

In this Health Affairs Blog interview, senior management and architects from the partner organizations of this ground-breaking commercial ACO-like project discuss the affordability crisis that precipitated a new price-based premium setting strategy and an unprecedented commitment to CalPERS for a zero percent 2010 premium increase for pilot participants. To achieve a $15 million savings target, the project’s three-way global budget honored existing hospital and medical group payment mechanisms and contracts while incentivizing the partners to aggressively drive out costs. The health plan and providers assumed any down-side risk; they also had the upside potential to share excess savings commensurate with the premium share and cost levers each controlled.

The conversation offers insights into the novel partnership’s data-driven project planning, goal-setting, and ongoing benchmarking. Discussants also describe the virtual health network’s five key strategies and their experiences with implementation initiatives such as integrated discharge planning, care transitions and patient engagement, creation of a health information exchange, patient repatriation, and evidence-based variance reduction. The project has produced impressive year-one clinical, quality, customer and financial outcomes, including reductions in readmissions, inpatient utilization, and cost savings of approximately $20 million, $5 million above the target.

Finally, the partners illuminate marketplace factors and provider community conditions for success, offer lessons for other organizations considering development of an ACO-like delivery system for the commercial market, and make the business case for forging such shared-risk arrangements, regardless of the outcome of current health care reform challenges.

Glenn Melnick: As I understand it, the story of this innovative ACO-like project began with Blue Shield of California in 2007, well before ACOs had become a feature of the U.S. health care system under health reform. What prompted the discussion?

Juan Davila Senior Vice President, Network Management, Blue Shield of California:  As part of our strategic planning we forecast health care costs. At the time, our projections showed that premiums would double within 10 years. This would make health insurance increasingly unaffordable. If we were to solve the affordability crisis—and we knew it had to be solved—we would have to do something different, have new methods. The issue would not be solved by using traditional methods and expecting a better result.  We had to find how to work with physicians, hospitals and members in a different way.

Melnick: Describe the thinking or approach you took that led you to the new model.

Davila: Historically, most insurers do cost-based premium pricing. In other words, they look at the pieces, what it costs for hospitals, for physicians, and so forth, add up all the pieces, and then figure out what needs to be charged to cover all the costs. We tried to turn that idea on its head. We started to look at our healthcare premiums by doing price-based costing. Let’s make necessity the mother of invention. If we aim for a price point and figure out how to get there and improve quality at that price, then we really have something. So, we decided on what the premium should be and worked backwards from there.

Melnick:  CalPERS is one of the largest (1.3+ million members) and most visible purchasers in the country. How were you able to gain their participation?

Davila:  We went at risk with CalPERS. We guaranteed them that Blue Shield of California would not raise their premiums in 2010 for members participating in the pilot, but that we would still maintain quality. We estimated that it would require about $15 million in savings to keep 2010 premiums at 2009 levels. So we had a pretty strong financial motivation to produce efficiencies.

Melnick: What were the key elements of the shared risk arrangement among the health plan, the medical group and the hospital? 

Davila:  Without question, the most important element was that all the organizations had to have “skin in the game.” We agreed that everyone needed to have an incentive to drive out costs. If those savings were not achieved, we all would share the risk to honor the zero premium increase commitment to CalPERS. If we exceeded the target, we’d share in that success. We set a global three-way budget, but did not change the existing payment mechanisms or contracts. The hospital was still paid fee-for-service and the physician group was paid capitation. We negotiated a formula for how those savings would be distributed among the partners. We worked together to find initiatives and “levers” that could be used to produce the needed savings.  A unique and important element of the risk-reward formula was that each organization had certain levers under its control that could be used to effect costs and efficiency and savings or losses would be shared based on the specific levers under their respective control.

Melnick:  Why was CalPERS willing to take a chance on this new and unproven approach?

