SDS-CSM - Statement of Purpose

The purpose of Section of Decision Sciences and Clinical Systems Modeling is to enhance the analytic research capabilities of the University of Pittsburgh in the following core areas:

Decision Analysis: The use of decision analysis as a technique for combining data from multiple sources as well as a methodology for conducting cost-effectiveness analyses is increasing. Formal decision algorithms appear in protocols, pathways and computerized decision support; the use of decision models has increased substantially in current scientific literature. We will continue development of our work in large scale Markov Processes (to model events over time) and in linking decision models to empiric datasets to develop quantitative models of natural history.

Cost Effectiveness Models: Although there is significant economic expertise throughout the University of Pittsburgh, the Section has significant expertise in building models for conducting cost-effectiveness analyses. We serve both as investigators and collaborators to other studies. We will continue our work to develop integrated models of resource use and effectiveness that are relatively easy to understand and to expand our expertise in the creation and analysis of cost models constructed from administrative databases.

Utility Assessments and Quality of Life: One of the hallmarks of patient-directed research is the ability to directly assess patient’s own preferences for outcomes and treatments and include those preferences in therapeutic decision. The section has expertise and ongoing research activities in standard methods of utility assessment, the theoretical foundation of utility measurement and quality of life, and the use of preferences to develop quality-adjusted outcomes.

Mathematical Simulation: The Section is actively involved in the applications of multiple types of simulation techniques to problems in health and medicine. We have expertise in discrete event simulation, which provides tools that model specific characteristics of real systems such as distance, resource constraints, queues, and bottlenecks. Designed to analyze problems of optimal throughput under various constraints, discrete event simulation is an excellent tool for examining many health policy and resource use questions. For example, we feel it is the only methodology able to inform resource allocation decisions (such as allocation of donor organs), where resource competition and queues are integral to the allocation process. We have used integer programming as a tool to evaluate the optimal organization of Organ Procurement Organizations into Regions, and have used Markov Decision Processes to analytically evaluate the optimal time to accept a living donor for liver transplantation.

Activities

The Section of Decision Sciences and Clinical Systems Modeling currently engages in three primary types of activities: methodological research, decision sciences training, and collaboration with other research projects to provide cost-effectiveness and decision modeling expertise.

Methodological Research: The methodological research agenda of the section focuses on expanding and extending the methods described above, to improve the application of mathematical modeling tools to biological problems. The Section collaborates with the Center for Biomedical Informatics (Bayesean inference and networks) and the Department of Industrial Engineering (simulation modeling and operations research) and the Department of Biostatistics (developing robust survival estimates for simulation models) . Faculty will advance current knowledge and methodological capabilities in terms of the operational characteristics of these models and their application to practical policy questions.

Decision Sciences Training: Section faculty have developed and teach several courses with the Clinical Research Training Program related to decision sciences. Independent study with members of the section is available as well.

CLRES 2120: Cost effectiveness analysis

CLRES 2121: Clinical Decision Analysis

CLRES 2122: Advanced Methods for Cost Effectiveness and Decision Analysis

In addition, section faculty teach national and international course in decision sciences for the Society for Medical Decision Making, The International Society for Pharmacoeconomics and outcomes research and other organizations.

Modeling expertise for Health Services Research: With appropriate faculty resources, the decision and mathematical modeling group will provide cooperative and/or consultative expertise to research groups throughout the university to build cost-effectiveness and decision analytic models in conjunction with their research.