Demand inducement as cheap talk

Abstract

The doctor-patient interaction is analysed in a game of cheap talk. Causes and consequences of imperfect agency are examined. One form of imperfect agency, supplier-induced demand, is a feature of neologism proof equilibria with some parameter values but not with others. The model is used to evaluate two tests that have been used to test for the existence of supplier-induced demand. The analysis suggests that the two tests, which compare the medical utilization of informed and uninformed consumers, are not valid.

The physician–patient relationship as a game of strategic information transmission

“…It follows that, for a communication equilibrium to exist, the E-signal must induce the patient to randomise between buying the E-treatment and buying no treatment at all. By the same reasoning, there cannot be any communication equilibrium in a variant of our game where the efficient action in the less severe state C is to buy no treatment at all; this explains why in Calcott's game [13], where the patient either does or does not need treatment, there cannot be any communication when there is a conflict of interest between physician and patient. Knowing which form the mixed equilibrium should take, we now show that such a mixed equilibrium can only exist in our game for certain physician and patient payoff levels.…”

“…To reflect this focus, we take the following four initial assumptions. First, we assume that the patient has already decided to consult the physician at the beginning of the game, and second we abstract from any information that physician (she) and patient (he) both observe (such as, in Calcott [13], commonly observable diagnostic results or symptoms). There is no loss in these initial assumptions, since our game may be considered as a subgame of a more extended game, which is played after the patient has consulted the physician, and after commonly observable events have occurred.…”

“…in questioning whether the first intuition underlying the SID hypothesis stated at the start of this paper applies for any treatment decision. In constructing a simple game (similar to the games of Calcott [13] and of Pitchik and Schotter [17]) where this intuition applies, we find that for rational patients, the SID hypothesis is only relevant for treatment decisions involving an expensive treatment that is equally effective in curing several patient states, but efficient in curing only some of these. This is because the patient is unable then to tell how this expensive treatment contributed to his utility, once he has received it.…”

“…Our first contribution lies in showing that there are literatures in the economics of information and in game theory which are directly relevant to the neoclassical approach to SID, and which have received little attention in this approach. An exception is Calcott [13], who realizes that neoclassical SID models resemble cheap-talk signalling games [14] with a conflict of interest. In such a signalling game, a sender possesses private information that is relevant to a receiver, in that a receiver who would know this information would let his actions depend on it.…”

Why prevent when it does not pay? Prevention when health services are credence goods

“…Our credence goods model is based on the theoretical framework of Dulleck and Kerschbamer (2006). The importance of the credence goods setting for the PID hypothesis has been recognized by Calcott (1999) and De Jaegher and Jegers (2001), who show that the PID hypothesis is rooted in the cheap-talk literature from game theory and in the literature on credence goods. The issue of physician reimbursement in the context of credence goods has been analyzed by De Jaegher (2010) and Emons (2013).…”