March 20, 2015 | Eric Liederman has spent his career as both a physician and a clinical informatics systems director for groups including the U.S. Navy, Catholic Healthcare West (now Dignity Health),the University of California Davis health system, and now Kaiser Permanente, a large capitated integrated delivery system.  

At the Medical Informatics World conference in Boston in May*, Liederman will speak about balancing privacy risks with patient care, but he’s also spent much of his career weighing external and internal incentives for decision making and the delivery of care.    

He spoke with Clinical Informatics News editor Allison Proffitt about the economic differences between fee-for-service and capitated models of care and how cognitive dissonance changes the care patients receive. Here’s an edited transcript of that conversation.  

Clinical Informatics News: How does fee-for-service work, in your view? What are the incentives?  

Liederman: If you apply incentives, you will generally get the behavior you’re incentivizing. The behaviors that fee-for-service incentivizes and the behaviors that capitation incentivizes are very different behaviors and in some cases they’re in effect opposite behaviors. That’s the real challenge.

Pure fee-for-service, which has been practiced in the United States for a very long time, really took off with the explosion of health insurance with World War II and then Medicare and Medicaid in the ‘60s. What it incentivizes is a number of behaviors. One is that you want to actually see patients in person rather than in some other non-in person venue because you get paid when you see the patient. And it incentivizes doing as much as possible, especially expensive interventions because you get paid more when you do those things. And there’s just no other source of revenue.

There are examples of practices trying to do "the right thing" in their patient population in a fee-for-service environment and going bankrupt. There was a practice in Minnesota back in the early ‘90s that was an eager early adopter of practice guidelines and evidence-based medicine and they basically emptied out their waiting room and went belly up financially because what they did worked. But they weren’t being paid to keep the patients out of their waiting room. They were being paid still to see the patients and give them their interventions, not keep the patients at home and healthy.

How does a capitation model work differently? 

With capitation, an organization is paid upfront to take care of the healthcare needs of a population. The only way you can really make money and stay in business is if you put into place wellness and healthcare management interventions that keep people home and healthy to the extent possible. When they do come in sick or injured, do the most effective thing right away or as early as possible and don’t mess around and do lots of other things. Make most use of dollars you’ve already been paid, because if you don’t do those things and the patients don’t stay healthy, they’re going to come in sick or injured more often and you’ll spend more money than you’ve been paid.

The incentives for capitation are to keep people healthy and to practice evidence-based medicine and to try to create an effective and efficient system that does the optimal intervention, not the maximal intervention. Fee-for-service, on the other hand, is incentivized to do the maximal number and type of interventions, and to avoid anything that would keep people out of the hospital or out of the clinic.

How hard is it to move between a fee-for-service and capitated model of care, and how does shifting between the two models impact care? 

It’s not impossible but it’s very, very challenging. The problem is if you’re in a fee-for-service world now and you’re finding that your incentives are shifting more towards capitation—which they seemed to be at least at the federal level—then it becomes really challenging to know what to do because if you’re somewhere in the middle. Some of your patients will be fee-for-service, some of them will be capitated perhaps, or maybe you’ll have a mix of sort of specialized quality programs where you get paid more for this or paid less for that. How do you know what to do becomes really, really hard. The valley between pure fee for service and pure capitation becomes a place of perverse cross-incentives and if you try to actually translate all of this into strategies and into clinician payment schemes, how you’re going to pay your doctors for instance, it becomes terribly, terribly complex to the point where if you were to actually try to do it where one-tenth of one percent of your bonus is for this and one percent is for that, nobody pays any attention anymore because they can’t keep track of it.

Imagine an organization that’s half capitated and half fee for service. So let’s take the highest paying diagnosis-related group (DRG) that Medicare pays nationally which is congestive heart failure. That’s the big money for an organization if it’s fee-for-service and it’s a big risk for a capitated organization. How do you manage these CHF patients so that they optimize your financial situation in either case?

The way the Medicare are set up is that if a patient is admitted with heart failure and they’re readmitted within 30 days of discharge then the government doesn’t pay anything for that second admission. But if they’re readmitted after 30 days then it’s the clock is restarted and it’s a fresh admission that they pay for.

If you’re to respond to these incentives, care management programs that targeted heart failure patients who were discharged from the hospital would intensively manage them for 30 days and on the 31st day, call them up, tell them they’re doing so well they don’t need to take those pills anymore and here’s a coupon for a free pizza. Then maybe you can get 12 admissions a year.

That sounds crazy! I mean you think nobody could ethically do something like that; we’re in healthcare. But it turns out that if you pay someone to do something then they will. The Wall Street Journal recently had an article about long-term care facilities. They were discussing how Medicare pays a lot more money for a patient who has been in the facility for 23 days than one who has only been there for 22 days.

In a pure economic world what would you expect to see? You’d expect to see a relatively low number of admissions for less than 23 days, especially for days like 20, 21, 22, then a huge spike for days 23 or maybe 24 and tail off after that. And that’s exactly what the Medicare data shows. Now is that likely to have any correlation with healthcare needs? Obviously not. People don’t need long-term care facilities for exactly three weeks and two days. I mean, how can you explain that clinically? But you can easily explain it economically.

