In medical systems that reward doctors for patient outcomes, how to avoid cherry picking patients

In the US, one of the reasons our system is so expensive is that we have a fee for service system, so the more procedures that are prescribed, the more money that can be made. An MRI makes more money than no scan, surgery makes more money than physical therapy, etc.

But in alternate systems like systems that reward doctors for making their patients healthy, how do they avoid doctors cherry picking the healthiest patients and the patients with the easiest to treat diseases, and shunning all the patients with complex diseases?

If a doctor is paid based on making patients healthy, it seems like that doctor would be incentivized to avoid patients with complex medical problems that don’t respond well to treatment.

How is this incentive avoided in medical systems with this kind of reward structure?

By using a patient-selects-doctor system, or Doctor is paid for a defined group: (a medical union or a common employer)

I’d imagine that another way to do it would be to pay the doctor a more or less flat rate per patient engineered to cover costs overall for both the folks who show up once a year for a physical and are otherwise in good health, and for the physical wrecks who probably visit the doctor more than a dozen times a year.

One thing to consider is that doctors don’t get kickbacks AFAIK, for sending patients out for procedures. If you go to the doctor and say your crotch hurts, and they send you for an ultrasound, they don’t get any money for that, other than the money for the office visit and the followup to interpret the findings. Some things are kind of gray, in that blood work is often done at the office, and then sent to the lab, and you may be charged by either one, depending on what it is, and how things are set up.

Genuine question, does such a medical system exist?

It strikes me that doctors shouldn’t be rewarded either way. Pay them a salary, then they have no incentive to either over or under service a patient.

I believe this is how it works in the NHS, at least for GPs (who are self or privately employed).

Pretty much. GPs are paid a capitation fee per registered patient, and most of us just take it as read that we register with a local GP as soon as we move into an area.

Hospital doctors are salaried (but the most senior are on contracts that allow a proportion of their time on private practice, or research or other outside professional interests).

Hospitals and other specialist services, both NHS and private providers compete for NHS contracts, but there is a fixed national tariff for each form of treatment, so they’re supposed to be competing on quality of care. The payment mechanism is a fixed total budget for the year.

So no-one stands to gain or lose by treating any one patient rather than another, and they can prioritise according to clinical need

the currency for kickbacks is in the form of referrals. you refer to me, i’ll refer to you. a nice xmas present may also suffice.

What are you talking about? If a doctor say… refers a patient to an ultrasound imaging clinic, there’s not any quid pro quo there. It’s a one-way thing, and the patient ends up paying the ultrasound clinic for the procedure, not the doctor. In general, if they’re recommending a particular place, it’s because it’s part of their larger umbrella practice/group, and they’re integrated with them in terms of getting results. For example, my doctor is part of Baylor Scott & White, so if he has to refer me out, he does it to affiliated facilities because he gets the results right away because they’re all integrated with each other. From his standpoint, it lets him do his job easier.

But if your doctor refers you to a specialist - for instance, an internist sends you to a cardiologist, then that cardiologist might be more likely to refer a different patient back to your internist. That’s the quid pro quo - that the doctors are generating new patient referrals for each other.

It might not happen often, but if I’m unhappy with my internist, but really happy with my cardiologist, I’m more likely to ask my cardio about a different internist. Especially if I’m already at my cardio

That cancer treatment center that advertises on TV a lot gets called out every so often for stuff like this because they tent to pick easier to treat patients to boost their success rates so they can brag how great they are

Its one reason they say “call us yourself” because the oncologists were getting annoyed with how they treated patients who were referred to them

The physician that performed my brother-in-law’s heart transplant surgery was barred from doing any more transplants because he was gaming the system by making his patients seem sicker than they were to move them up the transplant list. He did have a good record; but it’s hard to tell how much of it was due to him getting healthier recipients than the average, assuming the average doctor was playing by the rules.

On the other hand, my brother-in-law is still alive after close to 20 years, and the over/under for lifespan after a heart transplant is about half that, so there’s that. Maybe waiting until the recipient is at death’s door is not the best strategy, but practicing triage and letting the sickest die so the healthier ones can get the scarce organs seems callous.

I don’t know… it’s entirely possible that the best outcomes may come from a transplant well before things get to the pass where people are on death’s door.

While I agree that we shouldn’t necessarily let the sickest die, it’s also irresponsible to give organs to people who aren’t liable to stay alive long anyway, transplant or no.

Is there a scoring system to determine the suitability for transplantation in terms of health, social support, etc…?

There sure is. I pulled this off Tampa General Hospital’s website:

Patients must be younger than 69 years of age. 
Patients must have a diagnosis of end-stage heart disease, such as advanced cardiomyopathy, pulmonary hypertension, or significant heart failure. 
Patients must have a prognosis that indicates significant risk of mortality within one year if a transplant is not performed. 
Patients may not have an active infection, a cancer diagnosis, organ impairment other than impairment in the heart, or a significant arterial disease that affects circulation to the brain or legs, which could reduce the chances of long-term survival. 
Patients must be psychologically stable, must understand the risks and requirements associated with the heart transplant process, and must be actively committed to post-operative rehabilitation. 
Patients must not be considered medically obese. 
Patients must abstain from alcohol and tobacco for at least three months prior to transplant listing and refrain from using throughout the entire heart transplant process. 
Patients must have adequate financial and social support to assist with post-transplant medications and care.

