Providing Insights that Contribute to Better Health Policy
Insurance Coverage & Costs Costs The Uninsured Private Coverage Employer Sponsored Individual Public Coverage Medicare Medicaid and SCHIP Access to Care Quality & Care Delivery Health Care Markets Issue Briefs Data Bulletins Research Briefs Policy Analyses Community Reports Journal Articles Other Publications Surveys Site Visits Design and Methods Data Files
Patient Cost Sharing: Promises and Pitfalls
Paul Ginsburg: If you could take your seats? Im Paul Ginsburg and I want to welcome you to the conference, "Patient Cost Sharing: Promises and Pitfalls."
By way of background, there are very important trends going on now in employment-based health insurance towards greater patient cost sharing. Weve seen substantial buydowns by introducing more cost sharing into the benefit structure, and people Ive talked to expect that this trend will only increase.
In addition, just last week in the Medicare legislation, theres a very important provision on health savings accounts that will likely give strong encouragement to increasing the use of deductibles in employment-based health insurance as well as in individual coverage.
Now, this movement towards more patient cost sharing really is a reflection and a reaction to some major trends concerning managed care. Think of the unpopularity of restrictions on care by insurers, the emphasis on broad provider choice, and the accompanying trends, the premium trends, are way in excess of income trends, so that health insurance is becoming less affordable to many.
Now, there are a number of issues around patient cost sharing. One is whats the potential for cost control, how much can be achieved through this tool. Another is how often is it a barrier for people getting the care they need. And, finally, does it cause a great deal of financial disruption for families?
Now, if cost sharing is to play a large role, it probably needs to be refined; probably needs to apply more heavily to discretionary care; probably needs to apply more heavily to less effective providers or to patients that choose less effective providers; needs to protect low-income people from serious financial burdens. And these questions are what this conference is about.
This is not about whether cost sharing is a good or bad thing. My perspective is its happening, whether we like it or not, and this conference is really what are some of the ideas, what products might be rolled out, the thinking of leading-edge employers about how to take this cost sharing and make it a more refined instrument for the control of costs and maybe even the improvements of quality.
We have a number of terrific speakers here today, and heres how the conference will go.
First, two researchers from HSC will set the background for the conference. Joy Grossman will introduce the overall concept and issues in patient cost sharing, and Sally Trude will present results from an HSC study being released today on how cost sharing impacts different types of individuals.
Then were going to hear from John Bertko, who is an executive at a large insurer specializing in fully insured products for small employers. I think Ive got that right.
John Bertko: Yes.
Paul Ginsburg: Good. And then were going to hear from Arnie Milstein, who is a consultant to very large employers that tend to be on the leading edge in their health benefits programs.
Then we will have a break, and then a reactor panel, which will represent four different perspectives, although the panelists dont have to be locked into the perspective that--the perspectives of consumers, of employers, of physicians, and those who are leading the efforts to improve quality of care.
We will have question-and-answer time for the audience, and were going to continue to use the technique that we adopted from the Alliance for Health Care Reform of having a mix of oral questions from the microphone and written questions from the cards that you have in your packet. And well be collecting cards at two occasions: once during the break and then after each of the reactor people has made their initial remarks.
I want to thank the Robert Wood Johnson Foundation for their support for this conference and for all the work of HSC, and also point out that they are webcasting this conference on their website, www.rwjf.org.
Another announcement is that, as some of you know, HSC is transitioning its publication dissemination process to an entirely electronic process, and if you havent already done so, please go to the website and sign up for e-mail alerts so that you can be alerted when each publication is being released. This is the last month for mailing of publications.
So let me begin by introducing Joy Grossman, who is a senior health research at HSC.
Joy Grossman: Thanks, Paul.
I see it as my job to try to help everybody switch gears from thinking about Medicare legislation to thinking about what Paul noted is, of course, a very important trend in the employer-sponsored insurance market. And as he said, faced with rapidly rising health insurance premiums and a continued backlash against managed care, cost containment efforts have really shifted from focusing on trying to change provider behavior to looking at changing consumer behavior.
And employers have a number of levers they can use to try to offset at least some of the premium increases that theyre seeing: first, they can try to increase the proportion of the premium paid by the employee; they can reduce the services that are covered under the benefit package and ask the employee essentially to pay for those services out-of-pocket; or, lastly, they can think about increasing the patient share of the cost of services through deductibles, coinsurance, or copays.
And, to date, employers have really focused mostly on this last option, and its particularly attractive to them because it lowers the total premium, and that means that both the employers share is lowered but also the employees share of the premium or contribution to the premium is lowered as well.
This reduction in premiums comes from two sources: one is shifting costs onto employees that use services, and the second is providing incentives for patients to use less care.
During our site visits, we really found a lot of diversity among employers in how they are increasing cost sharing over the past few years. Essentially, theyre making incremental decisions as they look at their existing package and try to figure out how to reduce the premium increases theyre facing. And they have a lot of different options that they can play around with, so to speak, in adjusting their cost sharing.
First of all, at the most basic level, weve seen increasing copays and deductibles. For example, the Kaiser Family Foundation Employer Survey showed that there was about a 60-percent increase in the PPO deductible over the last several years.
Flat copays are in some cases being replaced with coinsurance, which is a percent of the bill; so obviously for more expensive services, such as inpatient services, this will have a larger impact on patients.
All three types of this cost sharing—copays, coinsurance, and deductibles—are being applied to a broader range of services, for example, inpatient care.
Were seeing a lot of enhancing of the prescription drug tiering, and this is an example of sort of incremental approaches. Some employers are still moving to the three-tiered approach; others are adding a fourth tier for lifestyle drugs, adding coinsurance or deductible or otherwise playing with the levels of copayments.
A really critical component of cost-sharing design is the annual out-of-pocket maximum. Typically, theres a cap on what patients are required to pay out-of-pocket in a given year, and so far the Kaiser Family Foundation Employer Survey hasnt shown a lot of change in this, but this is an area to watch where employers have an option of increasing the maximum out-of-pocket that patients are required to pay or could change whats included in that and start excluding services from that maximum.
And as everybody is aware, there are a lot of new products being marketed out there--consumer-directed health plans and tiered networks--that kind of incorporate more sophisticated cost sharing.
And the bottom line really is that cost sharing is increasing, but its also becoming much more complex as employers choose from among these different options.
As I said, cost sharing lowers premiums in two ways: by shifting costs onto those who use care and by creating incentives to reduce use of services. And, ideally, youd like to reduce the use of inappropriate care, but research has shown, particularly the Rand Health Insurance Experiment has shown that patients faced with higher cost sharing typically cut back both on needed care and more discretionary care. And this suggests that traditional cost-sharing designs are blunt instruments for trying to target reductions in use. For example, with drugs copays, it is just as likely to reduce necessary utilization by people with chronic conditions as it is to address overuse of antibiotics.
And theres a particular concern that low-income people or those who face higher health care costs because theyre ill will forego needed care because of these higher costs. And Sally is going to, as Paul said, talk more about the financial impact of particular subgroups of individuals.
But really what we want to know the bottom line is, whats this impact of foregoing care on health status? And theres some research from the Rand Health Insurance Experiment and other studies that has suggested that there isnt much short-term impact on health status for the average patient; however, we dont really know much about the long-term effects.
Today were going to be talking about ways to help try to refine cost sharing that helps steer patients towards more cost-effective options for care while trying to avoid the financial barriers to needed care. And one way to think about this is giving patients choice in the type of care that they get that would require them to pay more for less cost-effective options, and obviously this is the focus of Johns and Arnies presentations. But just sort of on a broad brush, as Paul mentioned, you can think about patients, for example, for particular diagnoses or facing particular treatment options that they be given incentives to choose the most cost-effective option by facing lower copays for those. Patients with certain types of chronic conditions might face lower cost sharing if theyre willing to participate in a disease management program. If you try to steer patients toward more efficient providers, for example, were already seeing plans that have lower copays if patients are willing to go to centers of excellence for particular major procedures.
And, ideally, in these designs you want to limit cost-sharing cases where care is necessary and theres not a lot of discretion in treatment options and where you want to make sure that the patients do get the care. And as an example, if youre somebody who has diabetes, you want to reduce cost sharing for preventive services so that they have an incentive to get that care.
And in the case where theres discretionary choices to be made, you want to minimize cost sharing for the patients who choose the most cost-effective option.
Obviously, there are a lot of issues to explore, which well be doing today, related to designing and implementing these types of cost-sharing arrangements and what the bottom-line impact is of them. And Im just going to kind of walk through a few of these, and hopefully well have more discussion about some of them today.
First of all, do we have adequate information about cost-effectiveness and the like in the clinical literature? Do we know what patients responses will be to different kinds of cost-sharing mechanisms? And assuming we can design these kinds of cost-sharing arrangements, will employers be interested in pursuing them? And large employers may be interested, but what about small employers having these options?
In terms of implementing them, we already know that patients dont know a lot about their health plans. Can we give them the right tools and information to both assess their cost-sharing options and treatment options that they might face? And we know that physicians are still the ones who ultimately make the clinical decisions. Recent research has showed that physicians really dont know much about cost sharing and that they dont talk to their patients about it. So would they know about it? Would their treatment preferences align with the incentives that the patient faces? And, more broadly, will the quality improvement efforts, for example, pay for performance, would those incentives align with the incentives the patient faces? And, ultimately, are these things too complicated to administer? You know, if you have to update--if medical effectiveness changes over time, do you have to update these cost-sharing arrangements? We already know providers are having trouble collecting copays and the like from patients now. What would happen as these things increase over time and are more complicated?
And as Paul said, the bottom line is how much money can we actually save from these types of options. We know that the bulk of expenditures comes from a very small proportion of the population. It raises a lot of issues that we have faced with the managed care backlash, issues about accountability for determining medical effectiveness and steering people toward certain types of care. And, lastly, theres a potential issue for risk selection. If employers offer employees different plans with different types of cost sharing, might you see the riskful separating between the healthier and sicker employees?
So, with that, Im looking forward to our discussion about these issues, and Ill turn it over to Sally.
Paul Ginsburg: Thank you very much, Joy.
Next, Sally Trude, whos also a senior health researcher at HSC.
Sally Trude: As Joy mentioned, theres been a universal trend of employers towards increasing cost sharing, and this raises the question of how far can this trend go and what will be the ultimate effect on the out-of-pocket costs that patients pay.
We have new research that translates the employer benefit designs into what the expected out-of-pocket costs are likely to be under the various benefit designs. We also examine the effect by health status and income. And I should also mention that you have an issue brief in your packet which describes this study in more detail if you want to get into some more of the actuarial points behind it.
And while theres been a universal trend towards increasing patient cost sharing, the employers are at different levels of cost sharing. So you have some employers who are still sort of at the low copays and are just now increasing the copays or introducing new ones, like an inpatient hospital per diem and/or an emergency stay. Others have sort of maxed out what they can do with copays and are now introducing coinsurance. And then theres sort of the buzz out there about all of the high-deductible product offerings.
The work that Im going to be presenting today is based on some actuarial models developed by Jim Mays and Monica Brenner at the Actuarial Research Corporation, and they used a 1997 Medical Expenditure Panel Survey to estimate 2003 average out-of-pocket costs.
And following along that scenario or the continuum, if you think of it as a continuum of employers incrementally increasing the cost sharing, our scenarios follow that continuum. So what we call the baseline option is the low copayments of that $10 per physician visit, and then the next level of copays doubles the copay for physician visits and introduces the per diem hospital stay and has about a $150 emergency room copay.
Then the next four scenarios are raising coinsurance rates and the deductible from a $100 deductible on up to a $2,500 deductible. And I should also mention in this that the models assume as part of it that, as the costs increase for people, theyll reduce the amount of care they seek or that they get.
We also assume in this that if an employer is going to go for a high-deductible option—which hasnt really happened yet, but if they do, we expect that they will also increase the out-of-pocket maximum cap. And so in all of these models we assume that that cap is $1,500 over the deductible.
So, not surprisingly, we see that the average annual out-of-pocket costs increase under this continuum of higher cost sharing, so you have about $50 under the lowest, up to about $600 average out-of-pocket costs under a $500 deductible scenario, and it goes all the way up to just over $1,000 per year on average with the $2,500 deductible option. But this, of course, includes people who arent seeking care, and what we also see is that the risk of facing large out-of-pocket costs also increases along this continuum of increased cost sharing.
So in the lower ends, at the baseline and higher copay, no one is paying over $1,500 because of their out-of-pocket maximum limit.