Ann Boynton, Deputy Executive Officer, Benefit Programs Policy and Planning, CalPERS:  CalPERS has a significant history of pursuing strategies that will reduce cost to our employers and members and improve health outcomes. We agreed with Blue Shield that we were facing an affordability crisis. We looked together at industry trends: California HMO rates were increasing an average of 11 percent per year. Blue Shield projected that even if they were able to reduce that trend to 8 percent in the next decade, extrapolated to 2020, its Access+ HMO family product for CalPERS members would cost about $39,000 annually. That would be equivalent to nearly 50 percent of the median income of a Californian in 2010. This, and other cost increases across the State, called for major innovation; small steps were not going to get us there. The project provided a way to test the promise of major redesign of system incentives.

Melnick:  So you set a very ambitious goal and essentially had to reverse engineer to accomplish it.  How did you go about moving from idea to implementation?

Davila: Having the will and good intention is necessary, but doing the upfront work was essential. We initially met at the end of 2007 and began working in earnest in early 2008 to be ready to launch by our January 2010 target. It was intense. Our strategy development was very data-driven. We formed integrated teams to work together, and we compiled and analyzed three years of data from disparate sources. We looked at the cost drivers, who is driving those costs and for what. For example, we analyzed chronically ill members and found that the top 5,000 members accounted for 75 percent of the total pilot population spending. We identified utilization outliers at the MSDRG level and established benchmarks for improved care in key areas. We worked to identify opportunities to hit budget and improve quality as a result. The process enabled us to focus on opportunities to expand existing care programs and develop additional ones. It was a very collaborative, transparent process that helped to create a dynamic for improving both quality and cost.

Ultimately we agreed on five key strategies and then built interventions to actualize them:

  1. IT integration to facilitate rapid and efficient communication (which was harder than we thought technologically, but also involved physician acceptance and use);
  2. Reduce drug costs through member outreach, drug purchasing and contracting. For example, we reduced oncology injectable costs with a cohort case rate;
  3. Improve physician clinical and resource variation through quantitative analysis and targeted interventions. This involved implementing unique procedures around OB/GYN, knees and hips, bariatric surgery. It also entailed increasing physician understanding of why affordability is as important as quality care delivery;
  4. Eliminate unnecessary utilization and non-compliance through personalization of population management, such as chronic care management, patient education, palliative care, pain management, and making professional home visits; and
  5. Develop CalPERS-specific utilization management through a coordinated operational infrastructure. We also developed a comprehensive dashboard of leading and lagging measurements so everyone could see our progress.

Melnick:  CMS estimates that start-up costs for an ACO are between $130 million and $263 million.  How does this square with your experience?  

Davila:  We adapted and built upon elements of a delivery system already operating so it really wasn’t a start-up. Plus, much of the extra work it took to launch the demonstration project was covered by existing staff at the three organizations. The partners spilt the additional costs, which were far outstripped by eventual savings. It was nowhere near $130 million. But it does require investment.  A successful ACO will return multiples in improved affordability. The key is to match return to the source of investment so there is a business case to make the investment.

Melnick:  What was it about the region of California (Sacramento) that led you choose it for the demonstration project?

Boynton: CalPERS had some 207,000 members in the Sacramento area in 2009, about 90 percent of whom were enrolled in an HMO. But, as is typical, only the physician component is paid on a risk basis, with hospital care paid on a fee-for-service basis. The incentives weren’t aligned appropriately in the standard contracting structures. Under the integrated health organization, all patient care would be part of the shared risk model. We were able reach agreement with Hill Physicians Medical Group and Dignity Health so that the project was seamlessly overlaid onto the provider network those members were already utilizing. For the most part, the demonstration project was transparent to our members, except to the degree that some saw a higher level of care coordination or outreach regarding chronic care conditions.

Melnick: What was the thinking within your hospital system that led you to become a key partner in this new model?