How do clinicians experience that disconnect?  

Clinicians and healthcare provider organizations, with probably some small exceptions, want to do the right thing. They’re professionals or professional organizations and they want to take care of patients and be as effective as possible.

But of course, everybody’s operating within incentive systems and economic systems that they have to account for if they’re going to be able to continue to be around. And so what happens when people are faced with two competing and contradictory thoughts at the same time? Cognitive dissonance. It’s very unpleasant to have cognitive dissonance. You have to resolve it; you have to decide which of those two contradictory thoughts you’re going to go with and then rationalize the other one.

There are lots of examples of that over the history of medicine. One example is that up from the start of Medicare in the mid-’60s to September 30, 1983, a hospitalization was paid for on a per diem basis, on a day by day basis and actually in earlier Medicare the hospitalization had to be three days or longer. And then midnight, October 1, 1983, they switched to the DRG system. The DRG system is a type of global payment and it basically said whatever the diagnosis for the group is that the patient ends up being discharged with that’s going to determine how much money you get paid.

And so first what happened was that in the early days of Medicare everybody got hospitalized for at least three days because that’s what they got paid to do. And so that resulted in some pretty bizarre things. For example, colonoscopies. Believe it or not, people would be brought in the first day for the prep in the hospital. The second day they’d do the colonoscopy and then they’d keep them overnight for observation just in case there were problems.

As this DRG deadline was looming, the physician community was up in arms in this country. They’re talking about how hospitals were going to require patients to come in the same day they were going to have this surgery. Doctors were saying, ‘This is dangerous. We can’t have people just showing up and then having their surgery. People will die.’

Then within a few months after the deadline in the land of DRGs all that talk disappeared and everybody was talking about how it was unethical and dangerous to keep people in the hospital longer because they could catch hospital acquired infections or bedsores, whatever, and that you had to bring them in same day because that was the safest and best thing to do for the patient.

Well, how did that happen? How did these changes happen across the medical community? It happened because you’ve got to resolve your cognitive dissonance. If you’re getting paid to do something then you have to find a way to decide if that’s the right thing to do.

Getting back to that heart failure example, if half your patients are capitated and half are fee for service, then what’s your strategy for dealing with heart failure? If you implement a robust long-term heart failure care management program do you only enroll the capitated patients? Is it ethical? If you enroll all patients, are you willing to risk going bankrupt? I don’t blame any individuals in that space. It could be anybody. I’ve been there. I’ve worked for organizations that have been there and it’s a very hard place to be because you have opposite incentives playing on two different substantial portions of your patient population.

I’m hoping ultimately more and more parts of the healthcare industry in the United States—particular the folks who are paying for it because they’re providing the incentives—may want to move toward capitation. But again, as we talked about earlier, if you don’t make the switch completely from one to the other and you’re in the middle somewhere that is a really hard thing to do.

At what level can a healthcare business actually do something about this? Can you change your payment structure?  

Physicians and hospitals and certainly integrated delivery systems have potentially, at least from what I’ve seen, the power to directly intervene in the kind of payment incentive program that they get. For capitation to work you need to have a large enough and well-defined population of patients. That’s critical because if you don’t have a large enough population of patients then a few really sick outliers can bankrupt you.

If you get into a situation where you think you can manage the risk as a healthcare organization, then you can, at least in my experience you can go to payers and negotiate with them, essentially [negotiate] capitated arrangements, risk-sharing arrangements. Since the payers, the insurance companies, are generally holding all the risk now, they are generally very open in my experience to these ideas. But of course, you need to be sophisticated about what you agree to. You want to make sure you don’t get all the overlay of fee for service while taking on the risk. If you’re taking on the risk, then nobody should have to go to the insurance company to ask permission for things. That should be managed internally all of a sudden. You shouldn’t have the insurance company managing prior authorizations if your organization is taking the risk. And so it does take a level of size, integration and maturity to move in this direction. But I think that as you have more and more of these accountable care organizations, at least you have the size and structures that are starting to come into being where they could negotiate these kinds of reimbursement arrangements.

If you can negotiate capitated arrangements, where does that leave your physicians?  

The hard part is that by necessity the clinicians have resolved their cognitive dissonance in the direction of fee-for-service incentives and it’s a huge cultural shift to get them to move in the other direction.

An incentive system creates a culture and that culture can be very strong and resistant to change because it allows people to go about their day to day and feel like they’re doing the right thing because they’re doing the same thing everybody else is and we’re all in this together and this is what we do.

Then all of a sudden say, okay, we’re switching over to this capitation and to be successful in capitation here’s what we need to do. We need to start collaborating on the telephone. We don’t necessarily have to see patients anymore. If your colleague calls, you just give him your best advice. You’re not going to get paid extra anymore if they come in and see you and we’re going to do telephone visits and we’re going to have these care management pathways and we’re going to have para professionals like nurses and pharmacists doing outreach to patients. All of that is going to be a huge change to the physicians especially, and many of them will be highly resistant to that and highly threatened by the idea. That’s really probably the hardest part at that point.

