Transparent Models

The July/August 2011 issue of Medical Decision Making is devoted to simulation modeling.  The models are mostly about diseases, including colorectal cancer. 

David F. Ransohoff, Michael Pignone, and Louise B. Russell have an article about using models to make policy. They stress the need for transparency in the following process:

1. Evidence is gathered systematically, evaluated for quality, and selected on the basis of strength for use in quantitative analysis.
2. Analysis of the evidence is conducted quantitatively, often through modeling, to show the likely outcomes of different intervention strategies.
3. Using the results of steps 1 and 2, guidelines groups then decide among different strategies, based on the decision makers’ values and decision thresholds.

They argue that neutral, experienced, professional model-builders are increasingly doing steps 1 and 2 (but not step 3), and modeling experts are also collaborating to understand the differences between their models, which increases transparency and helps guidelines groups explain their recommendations.

The complete citation is David F. Ransohoff, Michael Pignone, and Louise B. Russell, Using Models to Make Policy: An Inflection Point? Med Decis Making July/August 2011 31: 527-529, doi:10.1177/0272989X11412079

Two papers in the issue are freely available:

Clarifying Differences in Natural History between Models of Screening: The Case of Colorectal Cancer” describes a measure for comparing models of screening.

Accounting for Methodological, Structural, and Parameter Uncertainty in Decision-Analytic Models: A Practical Guide” describes the process of representing scientific uncertainty in a model.