John Wray, Senior Vice President, Managed Care, Dignity Health: We all absolutely recognized the issue of affordability. But all of our organizations also are non-profit, with deep ties to the community. We shared the belief that in order to avoid a sustainability crisis, we must also improve the quality of health and healthcare. It was automatic to think, each in our own way, about our version of Dr. Donald Berwick’s “triple aim” in healthcare: better community health through better preventive care; better patient care so everyone gets the right care at the right time in the right setting; ultimately lower cost through improvements. The partners were all big players in this competitive market, with strong market penetration. For Dignity Health, it was a natural. We have four hospitals in Sacramento County and were already the inpatient provider of choice for a large segment of CalPERS members. Also, all of the partners had been working together for awhile—our system, Hill Physicians and Blue Shield—so we had a foundation to build on.

Melnick: We have all heard that one of the barriers to gaining hospital participation in a shared risk ACO model is that much of the savings comes from the hospital side. Is that your experience and, if so, why would your organization participate?

Wray: It is true that one of the effects of an ACO-type arrangement is a reduction in inpatient utilization. However, we adopted a long term view that included several elements. First, we recognize that the health system needs to become more efficient and that if hospital care can be used more effectively, we have an obligation to seek to achieve that. In addition, reduced utilization leads to reduced costs, which lowers our cost base. Further, part of our agreement was to identify other services that we could cost-effectively provide to CalPERS members in the future. This resulted in an increase of our ambulatory surgery volume, for example. Finally, in a highly competitive marketplace such as ours, opportunities to grow or preserve market share were definitely a consideration.

Melnick: What was the thinking within Hill Physicians Medical Group when you agreed to become a project partner?

Cardoza:  Even at the start of the project, Sacramento already had among the lowest medical utilization in the country. When we explored launching this virtual ACO-like project, we really doubled down. It is virtual among the three of us–Dignity Health, Hill Physicians and Blue Shield—but it is also virtual because Hill is an IPA, not a staff model. Hill Physicians Medical Group has been a 25 to 30 year experiment to demonstrate how care can be delivered while organized in the same way as physicians in the rest of the nation practice, which is a network of small and independent practices. At the start of this project, Hill had 3,500 physicians system-wide, mostly in one or two-person practices.  Of them 800-1,000 were involved in the Sacramento project. If this could work with our group, it should be transferable elsewhere. We started with a target: not 8-9 percent year to year with an increasing trend.  We said 0 percent–$14-15 million. We didn’t think we could hit that. Everyone gulped.

Melnick:  From early results that have been reported, it appears that the shared savings model financial targets have been achieved.  What has the experience been, and what other key accomplishments has the demonstration project produced?

Davila:  Financially, we exceeded our goal. We estimate that the project is resulting in savings of about $20 million. The $5 million over our $15 million target is being shared among all three players based on our pre-negotiated agreement. From a market standpoint, Blue Shield and its partners gained more than 1,500 members in the Sacramento market, which we attribute to the success of the integrated model.

Wray:  Clinically the results have been very exciting. In 2010 alone, we saw a 22 percent reduction in patient re-admissions compared with 2009; a half-day reduction in inpatient length of stay; a nearly 14 percent drop in total inpatient days per thousand; and a 50 percent reduction in the number of inpatient stays per thousand of 20 or more days. We achieved a 13 percent decrease in bariatric surgery, while helping members shed about 1,500 pounds through our Lose to Win program in three CalPERS sites. Translated into real dollars, we achieved savings of more than $400 per member per year. Inpatient costs per day declined about $240 for the pilot population, compared to a $200 increase in cost per day for the control group.

Cardoza:  Our target was affordability. Improved affordability requires engagement with the most vulnerable and high cost patient population. As a consequence of that more intense and sustained engagement, patient satisfaction goes up. Not necessarily an intended consequence, but one we have pleasantly found to be true. That we improved quality and service is wonderful. We improved care processes. What we wanted to do is get away from the reactive—to be proactive. I think we were successful in that. One of the most important achievements of the project has been implementation of the best practice discharge planning process. But we also saw major accomplishments in the areas of using benchmarking to improve processes; development of a shared, high risk patient tracking tool; readmission reduction strategies, such as home visits and follow-up appointments; variant day analysis; and process improvement for coordinating patient transfers.

