When someone is arguing that the health system needs an overhaul, one of the most common reasons they cite is that “our health system is built on a flawed foundation of fee for service.” Arguments like this always sound so bulletproof when they rely on vague yet widely accepted assumptions, but let’s think clearly about this, just for a moment.
First, let’s go back to Econ 101 to recall that association does not imply causation. So, when you see a health system based on fee for service that is delivering surprisingly low value, you are seeing an association. It starts to look a lot more like causation when you compare this health system to other health systems in the world that are not built on fee for service but are all delivering much higher value. But it starts to look a lot less like causation when you compare this health system to any other industry and see that nearly all of them are based on fee for service yet somehow delivering excellent value. (I’ll come back to this later.)
So what is fee-for-service reimbursement, really? It is simply one extreme end of a spectrum, and at the other end sits capitation:
This spectrum is not well understood. People always think of it in terms of incentives (“bad” incentives at the fee-for-service end, “good” incentives at the capitation end), but what is actually being varied as you move from one end to the other? I’ve never heard anyone talk about that. So let’s talk about it; the conclusion will be surprising.
I will say that fee for service is the purchase of a narrowly defined service (e.g., a single doctor visit, a single operation, a single medication), whereas capitation is the purchase of a broadly defined service (e.g., all health care you need for a year). So, breadth of service purchased at one time is really what varies on this spectrum. But in addition to the breadth of the service being purchased, there is another important difference: risk. When you pay for narrowly defined services one at a time, you have all the risk (meaning, if you get sick or break your arm, you’re the one who is financially accountable). And when you pay for a broadly defined service, the party you’ve paid has all the risk (meaning, if you get sick or break your arm, they’re the one that is financially accountable). Note that this is primarily where the “accountable” comes from in accountable care organizations–they have the financial risk, not the patient.
This is a spectrum because a service could sit at any point along it. For example, Qliance offers flat-rate, no-limit primary care; they are financially accountable for anything that could be considered primary care. Another example is episode-based billing, where the patient pays a single lump sum in exchange for a guarantee that the provider will do everything necessary to treat the specified medical condition. An interesting side note is that even a typical fee-for-service doctor visit is not truly fee for service in the sense that the patient has all the risk; at least, last time I checked doctors don’t make patient pay larger copays any time appointments run longer than the allotted 15 minutes.
So what can we conclude from this discussion? Is fee for service actually bad?
Being too far at the fee-for-service end of the spectrum can definitely be bad if it means patients are expected to coordinate complex care by themselves. But being too far at the capitation end of the spectrum can also be bad if it means patients are not at all financially responsible for the services they choose to consume and are also stuck having to get all their care from one source that will inevitably provide less-than-exceptional quality for some services.
This means we need to find the perfect point somewhere in between the extreme ends of the spectrum where our health system will deliver optimal value. This point obviously depends on the service and the individual involved, as well as who can bear risk most effectively, but the way to think about it is using the “jobs” principle as taught by Clay Christensen. When a person enters the health system, it’s because they have a “job” they want to get done. That job generally isn’t something as narrow as to get an x-ray; more likely, their job would be to fix a suspected broken arm, so they’re looking to purchase all the services together that can fulfill that job. The job could also be broader, like to have the peace of mind that they have little to no healthcare-related financial risk and can just go and get care from one source no matter what comes up, or also to have help to keep healthy and thereby reduce care-requiring episodes (think: Kaiser Permanente and similar integrated delivery systems).
If we want to successfully redesign our healthcare delivery system (isn’t that all the talk these days?), we need to understand that fee-for-service isn’t intrinsically bad; the only bad thing is missing the sweet spots on that spectrum.
One final, crucial point: Yes, overhaulists* properly attribute many of the problems in our system to it sitting too far at the fee-for-service end of the spectrum (e.g., overtesting, overtreating, fragmentation), but they have overlooked what has caused the system to fail to adjust itself to a more optimal, jobs-focused point on the spectrum, as other industries do. This has everything to do with value not being financially rewarded in our system, and it is the topic of the publication I’ve been working on, as well as many future blog posts that will start to lay out the solutions more cohesively (after my publication comes out). More to come!