Then we find that although the average out-of-pocket costs increase for the chronically ill, its very striking for those who report themselves in poor health or who have been hospitalized during the year.
In fact, we find that for those that are hospitalized, over 90 percent of those hospitalized will exceed $1,500 in out-of-pocket costs under the highest-deductible and coinsurance options that we modeled.
But this also depends on how much income a person has, and so we find that for those in poor health or who are hospitalized, one in four will end up spending more than 10 percent of their income in out-of-pocket costs under the higher cost-sharing options.
And the financial burden also increases for low-income workers. For those between 125 percent and 200 percent of the federal poverty level—thats between $11,000 and $18,000 per year for a single person—16 percent would spend more than 10 percent of their income on out-of-pocket costs for the $1,000 deductible option, and that would go up to 23 percent under the highest option, which is the $2,500 deductible.
So we see from all of this that the financial burden increases the most for the seriously ill and for low-income workers. And while increased cost sharing reduces discretionary care, it also reduces needed care. Employers are, therefore, probably not going to be going into the far extremes because theyre going to want to target the discretionary care and not necessarily set up a financial hardship, financial and medical hardship for their workers. But increased cost sharing from the lowest levels to more moderate levels is going to raise patients awareness about the costs of care. But the part of it that will be most interesting is what part will the physicians play in this debate because up to now costs have not been part of the patient-physician relationship, and so that will be the dynamic that well have to see change, because although consumers will become more aware of the costs of care, we need to figure out how that will translate into a dialogue about cost-effectiveness within the doctors office.
Paul Ginsburg: Thank you very much, Sally.
Id like to introduce our next speaker, whos John Bertko, who is vice president and chief actuary at Humana, Incorporated, whos going to talk about the innovations in cost sharing that companies like Humana are envisioning to roll out to their clients.
John Bertko: Thanks, Paul, and good morning. Id like to just describe a little bit of the perspective Im bringing today.
First off, like Paul and maybe Arnie, weve been around for a while. I started my actuarial career in the mid-1970s, and some of this I think is the pendulum swinging back and forth on cost sharing. Growing up, I was covered by a steelworkers plan. That was a base plus major medical, if anybody else here remembers that, and things changed to PPOs with deductibles in the 1980s, the deductible in front of everything. And then as HMOs became popular in the 1990s, PPOs tried to look like HMOs and went to low copays. And here we are now saying, Hmm, maybe we need to swing back the other way.
Secondly, just compared to the researchers here, I have to operate in the real world, and so--
Thats my usual jab at everybody here with a Ph.D. But the second part of that is, as the chief actuary, I dont do much real work anymore, but I do work closely with our product development folks and monitor my teams of pricing actuaries who actually put numbers from our experience on those. We also have a competitive intelligence unit which monitors everything kind of across the spectrum of major players, like Humana, like the Blues, United, Aetna, Anthem, and all the folks like that.
So today Im going to speak from an industry perspective in terms of what I see going on basically today on the leading edge and with a little bit of speculation on where stuff might go off into the future.
Heres what companies like ours, I think, see as the goals of these increases in cost sharing, and most importantly is to make costs and choices of treatments more visible to consumers. With low copays, weve done enough consumer research to know that most people think an office visit costs $10. It costs either $50 or $75 or more, depending on what kind of specialist you go to. Arnie is going to talk more about this, but its very important that we channel individuals seeking care to the most efficient providers and the most appropriate site of care for these.
That follows to number 3 here. We need to reduce the inappropriate use of new technology. New technology is great. A relatively recent study for the Blue Cross/Blue Shield Association said that perhaps as much as 60 percent of health care trend in a high-trend environment may be attributable to the introduction of new technology, and thats not only the direct stuff but its use of new services, new devices, imaging in particular, which well talk about, and a whole variety of things.
And then, lastly, to get to a point where we introduce positive incentives, and in particular for disease management and prescription drug compliance, weve found in some of our studies that, for example, lack of compliance on drug therapy is actually an early indicator of a future catastrophe. And the simplest one, if you stop taking diabetes drugs, insulin, et cetera, you might be on the road to other fairly dramatic body system failure.
Okay. So back to the 1990s versus 2003, weve moved from what Ive described as old cost sharing--that is, we went to low copays from deductibles and coinsurance, which were very, very prevalent. I have lived in California for 30 years and remember the days when being a Kaiser member was thought of as being in a cult.
You know, there were those that were in Kaiser, and there was the rest of the world. That changed dramatically.
Many covered services went from having a deductible in front of them, in particularly hospitals and prescription drugs. In the mid-1980s, there was almost no prescription drug plan that had copays. You had a $100 deductible, nominal dollars at that time which would be worth much more these days. And they became exempt from deductibles.
Mental health services used to have those kinds of things, plus all kinds of inside limits. Those now have moved to a completely different kind of management with the managed behavioral companies.
What kind of new cost sharing are we looking to? In particular, one of the big trends is tiers to aid channeling to these efficient providers; secondly, site- and service-specific deductibles and copays. I think one of the things that Joy mentioned here is that this is going to become more complex. Absolutely true. Its going to be more difficult for consumers. Its going to be somewhat more difficult for providers, although one of our competitors has introduced swipe-card technology where you go into a doctors office, he takes his Visa machine, swipes the card through it, and up pops up, you know, what the deductible, copay, or thing is like that. Were experimenting with that as well.
Those technologies are not all here yet, but they offer the promise of instant access to some of us.
There are fewer requirements for approvals. Paul cited this in his introductory comments. There were lots of barriers, checks and balances to getting care of various kinds during the 1990s, and a lot of that has gone away. Companies like ours and most others in the industry have responded to that by reducing the barriers in that form.
What are some of these current trends? Okay. More layers of deductibles and copays. We are no longer going to have--go back to the days of a single large deductible. Sallys comment here about not many employers having big deductibles is true. Paul cited the new legislation which now allows for offering health savings accounts for high-deductible plans, and as I recall, thats with $1,000 or more. In our larger group segment, which are those that are above the 50 level for small group rules--I just did a check the other day--we have about 2 percent of our members covered by $1,000 or more deductible plans. So theyre not very common yet. They may become more common, but those kinds of high deductibles are relatively rare in that segment. I dont have the numbers for our small group segment. I know its more. But it still is not very common.
Well have both tiered hospital benefits that may have copays and coinsurance. We are already offering tiered physician networks, and we offer two kinds and many companies do: primary care versus specialist, or within the different tiers, youd have perhaps—I think, Arnie, this is your favorite phrase—high-performance networks versus regular networks very out-of-networks.
Consumers are going to have to do more work to find out where the best site for them is and where the most affordable care is.
Prescription drugs are returning to coinsurance. I think Joy cited this one, that fourth tiers are added. In some cases they are for the most expensive high-tech drugs. Probably the worst abuse that I read in the newspapers is human growth hormone for people wanting to take it to become better athletes, or if theyre aging athletes like me, maybe I should take that to help me jump better when I play basketball. And then some products now being offered with drug copays that are only having coinsurance tiers and no longer copays.
A few other comments here. Use of cost sharing to change site of care. I think one of the centers recent papers talked about the increasing use by insured patients of emergency rooms, and that is probably coming from a variety of reasons, one of which being the prudent lay person rule, and the second, I think, Paul, you guys cited as the lack of availability of office visits on an easily-scheduled basis. Well, an emergency room basically costs an employer, a health plan or an individual 500 to $700 the moment that you walk in the door.
And if you can exchange that or lead to an exchange of that for a $50 office visit, its a great trade. You probably have better care, and because its more coordinated, the doctor knows your history and your prior lab results and of course, obviously, much cheaper. An in-between level of care, and again the copays are frequently set up this way, $150 for an ER visit, $75 for an ambulatory surgery center, and of course, between a 10 and $20 copay if youre just going to visit a physician.
Much higher cost sharing for certain imaging services. We have found this in that whole spectrum of outpatient treatments, imaging being one of the ones that is running up the fastest. So the inappropriate use of MRIs and CAT scans with higher deductibles, versus, say, those for regular x-rays and other simpler forms of radiology.
Health reimbursement accounts to modify medical consumption, and here Im talking about the part that is within what is sometimes referred to as consumer-directed health plans. They start with an account or an allowance of 500 to $1,000, and then after that is exhausted there is a large deductible usually, frequently 1,500 to $2,500. There are other conferences talking about that. It is certainly one of the parts of the health insurance spectrum that is most discussed today. Again, enrollment in those is still relatively limited. I think estimates for this year are running to between a million and 2 million people out of 160 million or so insured this way, as being in those consumer-directed health plans.
When doing this, other tools come into play, cost calculators, various decision support types of tools. In some cases, trying to indicate the cost levels, many of the companies like ours cant just say Hospital X costs $1,700 a day, and Hospital Y costs $3,800 a day. Because the moment we would announce that, our contract are invalid, and our competitive position goes away.
So instead, look for the little restaurant guides, you know, one dollar symbols, two dollar symbols, 4 dollar symbols, 10 dollar symbols.
I would say that based on our relatively limited experience on about now 25,000 people, it shows evidence of cost savings, not so much in terms of increased cost, but reducing the actual use of services and in particular inappropriate services like the emergency room and imaging.
How about incentives for better behavior? And some of these have been reported in the newspapers about giving—and Ill call them "frequent flyer points"—frequent user points, or good user points. Some could be applied for getting, you know, refrigerators and TVs, others for being applied to perhaps reduced cost sharing. Disease management here is probably one of the key ones. We can identify that spectrum of people who are good candidates for disease management. However, when we go out and carefully ask them whether theyd be interested in joining, only about 20 percent of them respond positively. The other 80 percent are a challenge to outreach by us, so can enrollment and then compliance with disease management programs be helped through these?
I think were going to see those experiments. In particular, rewards for prescription drug better compliance might be there. We know how to track these. We can see where people have gaps in coverage. Somebody takes their drugs for 2 months and then stops ordering drugs. I mean we cant follow them day by day, but we follow the pattern of whether theyre having their prescriptions refilled.
So, Paul asked me to speculate a little bit about what comes next. Many of us are using episodes of care to evaluate how efficient or affordable providers are. There is one company out there that is actually offering insurance based on purchase of episodes of care. This has a couple of problems with it. Its very easy for a few episodes, like normal delivery, to say what thats worth. Its very difficult for others. There are 15,000 CPT-4s, and when you combine those with treatment options you could have literally millions of different ways of episodes here.
The other thing about episodes, at least the way that we do it, you have to have sometimes a retrospective look at that. You have a couple of small office visits. You have some drugs prescribed. And then you have a major event. Well, the way we look at it is we take a backwards look to create the episode, and thats probably impractical from a cost-sharing perspective. Theres going to be much greater specificity for copays, on imaging here, for example. Not only would you have low copays for x-rays, but you might have medium copays for MRIs and high cost things for medically indicated things. I mean if I fall off my bicycle and hit my head and go there, I would hope that theyll do an MRI on me. Now, more likely, I will have jammed or dislocated my little finger, like I did about 3 months ago. Dont need even an x-ray for that. They know whats wrong with it. They dont need an MRI. So higher copays might be indicated for things that are more discretionary.
And then the last thing here in my list of very much future is right now most companies like ours have contracts for hospital systems for all care. Ill let Arnie address that, but I would speculate that from our research, that many hospital systems are good at one, two or three things. Hospital System B is good at four, five and six, and in the perfect world we would buy from those hospital systems that deliver the best services both in quality and efficiency and not just slug it out with everything from a certain hospital system.
Again, what comes next? Heres one that I think is problematic. Again, back to my comment about being insured in the 60s and 70s here. We had plans which basically said you can have $1,000 worth of hospital care at the time; you can have surgical and professional care covered up to a certain amount. Those are being offered again by what I would call boutique health insurers, and theyre very low premium, and some people might suggest that theyre relatively low value, and the question is whether its better than none.
So I guess Ill sum up mine by saying that more cost sharing is likely in the future and its likely, at least from an employer and I think an insurer perspective, for getting better value from purchasing. The world, our consumers out there have told us they dont want managed care to tell them whats good value. Instead we seem to be heading down the path of having to have consumers decide whats better value. So this high continuing trend being partly at least as a result of lots of new technology.
David Wenberg [ph] presented a couple of months ago some information about the very, very high variability of supply-sensitive services. I would suggest you look at some of that research, and that using better and appropriate sites of care is important. So better cost sharing might promote centers of excellence, give better quality and at least for actuaries, you might be able to reduce trend.