Melnick:  Would you describe some of the specific initiatives in a bit more detail? Talk about the integrated discharge planning process, efforts to reduce readmission rates, and clinical data-sharing.     

Cardoza:  To drive the exchange of clinical data, we developed and implemented a Coordination of Care Document (CCD), enabling Hill EHR-enabled practices to share clinical data with Dignity Health hospitals through an expanded Health Information Exchange (HIE). The HIE facilitated automatic exchange of clinical and discharge summaries, and other information. But interpersonal communication was still essential. Two key tenets of ACO development are the importance of coordination and the integration of care. On the front end, physician communication with the hospital in advance of a patient arriving gave everyone time to prepare. There was communication from when the diagnosis was made through the transition to inpatient.

Wray: To drive reductions in length of stay, we defined enhanced evidence-based guidelines for surgeries, targeting high-volume, high-cost MSDRGs requiring more detail pre-authorization, and we used hospitalists and cross-hospital standardization to establish consistent processes to manage inpatient stays. The hospitals followed evidence-based protocols, such as for sepsis, and we kept looking for more, like pneumonia, knee/hip total joint.

When a patient was ready for discharge, a problem-oriented post-discharge needs assessment and summary of key medical issues was developed. An analysis of the clinical course and major events of the inpatient stay, integration of labs, identification of discharge clinical diagnoses, and review of medication errors and interactions all were conducted. A discharge planner initiated this process on day one or 48 hours after admission to coordinate services needed on discharge. We worked with nurses in the discharge process to do teach-back so that when patients went home, they were fully informed about their care and could be active participants in their recovery. The discharge planning process entailed a scheduled follow-up visit within 8-10 days, including measure adherence. We shared a written discharge plan with the patient in lay terminology, and forwarded it to the PCP and care managers within 24 hours so that, post hospitalization, everyone had the key pieces of information they needed to help provide the best care. And if multiple facilities were involved, we identified a point person at each and a designated manager over all the facilities who could handle exceptions.

Cardoza: These clinical integration activities allowed physicians and caregivers to improve quality, provide more patient centered care, and increase efficiency while by decreasing unnecessary or avoidable readmission. But sustained engagement with the patient was probably the single most important thing we did to reduce readmissions. Making a call to the patient’s home after discharge, for example, may be how we reached them initially, but what was pivotal was that we sustained engagement with the most fragile segment of the patient population.

Wray:  We also had a process to repatriate our patients who were admitted to other hospitals, where clinically appropriate, so that we could maintain the coordinated care process. Keeping patients within our system improved our ability to coordinate care, but it also helped Dignity Health make it work financially and maintain market share. If we had to have a “head in a bed,” we wanted it to be our bed. In addition, we developed a new case management program focused on chronic pain to reduce ER visits and pharmacy costs.

Davila:  In terms of project implementation, we learned three important lessons. First, don’t ignore low hanging fruit—implement changes in small increments as soon as possible. Long term initiatives can run concurrently. Second, have clear targets. Ensure accountability. Share the targets so everyone knows what the goal is. Finally, prioritize. You can’t do everything at once or your people will explode.

Melnick:  What are some additional challenges that had to be overcome, or ‘take away’ messages that were garnered from the virtually integrated care system project?

Wray: Given hospital costs and the critical role of care transitions, it is important all parties are at the table, working together, sharing information, developing trust, and realigning incentives so they produce the best result for all, most importantly the patient. While this pilot was being designed, Dignity Health was at the table so we had an opportunity to shape and buy into the project at all levels.

