Applying Game Theory Models to Inpatient Medicine: Opportunities to Improve Care

Abstract

Inpatient hospital costs have been increasing exponentially in the United States. Part of this increase is attributed to over and undertreatment, leading to higher costs and potential patient harm. Research improving clinician-patient interactions can help minimize and optimize the costs. Game theory has the ability improve clinician-patient interaction by modeling outcomes. Using variations of game theory, the bad doctor bad patient stigma can reframed to incentives. We believe the use of different models (prisoner dilemma, centipede game, assurance game, and chicken game) can outline the challenges faced during common inpatient scenarios, including end of life conversations and aggressive procedures. Applying game theory to multiple inpatient scenarios may also assist with analysis during morbidity and mortality conferences and quality improvement projects.

Keywords: Decision-making; Game theory; Inpatient medicine; Patient-doctor interactions; Prisoner’s dilemma.

Fig. 1

Prisoner’s Dilemma. Players choose a strategy to trust or betray. Numbers in each cell represent individual outcomes. Although the best collective strategy is to trust, the most rational individual strategy is to betray, thus yielding the worst collective outcome.

Fig. 2

Centipede model. Each player takes turn to decide to cooperate, thus continuing the game, or defect, thus terminating the game. The best collective outcome is to reach the last node; however, the incentive to defect is greater than to proceed to the next node. Hence, mutual cooperation involves taking mutual risks. Dashed lines are clinician’s move, while solid lines are patient’s move.