The New England Journal of Medicine has a short piece on a physician lead Accountable Care Organization. Within that piece is an incredible description of how risk adjustment optimization programs actually work:
For Medicare ACO and Medicare Advantage patients, it is essential to code health risk accurately so the Medicare calculated benchmarks reflect the likely cost of caring for patients…
When risk-based practices redesign their care processes to improve health and reduce hospitalizations, they also need to achieve more accurate coding, which can lead to improved revenue. Specifically, providers need to make concerted efforts to complete annual wellness visits for the majority of patients, typically more than 70% of patients, to document all diagnoses annually. An effective process is to begin outreach for wellness visits in January and work through their rosters by summer. Once wellness visits are complete, they then spend the remainder of the year managing those diagnoses aimed at secondary prevention and closing gaps in care to improve quality metrics.
This is what I did for several years at UPMC Health Plan. I mined claims data to identify highly probable recurring, persistent diagnoses to support an outreach team of clinical coders who encouraged primary care providers to get the highest value patients in for visits so that the right diagnosis would register on a claim so that we could get a big revenue bump.
In Medicare Advantage, the incremental increase in risk score is all revenue for the insurer. The immediate underlying claims experience does not change but the revenue changes significantly. If the new information of well-coded claims leads to changes in clinical decision making, there is some clinical value in this process, otherwise it is a revenue optimization process. This is more likely to be true for individuals who are already receiving treatment for a disease but are encouraged to have a work-up so that they can be coded at a higher level while receiving the exact same treatment. Some evidence shows that these visits don’t incrementally increase risk scores .
For an insurer to run a competent risk adjustment optimization process requires managerial sophistication, technical chops on the data geek side, good relationships with doctors as well as large data sets. Insurers with broad and deep data sets have an advantage over insurers with narrow and shallow data sets:
Risk adjustment optimization that can capture “bonus” revenue that does not change the underlying claims experience is a significant barrier to entry for new risk bearing entities that operate in risk adjustment lines of business. Entities that know how to look at previous claims and then get the high value codes entered on the current year claims will have a significant competetive advantage over firms that either don’t know how to do this or don’t have the data depth to optimize their risk adjustment scores.
Yarrow
I don’t understand this. Someone has a diagnosis and they’re being treated, but doctors are encouraged to bring them in, write up codes differently so the insurance company gets more income but for the patient nothing changes? Is that what your’e saying? Why would the patient want to do this? What’s in it for them? How do they sell it to the patient to get them in to the doctor? It’s a pain to go to the doctor and most people don’t want to go if they don’t have to. What’s in it for the patient? What sort of diseases or issues trigger this sort of thing?
If my understanding is correct, something seems very wrong about this. It’s not about improving things for the patient. It’s all about maximizing profits.
David Anderson
@Yarrow: Basically yes… Type 2 Diabetes not otherwise specified may have a different risk adjustment co-efficient than Type 2 Diabetes with peripheral neuropothy.
The patient may get a small incentive to show up for a wellness visit (gift card, lower deductible etc) and the doc might get some cash (either as a straight incentive or as a change in shared savings revenue baseline)
And yes, risk adjusting has some gamability where it is about revenue optimization.
Yarrow
@David Anderson: Companies often offer “incentives” to do wellness activities–like do a health fair to get weight, cholesterol, etc. checked, do fitness activities, participate in coaching, etc. Sometimes it’s a discount off your premium, sometimes gift card, etc. Is that the kind of thing your’e talking about? The ones I’ve seen are available to everyone not just people with illnesses.
In your example, would it be the doctor calling the patient in or the insurance company calling the patient to encourage them to go in for a visit?
Roger Moore
@Yarrow:
Why do patients ever get tests? It’s because their doctors tell them their needed as part of their diagnosis. You could probably do this just by telling the patient the test is necessary and even justify it medically, given that these tests are intended to affect a patient’s diagnosis. Most patients would never be able to tell the difference because they never understand the exact reason why they’re getting each test in the first place; they simply trust their doctors.
There’s a reason so many of us say the profit motive has no place in medicine.
Yarrow
@Roger Moore:
I’m one of them. I know maximizing profits is the goal, I just haven’t seen an example like this explained so clearly. It actually turned my stomach. The patient gets nothing but a gift card. Everything else for them is the same. Except their time is wasted because they go in for an appointment they don’t need.
AnderJ
@David Anderson: “In the New England Journal of Medicine is a short piece”
I am not a native English speaker, but shouldn’t this read “In the New England Journal of Medicine THERE is a short piece” or something to this effect?
Central Planning
@Roger Moore:
I agree.
However, aren’t there rules and regulations in place that specify a certain percentage of premiums has to go towards patient healthcare? It would seem to me (if that assertion is true), that by maximizing revenue to the insurer, premiums should go down in the following year (or increase is minimized) since the insurers are now more profitable. That idea also implies that there is no change in treatment of the patient (unless treatment becomes better for less).