Cardoza:  In a virtually integrated delivery system or ACO, it is inevitable that hospitals are going to experience the greatest stress, especially financially. Most savings come from the hospital side and that’s the stretch—reducing unnecessary hospitalization.

Davila:  Absolutely, and that’s why each partner got a share of the savings proportionate to their share of the premium dollar. If the hospital benefits from overall savings, they don’t need to be holding onto patients to keep them in the hospital in order for the economics to work.

Cardoza:  It also made a difference to have Blue Shield as an active partner to help mediate among the various provider issues and interests. Currently the medical group must do a lot of the heavy lifting, so you must have capability at the IPA or medical group level, as well.  There are some other realities to pay attention to, and a big one is financial integration. Hill was capitated and financially integrated within Hill, and we still had our own silos. This is integration is across all partners.

Wray:  Some 22 of about 100 initiatives that were proposed were initiated and had a return on investment. Everyone helped to drive solutions towards seamless support for the patient and physician. Clinical or financial sub-teams continually met to address problems and find answers. It was essential to employ matrix management among the players and to regularly examine dashboards and data. Among the key levers that made a dent in costs were pre-admission discharge planning, and transparent sharing of data.

Davila:  And on that score, one of the greatest challenges we faced early on was sharing information across disparate platforms. Connectivity and access to the same data was much harder than we expected it to be.  But fundamentally it’s about intention and working towards a common purpose. It’s about culture. One of the biggest changes has been how we do business with one another. Instead of being in silos and separate—whether financially, IT or culturally—we are all now meeting and striving at all levels of our organizations to deliver service, higher quality and drive costs out of the system. It’s one thing to talk about doing that; it’s another to actually do it. Though we may have worked together for 15 years, we really didn’t work together. Things could get contentious at times. Before, one party had the incentive to shift costs to another. There were different views about how to get things accomplished.  It hasn’t always been easy, but our biggest lesson learned is the importance of being willing to take a risk, break old patterns, and build trust by focusing on win-win for all, especially the patient. That part actually turned out to be easier than we thought.

Melnick:  How can the Sacramento experience inform ACO development? Do you have any additional thoughts about how this commercial model can be related to the Medicare market nationwide? 

Davila:  The Sacramento experience offer hope for the future of ACOs as an important strategy to improve affordability and improve quality. We need to look at virtual integration of care because physical integration is not there. Even if we each work well in our respective silos, we’ll go down the tubes. But, it is hard work that never abates. It requires committed partners bound together through a shared business model that aligns incentives among them. We are pleased with what we’ve accomplished, but there is still much to be sorted out and improved. It’s also important to continue to experiment and test new things. Every market is different. Nothing is cookie-cutter, though we’ve shown that the basic principles can be put into play elsewhere, and for any patient population, not just a commercial HMO.

Cardoza: I agree. I’d add that the physician delivery system is necessarily the engine for the ACO. Hill Physicians is an IPA and that is an important and deliberate distinction to make as it relates to the transferability of our experience to ACOs. We weren’t successful because we’re here in California or because our medical group is a delegated model. We’re an IPA, which demonstrates to the rest of the nation that this can work within the framework of how most medical care is provided in this country. Most care is not delivered in a staff model. It is possible to engage physicians of all kinds—solo, small groups, whatever–over a geographic area and make this happen. Experience does matter, though. Hill Physicians has 25 years of experience in organized delivery. The project allowed a focus of resources to expand and improve what we do, but we had the structure, know-how and operating capability to do it.

Boynton:  Putting money at risk for purchasers, payers and providers is a fundamentally important way to understand whether or not this concept works as a means to reduce costs. The Sacramento integrated delivery system project also gave us a great way to focus on what it takes to develop virtual clinical integration so our members can concentrate on taking care of themselves and their families.

