Modelling uncertainty in environmental health impact assessment

Abstract

Quantifying uncertainty in environmental health impact assessment models is important, particularly if the models are to be used for decision support. This thesis develops a new non-probabilistic framework to quantify uncertainty in environmental health impact assessment models. The framework takes into account two different perspectives of uncertainty: conceptual and analytical in terms of where uncertainty occurs in the model. The first perspective is concerned with uncertainty in the framing assumptions of health impact assessment, whereas the second perspective is concerned with uncertainty in the parameters of a model. The construction of the framework was achieved by focusing on five specific objectives: (i) to describe the complexity of how uncertainty arises in environmental health impact assessment and classify the uncertainty to be amenable for quantitative modelling;(ii) to critically appraise the strengths and limitations of current methods used to handle the uncertainty in environmental health impact assessment; (iii) to develop a novel quantitative framework for quantifying uncertainty from the conceptual and analytical perspectives; (iv) to formulate two detailed case-study examples on health impact assessment of indoor housing interventions; (v) to apply the framework to the two case-studies. After critiquing the uncertainty quantification methods that are currently applied in environmental health impact assessment, the thesis develops the framework for quantifying uncertainty, starting with the conceptual uncertainty (uncertainty associated with the framing assumptions or formulation of the model), then quantifying the analytical uncertainty (uncertainty associated with the input parameters and outputs of the model). The first case-study was concerned with the health impact assessment of improving housing insulation. Using fuzzy cognitive maps, the thesis identifies key indoor factors and their pathways highly sensitive to the framing assumptions of the health impact assessment. The second case-study was concerned with estimating the uncertainty in the health burdens in England, associated with three ventilation exposure scenarios using fuzzy sets and interval analysis. The thesis presents a wider uncertainty framework as a first step forward in quantifying conceptual and analytical uncertainty in environmental health impact assessment when dealing with limited information.