JCJ asked an excellent question:
Do you think in the future insurance companies will offer a single payment for services? My question has to do with my specialty – radiation oncology. If a woman undergoes lumpectomy for breast cancer and meets criteria for hypofractionated whole breast radiation (age > 50, no previous chemotherapy, separation (sorry, technical term) < 25 cm) yet still receives a more protracted course of treatments might insurance companies be justified in paying for the only for the shorter course and let the doctor explain the difference? Similarly a man with favorable risk prostate cancer could undergo a course of treatment with radiation receiving 28 doses while at another hospital seven miles away might receive 40 or 43 treatments….
Death panels!
A really good jumping off point for discussing payment models?
You decide!
There are three major types of payment models insurance companies (including Medicare) use to pay providers. Each have their advantages and disadvantages. There are two additional models that are in pilot testing right now that should be discussed as well.
Right now, the typical insurance company pays providers on a “fee for service” model. Under this model, volume of proceudres is what gets paid. In JCJ’s scenario, the radiation oncologists gets paid for either 28 doses or 40 doses. And he’ll get the same rate for each dose. This model is administratively simple as the claims systems just needs to see procedure code X and pays out rate Y. However, this model encourages both overtreatment (40 doses instead of 28) and treatment using more expensive options. For instance, there is significant research that shows physical therapy offers as good or better pain relief than surgery for certain types of back pain. However, PT pays out at a much lower rate (and to a different provider) than surgery pays to the surgeon making the recommendation for surgery. Effectiveness of intervention is roughly the same, but the cost efficiency of the surgical intervention is significantly lower.
The next model has been around for a while in a variety of guises and names. It is a capitation model where a provider group is given a fixed sum of money each year/month to take care of a roster of patients. This is the old HMO model, this is a part of the Accountable Care Organization (ACO) model, this is a “global budget” model. The idea behind this model is to get providers to be much more cost sensitive because their profit margin is dependent on not spending money.
In JCJ’s example, a capitated provider group would have a strong incentive to run a standard 28 dose protocal and then follow up with testing to identify which patients the protocal was successful and stop any additional dosing for them while continuing dosing/treatment for the patients who were not responding as well to the initial 28 doses. In a back pain example, the incentive would be for physical therapy (an effective and comparatively cheap intervention) to be tried first before surgery as the money is coming out of the provider group’s pocket in either scenario.
The upside of a capitated payment schema is the incentives are aligned to try the cheap but effective strategies of intervention at the first level. Groups that truly understand both the financial impacts of capitation and their patient population will try to intervene non-medically before they have to intervene medically as this tends to be much cheaper. Ezra Klein at Wonk Blog pointed out a good example of this in April, 2013:
Health Quality Partners is all about going there. The program enrolls Medicare patients with at least one chronic illness and one hospitalization in the past year. It then sends a trained nurse to see them every week, or every month, whether they’re healthy or sick. It sounds simple and, in a way, it is. But simple things can be revolutionary…
Health Quality Partners’ results have been extraordinary. According to an independent analysis by the consulting firm Mathematica, HQP has reduced hospitalizations by 33 percent and cut Medicare costs by 22 percent.
Using an in-person visit instead of the MBA approved high contact/low effectiveness approach of call center nurses dramatically increased both social bonds (those are pretty powerful disease fighting tools), medication compliance, and preventative risk maitenance. If HQP was a capitated group getting the average Medicare spend per patient, their profit margins would be obscene even with their higher labor cost. Furthermore, their patients would be in better health and spending way less time and risk in the hospital. Double win.
There are two major downsides of a capitated payment system. It is administratively complex, The complexity comes from risk adjustment payments as not all medical groups or medical homes will have identical risk pools. One group might have 90% 30 somethings and the other group will by 60% 50 or 60 somethings. The younger group will get less money per patient and send some money to the group covering the older adults. Additionally, there is significant back-end reinsurance. For instance, if there was a single medical group in a capitated model covering 30 people wounded in the Boston Marathon bombing, that is an extremely unusual event, and that group will get back end payments to make it whole.
The bigger “flaw” in a capitated payment system is that it is guaranteed to produced sobbing news conferences as the doctors denied an unproven and/or extraordinarily expensive course of therapy to a patient who died but whose family saw an article quoting Jenny McCarthy about this treatment for that particular disease. Any system that discourages overtreatment or treatments that have low effectiveness will hit outlier cases and in a population of 315,000,000 at least a few of those outliers will be well-connected and telegenic.
The last major model is bundled payments. It is a hybrid between fee for service and capitation. It is a fixed sum payment for a given diagnosis adjusted for a couple of variables. For instance and speaking hypothetically, a bundled payment for favorable risk prostate cancer might be large enough to cover on average 33 radiation doses, regular follow-ups and screenings for those 33 doses, and all the other things that occur within that course of treatment. On any given patient, the providers may profit or lose money, but the goal is to provide a bundle of payments that covers average treatments. Like capitation, there is a similar (but softer) incentive to minimize overtreatment. Like fee for service, the administration of these payments is “reasonably” straight forward. During the course of treatment if the diagnosis changes or the initial course of action is not working, the payment bundle can also change.
The fourth, and much smaller payment model that could be absolutely interesting directly addresses the central problem with US health care spending — our prices are way too high for a given service when compared to OECD norms. This model is called “reference pricing” where an insurance company generates a list of fairly uniform, non time critical services and the range of prices that providers charge. Most blood work, most MRI and other advanced imagining, most elective surgery would fall under standard, non-time critical services. From this list, the insurance company would say for a MRI of the left knee has contracted rates of $375 to $1700 to providers. The insurance company using reference pricing would then find a price where there is a good access to quality providers and declare that they’ll pay up to $900 for a left knee MRI (a payment level sufficient for to cover 100% of rates for 80% of the providers) and anything over $900 is member responsibility. The goal is to drive down unit cost of services by making the very high price providers lower their prices.
The final payment method for health care services by health insurance companies is to trade chickens for services — my company is a major stockholder in both Tyson and Perdue just to maintain our access to a strategic chicken reserve.