WARNING: THIS IS GOING TO BE A VERY GEEKY POST AS I’M TRYING TO FIGURE OUT WHAT I THINK I AM THINKING
Claims based risk adjustment is on my mind a lot this week.
I’m thinking that my dissertation is likely to be a Defense Against the Dark Arts dissertation on the ways that insurers and other risk bearing entities respond to incentives that are skewed. Risk adjustment is critical to aligning insurer incentives with societal incentives. I think we have a problem. I need to think some things through.
Claims based risk adjustment is quite common. It is used to move money to insurers that attract and insure individuals with high, expected and plausibly predictable future medical expenses. Some claims based risk adjustment systems are zero sum like the ACA where the money moves from insurers with low coded risk to insurers with high coded risk. Some risk adjustment systems are externally funded like in Medicare Advantage where CMS pays out risk adjustment. Risk adjustment is critical for any guaranteed issue, community rated(ish) products otherwise insurers will only compete on being as unattractive as possible to people with likely high medical expenses while also being attractive to people with few if any probable expenses. This means narrowing of networks, increased gatekeeping, prior authorization and administrative burden run-around for everyone as the game to play is hot potato with short fuse hand grenades.
Claims based risk adjustment requires claims. Claims are a function of utilization. Utilization is a function of actual medical acuity, access to medical care (cost-sharing, transportation, administrative burdens are some potential barriers that vary based on both individual, cultural and geographic basis), and willingess to use medical care. At a given level of acuity some people are more able and willing to go to a clinical environment for care than others for all sorts of reasons. Any given moment of formal utilization is a non-zero chance that a risk adjustable diagnosis is entered into the formal bureaucratic system of accounting and creditation. Any moment of non-formal utilization or outright care avoidance (rub some dirt on it medicine…) is purely a zero probability chance of creating a risk adjustable diagnosis.
Claims based risk adjustment relies on diagnoses. A claim can have 25 diagnoses (Dx) on it. Each Dx slot can either be empty or filled with a Dx Code. Conditional on it being filled, there is some probability that the code is a risk adjustable code and some probability that it is not a risk adjustable code. The number of codes that are filled in is a function of the interaction of the individual patient, the clinician, and the broader, local medical systems. Some patients may be deemed to be more credible in reporting their symptoms that leads to a diagnosis being entered onto a claim while others may not be able to either report at all or have their reports discounted. Some clinicians will code everything and anything, some will only code the immediate problem. Most are somewhere in between. Coding environments like Critical Access Hospitals will have different coding practices than tertiary academic medical centers. The distribution of coding environments is likely not random. It could be associated with various metrics of social vulnerability.
Claims based risk adjustment relies on the accumulation of novel diagnosis groups. Any diagnosis that is risk adjustable only adds incremental value to the risk adjustment transfer if it is unique. The probability that any new diagnosis slot is filled with a unique to the individual/time period dyad diagnosis grouping is a function of everything above plus the number of interactions. An individual with no utilization who then has one visit will have a fairly high probability that they are adding to risk adjustment value. An individual with 100 clinical encounters in the year having one more encounter is very unlikely to have a new diagnosis that has not been coded before on the 101st claim of the year.
If we think about the construct of “Predictable Healthcare Spending” as quasi-latent and each claim is a quasi-random draw into that well, some draws will come up with no risk adjustable diagnosis, a lot of draws for a lot of people will come up with diagnoses that were previously drawn before, and a few draws will come up with new stuff. The new stuff is likely to happen early rather than late for prevalent conditions and quasi-randomly for new conditions. As I’ve been thinking about this, I just keeping on thinking about the Rubber Ducky Game at school carnivals where every rubber duckie has a prize but almost all of them are small prizes but every now and then there is a rubber ducky with a big prize…. risk adjustment is not quite this random as people go to specialists because there is a prior belief that there is something big/weird happening but this is how I’m visualizing the draw process at the moment.
So where am I going with this?
If we think that the probability of a risk adjustment score is a cumulative interaction of the probabilities of an actual problem (or at least a medically defensible codeable problem) plus the probability of a patient actually having at least one clinical encounter plus the probability that a risk adjustable diagnosis is coded plus the probability that the diagnosis grouping is unique for the patient/period dyad then we need to think that these are socially constructed scores. People who are likely, all else equal, to seek care and who are likely to get a risk adjustable diagnosis are probably meaningfully different than people who either are not able/willing to see care or conditional on seeking care, less likely to get a diagnosis that is risk adjustable for the same fundamental condition.