Paul Ginsburg: Thank you, John. Could you just stay for a second? I have a question for you. I was wondering if you could comment on when these ideas are presented to the small employer, the person or the small employer that buy health insurance, or the brokers that interact with them, whats their reaction to that? How receptive are they?
John Bertko: Well, I will be overly-candid and cynical. The three most important things to those small employers and the right broker is price, price, price. If price comes down and its not overly complex, theyll buy it. If price does not come down and youre just trying to sell it on better quality of care, you might as well give up.
Paul Ginsburg: Thanks. Thank you, John.
Id like to introduce our next speaker whos Arnie Milstein, and I like his title so it want to say it to you. Hes Medical Director of the Pacific Business Group on Health, and the National Health Care Thought Leader of Mercer Human Resource Consulting.
Great to have you here, Arnie.
Arnold Milstein: Thanks, Paul.
Im going to switch the frame in two ways. Im going to talk about the thinking of large employers with respect to cost sharing and also try to move the frame for describing whats happening out a few years. A lot of my work involves benefits, strategy, planning exercises with large employers, so Im going to talk about what I see crystallizing on a 3- to 5-year basis, understanding that forecasts are dangerous.
Why are large employers reexamining cost sharing? Well, the reasons are self evidence. Every year the Business Roundtable, Americas Club of Fortune 500 CEOs polls its members on what is the number one cost problem in the entire corporation? And this part year it wasnt government regulation, it wasnt unions. It wasnt trial attorneys. It was health benefits cost, and nothing was within 20 percentage points of the rating that they assigned to health benefits cost.
Simply getting rid of it is obviously impossible because if you also--MetLife, for example, periodically surveys Americans about whats their number one financial security concern? And this year, again, its been going on for several years now, the number one rated item was continuity of health insurance, and that was rated above continuity of job and having enough money to make ends meet. So as hot as the potato is, its hard to let go.
Do the employers want to be in the business of kind of economic engineering and trying to figure out how you come up with consumer incentives to improve the value of health benefits? They dont want to do that at all, and if they thought that they could rely on professionalism or tighter managed care or government regulation to make this happen, they would happily do so, but theyre not optimistic about being able to rely on these alternative solutions.
Let me say that what Im about to talk about is in no way attributable only to large Fortune 500 companies. I do plenty of work with large state employee benefit plans and their boards, and with large Taft-Hartley operated by unions, and this perspective is widely shared.
With due apologies to Carl Jung, Ive tried to abstract my planning work with large employers into three kind of dominant—Ive called them archetypes. Lets just call them, you know, Three Face of Cost Sharing. And these are very rarely visible in pure form.
But the first archetype are the restorers that seek to restore what historically 30, 40 years ago was a much higher percentage share of beneficiary health benefits cost sharing. And they define the problem as an unintended or imprudent purchaser of cost share creep over the prior 30 years. In other words, that theres been a major shift in the percentage of the bill paid. And their preferred solution is to hollow out benefit coverage within all health plan options, and John has described that phenomenon.
Another common archetype is what I call the skin grafters, and here the notion isnt necessarily to transfer more cost, but even holding the plan, the value of the plan actuarially constant to so-called get more beneficiary skin in the game, their definition of the problem is the moral hazard of insurance, that if youre spending other peoples money, youre spending more of it than you would be spending if you were spending your own money. And a common solution to this is to offset a higher deductible with a portable spending account of equal or almost equal actuarial value so that you feel you own the dollars.
A third archetype is what I call the calibrators. They seek to calibrate the beneficiary cost share to traces by beneficiaries that involve discretionary inefficiency. In other words, for example, picking a treatment option that may have an equal expected outcome, but its much less cost effective. And the classic example of this, insistence on a brand drug when a generic drug is available, is a very clear-cut example of this.
They define the problem as that beneficiaries are day to day in choosing doctors and treatment options and hospitals, selecting inefficient options, and theyre preferred tool is to incentivize the selection of efficient—and well talk about this in a few minutes—and perhaps higher-quality options both before and then after on a continuous basis plan enrollment.
If I had to sort of forecast how this might crystallize over a 5-year time frame, I would say in the near term, as John has described, the restorers are predominating, the notion of simply lets so called buy down benefits, lets pay a smaller percentage, and thats obviously the simplest tool and thats predominating in the near term.
I think intermediate term, particularly with this new Medicare feature, new Medicare legislation, there probably might be a little bit more activity amongst skin grafters, but I think longer term, this more refined approach to cost sharing in which beneficiaries are being incentivized to select more cost effective, and I hope, higher-quality options is likely to predominate, because it prevents a lot of the--Ill call it the unintended negative consequences of higher cost sharing.
What Im about to comment on obviously doesnt apply to all large employers or to all employers. I dont have a good knowledge of the small employer market, and I really defer to John on that. It may apply less to large employers who view their labor forces as commodities rather than tight markets in which they have to compete, and obviously doesnt apply to desperate employers whose businesses are on the rocks, who are not constrained by labor agreements.
So what Ive done is, Ive tried to sort of bring you into my world and tell you that when you do benefits planning work on a 2- to 5-year horizon, and you being to focus on cost sharing, the questions you end up answering fall into six primary buckets, and Im just going to walk you through each of those six primary questions and show you where I think the large employers are beginning to land, not tomorrow, but on a 2- to 5-year going forward basis.
The first question is just the question of what fraction of average per-beneficiary-health care spending will employers pay? And I think the predominant answer that I sense crystallizing in the planning exercise I participate in, is enough cost shared to attract and retain the labor force they need, and clearly, as John and Paul and as Sally and Joy described, this fraction is decreasing in todays weaker labor market and economy.
You know, one question here is: well, in factoring in, in running this calculation, will employers--how much weight will employers attach to the so-called indirect cost savings? You know, in other words, reduced absenteeism and productivity losses due to health problems, how much will that be taken into account? It does get taken into account by the smarter, I think, larger employers, but it tends to be significantly discounted because it is not very visible in employers financial accounting systems.
Question two: how will a beneficiarys amount and percentage cost share be linked to individual beneficiary distinctions? This is obviously a very pivotal question. And I think the center or the answer that Im seeing beginning to crystallize now is that more will be paid by service users, by those with dependents—this has obviously been true for a while—by the more affluent—and Ill talk about that in a minute—and those beneficiaries making certain types of selections which increase spending for the insurance pool, and Im going to focus on that in a minute.
So lets take that last question and sort of say, exactly which beneficiary selections do I think will end up carrying a higher cost-sharing burden? And I think the answer is really split into—Ill call it pre-2004 and post-2004. Pre-2004, I think that the beneficiary selections that will carry a higher cost-sharing burden will be focused on bundled annual selections by beneficiaries. This will, for example, more cost sharing will go to beneficiaries who select a richer plan of benefits. A so-called less efficient health plan, "less efficient" being defined when you take a health care and its premium and you divide it by enrollees predicted cost, and you get kind of an efficiency rating. Those people selecting plans that score unfavorably on that health plan efficiency calc will end up paying more.
And there are some employers—I think this will increase—that are also beginning to charge more to employees that dont complete an annual health risk appraisal because of the utility of health risk appraisals in beginning to help people longitudinally manage their risk. This is a rare item. This is not a frequent item, but I sense its increasing. A good example of this is, for example, Caterpillar. Caterpillar, you pay substantially more out of pocket if you do not fill out a health risk appraisal.
Post-2004 I think the answer will begin to change a little bit. I think youll see more cost sharing going to beneficiaries first who do not participate in personalized programs to reduce the risk of illness and/or cost in the future, for example, declining to participate in a disease management program, not participating in a shared decision support program with respect to major therapeutic alternatives, et cetera. As John mentioned, I think youll start to see, again, on a continuous basis after enrollment, greater cost sharing if beneficiaries select a less longitudinally efficient—Ill talk about that in a minute—provider or a less longitudinally efficient or cost-effective treatment option.
There were a few concepts there that I mentioned. Let me talk about those briefly, a little side diversion. If you sort of say, well, how did you come up with variable cost sharing depending on choice of provider, choice of participation or non-participation in longitudinal health management program, or more or less cost-effective options? It really derives from current beliefs and near-term experience as to where the biggest opportunities for harvesting efficiency without jeopardizing quality lie. And this is an analysis of percentage point savings opportunities that was pulled together and peer-reviewed by a group of national health economists that I coordinated about a year and a half ago, trying to answer this question. And as you can see, the three items on my prior list dovetail pretty well with national economists estimates of where the primary opportunities for efficiency harvest lie. It is in the beneficiaries selecting more efficient providers to participate in longitudinal health management programs and to select more cost-effective treatment options.
Second point is that up until now is weve thought about linking our beneficiaries with more affordable doctors and hospitals. Historically the focus on defining affordable has been on unit prices, you know, what fee schedule are you willing to cooperate with, whats the average per diem being charged by a hospital? I think the large employers are beginning to wake up and realize thats a ridiculous metric. That is a crazy basis for deciding how to direct employees to certain doctors and hospitals. Theyre beginning to realize that its longitudinal efficiency over the course of an episode or a years worth of chronic illness that is the relevant metric for cost effective they need to be focused on.
I wont go into this Power Point because its pretty self evident, but youre going to start to see a shift as to which providers are defined as the desirable providers, whether its for purposes of network inclusion or steering through cost sharing within a current network that you offer.
In the time permitted I dont have—I love this slide, but I kept pushing a lot of carriers who I work with. I say, would you please—would one of you that has a lot of claims density in a given geography, please show me for doctors or small medical practices within your geography how these practices distribute with respect to current, first-generation very crude measures of longitudinal efficiency, typically using something like ETGs to a case mix adjust populations and quality of care, again, using very crude, first-generation methods of quantifying rate of compliance with evidence-based medicine.
This is a first return. I think Regence Blue Shield for producing this. The quality index is on the vertical axis. The longitudinal efficient index is the horizontal axis. As you can see—and this is I think probably not surprising to health services researchers—that we are today in a world in which there really is such a thing as a higher-quality, higher-efficiency provider. Thats what you see in the upper right-hand tier here. And researchers like David Eddy [ph] have said if you were to take treatment options and put them on the same grid, you would also see significant differences. And David has recently published some great research showing that different treatment strategies for managing diabetes land in very different places on this longitudinal efficiency quality quadrant.
Question four: How much of the estimated incremental cost burden from suboptimal selection decisions will be paid by the beneficiary making the decision? Sorry for that question, but it does turn out to be one thats relatively important. And theres to what degree do you want to fully transfer the cost, the incremental cost of inefficient selection decisions onto beneficiaries?
And I think the answer is increasingly going to be close to all of it, but I put here "subject to income tier limits based on the 20/20 ogre test." The 20/20 ogre test is the test, if you get a situation where a beneficiary cannot afford care due to your benefits design, and is interviewed, and you are interviewed on 20/20 as a result of that care--
--do you come across as a complete ogre to your coworkers? And this is the heart of the calibrator archetype, and I think the good news is as employers begging to think about differential cost sharing, I think at least the larger, more thoughtful employers are realizing that they have to begin to think about income tiering the maximum out-of-pocket limits.
And Ill give credit to illustrative company. Rockwell Automation, as they begin to move towards more cost sharing, has now begun to, for the first time, income tier their maximum out-of-pocket limits, which I think is a good idea. I think it will spare them failure of the ogre test.
Question five: Will employers transfer to beneficiaries the estimated incremental cost of selecting inefficient options in a bundled fashion via annual or maybe even multi-annual selection of plans, or using Vivius [ph] as the example, sub-plans, where Vivius is a health insurance solution in which you pick on an annual basis not just your plan and your plan design, but also which combination of providers you wish to use, and that then determines your premium. Or will it occur primarily in an unbundled fashion over the course of the insurance year via continuous beneficiary selections, you know, Doctor A versus Doctor B, Hospital A versus Hospital B, Treatment Option A versus Treatment Option B, after plan enrollment.
And I think the answer is both. You know, the first obviously is what Alain Enthoven had in mind in his vision for managed competition. But I think that the unbundled there are sort of so-called point-of-care opportunities to select more efficient options will receive more emphasis because first of all theyve been previously under used with the exception of three-tiered pharmaceutical, and secondly, I think psychologically they appeal because they are more trialable by beneficiaries. It isnt like an all or none decision. To pick a big network, pick a narrow network, you can have a full network during the course of the whole year and how much you pay in coinsurance will vary depending on which network tier you select from.
And I think this purchaser emphasis on unbundled cost sharing is reflected in early purchaser choices of so-called consumer-directed health plans. The research that were doing suggests that most purchasers are opting either for HRA type plans or tiered plans, both of which emphasize variable cost sharing at the point of care rather than a variable insurance premium based on decisions you make once a year.
