If a program charges an entire group of people to lose five pounds a piece, does the pre-exisisting case mixture matter?
The person who is one hundred pounds above their ideal who can start walking a mile or two a day and substituting water with lemon in lieu of full calorie pop will have a much easier time dropping the first five pounds than someone who is just slightly above their “ideal” weight and is working out four days a week. It really does not matter what incentives you tie to the first five pounds lost, the second person is much closer the the possibility frontier than the first person, so the marginal movement of five pounds is usually a lot harder.
The second person may try to game the system and do what every wrestler did before weigh-ins by running for an hour in a garbage bag and sweatsuit after fasting for the past thirty hours. They might hit the weight for the thirty seconds they need to hit the weight, but then they’ll rehydrate and gain four pounds in the next twenty minutes. The first person is so far from the possibility frontier, minor initial changes can produce reasonably rapid, sustained weight loss. Those gains will start to get a lot harder and a lot smaller after the first few weeks as the first person begins to get closer to the possibility frontier, so the performance gains go from two or three pounds a week to half a pound or a pound a week.
Why does this matter for a health policy writer?
If we substitute the costs for weights, and provider groups for people, we start to see the incentive structure for pay for performance Accountable Care Organizations (ACO).
McWilliams et al shows that the first year of Medicare ACOs delivered real cost savings (1.2%) while keeping quality either constant or improving quality. This is a real win. Usually, a policy that produces higher quality at the same cost is a good win, or the same quality at slightly lower costs is a win. Doing both at the same time is a significant public policy win. McWilliams et al also showed that the provider groups that are least efficient before they started their ACO year showed the biggest gains. Using the weight loss model, this is fairly intuitive, as organizations that have a lot of easy wins lying around are more likely to pick up a couple of them fairly quickly than organizations that already had integrated a lot of the easy wins and thus were running fairly efficiently.
This last piece, less efficient organizations pre-ACO getting the largest gains and thus the largest bonus payments, is important. Currently ACOs are scored based on their own history. Efficient organizations will have a lot harder time getting large gains and thus large bonuses than relatively inefficient provider groups. Is this desirable? The incentive structure would be for efficient organizations to either drop out of the ACO system, or never join the ACO system while less efficient organizations that know they have a lot of fat to cut, join ACOs or continue their membership because they’ll get good bonuses?
I don’t know. The argument for maintaining this set of incentives is that it would pick up some of the easiest fruit on the lowest branches by moving the least efficient provider frontier significantly inward over the next couple of years. It would reduce the cost variance by trimming the expensive outliers. They may still be less efficient and most costly than the best in class practictioners, but the gap would be a lot smaller.
The argument against this incentive structure is two fold. First, it is an equity argument that good practicioners should not be penalized for being good practictioners. Secondly, the goal of the ACOs is to transform how we pay for healthcare towards quality and outcomes instead of by the widget. The organizations that were already running fairly efficient, high quality practices are more likely to be the ones that can continue to be leaders of the pack as long as the financial incentive is to maintain high quality and low cost care. They’ll be the best in class and the source of the next round of lessons learned, so we should pay for the learning by doing and learning by obersation that the high quality ACOs are providing to the public.
The second post by Austin Frackt highlights some of the alternative payments methodologies and their incentive structures:
“[A] large organization may have the option of becoming an ACO or developing an MA plan. Such an organization, whose spending exceeds the local MA benchmark based on local FFS spending, would have an incentive to become an ACO. The more efficient organizations would have an incentive to create MA plans.” I had not considered the ACO vs MA tradeoff before. This is particularly interesting, and complicated.
Going to a Medicare Advantage (MA plan) for an efficient MCO would create the same basic incentive structure of cost controlling via improving quality and prevention intsead of widget paying. Speaking purely as a plumber, there is an insurance company admin gap where doing nothing makes sense as the ACO won’t pay and the cost of starting a Medicare Advantage plan, especially for a provider group that is not already part of an integrated payer provider system, is higher than the capturable benefits. The other option would be to change how incentive payments are calculated:
Basing payments on cost performances on peer groups has worked well in Medicaid payments to psychiatric hospitalsand psychiatric units in New Hampshire, accommodating systematic differences in casemix while maintaining incentives for cost-effective car
Under this scheme, ACOs would be compared against who they serve instead of their own past history. If a very efficient practice is delivering high quality care to a particular case mix that is 10% under national averages or national benchmarks but the ACO has not significnatly cut their year over year costs, the current system will not give them a bonus. The proposed system would give them a bonus. Inefficient ACOs that treat the same type of population for 4% more than national benchmarks but showing a 3% drop year over year currently get a bonus but under the new system, they would not get money for being less efficient than national average. This proposal would be an analogue to the second Silver pricing methodology on the Exchange where there should be constant jockeying towards a well defined but not always known baseline.
This is interesting stuff for if we can get the incentives right on ACOs to take them to national scale as the default payment mechanism for care, then the cost curve will be bent without sacrificing quality of care.