Risk adjustment as an economic and actuarial process moves money to insurers whose population codes as sick. Coding as sick is both a technocratic and a social process. Insurers will want to seek out individuals whose net revenue (premiums + risk adjustment) is greater than expected expenses. If we think that risk adjustment is socially skewed, it overweights people who can get coded “well”/”heavily” and underweights people who either have low utilization or have low probabilities of their underlying medical conditions being coded aggressively.
Steve in the ATL
On Balloon Juice? Heaven forfend!
narya
It sounds like you’re assuming that the social determinants of health (SDOH) are less a part of this equation because the folks who are most strongly affected by SDOH are either not getting care, or on Medicaid (which doesn’t seem to be part of this, because it’s talking about insurers who have some ability to “select” their insured population). Is it helpful to compare diagnoses in the populations you’re assessing with, say, the 30 million folks who got care at an FQHC last year?
Eunicecycle
The hospital where I used to work had employees whose jobs were to make sure docs were coding their patient encounters properly. Docs are pretty notorious for undercoding, I guess, because they are so busy. I don’t know how this plays into this; it probably doesn’t but I always thought it was interesting.
David Anderson
@narya: SDoH is a huge part of the implicit set of questions that I’m poking at. The challenge is that someone whose SDoH situation keeps them from getting care is effectively accurately coded as having no predictable costs because they don’t see a clinician when they really should be seeing a clinician.
I’m more in the line of thinking that if there are differential coding practices that are not randomly distributed (and we’ve seen that with Z-codes etc) and SDoH are part of that matrix of distribution, then people living in highly vulnerable environments are likely to be significantly underweighed in the risk adjustment formulas.
Another Scott
The social, regulatory, political, legal aspects of this stuff are huge and do not get enough attention.
I was at an online meeting yesterday where someone was talking about how applying industrial solutions to digital problems was doomed to failure (with the example given of the fall of the behemoth General Electric and the rise of FB, Google, Apple). A nice slogan, I thought, but far too simplistic. I wanted to interject that one has to consider the legal framework, patents and copyrights, network effects, the push for monopoly and monopoly rents, etc. While economies are always changing and there will always be old losers and new winners, we have the winners and losers that we do because of the legal, regulatory, political, etc., framework that we have constructed. MS would not have gotten huge and crushed all of its competition without per-processor licensing agreements. MS would not have gotten huge and crushed all of its competition if software was free (or if copyrights were substantially shorter, or if exclusive licensing and bundling were forbidden, etc.). Dean Baker beats this drum a lot.
Similarly with health insurance. As you know, a lot of these issues and pathological incentives go away when players in the system cannot collect the revenue but try to push the costs on to others. We’ve got a long way to go to get there, of course…
Good luck figuring all this stuff out! ;-)
Thanks.
Cheers,
Scott.
Brad F
David
I think your hypothesis is testable. Im using a GI condition as an illustrative example but it’s a dealer’s choice.
Think inflammatory bowel disease and severity. Perhaps SES factors will impact mild symptoms and presentation to a PCP. But this early in the diagnosis, its unlikely an IBD diagnosis will be made.
But look at new-onset IBD in a Crohns or UC database when a scope and biopsy are performed and a gold standard test is performed. THen look at the time of onset of symptoms, months or years elapsed until discovery, and its association with the type of insurance. Nausea, vomiting, rectal bleeding, etc., are agnostic to transportation and income. Folks will find their way to an ER.
The above could be extended to cancer, autoimmune, or other conditions if acuity of onset and diagnosis are closely correlated. The condition would also have to be a desirable RA code from the payer’s perspective.
