For decades, health policymakers considered Kaiser Permanente the lode star of delivery system reform. Yet by the end of 1999, the nation’s oldest and largest group model HMO had experienced almost three years of significant operating losses, the first in the plan’s history. It was struggling to implement a functional electronic health record, and had a reputation for inconsistent customer service. But most seriously, it faced deep divisions between management and the leadership of its powerful Permanente Federation, which represents Kaiser’s more than 17,000 physicians, over both strategic direction and operations of the plan.
Against this backdrop, Kaiser surprised the health plan community by announcing in March 2002 the selection of a non-physician, George Halvorson, as its new CEO. Halvorson had spent most of his career in the Twin Cities, most recently as CEO of HealthPartners, a successful mixed model health plan. Halvorson’s reputation was as a product innovator; he not only developed a prototype of the consumer-directed health plan in the mid-1990’s, but also population health improvement objectives for its membership, both firsts in the industry.
During his twelve year tenure as CEO, Halvorson not only guided the plan to solid profitability, but added a million members in California, its largest market, despite a devastating recession and a national retreat of commercial HMO membership. He invested over $6 billion in computerized patient care systems and population health management infrastructure, healed the breach with Kaiser’s physicians, and markedly increased its consumer satisfaction scores, earning 5 STAR ratings under Medicare Advantage. He left the organization at the end of 2013 with more than $53 billion in revenues and more than $19 billion in reserves and investments.
This interview covers Halvorson’s time at Kaiser, his views of health reform, including the unfinished reform agenda, and his public health activism. It was conducted by Jeff Goldsmith, a veteran health industry analyst, and Associate Professor of Public Health Sciences at the University of Virginia. Jeff is a member of the editorial board of Health Affairs.
The Evolution Of Managed Care
Goldsmith: During your time as CEO of Kaiser, HMO membership went from 29 percent of the total employed population in the country to 13 percent, yet Kaiser added a million lives in California alone. How did you do that?
Halvorson: Kaiser isn’t an HMO in the classic sense of HMO. Kaiser is a vertically integrated care system that uses a health plan as its sole financing vehicle. It’s a very different model that enabled us to do things that other settings didn’t: Focus on the patient, on prevention, on population health, on changing care delivery, and focus all our information flow on the patient. Almost every other care setting basically creates information so that it will generate a bill. Kaiser Permanente creates information to support care; a secondary function of that information flow is to create cash flow. The reason Kaiser Permanente went down a different path than other HMO’s is that Kaiser Permanente never was their kind of HMO.
Goldsmith: Well there has been a lot of diversity in the HMO movement. You had the community-based/group health movement; you had commercial carriers that created network style HMO products, as well as “integrated” care systems that developed their own health plans.
Halvorson: Some of those are still alive and growing. HealthPartners in the Twin Cities is still growing. Group Health of Puget Sound is still growing. The Wisconsin- based plans are still growing. The Harvard Community Health Plan has evolved a bit, but it’s still in existence and has a robust market presence. So the original model that was more tightly aligned around delivery systems survived the general shrinkage and didn’t go away. Some of the HMOs that did not survive were the ones that were purely contractual relationships.
Goldsmith: What happened to them?
Halvorson: Well, in a sense, managed care became the health insurance industry. Managed care has won. United Healthcare, WellPoint, Humana — they all deliver care through managed networks. So the old, classic, first-generation HMO model is gone, but the second- and third-generation managed care products are doing very well. They’re growing through greater involvement in public health financing. They’re taking over the Medicaid market entirely and they’re on a path, through Medicare Advantage, to taking over the Medicare market.
The Growth Of High-Deductible Health Plans
Goldsmith: If you look at the Kaiser/HRET survey, which focuses on the employer segment, the HMO wedge shrinks while the high deductible health plan wedge seem to grow in proportion. In 2013, High Deductible Health Plans (HDHPs) represented around 30 million lives. Are HDHPs “managed care” in your view?
