RECOMMENDATIONS FOR MODELING DISASTER RESPONSES IN PUBLIC HEALTH AND MEDICINE: A POSITION PAPER OF THE SOCIETY FOR MEDICAL DECISION MAKING

Purpose: Increasingly, modeling is being used to evaluate health sector responses to disasters. This paper provides an overview of exemplary and representative models that address the public health and medical disaster responses, and proposes guidelines for the design and reporting of such models. This is the first in a series of position papers by SMDM providing standards for modeling high-impact topics in health policy. Methods: We performed targeted searches of disaster response models addressing public health or health care delivery (published 1966-2008). For each model, we abstracted information about the type of disaster and response decisions evaluated, model methodology, and outcomes of interest. Through consensus, we created a set of recommendations for best practices for developing and reporting such models. We will solicit feedback on these recommendations at the SMDM annual meeting and during a subsequent 90-day comment period when the draft will be available on the SMDM website. Results: We identified 567 potentially relevant articles, reviewed 120 in detail, and found 48 disaster response models meeting our inclusion criteria. The included models varied in terms of the natural and manmade disasters considered, methodology utilized (e.g., simulation, mathematical/analytical), decision makers included (e.g., first responders, planners), decisions modeled (e.g., inventory/stockpiling, triage, dispensing, treatment), outcomes (e.g., mortality, cost), and quality both in terms of model design and reporting. We propose a set of criteria for model construction and reporting inspired by the most exemplary models. For example, good models should be designed to reasonably simulate the disaster scenario and real-world responses, and report detailed justifications of parameter/distributional choices and model validation. Models intended as planning tools should be customizable over a variety of regions/populations, often by end-users with little analytic training. Given the highly uncertain nature of many disasters, scenario planning is critical and should include a spectrum of realistic human behaviors. We advocate for an iterative approach to model development and reporting that includes early involvement from relevant decision makers and stakeholders and attempt to make the model available to end-users. Conclusions: Quantitative models can provide valuable input into the process of planning effective health sector responses to disasters. The guidelines we propose should increase the applicability and interpretability of models designed to improve strategic, tactical, and operational aspects of preparedness planning and response.