So if there’s not a significant risk of dying over the next 12 months no heart for you.

I hadn’t heard that but it makes sense.

What also makes sense is for cancer centers to pass off every lump they see as cancer, so they can claim a higher survival rate. If you pass off a bunch of benign lumps and growths that’d never become dangerous cancers, then treat them then you can say ‘our 5 years survival rate is 95%’ or something like that since those people would’ve survived anyway.

That’s more of a list of eligibility requirements; what I’m talking about is more of a scoring system where you’d have something like this:

[ul]
[li]Age less than 40 = 4 points[/li][li]Age greater than or equal to 40, but less than 62: 3 points[/li][li]Age greater than or equal to 62, but less than 69: 1 point[/li][li]Age greater than or equal to 69: 0 points[/li][/ul]

And extend that for all the questions, based on evidence and data. That way, if you had two people who met the criteria, it would prioritize them based on the best likely outcome.

Many doctors in the US are paid using the same system. At one point I did tech support for medical billing software so I became very familiar with the way doctors were being paid at clinics and hospitals all over the US.

This is exactly what happened to me, regarding spinal stenosis. After getting little help from several doctors, I went all the way to the top, to the head of the spinal surgery department. After reviewing my case, he informed me that he couldn’t operate on me because the surgery would be too risky. What he meant wasn’t that it would be risky for ME, but risky for HIM. If I died on the table, it would ruin his stellar performance record. The only patients that he’d operate on were patients who had absolutely no medical issues other than spinal stenosis.

AFAIK Canadian provinces don’t have incentive system as described, but…

Patients must go to a GP who refers them to a specialist. I suppose it’s conceivable that one specialist could refer a patient to another specialist for a different issue, but generally it’s the GP’s that do so. If a patient has multiple problems, odds are the GP has been apprised of both and has already started the referrals to both specialists.

Plus, funny thing about fully covered medical care - when it’s available to everyone, everyone uses it when they need it. As a result, the problem in Canada is lineups, not bankruptcies. A specialist has a long waiting list of people waiting to see them, they don’t need to quid anyone’s quo to drum up more patients.

Here’s the current way things are set up by CMS for payments to doctors working in nursing homes and rehab settings. The payment for each visit is determined essentially* by the level of complexity of the medical decision making at each encounter, from level 1 to 4, the higher numbers being reimbursed at a higher level. There is a long list of things that make any particular case less or more complicated. If someone is receiving IV antibiotics whose level has to be tracked by blood work and you are doing that during a particular visit, that would be level 4. If instead you are adjusting someones blood pressure pills because their BP is high, that would probably be a level 3 visit. Visiting someone who has high BP but whose BP is controlled would probably be a level 2 visit, and so on. In addition, each visit has to be “medically necessary” so a sicker patient will legitimately need more frequent visits than someone who is less ill, and the higher number of visits will of course lead to a higher level of reimbursement. Needless to say, there are ways some doctors can game the system, but from what I’ve seen the doctors who do so eventually get caught and have to pay back a whole lot of money, even if it takes years before they are caught.

Another part of the system is the larger level of hospitals and facilities. Again CMS looks at several things, but two of the biggest things they look at are length of stay (shorter is better) and rates of readmission to the hospital. In theory a nursing home with good scores will receive a percentage adjustment in their reimbursement. There is also supposed to be a system where the better nursing homes receive more referrals from the hospital, although in practice this doesn’t happen, and hospital case managers end up sending patients to whichever nursing homes their friends work at.

  • There are other factors, like how thorough a history and exam were done, but for the most part the complexity of the case ends up being the deciding factor.

I would need to see a better assertion for your statement.

The same system that promotes that will also deny it. For instance, my doctor has recommended a test, and my insurer, Aetna, has refused to cover it.

Aetna is paying the doctor but also has a vested interest in not doing everything he says.

This scoring definitely exists, but effects on score at the variable level are largely hidden / proprietary such that even doctors only have a vague idea how much a given factor will move a patient up or down the list, and they often just empirically plug things in to see what happens to a given score to get a real idea.

UNOS is the main body that dictates these scores, and they include factors like:
[ul]
[li]Age[/li][li]Ability of the patient to recover[/li][li]ABO (though very young recipients are often considered for ABO-incompatible listing.)[/li][li]Distance[/li][li]Height and weight[/li][li]Life support status[/li][li]Listing status[/li][li]Time on the waiting list[/li][/ul]

Empirically, despite all the folderol about age and patient ability to recover, the scoring tends to favor pretty sick older people with a lot of comorbidities (those folk will have the highest scores), ideally on oxygen or assisted by devices.

One thing they never take into account is QALY’s, or quality of life years. If you have a young and, aside from the one diseased organ, otherwise healthy patient patient whose transplant could give them another 20 QALY’s, tough nut, because Grandma is sicker, has diabetes and other organ problems, and is kept alive by various machines, so she’s top of the list and will get 3 QALY’s out of those new lungs. Come back when you’re a lot older/sicker!