You clearly believe a capitated model is better; how are you defining “better”? Better for the patient? Better for the practicing physician? Better for the business? 

In my personal experience you’re right, it depends on what you measure. But let’s say you measure quality outcomes, let’s say you measure patient satisfaction, whatever standardized survey metrics you can find. Let’s say you measure long-term quality of life stuff like life expectancy of populations or sub-populations. Then, generally speaking, systems that are incentivized to deliver on patient health—and capitation does that—have better, substantially, really substantially better outcomes there.

The United States is an outlier around the world even if you just compare it to other OECD countries in that our expenditures, per capita expenditures on healthcare are by far the highest in the world and yet pretty much any healthcare quality metric we’re pretty far down the pack.

Why is that? How is it that we could be spending all this money and coming up with pretty mediocre outcomes? We’re certainly not getting value for our money like others are. And that doesn’t mean we need to adopt health systems that others have. I’m just suggesting that at least in my personal experience the behaviors that you get are in a global sense relatively predictable based on the incentive system that’s in place. And from what I’ve personally experienced fee for service has an incentive system that results in, incentivizes and results in behaviors that are going to increase the cost of care without concomitant increase in the quality outcomes because the behaviors that fee for service incentivizes that I mentioned at the beginning of this conversation are to wait until patients are real sick and then intervene maximally, do as much as possible.

At a societal level and at an individual patient level that’s generally not what you want. I mean what you’d want would be a system of care that would provide self-care knowledge and tools and outreach where appropriate to keep people as healthy as possible for as long as possible, get them in to have targeted evidence-based interventions as early as possible in their illness or injury and as a result have much lower total expenditures and higher quality outcomes. And in my experience that’s what capitation incentivizes.

The flip side of capitation is you could have people doing just the minimum to make their money. You’re not necessarily insuring quality.  

You’re absolutely right. You’re absolutely right. So, if you have an organization that takes a short-term view and doesn’t really understand, or maybe they’re not in it for the long-term, maybe it’s an organization that’s trying to make a quick buck and get out, then you would put frictions in place to keep people out of your healthcare institution as much as possible. You’d collect your money upfront, but make it as hard as possible for patients to get in and get seen, you’d try to spend as little as possible and at the end of the year have as much leftover as you possibly can. That might work the first year maybe but in the longer haul that is not only not going to work but it’s going to drive your organization to bankruptcy because you’re not taking care of your patient population either proactively where they are at home or work or allowing them rapid access to care when they’re sicker or injured to get the best care as quickly as possible. They’re going to still cause expenditure for you. Either they’re going to end up in your emergency room or they’re going to end up in somebody else’s emergency room or hospital and you’re going to have to pay, your organization is going to have to pay because you’re capitated, you’ve already been paid. So now, you’re on the hook. You could try that approach but there’s no way that it would work for longer than a short period of time, at which point both your reputation and your clinical viability would go up in smoke.

You have spent much of your career managing clinical and medical informatics systems. Is there an informatics strategy to address these challenges? 

Informatics is a supporting methodology and system of tools. In both fee-for-service and in capitation you need certain things. You need a practice management system; you need a scheduling system; you need to have access to patient data; you need to be able to place orders; you need to be able to document actions; you need to be able to send messages to your colleagues. You need to be able to collect proper copays or deductibles from a revenue cycle perspective. All that’s pretty straightforward.

Then you want to use this platform to intervene to result in better care. It’s a powerful platform. If somebody’s already in there documenting care, already in there ordering, then you have this great opportunity to have the system say, hey, what about this. And so, there are some things that make sense across the board, whether it’s fee for service or capitation. Clearly, especially when Medicare is moving towards not paying for never events, you’re going to want to alert on any potential never events. I mean you don’t want to prescribe medication that the patient is allergic to and if the patient has a very high risk of developing bedsores, which could be easily calculated, then you want to alert to that and get some nursing intervention when in hospital and things like that.

But when you talk about interventions that will keep them out of the hospital, interventions that will reduce the likelihood of getting sick, then all of a sudden you run into incentive issues. All of a sudden the question isn’t whether your platform can do that level of decision support or disease management or protocols, it’s whether it should. That’s a strategic decision that your organization has to make.

That’s why you don’t generally see long-term use of interventions that are antagonistic to the payment system in place. And that’s challenging, I mean really challenging. If you know what the right thing is to do and for good economic reasons your organization decided not to do that, well that’s very hard. That creates right front and center cognitive dissonance, especially among the leadership and that’s a hard place to be.

I personally think that maybe eventually there will be a desire to move in the direction where there is a way to resolve that cognitive dissonance structurally. And in my experience, capitation can do that.

Editor’s Note: Medical Informatics World is a conference hosted by Cambridge Healthtech Institute, Clinical Informatics News’ parent company. It will be held May 4-5, 2015, in Boston. For more details, visit www.medicalinformaticsworld.com.