Comparative Risk Assessment and Systematic Analysis for Global, Regional and National Attributable Burden to Major Risk Factors
摘要
This chapter reviews the process of estimating the attributable burden of disease for nearly 100 risk factors under the Comparative Risk Assessment (CRA), based on 23 papers from the GBD study. The first section describes the elements of the Comparative Risk Assessment (CRA) conceptual framework for health risks developed by the GBD study and reviews the classification of risk factors and the estimating process of the attributable burden analysis, which has shown how the early GBD study performs imputed burden analysis of multiple risk factors within the framework of CRA. The second section reviews in detail the strategies used to ensure comparability of the attributable burdens for different risk factors in the six rounds of the GBD studies since the GBD 2010, including hierarchical categorization of risk factors and outcomes, the unified counterfactual scenario of theoretical minimum risk exposure levels, etc. The second section also describes the development of the Summary Exposure Values (SEVs) to compare the prevalence of different risk factors. The third section reviews in detail the approaches and data sources for estimating mediated effects in the GBD study. The GBD study quantified the mediated effects of metabolic risk factors for cardiovascular disease and diabetes mellitus through meta-analysis of data from large integrated aggregation projects and multiple prospective cohort studies. On this basis, the GBD group proposed the approach for adjusting for mediators and calculating mediation factors: comparing crude RR with RR adjusted for mediators and estimating mediation factors by effect pathways. The fourth section systematically reviews the methodological evolution on estimation of smoking attribution burden in the GBD study. Firstly, the GBD extended the Smoking Impact Ratio (SIR) developed by Peto-Lopez to estimate accumulated exposure of smoking for populations. GBD 2017 used cigarette equivalents per day, pack year, and years since cessation to estimate cumulative exposure to smoking in populations. The GBD then estimated dose–response risk curves for different health outcomes attributable to smoking and calculated the population attributed fraction (PAF) to estimate the attributable burden of smoking. Section 6.5 summarizes the characteristics of the attributable burden estimates for GBD and points out the similarities and differences between attributable burden studies for GBD and the traditional epidemiological attribution studies.