Last question: Will higher quality be subsidized by employers? And I think I would have put this question up here even if I had not known Mark was going to be on the panel, but I knew he would be joining me, so I wanted to address it, and obviously I think it would be fair game to say, Arnie, we know employers cards are on the table, we know that very few of them, for example, have sufficiently valued a very important source of quality improvement like NCQA accreditation. But that said, I think post-IOM report, a large employer constituency has become more quality sensitive, and so youll see the answer is something different, then it wont make any difference.
I think that it will be subsidized employers, but I think it will only occur probably up to the inflection point where higher quality unavoidably incurs higher longitudinal costs after, obviously, netting out indirect illness cost savings. I think once we get beyond the inflection point, that is where higher quality must cost more, I think at that point, though employers will make the higher quality opportunities transparent to beneficiaries, I think fewer and fewer employers will be interested in attaching any weight to quality. But fortunately, were a long way from getting across the quality chasm, and there are a lot of excellent near-term opportunities to both improve quality and longitudinal efficiency.
An example of this would be hospitalists, you know, where current research suggests that an innovation in care that improves both quality and longitudinal efficiency, at least over the course of hospitalization and a post-admission/readmission risk period.
This quality efficiency inflection point, this is a pure speculation on my part, but I think if you sort of think about what are the opportunities for near-term harvesting efficiencies by better engineered health care, I think that as Elliot Fishers research and annals has pointed out, there are some near-term opportunities for improving the amount of health benefit from spending by more physicians adopting more conservative utilization practices characteristic of geographies like Portland and Minneapolis.
In the intermediate term, as the Quality Chasm Report points out, theres a lot of additional opportunities to reduce cost and improve quality through waste reduction or what Ive called improved production efficiency, and the folks who write and talk about this are folks like Brent James and Don Berwick. We will get to a certain point--Ive said 2018; I dont know whether it will be sooner—actually, I hope its sooner—where we have sort of taken, weve harvested all opportunities to both improve quality and efficiency, and we will get to situations where the only choice if you want higher quality will be more cost, even over the intermediate term. At that point I think, as I mentioned, I think employers will be hesitant about subsidizing quality.
And just to give you a sense of where employers are coming out on this, a number of carriers have gone forward with these tiered-network products and have been faced with the challenge of well, how much do we weight available quality ratings versus longitudinal efficiency ratings? And those that have actually specifically thought it through have obviously talked to their employer clients, so thats how we have some evidence of employer values here, and the early versions of this have weighted quality ratings about 25 percent and longitudinal efficiency ratings of about 75 percent. So thats an early return of at least what some of the carriers that have pushed forward, how they rate the employers relative rating of quality and efficiency.
In terms of obviously going forward and trying to engineer, its obvious that cost sharing is not the optimal tool for pushing American health care across the quality chasm, and fortunately were not going to be relying on it alone. There will be regulatory reform. Professionalism I think is on the rise, in terms of professional responsibility for improving performance, and so cost sharing will not be the only leg on which improved health gain per insurance dollars spent will rest.
But if were going to move forward and make cost sharing work better, its obvious that certain knowledge vacuums will need filling, and if we want cost sharing to optimize American social welfare, number one, we need more health economists to research especially so-called consumer price elasticities beyond demand. We have good information on how, at least respectable information, on price elasticity of actual demand for care, but we have, just to give you a sense of it, we have zero published research that Ive been able to find in consultation with national economists on price elasticity of patient switching behaviors, particularly the willingness of a chronically ill patient to switch doctors or hospitals based on X percentage of price variation in their out-of-pocket. We dont know anything about that, and yet here we are on the brink of trying to implement refined approaches.
Were going to need a lot of help from health psychologists so we can begin to understand nonrational decisionmaking as consumers, facing anxiety-provoking decisions, very complicated information sets, attempt to optimize their spending.
Were going to need some health ethicists to help us fine-tune the ogre test. Its funny. Health ethicists have I think done a little bit more in thinking through how do American values align with current notions of what the insurance pool owes its sicker members. Health ethicists have written a little bit less on what sicker members owe others in their insurance pool, in terms of American social values. Its something I think needs more thought.
And then last, but not least, we will need health care operations engineers, people who can help us speed our way across the so-called quality chasm and perhaps extend that inflection point so that we can,in the near term, harvest more and more opportunities than we know about today to both improve efficiency and quality.
Why dont I stop there.
Paul Ginsburg: Thanks, Arnie. Could you stay there for a second. Ive got a question for you, but first an observation when I was listening to you talk about what the calibrators are doing and also some of the things that John Bertko was mentioning.
What you see is some of the ideas thats emerged in the better thinkers about managed care, which in a sense was focusing people on the more effective, higher value care, where the managed care notion was block them from doing the things they shouldnt do. Now, in a sense, the same thinking is being infused here, but always through incentives and always letting the patients choose, hopefully with the support of a physician.
The question I have, not related to that, is you mentioned going back to your first question about what really motivates employers as to what they have to do to retain and recruit the people they need, to what extent do you notice the degree to which, as the economy goes through the business cycle, that interest in ideas like this waxes and wanes?
Arnold Milstein: I think its very definitely business cycle related. I think when times are good and theres extra money, the last thing that most employers want to do is risk employee or labor relations. I think its only during hard times--
Paul Ginsburg: So I guess the question is, you know, given this recent spate of encouraging news about the economy, are some of these approaches in real jeopardy if the economy does well?
Arnold Milstein: I think so. Its obvious that there are two factors. One is the state of the economy and another is state of health insurance trend, and as that, you know, depending on--its the gap between those two things that will influence the speed and motivation of employers to take the kind of employee relations risk associated with experimentation and cost sharing.
Paul Ginsburg: Thank you.
Well, this is a great start to the meeting. We are going to have a break of 15 minutes. So please be back in your seats at 10:30, and well go to the Reactor Panel.
Also, if you have some questions that came into your mind during this first session, please write them on the cards and give them to some of the HSC staff.
[Recess from 10:15 a.m. to 10:31 a.m.]
Paul Ginsburg: Its time to get started. As I mentioned before, we have four reactors, and while each was asked to present their particular perspective, theyre obviously not limited and can say anything in reaction to what theyve heard before on their topic, and Id like to begin this segment of the session with a reaction from Karen Davis, who is the president of the Commonwealth Funds.
Karen Davis: Good morning. Its a great pleasure to be here with you.
I was asked to comment from a consumers perspective, but that we could try other perspectives. So I thought I would start with my economists hat on.
Most economists believe in the backward shifting to employees of the costs of health benefits. So if theres anyone in the room who is a true believer, they really probably should leave.
Because this discussion doesnt really make much sense, if you believe in backward shifting. Because if costs are really borne by employers, then employees shouldnt really care about the total costs, and they should really act as the employees agent in purchasing the kind of coverage that employees want.
The way an economist would think about this, suppose a firm has 3-percent productivity growth. Then, it cant afford a 3-percent increase in its total labor compensation package without having to raise prices or reduce profits.
And one would also assume that prices are already set at a profit-maximizing level. So, if the firm were to raise prices, there would be fewer products bought, and there would be fewer jobs, and they cant reduce profits or they would lose access to capital, and that capital is important for future productivity, and if you dont get future productivity, there wont be wage gains in the future. So the rational employer would just explain this to employees.
On average, total compensation would track whatever the growth in productivity is. Just to give you a concrete example, suppose there is a 3-percent growth in productivity and a 3-percent increase in compensation and assume health benefits are 15 percent of total compensation costs, labor compensation costs, and suppose health care costs are going up 12 percent, then, next year, if you want to have the same benefits, youre going to have health benefits go up at 12 percentage, wages can only go up at 1.4 percent. Trust me on the math of this. And that comes out to a total compensation cost of 3 percent.
Now, if you say, hey, we as workers are willing to have some higher cost sharing, if it will keep the premium next year only going up 10 percent, then the employer can afford a 1.8 percent increase in wages. And if they say, hey, were willing to take drastically higher cost sharing to keep the total premium cost to the employer next year the same as this year, then the employer can afford a 3.5-percent increase in wages. So the employer says, Do you want 1.4-percent increase in wages with the same health benefits or do you want 3.5-percent increase in wages with drastically higher cost sharing, figures out what that is, what the employees preferences are and acts accordingly. So thats the only discussion that is worth having if you believe in backward shifting.
But I dont really believe totally in backward shifting. I dont think it happens perfectly or automatically or simultaneously, and so total costs are important. Arnie Milstein said, employers are motivated by attracting and retaining labor force, and they will have whatever health benefits they need to attract and retain labor force.
Paul Ginsburg asked if this changes with the business cycle. I think I am always a glass mostly full kind of person, and I believe we will return to a tight labor market that we had in the late 90s. And the reason I believe that is that fertility rates went down by 40 percent in the 60s and 70s, and it just has a huge effect on the labor force in the coming decades.
So I think this issue of satisfying employees and really listening to them and giving them what they want in order to attract and retain the best workers will, in fact, be the conference five years from now, if not the conference today.
But employers are interested in getting value for their dollar. I think thats a little bit what we heard from Arnie, and total costs are important from a societal of view that wants efficient use of resources, it wants quality care, and it has a social value that everyone receive needed care. We dont want overuse, we dont want errors, but we also dont want underuse of needed care. I kept hearing that phrase this morning.
We also say we want equity in access to care and equity in the financial burden of paying for care. So Id really like to congratulate the center on the report thats being issued today by Sally Trude on cost sharing. I think its a very valuable contribution.
And now Ill shift to thinking about this from a consumers point of view. It shows that higher cost sharing, if not carefully designed, falls heavily on low income and the sick. And I particularly focused on her results on who would pay more than 10 percent of income under various cost-sharing scenarios. Pamela Short, you may recall, in JAMA a few years ago, talked about underinsurance, and she defined that as anyone paying more—having the potential of paying more than 10 percent of their income out of pocket business of the design of their health insurance.
What we know from this issue brief is, in the low co-payment options, only 1 to 2 percent, on average, pay more than 10 percent of their income. Its 3 percent if theres $100, 5 percent if theres a $500 deductible, $1,000—and I think thats particularly important, given the health savings account that are in the Medicare drug legislation—7 percent of people would be spending more than 10 percent of their income out of pocket.
But what you really ought to focus on in that issue brief is the row for hospitalized patients because any of us could get really sick and be hospitalized. So that gives you some indication of the potential for paying more than 10 percent of your income and the potential and therefore the numbers of people who are underinsured.
Under current practice, which is a deductible of about $300, I interpolate in that chart that about a fifth of Americans are currently underinsured. With $1,000 deductible, a third of Americans would be underinsured, and obviously there are also the breaks by income, and you see that those numbers are 40 percent, 50 percent for people below poverty.
So I think what you see embedded in that issue brief is, for the first time, and I congratulate Jim Mays and the Actuarial Research Corporation for generating those numbers, really good counts of what would be the magnitude of underinsurance under various cost-sharing scenarios.
Next, Id like to turn to the issue of appropriate care. This year, weve been thinking about what are the most important things that happened in 2003. I happen to think the Medicare drug legislation was one of them, so half of you will tune out, but I thought Beth McGlynns New England Journal of Medicine Article at the end of June was really one of the key research breakthroughs. What Beth and her team found is that only 55 percent of Americans get indicated care. Other people underuse services or they overuse services or there are errors being made in their care, and it was through a clinical chart review.
Again, some folks at the Mass General have taken that and translated it into population counts, and youre talking about, about 100 million Americans underuse health services, about 30 million Americans overuse health services.
What I included in a handout thats available for you at the desk if you didnt pick it up are several research studies that show the effect of cost sharing on underuse and overuse, on appropriate care, on inappropriate care. And the basic bottom line is that cost sharing reduces use of both appropriate care and inappropriate care. So, in other words, it would, if you increase cost sharing, you would have more than 100 million people underusing services, and you would have fewer than 30 million Americans overusing services, but you would affect both. Its not a fine-tuned tool, at least as currently developed. Also, you know the costs are concentrated in the top 10 percent.
Is there a way to design cost sharing to promote increased use of underused services? Well, you would have to vary it by income and by health status, you would have to vary it by clinical indication—the reason youre getting an MRI—consumers would have to have access to information that is not now available to anyone, much less to consumers. You would have to change physician-patient relationships with patients being in charge of decisionmaking.