Halvorson: Yes. I think they are a form of managed care. HDHPs get involved to some degree in delivery of care. They require care/data reporting. They have some care protocols or some care-related elements they don’t allow. So I would classify all those plans and products under the broad category of managed care. If employers buy a $500 dollar deductible plan from Kaiser Permanente — and about 10 percent of Kaiser members are now in plans that have deductibles that are high or higher — the employees still get their care at KP. When they pay the fee for the office visit, they pay KP as opposed to paying it outside, so that the cash flow for KP is basically the same. It just comes down two separate channels, copays and premiums, instead of coming down one prepaid premium channel.
Goldsmith: How hard was it for your colleagues, particularly the Permanente physicians, to accept the philosophical change that required? In the legacy Kaiser model, there was no cost to the patient.
Halvorson: Everyone at KP, including myself, strongly preferred the model where there was no cost for the patient. However if 10 percent of our employer customers want to buy a high deductible product and we don’t sell it to them, then they leave. Kaiser has hospitals, clinics, pharmacies, lots of care-related infrastructure that would not do as well if 10 percent of the patients were gone. It is important to point out that we made a very conscious decision to not change care in any way for those patients. That was a critical issue. It is easier to maintain the old approach for all patients because 90 percent of the KP’s members are still in the older benefit model.
Goldsmith: Of the million people that were added to Kaiser in California during your time as CEO in California, how many of them came because of this change in your benefits strategy?
Halvorson: Now about 10 percent of Kaiser’s members are in plans with deductibles of some kind. I personally went to California’s managed care regulators and worked to persuade them to allow us to offer the product. I had to give them assurances that we would not be changing the delivery of care for those members in any way. Those members would not be there if we had not offered a product with personal financial exposure in it.
Goldsmith: Did the high deductibles make your job as a care system easier or harder?
Halvorson: Because we had to bill the patient.
Goldsmith: What else besides the billing?
Halvorson: The billing was the major issue.
Goldsmith: It didn’t make it more difficult for you to get them to do things that were in their own health interest, even if it cost them money?
Halvorson: Because we are who we are and because we ask people to do what we ask them to do in a very consistent and reasonable way, members tend to follow. We didn’t see a lot of people refusing to get needed care based on the fact that they had deductibles. That was largely because we are so committed to a particular set of care concepts and protocols that the credibility of the individual caregiver for the patient triumphed over the financial exposure.
The Impacts Of Ethnic Shifts
Goldsmith: The other thing that has been happening in California, indeed for many years, is the shifting ethnic composition of the state. Did that pose a challenge to KP?
Halvorson: The changing ethnicity of California was a wonderful opportunity for KP. Members of minority groups who didn’t have longstanding care system relationships had the comfort of having all of the pieces made available immediately from KP as a package. We also knew that Hispanic families liked having family-oriented care and care teams. That’s a good fit for Hispanic culture, which we promoted heavily. It had great results with Hispanic patients. In addition, we ran Spanish language ads and had clinics full of caregivers with high level of language proficiency. We have one of the more ethnically diverse medical staffs of any care system.
Goldsmith: How hard was it to find Hispanics physicians?
Halvorson: Extremely difficult, as it was to find Spanish-speaking physicians and nurses; However, KP is so attractive to physicians and nurses as a place to work that we generally had ten applicants for every care staff opening.
Goldsmith: So you selected out of the queue of workers to try and match up ethnically?
Halvorson: Yes. During the hiring process, one of the things we looked for was matching the ethnicity of local membership to that of the care team. Our growth in northern California in the Asian-American population has been very robust. In southern California, the growing population was more Hispanic. In both Georgia and D.C., African-American patients were a bigger part of the growth opportunity.
We did well in Washington D.C., but we won on service. The wait time for an urgent care visit at one of our mega-clinics there from the time you came through the door until you saw the doctor was 23 minutes in December 2013. We knew that we had a patient population in D.C. that was ready for much better service experience than they could get from the alternative care systems, so we invested a billion dollars in building care sites to accommodate that kind of access. We targeted 3-5 percent membership growth per year. If we grew beyond that rate, we could outgrow our capacity, and service could suffer. If we grew less than that, we wouldn’t meet our needs.
EHRs And The Learning Health Care System
Goldsmith: In your book, Don’t Let Health Care Bankrupt America, you talk a lot about the powerful enabler of the electronic health record. You invested $6 billion in health IT during your time at Kaiser.
