PLoS Medicine
(Accessed 2 November 2013)
http://www.plosmedicine.org/

Guidelines and Guidance
Complexity in Mathematical Models of Public Health Policies: A Guide for Consumers of Models
Sanjay Basu, Jason Andrews
http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.1001540

Summary Points
:: Mathematical models are increasingly used to inform public health policy, but a major dilemma faced by readers is how to evaluate the quality of models.
:: All models require simplifying assumptions, and there are tradeoffs between creating models that are more “realistic” versus those that are grounded in more well-characterized data on the behavior of disease processes.
:: Complex models are not necessarily more accurate or reliable simply because they can more easily fit real-world data than simpler models; complex models can suffer parameter estimation problems that can be difficult to detect and often cannot be fixed by “calibrating” models to external data. Conversely, complexity can be important to include when uncertain factors are central to a disease process or research question.
:: In many cases, alternative model structures can appear reasonable for the same policy problem. Sensitivity analyses not only around parameter values but also using alternative model structures can help determine which factors are particularly important to disease outcomes of interest. Explicit methods are now available to transparently and objectively compare different model structures.