Brad
susanna
It’s always a responsible thing, to know what condition your condition is in….
narya
@David Anderson: I don’t know if this helps at all, but the yearly reporting that FQHCs complete includes a bunch of ICD10 codes for several tables (especially 6A), and we are required to report on %FPL (based on income & family size) and insurance type, including uninsured. Would it be useful to look at folks who get their care at “regular” institutions vs. folks who get their care at FQHCs? I will also note that the Bureau of Primary Health Care has made a huge push around diabetes, hypertension, and maternal health, and, more recently, HIV. They recognize that folks coming to FQHCs disproportionately experience the first two of those conditions, those conditions are expensive if left untreated, etc. I apologize if this isn’t any use to you! I just love the UDS dataset and tend to proselytize about it.
Lobo
I think here is when you say: A multi-player insurance system that looks to poach the least expensive patients and shed the most expensive to others, while others are trying to do the same does not have any kind of optimal outcome. The variability of being human can become just noise if the system is constrained to a limited set of players such that the action of one person becomes a rounding error.
Lobo
It is a Rube Goldberg system that only allows tinkering around the edges. The system reminds me of the clay soil around my area. Yes, you can try to enrich it various ways but it is still clay soil and darn hard to grow anything in it. The solution is raised beds with actual soil.(Learned the hard way. ;) )
The solution is keep fighting for improvements until we are ready to move a new paradigm.
David Anderson
@narya: Let’s talk as I’ve never played with that data set.
Can you e-mail me at dma34 DUKE
David Anderson
@Lobo: Maybe — my brain is more going along the lines of — we use risk adjustment to make certain classes of people (sick/expensive) more attractive to insurers to avoid or at least minimize insurers from running like hell. The way that we categorize people as sick and/or predictably expensive has a lot of minimally examined potential bias in it. If we think that there are predictable sub-classes of the “sick/predictably expensive” group that are over or underweighed in the aggregate analysis, then the incentives are to design plans that are attractive to some subgroups of sick/overly predictive expensive relative to other subgroups… and those disparities likely map onto other pre-ex disparities.
Aaron S
It’s reminiscent of the study released a couple years ago that found racial bias in an triage algorithm created by Optum to target patients for intensive care management. The algorithm used health costs as a proxy for health needs, but cost isn’t a race-neutral metric because less money is spent on Black patients at the same level of need. The researchers found that “at a given risk score, Black patients are considerably sicker than white patients, as evidenced by signs of uncontrolled illness.” (abstract in Science)
narya
@David Anderson: will do (from my personal email)
Nobody in Particular
@Lobo:
I dare say it is little more than a protection racket. I’m no expert, (as I’ve already made known). Most of them are full of shit anyway. But I’m a republican in the mold of Ben Franklin and many of the Founders. Not all. And we can never really know about Hamilton because he went and got himself shot. Who knows how his policies might have developed. Commerce is “generally cheating.”
I know what my Great Uncle thought of Insurance because he founded the American College of Underwriters. But that was mostly regarding life insurance. Neither he nor his brother, my grandfather, who also taught at Wharton and wrote over thirty books on Commerce and Transportation, testified before Congress often and set the Tariff rates for the New Panama Canal, (ironic, no?) would be happy to associate their names with any of this bullshit. And you have to admit that McDonald’s speech to the Economic Club was not screaming about “Socialism.” Although that site does apply the spin they prefer with this.
Wisconsin was a hot bed of progressivism back then. Most Americans, in fact the world, were supporting Henry George’s Single Tax.
I can read. My reading comprehension is just fine. McDonald’s point is clear: This is not what most of the founders envisaged. But that’s what happened. The left tends to forget that the compromises made for ratification were a matter of urgency, as was building an economy to defend the new nation in a hostile world. Jefferson’s DOI was heavily edited. Otherwise slaves would have been manumitted. He wanted to free the slaves. But national survival in a hostile world had to come first, as a matter of expediency. Read Franklin’s final speech at the convention. Ratification quickly was a matter of national survival. To say that all the founders were in agreement is risible nonsense. They were a contentious lot and rarely agreed on anything. I think I’ll leave the number crunching and bean counting to others more qualified, like David. Sol’s point back then was that human life had value. Nobody pays attention. I am firmly convinced we’d be better off without insurance in Medicine. And lots of the medical professionals and academics in my family agree
It’s the corollary to one of Arthur C. Clark’s laws. For every expert there is an equal and opposing expert. And after cogitating on it for the last ten years, I’d say we get the government we deserve, and we got him: Trump. He personifies late stage America, sadly.