Halvorson: We spent $4 billion for sure for the core electronic record, six if you count all of the support systems. When I got to Kaiser Permanente, one of the things that I told the board was that we were going to do what I did when (in a prior life) I helped build health plans in Uganda: We were going to have every single element of the care system connected electronically, so there would be no paper record.
Goldsmith: So you told them you wanted to catch-up to Uganda?
Halvorson: I’m not kidding. I actually learned in Uganda that to strip the whole process down to its most elegant essence was to have no paper anywhere. In Uganda, we couldn’t afford to pay a claim or for patients to show an ID card.
Goldsmith: But as I understand it, the electronic health record wasn’t sufficient. Didn’t you also have to build a health data repository on top of the record that extracted information from the record to support care?
Halvorson: Yes. We had to build patient information archives, and also connectivity. So, the pharmacy, imaging, etc. had to be connected to the medical record. We also had a very robust care research agenda. Last year, there were 1,400 medical journal articles published by KP doctors. The year before I arrived at KP, they published about 300 refereed journal articles.
Goldsmith: How strong is the efficiency case for the medical record installation? It’s no longer controversial in our industry to want an electronic record. What were the operational consequences for KP of this big investment?
Halvorson: Implementation was a painful challenge. You have to take people out of production and train them. But once people are up to speed they actually can work very well and deliver better care. So we were able to cut the number of stroke deaths by 40 percent, for example, by delivering computer supported care.
Goldsmith: Did the amount of time a KP nurse or KP physician had to spend with patients go up or down after the installation was completed?
Halvorson: Patient face time went up. But it wasn’t just the medical record. There was also an extensive agenda of process improvements. For example, we took the shift change time for nurses down from 40 minutes to 12 minutes by standardizing the process and related information flow. Shift change is one of the most dangerous times for patients during a hospitalization. If you have a patient that has to go to the bathroom during shift change, help is provided. The two nurses are no longer caucusing and sharing information in the nursing station. As a consequence, the number of patient falls is reduced.
Goldsmith: Was there also a shift in the locus of care during those 12 years?
Halvorson: The locus of care is evolving. The home is ultimately going to be the major site of care for all but the sickest patients. So we started building different ways of connecting the patients in the home: e-visits, e-connectivity, monitoring, iPhone apps, etc.
Goldsmith: What happened to clinician morale during this transition?
Halvorson: Clinician morale was actually very high. Patients really like the electronic medical record and patient portal, as well as the care improvements they made possible. So patient and physician satisfaction both went up. In our last survey, we had the highest level of physician satisfaction of just about any care site in our markets. Did you know that patients are more likely to thank the physician on an e-visit than they are on a face-to-face visit? It’s easier to type “thank you” or “you’re wonderful” than it is to say it face-to-face.
Goldsmith: I have heard that there was a striking increase in consumer satisfaction among Kaiser members and that it contributed materially to the Medicare Five STAR ratings KP earned in most of its markets.
Halvorson: If you look at Consumer Reports, KP had fairly average scores a decade or more ago. The last several years KP has been number one in every market.
Goldsmith: What changed?
Halvorson: We developed a culture of being the best. Caregivers take great pride in that. Being number one is very good for caregiver morale and patients notice.
Goldsmith: In your book, you talked about the idea of a “learning” health care system. How far did you push that?
Halvorson: When people begin to do data-based continuous care improvement, there’s some initial resistance. But once people start doing it, they love it and become addicted to it. It becomes a way of life — a way of thinking. For example, in hospitals in the rest of the country, about 5 percent of patients get pressure ulcers. So when KP started down the path of continuous improvement, we decided we’d cut the number of pressure ulcers down to 4 percent, and then to 3 percent and eventually down to 1 percent. We had a dozen hospitals last year that did not have one single pressure ulcer.
The care team loves being in the place where patients are absolutely safe. In some places, when a pressure ulcer actually does happen, members of the care team get depressed because they feel they’ve let the patient down. When you’re a vertically integrated care system doing data-based continuous improvement, it can be a stunningly powerful process.