Cooperation and coordination with payers makes financial sense over the long term. We are very pleased with the results thus far. Overall, our goals were to find a way to contain rising costs in a way that is sustainable and to improve integration of care for our members so they don’t have to worry about that aspect of their healthcare. This program is an example of how it can work and build trust among providers and everyone benefits in the end.

Davila: As a practical matter, success in the ACO model does require a critical mass patient base. It requires a competitive provider community with discreet delivery systems — or at least discreet allegiances if organized delivery systems are not present — and a willingness by the health plan and employer to choose between them. If providers are to sacrifice margin and do the hard work to improve affordability, then they need market share for it to make sense. This experience also shows that physicians are excited at the opportunity to shed some of the roles they’ve had to play in recent years and to be reimbursed for designing effective new care strategies. For hospitals, it’s a new profit paradigm if it is not about filling beds, so aligning incentives and sharing risk across all parties is fundamental to making the accountable care concept work. This project also shows that the care delivery system can be profitable and beneficial at the same time. This isn’t theory. It is real life and real people.

Melnick: What I’m hearing you all say is: it’s difficult, but it is do-able. Your experience with the Sacramento ACO-like model can help dispel the notion that it’s a choice between quality or affordability; it can be both. Call the structure whatever you want—an ACO or a virtual integrated care model—it can be done.  Getting providers hospitals, physicians and plans together, with patients at the center, can be done, and produces results at several levels.      

Cardoza: They can work together to deliver integrated and coordinated care and achieve these kinds of results. We were shocked to get the kind of results we got and in a market with some of the lowest utilization and higher prices. We lowered readmission rates by 15 percent, decreased days per thousand by 15 percent; also, length of stay and ER visits. These were significant reductions.

Melnick:  What’s next?

Cardoza:  We’re not done yet. Halfway through the second year we were achieving similar results in an environment where we thought we were already best in breed. This tells you that there is more to get and what can be attained by working in a care-coordinated system. We all have been trading partners for 15 years. The fact that we met and exceeded the goal and are continuing to follow on course lets you know how much can be done. But it gets tougher to keep getting more and more savings. Though we sustained our gains in the second year and actually improved upon them some, the big initial gains are tougher to achieve at the same order of magnitude in subsequent years. That said, we do anticipate sustaining a moderated trend.

Boynton:  These results show what happens when everyone works in a coordinated way. We cannot overstate the importance of this and we’re encouraged about expanding the model to other areas.

Davila: When we started this virtual integration project, Blue Shield of California’s aims were to achieve cost savings, grow membership, and create a sustainable model for expansion. We achieved these and hope to see continued benefits from this demonstration over time, even as we begin to replicate in other markets. Our goal is to have 20 such projects across the state by the end of 2015. Blue Shield has more than 3.4 million members in California, so imagine the potential. And we’re just one player of many nationwide.

The California HealthCare Foundation has funded a full evaluation of this demonstration, which should be completed in 2013. In addition, Blue Shield recently invested $20 million in 18 integrated health delivery system incubator projects throughout California. Many of them reference the Sacramento experience, while others are testing out new strategies to improve care and reduce healthcare costs. These projects launched in late fall of 2011 and Blue Shield plans to study these projects and share their findings.   


Glenn Melnick ( is a professor and BlueCross of California Chair in Health Care Finance, and Director of the Center for Health Financing, Policy and Management at the Sol Price School of Public Policy, University of Southern California in Los Angeles. Lois Green, President/Principal of The Performance Alliance in Woodland Hills, CA, co-authored the Q and A. Ann Boynton serves as the Deputy Executive Officer, Benefit Programs Policy and Planning for CalPERS. Juan Davila is Senior Vice President of Network Management for Blue Shield of California. Darryl Cardoza was recently appointed Chief Executive Officer of Hill Physicians Medical Group, having been serving in the role of Chief Operating Officer. John Wray is the former Senior Vice President of Managed Care at Dignity Health (previously Catholic Healthcare West).  This interview was supported by funding from the California HealthCare Foundation.