Title: Factored Stochastic Tree Modeling forArthritic Joint Replacement Decisions
Authors: Gordon B. Hazen, Department of IEMS, Northwestern University,  James M. Pellisier, Merck Research Laboratories, Blue Bell PA, and Rowland W. Chang, M.D., Department of Preventive Medicine, Northwestern University
Date: June 1998
Status: working paper

Total hip arthroplasty (THA) and total knee arthroplasty (TKA) have proven to be clinically reliable and durable procedures for the surgical treatment of severe osteoarthritis of the hip and knee. Although THA and TKA seem justified in terms of clinical success, they are particularly vulnerable to scrutiny in economic terms. We present here decision-analytic models of the short- and long-term consequences of knee and hip replacement. These models are constructed using factored stochastic trees, which combine features from decision trees and continuous-time Markov chains. A novel feature of the presentation is the use of influence diagrams to portray the relationships among the components (factors) of the model. The authors' software for factored stochastic tree construction is also presented. Although knee and hip replacement do not extend life, when quality of life is included, our results show that these procedures are very cost-effective, comparing favorably with such well-accepted practices as cardiac bypass and renal dialysis.

Click here to access a pdf version of this paper.

Back to the Decision Analysis Working Paper Index