Addressing Chronic Illness
Goldsmith: For years you have warned about the onrushing epidemic of diabetes. There are now almost 30 million diabetics in the US, a quarter of whom aren’t diagnosed. How much did you move the needle on diabetes for KP’s 9 million subscribers?
Halvorson: We reduced the number of diabetics at Kaiser Permanente by identifying people who are at risk for diabetes and then intervening individually with them to get them to be more active. The number one intervention that works is walking. Physical activity levels have a higher impact on avoiding diabetes than glucose control. We have multiple cases of early documented diabetes that we were able reverse just by walking. In a piecework payment system, caregivers have no incentive at all to prevent diabetes.
Goldsmith: Many physicians in private practices passionately want their patients to change their lifestyles — eat healthier, exercise, etc. — and discover they just don’t have the leverage.
Halvorson: I didn’t say that people aren’t interested. That’s not fair. But it’s hard for individual physicians to do something consistently to improve health across an entire population of patients.
Goldsmith: Were there reductions in prevalence of diabetes among KP members?
Halvorson: I don’t have quotable numbers. But the number of diabetics is down, as is the prevalence of heart attacks and strokes. Both are down by almost half.
Goldsmith: But those reductions in prevalence have been happening to the population as a whole, or close to them, in the last thirty years. One would expect that in an organized care system they would go down faster.
Halvorson: We’ve seen stroke deaths go down by 40 percent in five years, so the overall trend is better.
Goldsmith: You talked also about asthma in the book and that asthmatic crises could be reduced by caring for the patient in a different way. What happened at Kaiser?
Halvorson: You’re significantly less likely to have an asthma attack if you’re a Kaiser patient. So identifying which patients are at a high risk and making sure they get the right medications, coaching with them, working with them, has really helped. KP, HealthPartners, etc., are all managing care for those patients in essentially the same way (e.g. differently than the fee-based system). They identify high-risk asthma patients and get patients to change behavior, get people to not smoke and be active. KP actually has one of the lowest smoking rates in the country now.
Goldsmith: Could this have been a selection effect? That is, adding healthy high-deductible patients or marketing to people who want, as you put it in your ad campaigns, to “thrive.”
Halvorson: It’s pretty hard to quit smoking by selection. I doubt we drove away smokers by the “Thrive” campaign.
Goldsmith: Isn’t it brilliant selection strategy to recruit people who would come to you for the institutional support to live a healthier lifestyle? Wouldn’t this help offset the tendency selectively to retain the multi-functionally impaired folks?
Halvorson: We know that in a risk panel, you never lose the people who are really sick. They have a hard time leaving in any care model. People who have cancer almost never change health plans.
Goldsmith: That was certainly the case before the Affordable Care Act ban on pre-existing condition restrictions. . .
Halvorson: Even when they have a choice, when somebody has cancer — I saw it both in California and Minnesota — that’s a time where they really bond with their doctor. It’s not a time when people go shopping, when you have stage two or stage three cancer. But one of the things we did at KP over the last several years was to get really, really good with cancer care. People used to say that KP did a great job of population health, but worried, “What if I get cancer?” We talked about that very explicitly. We knew that we needed to take that fear away from people.
When you look at measurable cancer care outcomes, Kaiser Permanente is doing as well or better than other sites in our communities. When you look at colon cancer and prostate cancer, KP now has better survival rates than the rest of the country. That’s because KP oncologists examine the very best current science for a particular cancer, not what can be earned on the chemotherapy. Morale among the oncologists is quite high.
The Unfinished Business Of Health Reform
Goldsmith: Let’s switch gears and talk about health reform. You were in Washington a lot during the health reform process. What were you trying to accomplish?
Halvorson: Well we’re the only industrialized company that doesn’t cover all of its citizens and we spend twice as much money on care as any other country. I think the combination of those two facts is just absurd and obscene. So I wanted to make sure that we covered everyone. I believed you could do that by using health plans. Switzerland covers everyone, and they only use private health plans. Bismarck invented them; he did not want the German government in the health care business, so he set up separate health plans to cover sick Germans.
