Working Paper Series 3/2011 Performance-Based Contracts for Outpatient Medical Services

These papers are produced by Cambridge Judge Business School, University of Cambridge. They are circulated for discussion purposes only. Their contents should be considered preliminary and are not to be quoted without the authors' permission. In recent years, the performance-based approach to contracting for medical services has been gaining popularity across different healthcare delivery systems, both in the US (under the name of " Pay-for-Performance " , or P4P), and abroad (" Payment-by-Results " , or PbR, in the UK). One common element of performance-based compensation is the inclusion of patient service access metrics, in addition to the quality of clinical outcomes, in the process of performance evaluation for a provider of healthcare services. For example, the implementation of the " Payment-by-Results " approach includes appointment scheduling targets designed to shorten patient waiting time, and adherence to these targets is monitored through a dedicated online outpatient appointment system, " Choose-and-Book ". The goal of our research is to build a unified performance-based contracting (PBC) framework that incorporates patient access-to-care requirements and that explicitly accounts for the complex outpatient care dynamics facilitated by the use of an online appointment scheduling system. In our model, a service provider needs to allocate his service capacity among three patient groups: urgent patients whose service cannot be postponed, and two groups of non-urgent patients, dedicated patients who insist on getting served by their first-choice provider and flexible patients who will choose another provider if the online appointment system shows no available appointments with their first-choice provider. The principal wants to minimize her cost (payments made to the provider offset by the waiting-time penalty) of achieving the expected waiting-time target. We model the appointment dynamics in the presence of a mixed-patient population as that of an M/D/1 queue and analyze several contracting approaches under adverse selection (asymmetric information) and moral hazard (private actions) settings. We study the first-best and the second-best solutions, as well as their specific contracting implementation schemes. Our results show that simple and popular schemes used in practice cannot implement the first-best solution and that the linear PBC cannot implement the second-best solution. In order to overcome these limitations, we propose a threshold-penalty PBC approach and show that it coordinates the system for an arbitrary patient mix and that it achieves the second-best performance for the setting where all patients are dedicated. As the US healthcare system is preparing to face a set …