Jim Mangum points out these large expenses are people with heart attack, stroke, trauma, car accidents and that they are no longer in control of making decisions. The whole health system takes over and makes decisions.
Is there a better way of promoting cost-effective care? And I thought a supply side strategy, to me, really has a lot more potential than cost sharing to move us in this direction. John Bertko talked about frequent flyers in ERs and having cost sharing for emergency rooms. But Mass General uses a call bank of nurses that call up anyone who is making a frequent use of emergency rooms, screens them for depression, finds out if theyre complying with their medications, finds out whats going on, and thats the way they reduce use of ERs.
Inappropriate use of imaging. General Electric gives financial incentives to Mass General to reduce MRs, and so they give clinical criteria to order the test.
Disease management. You see an example in this handout by Mark Pauley of how use of advance practice nurses can reduce rehospitalization.
Paul Ginsburg: Karen, were running low on time.
Karen Davis: So, to wrap up, what would it take? We need research, we need data on quality, and we need to pay for performance.
Paul Ginsburg: I would next like to introduce Helen Darling, who is president of the Washington Business Group on Health.
Helen Darling: Thank you. I guess I would like to try to bring everybody back to the total dollars were talking about. I understand the importance of the topic "cost sharing." Thats the subject of the day. But we havent talked enough about the total cost.
And very often when you see data on cost sharing and employee out-of-pocket costs or beneficiary out-of-pocket costs, thats what you see, and you see it going up, but you very rarely see, in the same charts, the total cost. I would bring--I think John Iglehart is here--about a year or two ago he did an article in the New England Journal of Medicine, in which he put the data together so you could see the employee cost sharing or the beneficiary cost sharing going up at the same time that total costs were going up. We just have to remind everybody, on average, nationwide, family coverage in this country is going to be about $11- or $12,000 starting in 2004.
We have an average pay in this country of about $27,000. So weve got a really complex picture of lots of health care or health benefits, and that may be what they want. Karen may be absolutely right. If they had the choice between money, more health care, more coverage, they might choose that, and the data are a little bit mixed on that. But even if you assume its true, the question is, is that good for the country, and is this what we want to have happen in our country? So I would hope we could talk about that a little more.
Also, several times we heard needed care versus recommended care or even needed care versus unneeded care. For those of you who picked up my little handout, I did have Beth McGlynns table. Its on the second page. And I have relabeled it, "You need Cost Sharing to get Attention." This shows exactly what Karen said, and that is we talk about cost sharing as if something is going to be taken from people that is absolutely needed, wanted, desirable or recommended.
Whereas, with or without cost sharing, about half of the care in this country, on average, nationwide—it varies by area, and people, and circumstances and everything—but the question is are we getting recommended care, not just what somebody thinks, even if it is needed, are we getting what they should be getting?
We also would remind everybody that, in this town, the concept of spending money you dont have isnt something that seems to disturb anybody.
Helen Darling: But, you know, if you are GE, and you are using Karens math, which I thought were excellent, GE—Ill profile a refrigerator—it now has about twice the functionality that it did 5 or 10 years ago, and it doesnt cost any more; in fact, it may even cost less. And its very typical in this country right now--certainly within the last couple years--to actually have either flat or declining prices. Certainly, employers cant pass them on.
So the question becomes how are they going to deal with these total costs, and cost sharing is one of the ways, although we would certainly be the first to say wed like to change the system were buying into.
I would also recommend to anybody in the audience who did not see on Monday the op-ed piece in the New York Times by Elliott Fisher, which I thought did the best job of summarizing some of what Karen said and probably some of what Mark will say about quality and what were buying into. And as we spend more money, we have to be thinking about what were buying.
And then finally I would just say a couple other things. Cost sharing, from our point of view, isnt just about cost, and in fact what you get out of cost sharing is still pretty minimal, and I would urge anyone whos in the audience who doesnt work with this full time, when somebody says that a deductible is doubled or 50-percent increased, were talking about $100 deductible on possibly a $30,000 or $100,000 admission. So it goes to $200.
Well, maybe thats rough on very low-income people, but as a proportion of the total cost of what the people are getting, if we dont pay some attention to that, however we do it, and we dont at the same time make changes in the quality and inefficiency—also, like Arnie talked about, and Im sure everybody will talk about—into the system were buying in, we cant afford the system. So we will have ever more people with ever richer benefits and little or no cost sharing, and we will have a vastly larger number of people with no coverage because employers cannot afford to provide the coverage that theyve got without cost sharing.
Paul Ginsburg: Thanks, Helen.
One thing I thought you might address that I should just ask you about is the complexity, in the sense how much complexity are employers and employees willing to tolerate from some of the things that John Bertko and Arnie Milstein were talking about?
Helen Darling: Well, I dont think anybody likes complexity, and we recognize that with complexity theres always higher dissatisfaction. We would hope that we could get better at communicating the complexity, and also, I mean, getting peoples attention to realize that they have to be active consumers.
They dont have a choice, and its not just because of the money. Its because of safety, and quality, and quality, and a lot of other things, and efficiencies, and getting them to understand that it isnt going to be simple, and its never going to be simple, and thats hard, I agree, but thats where we have to get to.
Paul Ginsburg: Sure. Thank you.
I want to introduce Bob Berenson, who is senior fellow at the Urban Institute.
Robert Berenson: Thank you, Paul. I apologize for my voice. Its an allergy. And Im sure youre going to cut me off. I have a lot of things Id like to say, so just cut me off when you want.
Paul Ginsburg: Okay. Well note when you started.
Robert Berenson: For a few sort of introductory comments, I very much appreciate the more enlightened view to cost sharing that what John and I all experienced back in the 60s and 70s. From a physicians perspective, what we had were sort of traditional notions of insurance and moral hazard insurance, where there was full coverage for highly technical procedures and, in some cases, no coverage for prevention or office visits, and we ran into distortions and charges in Medicare with the RBRVS system, had to try to correct all of that, and 10 years later I dont think its gotten quite right yet. A lot of us have worked on those kinds of issues.
So Im happy that, in fact, new sort of ideas about cost sharing would, in fact, try to promote prevention and, in fact, apply cost sharing to things that might be discretionary or overused. In a few moments Im going to talk about some of the practical difficulties. So, in general, I think not being grounded in traditional notions of insurance is an improvement of where we are.
Nobodys really talked a lot about the issue of asymmetry of information. I think that still is a factor in where you would sort of apply cost sharing. As Karen said, when youre in a hospital with multiple conditions and the system takes over, even though that may be where a lot of the overuse occurs, I dont think we would expect an individual patient to sort of be making decisions on individual services. There may be a hospital deductible, but not sort of cost sharing related to individual services there, and yet a lot of that is where Elliott Fisher, whos gotten a lot of compliments today, points to where a lot of the probable waste is.
So theres a bit of a conflict between those kinds of services which a consumer really can make some informed choices, and I think tiered pricing for pharmaceuticals is a good example of that, where in other services where it might be a useful break, but where theres really an asymmetry problem.
The discussion about moving from co-payments to co-insurance, one of the things that wasnt discussed explicitly was the opportunity for co-insurance, rather than co-payments, to deal with the issue of waste and fraud in the system or at least sticker shock.
As most of you know, I keep changing jobs. Somewhere along the line I went from co-payment to co-insurance for prescription drugs, and it was only then that I really understood the cost or the price of the drugs that I was being asked to pay for. The problem, it works very well in a pharmacy. Everything is automated.
It hasnt worked well in doctors offices, and Im encouraged that John knows at least one insurer who wants to do that. It gets very complicated at the point of service, when you dont know what the insurer is going to approve and pay, to ask for that co-insurance, and so people then do co-payments, and I think theres some discussion around trade-offs between co-payments and co-insurance.
But what I really wanted to do today was take the prerogative of a physician and tell an anecdote. Now, I know that anecdotes are not appropriate because you can find an anecdote to demonstrate any point you want in health care, but Im going to demonstrate a point.
Robert Berenson: MRIs, I agree that that seems to be an area, imaging broadly, and MRIs in particular, seems to be an area of great concern about overuse. And it would appear that, in many cases, they are discretionary. Theyre done for quality of life, et cetera. I think implicit in that is some notion that they are technical commodity like services, where you sort of can go to any MRI place and get an equivalent kind of an outcome.
Two brief comments on that. A friend of mine has knee pain. We could call that a sports injury, but in fact hes a carpenter. His job is being seriously affected because he cant bend, he cant lean, he cant do a number of things. He goes to the orthopaedist. The orthopaedist naturally gets an MRI, which finds something about his cartilage, but also shows that he has a cyst in one of the bones next to the knee. So the next thing he knows, hes being referred to an orthopaedic oncologist, who basically looks at this and says, "I think its a cyst, but why dont we just repeat another one in 3 months or 6 months or something like that."
Now, I dont know, the first one would be called the sort of discretionary sports injury. Thats the one he probably needed to get his diagnosis and to get back to work. The next one is really sort of, whats he supposed to value? I mean, I might have cancer or I might not have cancer. A professional is telling me to get this in 6 months. Thats the one that probably is the discretionary one. Thats the one I think relies upon professional leadership to define some evidence standards and to take some responsibility and not sort of to rely on the patient to say, Im not sure I need the follow-up.
The whole point is these are not neat categories. If, in fact, we go into this area in a big way, I think there needs to be lots more informed advice and informed consent from physicians to patients about trade-offs. Prices would need to be much more transparent than they are now, in terms of permitting informed decisions.
And, in fact, Ill finish with this one. And where we assume these are sort of interchangeable commodities, like cell phones, in fact, they are probably not. I dont know how many people, when they were focusing on the Medicare bill, saw this Wall Street Journal article about shopping for MRIs, where the reporter—I think this was a very interesting thing—went to, I dont know, 8 or 10 places. The prices varied from $450 to $3,675. And just a couple of comments from some of the places she went to.
One of them: No prescription needed for the scan, and valium was free on demand. The receptionist finagled the discount at 20 percent for us and then made sure that contrast dye was used to get a clear image. This was the receptionist doing that.
Robert Berenson: And, finally, a couple of these places, their three expert radiologists could not interpret the scans. We have evidence that breast screening mammography results are not accurate.
So, whether we move towards more cost sharing or not, there is still a very large responsibility for whether its health plans or brokers or somebody to organize and provide information. And I agree with Karen, theres a lot of supply side responsibility, and where cost sharing can be appropriate, its clearly going to be used. I think its not going to be very easy to do what I agree is a better thinking about cost sharing than has happened in the past.
Paul Ginsburg: Ive got a question. Id like to pursue the physician/patient relationship. Clearly for this to work well, physicians are going to have to be talking to patients about what their values are and the trade-off.
And in that anecdote, maybe this can happen with someones regular primary care physician, but its much less likely to happen to some specialist you were referred to, which doesnt have the relationship. But even with the primary care physician, are physicians just going to say I ought to do this? Or are patients going to have to really push them hard? How is it going to happen?
Robert Berenson: When you say do this, you mean provide real sort of information about choices, informed choices?
Paul Ginsburg: Yes, thats right.
Robert Berenson: Well, physicians say that theyre ready to do this. It seems to be at least where the AMA is. They want to see a system that relies much more on out-of-pocket expenses and that theyll step up. I have some skepticism about the ability.
I guess what Id say now is for better or for worse we have a system in which most patients trust their physician to make good referrals. That may not be well warranted, in many cases. I think some of the findings out of the New York information, that Mark is intimately aware of, suggest that docs dont refer necessarily to the best quality places. And yet, physicians take that responsibility fairly seriously and just have a lack of information themselves.
If we go in this direction, and even if we dont go in this direction, I think there just needs to be much more transparency about quality and price so that physicians, in fact, under the current little cost sharing and in a new world of a lot of cost sharing, can be better information givers. I do think there might be new grounds for suits for lack of informed consent, as people have to navigate their own way with their own money. That you didnt tell me that if I went to that place it was going to have a different potential impact and you should have known. I think there might be some creative kinds of things.
For the most part I dont think physicians have really thought very much about their responsibilities under a system in which theres lots more cost sharing. Right now most of the people who have significant out-of-pocket expenses tend to be affluent. And so the tougher questions about trade-offs between quality and price havent really been forced on the physician and they would need to be brought into this discussion.
Paul Ginsburg: Thank you.
Next Id like to introduce Mark Chassin, who is the Edmond Guggenheim Professor of Health Policy and Chairman of the Department of Health Policy at Mount Sinai School of Medicine.
Mark Chassin: Thanks, Paul. Ill try to do this in five minutes and elaborate on some of the other points in the question period.