Betsy
The rationale behind risk adjusted payments for Medicare Advantage plans was that the MA plan required more money to care for sicker patients than healthier patients. It kept the plans from signing up on healthy patients to MA plans. Presumably, it cost more to care for them, so the money was spent on health care. With any coding system, there is the opportunity to learn the system and to use the system to collect revenue that you might not have collected if you didn’t understand the system.
The strategy suggested for the providers (do your wellness visits early in the year, and get as many of your patients in as possible) seems to me to be a sound strategy from the perspective of the care of the patient and for risk scores.
If I am understanding Dave’s excellent analysis it is that MA plans (and presumably other entities that take risk) can increase revenue for existing conditions by more accurately coding those conditions, and that this isn’t necessarily related to increased costs.
HILFY
Who are the “highest value patients”?
Thanks for all the info you’ve given us abt the health insurance industry, and how premiums are set. Could you someday write about how the Obamaites were able to create the ACA so cleverly that Trump’s vandals still have not been able to kill it?
TomatoQueen
@Yarrow: Managing Type 2 Diabetes is always with the goal of getting certain lab results over time, for example, HA1c below 6, BMI appropriate, and to monitor and control complications, such as neuropathies and other organ complications (kidney function, gastrointestinal function). It’s routine for docs to require patients to come in (copay, kerching!) for labs (lab tests, kerching!) every 90 to 180 days depending on results, and not refill prescriptions until the new lab results are in. So the incentive to the patient is to control the symptoms and even reduce or eliminate them (usually through very strict compliance with medications and diet). Anyone who has lived with Type 2 D for any length of time learns to get with the program at least to the point of taking the medications, if not following all the other guidelines, because the results are noticeable right away–stabilize your blood glucose and you will feel better. The issue of cost, therefore, is also almost immediate: prescriptions, office and lab visits, and testing supplies, and without good ins., it gets expensive quickly. If you’re one of those people who doesn’t get good results, all this is increased, as you’ll be coming in for frequent labs, paying for other medicines, and possibly seeing other specialists (kerching kerching kerching). Complicated diabetes is its own volume business generator.
So if you don’t have a chronic condition, what is your incentive to visit your doc more often? Annual checkup and flu shot should be the standard, shouldn’t they? If you’re under 30, maybe. But nobody is without some history or issue, and over 30 it just becomes more likely that more than minimal care will be needed.
David Anderson
@HILFY: In this post, the highest value patient is the one where the currently coded diagnosis value has the biggest gap to the “maximum likely/defensible code-able value”
When I worked risk adjustment, we routinely had individuals who had a major organ transplant in the past where the status code (V4x.xx) was not entered on any claim in the past year. We would see the problem diagnosis codes and the transplant procedure codes from 2013 but not the status code in 2016. Putting a four or five character diagnosis code on a claim could mean a $10,000 a year revenue bump because the history/status of a major organ transplant (heart/lung/liver etc) factored into every decision the doc and patient made together. The status is therefore medically relevant but the code itself was seldom the reason for a visit. The “V” codes were routinely “high value” codes for our program.
Julia Grey
@AnderJ:
Not necessary. The piece presently exists in the NEJM.
3 pieces, “it” “is” “in.”
“There” could actually be considered redundant, since you’ve named the “where” of the “in” already (the NEJM).
opiejeanne
Hi David,
I don’t know if this is a good place to comment on this but I just got online and saw that some 30,000 people in Louisiana are being kicked out of nursing homes due to cuts in Medicaid.
David Anderson
@opiejeanne: That is a notice that the cuts may happen, not a notice of actual cuts (yet)
Betsy
@David Anderson: And of course, these are Z codes now.
Mary G
My doctor has now hired a nurse named Rosie who calls me at the beginning to every month to schmooze about my health. She doesn’t work for him; this appears to be a full-time job she does for any number of doctors under a new Medicare program. I am pretty self-directed, so if I have a problem I usually just go to any one of the number of specialists I have to visit once a year anyway that it comes under.
This often results in the primary provider not knowing a thing about it, and since I am usually only there once or twice a year, information doesn’t always get remembered and passed along. Also, there is serendipity. I had a nagging cough all last year and nothing we tried seemed to fix it. In the end, the pulmonologist decided I was allergic to my cats, which was wrong. At a standard followup to the rheumatologist, she instantly said it was because of GIRD. I have never had a symptom of heartburn, because due to the massive amounts of ibuprofen I took for RA, I was put on Prilosec the minute it was approved by the FDA decades ago and have taken it daily ever since. She told me to triple the dose, and voila, cough gone.
It would be nice if computerized medical records could be standardized and go with the patient, so every doctor you see could see all your records. Instead, they stay in their separate silos. I have three or four patient portals. I realize this would need a major revision to HIPPA, but it would be so helpful.