Most of Europe uses this model. If you go to the Netherlands, there’s not a single person in the Netherlands that doesn’t have government-guaranteed universal coverage, but it is provided through private plans.
Goldsmith: Did you think it was even remotely possible that Congress would put the health plans out of business and just have a single, government payer?
Halvorson: Well I wasn’t actually being defensive about single payer. I was being proactive about universal coverage. I wanted to make sure we implemented a health care reform that would cover as many people as possible.
Goldsmith: But for all the anguish, according to CBO, at full implementation, the ACA will still leave 31 million Americans uncovered. We didn’t even come close to universal coverage.
Halvorson: I think of that as unfinished business. We didn’t get to where we needed to get.
Goldsmith: And why was that?
Halvorson: All of the countries in Europe that have universal coverage have gotten there by requiring everyone to have coverage, and that was a bridge too far for us. We did it with Social Security, but we couldn’t follow that model and do it for health care.
Goldsmith: Wouldn’t it have cost like triple what it ended up costing to cover all of the people who were left out? Wasn’t the fiscal constraint part of the problem?
Halvorson: I think we could make the existing amount of money go a lot further. What I argue in my book is that we could do a much more efficient job at delivering care if we could get everybody into care delivery macro-systems and do the right things relative to care improvement and prevention. There’s enough money in the system now to do everything we need. I know that to be true because all the other countries in the industrialized world cover everyone and spend less money for care.
Goldsmith: Don’t they do that in major part because they pay everyone a lot less? You talk about your book about how much higher our prices are. But, those high prices are really peoples’ incomes.
Halvorson: I think there’s enough money there now in total. We don’t have to reduce incomes but we need to redistribute the way we buy care. We don’t need to close hospitals. We already have the most efficient hospitals in the world but we need to do is step up one more level. We need everybody to have coverage but fewer people going to the hospital, and we end up spending the same money in other ways. So I think we should’ve been able to do it for what we are spending in total now and cover everyone.
Goldsmith: Wouldn’t you therefore have had to mandate payment reform instead of the “let a hundred flowers bloom” approach taken in the ACA?
Halvorson: I think we buy care in a very inefficient and ineffective way. We reward screw ups; we encourage bad outcomes; we reward duplication and waste by how we pay for care now. I don’t think we need to ration care. That is the wrong answer. Repackaging how we pay for care is the right answer and we’re not doing it. If we had hospitals in combination with teams of physicians and paid them flat amounts of money to take care of populations of people, we could do it both without reductions in income and a significant improvement in care. We can’t get there as long as we’re going to stay wedded to fee-based models.
Changing The Cash Flow
Goldsmith: There’s been a very high casualty rate in converting care systems built on fee-for-service to an integrated care/financing model.
Halvorson: Some have failed, but the reason is that they didn’t really change. I argued in the book that you cannot change the delivery of care until you change the cash flow of care. Cash flow first. All of those people who tried and failed did it by changing the care model without changing the cash flow at all. So of course they’re going to fail.
I was working in the hospital sector when DRGs [the payment groupings in Medicare’s hospital inpatient prospective payment system] hit and what happened — immediately — was people stopped doing preadmission x-rays, people stopped doing Friday admissions for Monday surgeries, etc. All of those things changed overnight.
Goldsmith: In 1988 ProPac (the forerunner of MedPac) proposed expanding the DRGs by adding the associated hospital-based physician fees to them, and the affected docs stopped it. In the 1990s, there was a proposal to pay a lump sum for heart care on a competitive basis. Rumor had it that several very influential large health systems prevailed on Congress to stop it. Hasn’t post-DRG history shown that we lack the political will to change the payment system fundamentally?
Halvorson: We do need the political wisdom to say that we’re going to pay for care by the package, that we’re going to very carefully monitor the quality of care and outcomes of care, but we’re going to make paying by the package a reality. I argued in the book that we need to give hospitals three-year rolling price guarantees. If you’re going to re-engineer, you need to have predictability to mitigate the risk.
Goldsmith: How do you define the package the system is going to pay for?
Halvorson: The total care of a patient for a fixed amount of money per year. Yes, it’s capitation — capitation with really good quality measures and appropriate risk adjustment. Kaiser Permanente knows how to do it.