I think there are four problems, only four, with relying to a significant degree on increasing patient cost sharing. Its wrong, it wont work, it will do harm, and its not necessary.
Its wrong because I think its wrong to push patients around with big economic incentives at the point of care when theyre sick and most vulnerable, making already very difficult decisions even more difficult. Its also, I think, especially wrong and borderline irresponsible if, at the same time were doing this under the guise of well, were really trying to improve quality.
I think that our model really here should be air travel. The public demands excellence in air travel and weve achieved it. Weve achieved remarkable safety in air travel not by individual air travelers choosing airlines on the basis of arcane and difficult to interpret crash statistics or maintenance reports or close call reports, but rather by transforming that industry by two mechanisms. One is highly effective regulation and the other is a very serious and aggressive commitment to safety systems. So thats why its wrong.
It wont work. It will clearly shift cost to patients and it will reduce utilization. And if thats the aim, fine, we should be up front and say that thats the aim. But it wont lead patients basis better quality. And thats for a variety of reasons I dont have a lot of time to get into, but just let me hit a couple of points.
First of all, our measures are just not that good if the aim is to pinpoint really good quality and lead patients to it. Even when the measures are good, theyre out of date. In fact, that can lead patients away from a really marvelous improvement story if providers have really recognized their poor performance and improved it, and there are lots of examples of that, and a couple that I can talk about later, if you want.
The third is it is really not feasible because quality is so non-uniform. Some hospitals are great in one, two, or three things; theyre mediocre in three, four, and five; and theyre really bad at five, six, and seven.
Now how on earth do we expect patients to choose hospitals at the point of care either at the time of plan selection, when they dont know what disease theyre going to get the next year, or at the point of care when they have to sort through arcane statistics that have different meanings for different diseases. And exactly the same is true for physicians. So it wont work.
It will do harm, however, because we do know that cost sharing, as others have commented, impairs health. The best data are from the Rand Health Insurance Experiment where basically a healthy population, if you had only a simple condition like hypertension or poor vision were worse off if they were exposed to cost sharing. And clearly, it has much worse effects on sicker and poorer populations.
Fourth, its not necessary. Other much more potentially effective strategies, both to reduce cost and/or improve quality, have simply not been tried. I think that economic incentives need to be directed at providers because the economics of quality are really quite simple. Overuse, provision of ineffective, unnecessary services impairs quality and it increases cost. The same is true of misuse, of avoidable complications, of errors that lead to complications. They increase cost and decrease quality. The problem is in underuse because underuse, the underuse of effective care, reduces quality but it also lowers cost. Fixing underuse problems always, nearly always, results in improved quality at a price.
The two major obstacles to using that quality driven approach to cost containment have been traditionally, number one, a lack of political will on the part of any stakeholder to tackle overuse or misuse seriously and aggressively.
And then second, within the delivery system there are enormous economic disincentives for providers to invent really wonderful ways to reduce overuse. Theres absolutely no reason in favor, and every reason against my hospital figuring out how many carotid endarterectomies or bypass surgeries are totally inappropriate and getting rid of it.
The same is true for the investment thats necessary to reduce errors in a serious way and improve safety. The same is true in underuse.
So I think that those are the two major obstacles to really aligning the very simple economics of quality and cost sharing really on the patient side doesnt play any role at all.
Paul Ginsburg: Thank you.
Mark, one specific question I had in mind, that you might address, is when you think about the ideas in the Cross the Quality Chasm reports, to what extent would increased cost sharing of the type that John and Arnie were talking about facilitate or impede steps taken in those directions?
Mark Chassin: I think again, patient cost sharing, if the idea is to try to steer patients toward providers who have better services, the problem is the fog of bad information and bad measures. We dont really know how to construct measures that really would allow patients to be directed toward better services at a global level. And its really impossible to imagine patients sorting through trying to figure out well, I think Ive got heart disease, but its heart failure its not coronary disease. And its really St. Peters that has the best record with coronary disease. But then if I have heart failure Ive got to go across town. I dont see how that works.
Paul Ginsburg: Thank you.
Let me tell you where we are. Weve heard from the four reactors. If you have additional questions, please hand them to people that will walk through the aisles.
One problem with this approach is that I have a number of really good questions that were handed during the break and they were all answered during the reactor panel. So I need more questions.
Anyway, while were waiting for them, and people can come up to the microphones too, I want to give an opportunity to John Bertko and Arnie Milstein to say anything that they have any reaction to these comments.
John Bertko: I guess I wouldnt disagree with Marks comments. But as I started off, living in the real world, cost is a constraint. Arnies firm and my firm with colleagues know how to measure one thing pretty well, which Ill call the efficiency based on episodes, and the quality component that Arnie referred to being like a 25 percent quality measure and 75 percent reliant on longitudinal type of stuff may be the only mix we can do right now. The quality measures are emerging.
But I suggest that in the near-term folks like Ellen Darlings constituency, and I certainly know our small groups who are interested in that price, are interested in getting more efficient lower cost insurance. And if we have to settle for that for the time being, quality being still nebulous, I think we need to move overhead.
In the best world, back to a comment that Karen made, health care benefit costs going up at 10 percent a year, I think, is still unacceptable, certainly unacceptable on the low end of the scale for small employers, and I think reasonably unacceptable even in good economic conditions for larger employers.
So rather than sit still, many of us are going to continue to push forward. And I agree its incremental.
Arnold Milstein: Theres a lot to respond to. Maybe I could just arbitrarily pick a few items.
First, I also agree with Marks points but I dont think I agree with his conclusion. I think performance failure in the airline industry we solved because performance failure in the airline industry is highly visible. I think, per Don Berwicks point, performance failure in the health care industry is extremely invisible, not just to the customers but also to the health care workers.
So it makes—Ill call it the overwhelming either political or business case for its solution, highly problematic.
I am a little bit more optimistic that Mark on whether or not we might find ways of people who are at high risk for needing certain types of hospitalization of beginning to clue them in before they actually need a hospitalization that there are differences in their community with respect to the efficiency and quality of different hospitals.
I think the forerunner for this has been the relatively successful Centers of Excellence Programs that many carriers have implemented over the last 10 years. The better of those programs have tried to incorporate measures of quality using the Medicares risk adjusted transplant outcome database, which was sort of the first of our better risk adjusted outcome databases to go public.
And then also most of the enlightened carriers are not using price per day or price per stay, but a longitudinal measure of costs associated with the pre-admission activity prior to a transplant, the hospitalization. And many of the contracts look at a 12-month post-discharge window for a total longitudinal cost picture.
If you talk to the folks that run these programs and say well, what difference does patient cost share make in inducing patients to select the Center of Excellence, which in almost all cases rate more highly on both risk adjusted outcome and risk adjusted longitudinal cost, they will say it makes a big difference and that those employers that allow the insurer to typically positively incentivize folks to use these so-called Centers of Excellence, it tends to increase the frequency with which those patients use the transplant center by about 30 percentage points. You go from about 50 percentage point yes, I will use the Center of Excellence, to about 80.
So I think we do have—its a cup half empty/half full. Obviously nobody prone in the back of an ambulance is a consumer that cares about cost sharing. But its also true that most of what ails America are chronic illnesses and its possible very far in advance of getting into trouble to I think select a hospital and a physician thats in a favorable tier with respect to either managing your particular chronic illness or, if you have multiple chronic illnesses, doctors or hospitals that more efficient and have higher quality longitudinally in managing multiple chronic illnesses.
I had a chance to talk with Mark about this at the break and I think it would be helpful for all of us to hear some of his thoughts on how, if patient selection of better performing providers is something we dont have to worry about, some additional thoughts from Mark on what would induce American doctors and hospitals thats politically credible to get across the chasm quickly. Because eventually well get across a chasm. The question is will it be in 10 years or will it be in 100 years? And obviously most of us here would like it to be in 10 years.
If patients arent selecting better performing providers, how are we going to--I know Mark has a good answer for this because he gave me one, but Id just like others to hear it.
Paul Ginsburg: Helen.
Helen Darling: Just two points. One, we talk about hospitalization and the cost sharing. In fact, thats at least in the employed world. The frequency and even the amount of money when theres a hospitalization is very modest. So its really on the outpatient side and prescription drugs and things like that. So I think its important to keep that in mind.
The second thing is that weve had many, many years to correct the system and change the supply side, which isnt to say that we shouldnt be doing that. But unless we can do it immediately, we feel we have to have cost sharing as part of it. And if the system gets reformed and the supply side turns out to be managing everything, which is obviously one of the purposes of managed care, then we dont need cost sharing.
But in the interim, whether its this year or next year only—and I wish we could change all of these things faster--we need cost sharing to at least change some of the incentives within the system.
Paul Ginsburg: Karen?
Helen Darling: Just a couple of quick points. Weve done a survey of physicians about quality of care. 15 percent of physicians say that they always or sometimes have information on the quality of a specialist they referred their patients to, 85 percent say they rarely or never have such data. In terms of quality of their own care, 7 percent of physicians think it should be made available publicly.
So theres where we are with both the availability and the willingness to make quality data available.
Now insurers have good information on total longitudinal costs and I love this concept of longitudinal costs over an episode or for a chronically ill patient over a year. So I think we need to really generate data on that. And insurers could do a lot better job doing that. But even that needs to be risk adjusted. So thats not a trivial problem.
But I think the real solution to this, to respond to Helen, is for managed care plans or insurance plans to create their own elite networks of providers in that upper quadrant in Arnies scatter diagram. So those that are in the top 25 percent of quality, those that are in the top 26 percent of efficiency, create a network of those. And as Mark says, a hospital may be in for cardiac care but not in for oncology care if they have different quality and cost performance on different services that they provide.
But I think for the patient to go and find the right hospital or the right doctor is really going to be less effective than for the plans themselves, and for Medicare, to identify elite networks of high performing health care providers.
Robert Berenson: I just wanted to enter into this a little bit.
I agree basically with Arnie, that the one sort of desirable area where I can see cost sharing use is where the health plan or Medicare does its work to identify high performing entities and then uses differential incentives to try to steer in some way.
Medicare has also had experience with what used to be called Centers of Excellence. Some of the hospitals objected to not being in the category of excellence so they had to change the name.
But just one of the realities in traditional Medicare, and if we still have traditional Medicare, is that a lot of people do have their cost sharing bought down through a variety of mechanisms, whether its retiree coverage or Medigap insurance. And sort of the traditional approach of giving a cost sharing break is less available in Medicare.
I would point out I was looking through the bill yesterday and there are two new Medigap plans that are being designated which essentially buy down part of the cost sharing, part of the 20 percent, with the idea that people might want to see that as new alternatives to first dollar coverage.
So in any case, that is happening there.
Paul Ginsburg: Actually, Id like to ask John and Arnie about whats the current state of the ability of health plans to make these identifications of who the high-quality efficient providers are? And if you could just talk about where we are now, where we might go, and the degree to which people changing around, the degree to which thats a real hindrance in doing that.
John Bertko: Let me start and Ill do it on a health plan basis and then Arnie can perhaps do it on the stuff that I know his company has done for employers.
Keep in mind that we are what is called a super regional plan. So in about 10 markets weve got pretty deep penetration. By that I mean we have lots of members covered. We cover fairly large networks. And we have a pretty decent number of—Ill use the term episode for the longitudinal description—episodes per doctor.
On that basis if you can remember back to that really cool scatter graph that Arnie did, I can probably identify folks in the top half in terms of cost, or maybe even better, pretty good numbers on that. I can probably eliminate some of the quality issues on one side of it, but I wouldnt even begin to say that I can squeeze into the quadrant, the highest quadrant that Karen says is the best. I mean, we all want to go there but I just cant get there yet.
The limitation that we have is there are a fairly significant number of physicians at least for whom I dont have enough episodes. Ive got to say not enough data. We have addressed that problem by trying to access other much larger databases with some success, and I think there is movement along there.
So we have a product that is both of a long care. So we have a product available in two of our markets right now. Its actually as of 1/1/04 its going to available. And were going to roll it out to another 10 markets during 2004. I would say for our competitors theres probably a dozen or so that are in a similar status of rolled out or in a planning stage.
So the good news is were on the way. I would give us perhaps collectively a B-plus for effort and execution on this but were not quite there.
Arnold Milstein: I think the current methods, particularly at individual physician level, for evaluating longitudinal efficiency and quality are at an extremely rudimentary state of development. Obviously the question in terms of going forward is, understanding theres a lot of noise around these measures, is there enough signal for them to at least be directionally correct? I think the jurys out on that. Thats what I hear from the health services researchers who are working on longitudinal efficiency measures.
