Over the weekend I was browsing through google scholar and came across an interesting concept in economics called a Credence Good Market; a place where asymmetric information exists between buyers and sellers that may give rise to inefficiencies such as under- and over treatment, or market breakdown.
Put simply, the doctor as an expert often has more information than the patient, and the patient is therefore not always able to make a reasoned assessment about the nature of treatment they should receive. This obviously varies by situation, as it is often possible for an intelligent patient to gather sufficient information to participate in the decision making process. This does not necessarily level the field between buyers and sellers.
The question that rolls through my mind is how will this change as value based reimbursement rolls through CMS and on into the health system? Will value based reimbursement payments change the nature of the credence market? Does a patient gain more knowledge and hence more ability to participate in a system where value based reimbursement is in use?
To take a close look at this we need to consider the four basic models of VBR; pay for performance, patient centered medical home, bundled payment, and shared savings/accountable care organizations.
Pay for performance links provider revenue to payer set standards that hopefully reflect best practice. Providers who perform poorly in term of quality outcome are paid less, and those who outperform stand to gain more. Is this still a credence market? Why yes, of course it is, and since the provider is incentivized to perform the service it is likely they will, resulting in over-treatment. It’s also possible the patient will be pushed through the system so quickly that they are under-treated.
In the patient centered medical home (PCMH) one or more physicians elects to take on a community of patients and covers all their services for physical and mental health. This is somewhat reminiscent of going to a car dealership that provides every service possible for your car. By participating you are assuming that the set of physicians you are working with are all of equal stature to those you could potentially find outside the PCMH. Worse still, the credence market effect is greatly increased for now the patient has multiple opinions to select from with no real basis on which to evaluate them. They may be both under- and over- treated in such a situation, and probably must rely on their primary care doctor’s opinion as a tie-breaker.
Bundled Payments are an interesting form of price management where a fixed rate is paid for an acute episode. The payer sets the rate and the physician sucks it up, or only takes patients who avoid that payer. The credence effect remains unchanged, and under treatment may be a real possibility as the physicians must stay in business as the bundled payments decline.
The accountable care model defines a population of patients, which under Obamacare must contain a minimum of 5,000 medicare beneficiaries, that will be collectively served by a provider organization. There’s essentially a bonus structure; good results equal additional pay, whereas poor results directly affect the ACOs bottom line. There is significant concern within the industry today that ACOs may lead to mass consolidation that will further reduce patient choice because corporations are allowed to participate.
From a credence perspective the ACO works something like the bundled model where a panel of experts guides the medical treatment provided to the patient. Each expert could over or under treat, and the hope is that it all comes out in the wash.
My assessment is that all known forms of VBR are credence markets designed to control cost and verifiability, but not to decrease liability, and since liability is perhaps the most important factor in controlling behavior in credence markets, it is unlikely that these new models will massively change the underlying economics of healthcare.
The real change needed in healthcare is far harder. First the patient and the physician must be placed on a more level field where the patient has access to other data sources in digestible form and can judge if over or under-treatment is occurring. Second, a good database of local pricing for the services being provided must be available. Third, there must be an accessible way to change between payer networks when required, without the need for a “major life event”. Fourth, and most importantly, medical liability must be statutorily controlled as it has the greatest possibility to drive inefficiencies from the market.
Free Competition and the Optimal Amount of Fraud, Michael R. Darby and Edi Karni, The Journal of Law and Economics 1973 16:1, 67-88
Dulleck, Uwe, Rudolf Kerschbamer, and Matthias Sutter. 2011. “The Economics of Credence Goods: An Experiment on the Role of Liability, Verifiability, Reputation, and Competition.”American Economic Review, 101(2): 526-55.