Goldsmith: Kaiser has also worked 70 plus years to figure out how to do it. . .
Halvorson: The vertically integrated model didn’t spread to other sites because nobody else got paid for packages of care.
Goldsmith: The Mayo Clinic is even older, vertically integrated and incredibly successful.
Halvorson: They still get paid by the piece. They deliver great care, but they’re not bringing down the total cost of care. Virginia Mason and half a dozen other first class group practices have great vertically integrated infrastructure. If they were flipped from the fee-based model to a capitated model, they could take their game up a notch and make significant improvements from where they are now very quickly.
Goldsmith: There are profound philosophical differences in the health insurance community about the advisability of paying health care systems by the package. A significant number of health plans that don’t want to do it.
Halvorson: Well, at least with Medicaid managed care, the plans are controlling care. Then you don’t have to pass the risk on.
Goldsmith: They are also using 1970s managed care tools like prior authorization. Do you really think these plans are adding a lot of value?
Halvorson: Are they the very best models? No. But they are going to change care significantly for Medicaid beneficiaries for the better. The non-managed care Medicaid programs had no quality practice metrics, no monitoring, no accountability. Even using these venerable tools, it is still an important step in the right direction. Medicare is the biggest player. Medicare needs to make a smooth transition to a new way of paying for care.
Goldsmith: Are you optimistic based on what you’ve seen from the CMS Innovation Center that Medicare is prepared to do it?
Halvorson: I think Medicare will need to go through another major crisis first. When the next Medicare crisis hits, if you already have half of the patients in the country in prepaid, coordinated care models, it is going to be an easier transition for Medicare to the next model.
Public Health And Disparities
Goldsmith: Let’s talk a little bit about the public health projects that you’ve decided to take on post-Kaiser.
Halvorson: The number one issue that I’m working on right now is the fact that the first three years of life for children are the years when their brains develop. If you exercise children’s brains in those first three years, the children have a total different life trajectory. They are much more likely to learn to read. If the kids aren’t able to read in school, they are 40 percent more likely to get pregnant, 60 percent more likely to drop out of school, and 80 percent more likely to go to jail. Look at the jail population. Right now 70 percent of people in jail either read poorly or can’t read at all — that’s basically a public health issue. It is linked to the fact that those kids did not get neuron development and brain exercise in those first three years of life.
So I’m working in California right now to make sure that as many kids as possible get the right brain exercise in those first three years. I chair the First Five Commission for the State of California, appointed by Governor Brown. I do believe that if we don’t fix this issue of early brain development, we’re going to doom this country to perpetual failure for major segments of our population.
Goldsmith: You’re also written a book called Ending Racial, Ethnic and Cultural Disparities in American Health Care. What’s your argument in this book?
Halvorson: The disparities book is both candid and directional. Disparities in health derive from three sources: biology, behavior, and bias. For example, African-American males who have a heart attack are 40 percent less likely to get the right treatment for a heart attack. Peoples’ behavior — the ability to eat well and live an active life — is a major driver of disparities. The net effect is that African-Americans and other minority Americans live several years less than White Americans. So the book talks about the care gaps that exist and identifies several things we need to do to close them.
Goldsmith: Isn’t the political condition of the country a major constraint here? How do you make the resource allocations decisions that need to be made to reduce disparities in a highly polarized political environment?
Halvorson: One thing we need to do is have both Medicare and Medicaid be accountable based on race. We need to collect race and ethnicity in the reporting systems and we need follow up so we can focus on patient needs that may differ by race or ethnicity.
Goldsmith: I’m trying to remember when I enrolled in Medicare whether I was asked about my ethnicity? I don’t think I was.
Halvorson: That information should be known because that’s a legitimate health care issue. Anyone who reads my book on disparities will see why that’s not an intrusive thing to ask. It’s actually a helpful thing to know. The book basically says that we need to do what we can to close the gaps that exist today. We closed some of these gaps at KP and others can close them as well.
Goldsmith teaches at the University of Virginia. He is a frequent contributor to The Health Care Blog and also at The Health Affairs. This post first appeared at The Health Affairs Blog.