With respect to quality of care, I think things are equally rudimentary. Beth McGlynn, whose research has been cited here, has been working on testing a version of her QA tool method of quantify overall rate of compliance with evidence-based guidelines that could run on claims and enrollment data only, including pharmacy claims. Thats obviously going to be a huge step forward because right now the only way of reliably evaluating quality would be a medical record review. Thats what Beth used in her research but thats obviously very likely going to be deemed unaffordable.
So what are we stuck with in the meantime? We have early attempts to take evidence-based medicine and pass that through claims databases. The Hawaii Blue Cross plan has been using this for about five years to differentially pay primary care physicians. Theres something like up to 10 percent now variation in what physicians get paid based on their overall rate of compliance with a basket of evidence-based guidelines that can be tested using claims data. But were clearly at a very early stage.
The other comment that John made I very much agree with, that as you begin to try to apply current first generation methods of quantifying longitudinal efficiency and quality of care compliance with evidence-based guidelines, you do face this terrible problem of adequate cell size, particularly at the individual doctor level.
I think that does have the interesting effect of tipping the initial ability to do this more toward regionally dominant plans. Its not by accident that the plan that I used in my scattergram was one that is regionally dominant in the Pacific Northwest. They have plenty of denominator size per doctor to run the calculation.
But if were going to move forward on a national basis, its going to either require intercarrier cooperation or its going to require CMS to reevaluate its long-standing policy of refusing access to its Medicare fee-for-service database, which obviously has a lot of denominator size in it, in a way that would allow identifications of individual physicians. By policy decision CMS decided in 1985 to stop doing that. And that is a policy decision thats not locked into regs. But whether its a changeable policy decision is unknown.
Paul Ginsburg: Mark, is it on this point?
Mark Chassin: A couple of points that have been raised that I wanted to comment on.
I wanted to start with the question of measurement of quality and where the current state-of-the-art is and where it might go if measures were really good.
I think that Arnies right, that the current state-of-the-art in measurement is rudimentary but very uneven. Its much easier to measure underuse. And in fact, virtually all of the activity around measurement as it relates to almost any kind of pay for performance is around underuse. The research that was cited, the Rand research, 85 percent of the measures were measures of underuse problems.
NCQA, CMS all focus on measures of underuse, in part because its politically easy. Its politically easy because everybody wants to get appropriate care and doesnt like when they dont get appropriate care, and thats very nice. And if you put incentives in place to make that happen, that improves quality. But it also increases cost and it will make Helens problem a lot worse.
Disease management was invented by pharmaceutical companies to increase pharmaceutical use, and it works.
The problem with measurement is that it gets much more difficult to measure overuse. Unless youre talking about antibiotics for colds, which a few years ago consumed one-fifth of all ambulatory antibiotic use—one-fifth. The likelihood of getting an antibiotic prescription for a cold if you went into a doctors office and you were an adult was 50 percent if you went in with a cold.
That measurement is much more difficult because weve got overuse and underuse of exactly the same service in the same geographic areas. It takes much more clinically sophisticated measures.
Measuring misuse, which is what kills people in hospitals, is even more difficult. We havent really made many serious efforts to do that.
When you do have good measures, and the best and longest continuing program that I know of that produces really good measures on quality is the New York State Cardiac Surgery Reporting System, which has been producing surgeon specific and hospitals specific risk-adjusted death rate data following bypass surgery for more than a decade.
The way that information has been used, and its been used with amazing effectiveness to reduce mortality statewide. New York achieved the lowest rate of bypass surgery mortality four years after that program was established. And the study just published a couple of months ago showed that that standing has been maintained through 1999, not through any use of those data that anybody can measure, by consumers, by managed care plans, or any other stakeholder to move patients around in a shell game, but rather with the help, enthusiastic cooperation, sometimes coercion of the state health department, to force hospitals with really poor performance to improve. And they did. And that, I think, is a model thats not seen any expansion whatsoever.
The last point that I wanted to make is yes, Arnies quite right that air travel failures are quite visible. And thats what reinforces the publics amazingly high demand for excellence. That in fact, in my opinion, is the missing ingredient in health care. The public does not demand excellence. So none of the folks that depend on the public, whether its employers or health plans or hospitals or doctors are under a whole lot of pressure to produce higher quality.
Paul Ginsburg: Thanks.
I had a point that I wanted to make in this discussion and then I got a question shoved in front of me that says it better than I. Or at least introduces it better than I. So let me read this.
I found a quality low-cost auto repair shop. I recommended it to friends. They went, too. Now I have trouble getting appointments.
Paul Ginsburg: Access to this shop--well, I dont think I have to finish it.
So this brings up the issue about what if we have some success in identifying the high-quality providers and success in motivating people, whether its through the information or through the incentives of refined cost sharing, to go to them, how much capacity do they have to do more?
I think the answer is that for the quality of the system to improve, the real improvement is going to come from the physicians or hospitals that lose patients, that see as a result of their loss of patients--or maybe, in the case of the CABG data, just are embarrassed because of the low scores theyre getting--to change how they practice. Which in a sense implies that for system quality to improve, its going to take a long time because its going to have to be based on the perceptions of lower quality providers that their low quality is hurting them, or that they are low quality, and then doing something about it.
Karen Davis: Again, I dont know that the only way this can work is through market incentives and losing patients. Its interesting, the Wall Street Journal today the write-up of the National Health Services and Englands adoption of IT, so that by 2008 they will have electronic medical records that are standardized across all of England.
So I think, first of all, you need, physicians and hospitals need comparative data on their quality performance thats comparable so that they know whos doing well, how are they doing, where do they stand relative to that. And theres a pretty competitive drive, because thats how physicians made it through medical school, to be in the top half of the class. Theres only going to be 50 percent in the top half, but the whole group can move up if they know what the best practices are.
So information is one thing.
The second thing, Ive really been impressed, again looking around the world, at how well learning collaboratives are working to improve quality of care. Once youve got a best practice in New Zealand for primary care physician management of chronic obstructive pulmonary disease in Australia, learning collaboratives for emergency rooms on effective time to treatment and pain management, in the UK learning collaboratives on both primary care and cancer care markedly reducing waiting times and improving quality.
And its interesting, even though this started, in my view--I know others did it--with Don Berwick in the U.S., these are really taking off like wildfire in these other countries and less so here.
So I think its really going to take a boost--give providers the information, the tools they need to do better. Its not just financial incentives or losing market share that motivates them. They do want to do better by their patients. And we need to do a better job of giving them the tools, giving them the training, identifying the best practices that they can incorporate.
Paul Ginsburg: John.
John Bertko: I guess Id like to address a part of your worry and give you some what, to me I interpret as good news. This is in the kind of network analysis that weve done for longitudinal studies.
The first part of the good news is that when you look at specialists, most specialists—and Ill say thats in the 70 to 80 percent range—are clustered pretty tightly around the mean of what I would call acceptable cost or good cost for care. The outliers are a relatively small part.
We did this step-by-step by specialty, and as you try, you get to a point about 20 percent or so. you get almost no more dollar oomph or reaction diving below that 20 percent level. If you believe that in most urban areas at least that specialists by many measures are abundant, then you should probably think you have less worry that youd run out of spaces, in the analogy for the auto care.
Primary care physician is entirely different. Many primary care panels are extremely full. So Id say for a while at least we have the ability to shrink while were in the game. But the learning cooperatives and the other things that Mark Chassin said could be done to, in fact, improve individual physicians.
Robert Berenson: This is a time when Id like to sort of cite an insight I heard at a recent conference by David Lansky of the Foundation for Accountability who clearly doesnt--they do a lot of focus groups and surveys on the consumer side of what theyre looking for.
And he had this very interesting point, which is that the focus groups that they did, the consumers wanted information—basically, they didnt want the information to switch doctors. They wanted their own doctor to do better. I think that is consistent with what Karen said and some things like that.
People dont want to—it is so hard to get the right relationship. The doctor/patient relationship is somewhat unique and special. And things about comfort and trust and things like that sort of dominate over objective measures of quality. So people dont want to—when theyve finally found the doctor theyre happy with, its sort of oh, he or she is not so good on such and such, Ill move. Rather, they want that doctor to do better.
I think we need to sort of take that in mind, whats the information for? I think we need the information. But I dont think it largely is about having this open market where people are just sort of moving at the point of service to different providers.
Clearly, it varies, the kind of relationship you have with your primary care physician that ideally goes over decades is different from when you need a heart bypass or an organ transplant, which is a onetime event presumably. And so I think theres different appropriate uses of incentives and management. But a lot of what needs to happen really is on the provider side.
Arnold Milstein: Just elaborating on the question about capacity constraints and the autobody metaphor. First of all, I agree with Johns point about their major additions in capacity by the better performing providers are probably not necessary in the near-term.
And secondly, I think its the point one can pull out of the Quality Chasm Report is that it isnt like were talking about--the situation were currently facing are not well--the best performing providers on quality and efficiency still have some major opportunities to improve throughput. And one of the interesting aspects to some of the learning collaboratives has been shared opportunities to improve throughput both in physician offices and in hospitals.
Even in relatively whats viewed to be kind of a maxed out, in terms of current capacity, subsystem like primary care, people who have begun to think about applying engineering principles to primary care and have looked at early experiments—Im talking about people like Tom Bowdenheimer and Don Berwick—they believe that there are very substantial opportunities to improve throughput while improving quality and longitudinal efficiency in primary care.
I love the term theyve come up with, opportunities for reengineering primary care offices. They refer to it as the de-hamsterization of American primary care physicians. I think the opportunities are profound.
Paul Ginsburg: Thank you.
This is probably a good time to turn to the audience for questions. I will ask a couple but people that would like to come up to the microphone can start doing that.
Heres a question about income tiering cost-sharing. Milstein mentioned different out-of-pocket caps by income. But lower income families will face larger relative initial costs of deductibles or coinsurance payments. Are businesses varying deductibles or copayments for lower income workers to ease initial access?
Arnold Milstein: Not yet. Most of the income variation that Ive seen--and I hope Helen will join in on this--have been focused around, as these so-called consumer directed plans are implemented, its focused on variation and maximum out-of-pocket.
Helen Darling: First of all, you still have companies left that from the old days had a differential. Typically it was between hourly and salaried, so that usually the hourly workers paid less, whatever their contribution would be lower, their cost sharing, that kind of thing.
Thats really as much of a leftover. Some of those things went away because in managed care there was no cost sharing. So now companies are going back and saying if were going to put them in, then shouldnt we be doing something thats different for the people at a lower wage?
Its still kind of controversial because their questions, a lot of questions especially in households where usually two people work, how do you do it? Because its done on wages. And yet you have somebody sitting next to you who has a working spouse and theyre living up in the McMansion and all that usual stuff.
But we do see more interest in that. We also see the possibility of things like the out-of-pocket limit being a percentage of wages. So for example when I was at Xerox, the deductible was 1 percent of pay and the out-of-pocket maximum was 4 percent of pay.
So those things have been around. I think they will become more important if the cost-sharing momentum continues.
John Bertko: Id just like to add here that theres clearly a gap on this between large employers--in fact, what we call jumbos, 5,000 or more--and the small employer group. We are certainly aware of some large employers that have various kinds of income related things.
On the small employer stuff, a 10-life employer, the purchaser is the owner and he or she has virtually no incentive to have these kinds of things. On top of which they have rudimentary if any kinds of payroll systems. Theyre probably using something like Quickbooks to write it and are not The going to be able to administer anything like this.
So its a good idea. An d in the interest of equity in particular, the maximum out-of-pocket lids are probably among the very best ideas in order to keep what might be a $2,500 or $2,000 max out-of-pocket at that level for the people least able to afford it without creating too much controversy.
Karen Davis: Since weve got a Washington policy audience, I just wanted to toss up one thing I think we should be thinking about, tax credits rather than tax deductions for high out-of-pocket expenses related to income. In other words some percent, not 100 percent, but 75 percent, 90 percent of expenses over 10 percent of income being a tax credit through the personal income tax system.
That is a way in which instead of expecting Medicare, expecting employers, to protect the financially most vulnerable, where you could provide at least some threshold of financial protection for people with high out-of-pocket expenses in a way thats better than the current deduction for expenses in excess of 7.5 percent of income. So its the difference between the credit and the deduction.
Paul Ginsburg: Also, your point might be relevant to the new health savings account, where an employer or an employee can contribute to an account to draw out to pay the deductible is, in a sense, an exclusion. And if that were a credit, that would become more powerful.
