As is often the case, work projects steer me away from my primary research interests and bring me into contact with new and often surprising problems.  While on the surface, these problems may share little with the more familiar terrain of international relations, national security and intelligence analysis, often times a surprisingly deep conceptual connection can be found (or at least constructed).  As it happens, our Advanced Analytics practice at LMI is working on extending our reach outside of national security into health care, which makes us typical for many firms that see the long-term budgetary trends and opportunities to transition technology, analytic skills, and other offerings into new markets.

In my case, I’ve been learning more about the emergence of Accountable Care Organizations (ACOs) that are one of the major changes introduced by the “Obamacare” legislation.  ACOs have existed for quite some time as pilot projects and regional efforts, but are now going to be the centerpiece of the healthcare system—at least in theory.  Having just started my research, I’m trying to avoid reaching any conclusions on the wisdom of these ventures and am instead trying to understand how they are supposed to work, and their respective strengths and weaknesses given current and potential alternatives, particularly over the long-term.  There appears to be a good track record for these organizations, but nothing seems to exist at the scale that the legislation has encouraged.

The basic premise of the ACO is that a group of medical practitioners, whether a hospital, collection of doctors, nursing specialists, physical therapists, etc. take on responsibility for the health of a group of patients.  The ACO is then compensated based on two factors—first, they must meet certain performance metrics that include quality of care, information security, and others (I think a total of 60+ metrics exist that cover 5 or so topics); second, if they provide quality care for less than an estimated cost (based on the composition of their patient population) they can receive a portion of the savings as a bonus.  Thus, the combination of performance/quality metrics and incentive to share in overall savings are intended to encourage innovation and preventive measures that emphasize reducing costs at the population level.

It appears that several pilot efforts to do this have performed quite well and have saved Medicare and private insurers that have used ACOs significant amounts of money given what they had spent historically on a given population or concurrently on similar groups being treated under the traditional fee-for-service models.  The interesting part, at least to me, however, is that there exists a long-run counterfactual problem with this model that is not unfamiliar to international relations scholars that have struggled to measure the value of deterrence, both nuclear and conventional.

One highly optimistic op-ed posed the interesting question by noting that by 2020, the ACO model will outperform conventional private insurers, thus eliminating the need for private insurance as we understand it today and replace it with an ACO-centric model that does a better job of managing risk.  If this is the case, the health care ecosystem will be transformed and interesting problem will emerge—the population of patients being served by traditional fee-for-service medical programs being reimbursed by private insurers will be gone, leaving behind no group against which the savings provided by ACOs can be compared.  While this is far off, and may perhaps never occur, it strikes me as interesting that a model predicated on rewarding medical providers for saving money could eventually lose any sense of what is being saved once everyone is treated under the same rules and incentives.

For example, if a patient currently costs an insurer $10,000 per year in treatment, an ACO may take them on for $7,500 and then keep a fraction of the $2,500 in savings as profits assuming all the associated quality of service metrics are met.  The problem comes there is no basis for knowing that a patient would cost $10,000 rather than $7,500 to care for under a system of providing health care that ceases to exist.  How then will it be determined what amount of additional profit should be available to the ACOs?  Will there be any memory for the “old days” of expenses against which the bonuses or profitability of ACOs will be determined, or will new metrics based on the counterfactuals of alternative, less efficient care practices be introduced, etc.?  In short, the information that serves as the basis for determining the size of the savings ACOs will share is itself likely to be destroyed as ACOs become increasingly common, thus making the system increasingly dependent on counterfactuals of worst-case estimates of costs.

Returning to international relations, this problem reminded me of deterrence in two ways.  First, the long-term prospects of ACOs and their savings rests on imagining alternative health care models that provide a basis for determining that the ACOs have provided savings and are therefore preventing the occurrence of something worse (higher treatment costs and lower quality of service based on the traditional fee-for-service model).  This is like Philip Bobbitt’s arguments regarding deterrence being possible because the party being deterred can imagine a world worse than the one they are in.

Second, more concretely, the US defense budget and the extensive military assistance provided to Israel and Egypt show how people’s expectations change overtime and therefore cease to value the contributions of actions that, at least in theory, are in place to keep events from occurring.  Thus, the fact that the US has largely subsidized the national defense of the industrialized world for several decades in order to prevent them from rearming and diverting resources away from developing competitive capitalist economies embedded in a system of global trade rather than national armaments that would have been a drag on their economies and likely elevated regional tensions is forgotten, and the image of US as wildly outspending its rivals and allies and being overly militaristic is remembered.  Likewise, the tensions that resulted from the 1973 Arab-Israeli War, which nearly led to a nuclear war between the US and Soviet Union, led to the belief that giving Israel and Egypt lavish security assistance was cheaper than rebuilding the Middle East after another war is now a long forgotten memory and in its place is the contemporary image of two countries receiving vast sums of military assistance that coddled and even encouraged regimes whose democratic commitments were (and remain) suspect.  In both of these cases, the logic of the counterfactual that motived initial policy-decisions was eventually forgotten, and the perverse consequences of well-intended actions left observers suspicious, while the notion that something far worse was being kept in abeyance simply vanished from the debate.

While I don’t wish to equate ACOs with nuclear deterrence (although the rhetoric around the health care legislation makes it seem as if we are dealing with the future of human life as we know it), I do see something in common regarding the structure of the problem and the long-term of use of counterfactuals as metrics for evaluating policy and governing decisions, whether we are thinking about savings accrued by keeping sick people healthy and out of hospitals or preventing nuclear wars.