* My term for someone who believes the entire health system needs to be overhauled
UPDATE: Added a paragraph or two for clarity.
Clear Thinking on Healthcare 
Reblogged this on Carlo Favaretti.
I always appreciate the reblogs, thank you Dr. Favaretti.
Well done, Taylor! I really enjoyed this one. I was skeptical when I started reading it–thinking you were advocating ‘no change’.
Very interesting post. I’m curious about how you think a clinically-nuanced value-based insurance design scheme would, or if it even should, fit in to all of this. Thoughts?
Good question. VBID doesn’t exactly map to the spectrum above because it doesn’t deal so much with varying the scope of services purchased at a time.
However, the further toward the capitation end you get, the more risk is shifted to the payer, and the more likely they’ll be to implement VBID as a cost-saving measure (paying 100% of the cost for high-risk patients to take a pill every day for 5 years is cheaper than paying for an ED visit for a heart attack).
Of course, even a patient with all the risk on himself can make the same choices VBID encourages him to make (he can choose to pay for and take the pill as his doctor recommends), but patients usually aren’t as good about taking the long view when making decisions (“Do I really have to pay for that prescription? I don’t feel sick.”).
I guess my point is that this question about VBID deals with who has the incentive to do cost-saving prevention, and the answer is whoever stands to save the money. This is all part of my model that I’ll be blogging about soon enough!
You mention in your final addendum that ‘overhaulists’ too often blame the pervasive fee-for-service model for our healthcare woes when in reality it isn’t an inherently bad model; there is just no economic impetus (i.e. no ‘quality control’ measures to ensure patient value), but I’m curious about how you think a fee-for-service model may adversely affect continuity of care. As you mentioned above, a move towards full capitation might unveil a powerful incentive to keep patients healthy (perhaps through more robust preventative care), but reaching full capitation would encourage cost-cutting measures, thereby degrading quality. I find this argument powerful but am particular curious about what you mean by the following statement:
“This means we need to find the perfect point somewhere in between the extreme ends of the spectrum where our health system will deliver optimal value. This point obviously depends on the service and the individual involved, as well as who can bear risk most effectively…”
Does this imply that there should be multiple models depending on the patient, disease state, etc.? How do we effectively balance economic incentive for payer, patient, and provider while still ensuring quality and continuity of care? Looking forward to reading your paper!
I see you found my blog! Sorry we didn’t get to talk more.
Your question is an important one, and the answer is yes, there should be multiple models depending on the patient, disease, etc. Some patients will really want/need to be at the capitation end of the spectrum so they have tons of help staying healthy. Others are already generally healthy but have different care episodes come up (the broken arm, the infection, the malignancy, etc.), so they will be mostly purchasing at the FFS end of the spectrum. The important point is that, for example, if they get diagnosed with cancer, they will be able to buy (all at once) all the services needed to treat the cancer until it’s in remission or they pass away.
Why is that so important? (This gets to the balancing of economic incentives part.) If multiple providers are all offering that one-payment cancer care service, the patient will be able to choose right up front the provider who is likely to do the best job for the best price. Providers who consistently achieve this (i.e., are the highest value competitor) will be rewarded financially with full patient loads. Thus, the economic incentive will be to become the high-value provider. In summary, when the scope of service purchased is matched up with the jobs the patients have, they can more easily compare the relative prices and reward higher-value providers with business.
Many things get in the way of that process actually happening, including insurance plan design and a lack of usable quality information, among other things. These represent the barriers to patients getting to figure out the highest value providers and choose them, which, consequently, gets in the way of value being rewarded financially.
That might not make sense how I explained it, but that’s the basic idea of what I’m planning to blog about over the next year or so with hopes to lay everything out in a more systematic way.