Karen Davis: And a credit tied to income.
Paul Ginsburg: A credit tied to income would be a more powerful thing for low income workers.
How do employers and plans intend to get the right information to employees and members in order to facilitate more informed decisionmaking? Do you want to start that, Helen?
Helen Darling: Yes. Actually most large plans now, and most large employers, I would agree its large, but even for small employers, they can get it through health plans, are getting information online available to them. I actually find it fascinating, some of whats available. If any of you have seen SBEMO or Aetnas Health Navigator. And every plan—Im sure Humana has one. Feel free to name yours.
John Bertko: Ours is called a "Wizard."
Helen Darling: Basically they all have them. And Ill tell you most people, if you havent seen them, you will be amazed at how much very information is on it. We can debate about its lack of perfection. But whatever is there is now being fed into these systems.
For example, and maybe the Wizard can do this too, but Ive seen the Aetna Health Navigator. You can put your ZIP code, you can put in how many miles you want to have for say which hospitals. If youre going to have a baby you can specifically say I want to know the delivery hospitals within this area. What do I care about? I care about complications rate and nosocomial infection rates and things like that.
Then you hit a button and it ranks and spits out the hospitals within that mileage distance from your ZIP code. I actually sat there and watched—for those of you who know the Wall Street Journal and Barbara Martinez, who has had her baby by now, but she was just about to deliver. We did it for her with maternity in this big room. She was gutsy to do that.
And it turned out that the hospital she was going to was like number seven out of eight. So I said, is this going to change your plans.
Theres a lot more information now than ever before and I think that is not perfect, but it is very useful. And its not for everybody. There will be people who dont understand it. But an amazing number of people do understand it. And we tend to underestimate what consumers are interested in using and theres more data to support that.
Paul Ginsburg: Helen, do you have a sense from employers as to how extensively the information theyre putting out is being used?
Helen Darling: No, we dont because the employers may not necessarily even measure it and wouldnt have a way. The plans can tell you how many people have gone to the site, and how long theyve stayed there, and what they have gone through and that sort of thing. But I havent seen the data.
John Bertko: I would only mention here that one of the things that we are trying to push very, very hard, including price discounts, is positive enrollments. What that means is when you sign up you dont get to just default into the same plan you were in last year. You have to sit and think about it.
When we tested this stuff out, we got pushback saying youre going to make me spend 20 or 30 minutes thinking about my health policy choice. And we go yes, how long did you spend looking for that new car? 16 hours.
Arnold Milstein: I do some work for large labor unions who are operating their own health plans, and have a chance to hear from people representing less educated and lower income consumers. And what Im hearing from them is that they want decision optimizing help that is active rather than passive. What theyre saying is please dont force my people to have to go and navigate a website.
Their metaphor is more analogous to like a Kmart blue light special that would also address--they say quality opportunities, not just cost saving opportunities. That would, on a continuous basis, mine information from the claims database, from nurse line interactions, maybe from a health risk appraisal that was filled out upon enrollment, and continuously identify opportunities for enrollees to make different choices, whether its something simple like switch to a generic drug or whether its a link-up, now that you have your first diagnosis of heart disease heres a list of hospitals in New York that score very well in your neighborhood on risk-adjusted mortality.
What Im hearing from people—Ill call it the worker or the consumer side of the market—is give us active decision support tools. Dont make us come to you.
I think an illustration of this is this support tool being offered by—its called Resolution Health. Its called Direct to Member that on a continuous basis flags opportunities for either affordability improvement or quality improvement or both, depending on what youve designated with your fax or e-mail, et cetera.
Helen Darling: Just to build on that point that Arnies making, we do know that companies like General Motors and Verizon and others, who have used a sort of designated blue ribbon plan, in fact, there is migration to those plans. So employers are doing that. Theyre making that easier to choose.
Question: Hi, Alwyn Cassil with the Center for Studying Health System Change. I cant pass up the opportunity to have three physicians on a panel without asking this question.
That is that it seems to me if were ever going to get this right, that we have to get the provider incentives right. Whats going on out in the marketplace right now is moving in the opposite direction, in most cases, from where we should be going. With the exception of some of the very fledgling pay-for-performance initiatives, we have moved for all intents and purposes back to fee-for-service payment.
In our recent site visits to 12 nationally representative communities around the country, we have detected this increase in what I will use euphemistically refer to as physician entrepreneurship where they are adding more and more capacity to provide services within their office-based practices to the degree of adding MRI capability in a large group.
What do you do about that? What kind of payment mechanisms, if were going to I guess kind of bribe the providers to do the right thing, how do we design payment systems to keep them from doing the wrong things?
Mark Chassin: Ill start. I think our track record in designing economic incentives for any provider, whether its physicians, hospitals, physician groups, that really rewards excellence, that lines up the economic incentives with quality incentives, is absent. Its really dismal.
I looked at this a couple years ago and could find almost no research whatsoever on developing models like that. Weve got great incentives for overuse. Its called fee-for-service. Weve got great incentives for underuse, which is called capitation. Weve got great incentives for no use, which is called salary.
But what we dont have is any even experimentation or ways of thinking about organizing economic incentives that reward excellence. Thats across all three of the quality domains.
So I agree. I think its an enormous problem because I think were really foregoing the enormously powerful impact on action that economic incentives have. Because not only do we not have economic incentives that reward excellence, but weve got a completely fragmented and multidirectional set of incentives that act on very different ways on different parts of the delivery system. The incentives on hospitals are different for different payer groups. Theyre different for the same patient, for the physician and whether the physician is outside or inside compared to the hospital.
So there is no way of taking advantage now of the very powerful effect of economic incentives in aligning them to produce better quality.
Karen Davis: Theres a sense that we cant do anything about overuse in a fee-for-service system. The Japanese systematically reduce the fees every year on those services they think are being overused. And they get immediate reductions in those services. In Medicare were to move away from strictly a resource based relative value schedule and use that power of juggling the fee schedule, you could make a difference.
And the same way with the DRGS. Were talking about cardiac surgery hospitals, orthopedic hospitals, oncology hospitals. Were not talking about trauma care hospitals, burn care hospitals. Its clear that certain things are profitable and certain things are not profitable. And the things that are profitable, theres an expansion in capacity to provide them.
If you really want to do something about it, youve got to look at your rates in the raw. And managed care plans use the Medicare fee schedule for physicians as a basis for which they base their payments. So if Medicare would tackle this, you could see a ripple effect through the private sector as well.
Arnold Milstein: I think this could be a major source of accelerating progress in both quality and longitudinal efficiency. But the resistance to this is not simply unenlightened insurers or purchasers. There is--in California, where we finally were successful at getting, we hope now, up to 10 percent of HMO premium to vary based on quality scores in California medical groups. That was a six-year battle.
And among the most aggressive opposers of that initiative early on were physician groups who were very wary about both performance transparency, and most importantly, potentially not being in the favored group.
So it would be, I think, a very important method of inducing faster progress in health care quality and efficiency, but the resistance is not only on the buy side of the market. Theres also great reluctance on the sell side of the market.
So I m very much supportive. Lets use the Medicare demonstration with Premier. Thats an example of a quality based—I mean, there are 1,500 Premier hospitals. Something like 220 have signed up to be part of this tournament and potentially risk being paid less if they dont improve their performance. So its a very good solution, but at this point the reluctance is not only on the buy side of the market.
Robert Berenson: Let me just make a couple of points. I think Karen is exactly right. I think we just need to be much more creative with the basic payment systems, not just the marginal incentives for quality.
I interviewed some of those groups that Alwyn was talking about. I guess I cant be very specific here but there was one cardiology group that I talked about who sort of do a very good job of being entrepreneurs around stents and other activities in hospitals. Lets just keep it vague like that.
I talked to them about what their investment was in disease management to keep people out of hospitals and they looked at me like I was crazy. That wasnt what they were doing.
Clearly, the DRG system is rewarding—I mean, I think theres been increasing recognition that it seems to be skewed in favor of procedures away from medical DRGs. I think theres a fundamental issue of paying each incremental service at the average rate so that you develop winners and losers. And I think Im encouraged to hear about the Japanese experience with the margin, changing rates and getting behavior change. I think thats what we should be doing with the basic payment systems. Im all for, at the margin, giving incentives. But they may fly in the face of how you keep the doors open.
PARTICIPANT: Leighton Ku with the Center on Budget Policy Priorities.
I have a question about how good the evidence is, particularly from larger scale studies, about whether cost sharing really makes consumers smarter and more selective. Certainly the Rand study suggested that in general cost sharing did not make people choose more appropriate care. Studies that Ive read recently on prescription drug tiering, which is I think the main area where weve seen sort of the selective cost-sharing, indicated that—and I particularly remember a Rand study that came out about a year-and-a-half ago—suggested though tiered cost sharing was associated with reduced reduction of drug use and more unfilled prescriptions, it did not increase generic drug use. That in fact, everything fell, basically proportionally.
So that people didnt respond to the incentives the way that we expect them to respond to. My interpretation is that a major reason is because its the doctor whos choosing the drug, not the consumer in the first place.
Weve talked a lot about the ability to use cost-sharing in copayments or coinsurance to make people more selective for higher quality or more efficient care. I guess I dont see the evidence that it actually works.
John Bertko: Part of this goes back to what information is available? And Ill just keep you one of the things that were doing.
We have, for drugs, we scan through the database and look for people who would have a choice between either a lower cost brand or generic drugs. And through interactive voice response, a computer calls people up and very anonymously says gee, youre taking this. Perhaps youd be interested in saving some money by taking the substitute.
Believe it or not, our feedback has been they like anonymous calls and I think its up to about 19 percent of the calls made show up in the next script as having the lower cost drug. So I think its a mixed bag here of what people know or dont know.
The second part of this—first of all, the Rand experiment is now 23 years old, something like that, so its almost ancient history. On the consumer directed parts there is slowly emergency evidence on a relatively small basis, the 25,000, 30,000 people that we have and perhaps the 100,000 that Definity Health has, that people in those plans change some of their behavior. And particularly in outpatient, particularly on site of care.
I wouldnt hold that up to be anything close to the strength of the Rand report. I think if were looking for places to do further investigations, this is one of them thats out there. And folks like us would be happy to cooperate.
Helen Darling: I think a lot of the information that people work with, like us in the real world, isnt published because theres no reason to publish if youre working in a PBM. If you look at the PBMs data on changes, when they change cost-sharing say to mail-order, generic, anything, they have actually pretty dramatic changes in behavior, if its financial and theres an option. I dont have the data on the top of my head. But theres a reason that everybody is going to these things. It isnt just to get a little more money. Its actually to change behavior.
In prescription drugs in particular is where you see the strongest Thomas relationship.
Paul Ginsburg: I just have one comment I wanted to make about the Rand study. It is old, but I dont think thats—theres another significant issue in its relevance today, which is the fact that it was a randomized controlled trial. They selected individuals, gave them different health insurance plans.
Today, its a very different situation. An entire company changes its cost-sharing. So all the people, a lot of peoples neighbors, have higher cost-sharing. cautionary. Its possible that well get a very different response because its so uniform or so universal, for good or for bad. In a sense, getting physicians attention a lot easier when a lot of people have more cost-sharing than when one person who has the good fortune to be selected for this clinical trial.
Weve reached the end and I just want to say that the speakers and the reactors have done a marvelous job. Its been fascinating just listening to them talk about these issues. I want to thank them.
Paul Ginsburg: I also want to urge you, I think those green forms in front of you are evaluations. We do make good use of them and theyre that much more valuable if we get a high response rate.
Please sign up for—if youre not on HSC Alerts for our publications, please do that.
Again, I want to thank the Robert Wood Johnson Foundation for its support of HSC and this conference. And at RWJF.org is the webcast of this thing.
Our next conference is going to be on March 18th and its going to be a conference about market change with results from our round of site visits. And this conference is being put on in partnership with Health Affairs who, in March, is also releasing a special or thematic issue on market change.
Thank you very much.
[Whereupon, at 11:58 a.m., the meeting was adjourned.]
MARK R. CHASSIN, M.D., M.P.P., M.P.H. - Edmond A.
Guggenheim Professor of Health Policy and Chairman of the Department of Health
Policy, Mount Sinai School of Medicine
KAREN DAVIS - President, The Commonwealth Fund
ARNOLD MILSTEIN M.D., M.P.H. - Medical Director,
Pacific Business Group on Health; National Health Care Thought Leader, Mercer
Human Resource Consulting