A Causal Model of COPD

Chapter 1 suggested that good causal modeling is at the heart of good risk analysis. Some of the most important and enjoyable challenges for risk analysis practitioners are understanding, validating, modeling, and documenting causal mechanisms in sufficient detail to predict how changes in controllable inputs, such as exposures, will affect changes in outputs of concern, such as risk of disease and mortality. Typically, a risk analyst needs to learn only part of what scientists, who study causal mechanisms in detail, know: the part providing valid causal relations between changes in inputs and outputs. The deeper questions of how and why these relations function are not needed for purposes or accurate risk modeling, although they may be of great (or even primary) interest to the scientist.