In this study, as in GBD 2015, we have included risk-outcome pairs that we have assessed as meeting the World Cancer Research Fund grades of convincing or probable evidence (see appendix 1 p 10 for definitions of these grades).15 Table 1 provides a summary of the evidence supporting a causal relation between a risk and an outcome for each pair included in GBD 2016. For each risk-outcome pair, we used recent systematic reviews to identify independent prospective studies (randomised controlled trials, non-randomised interventions, and cohorts) that evaluated the putative relationship. For risk-outcome pairs with fewer than five prospective studies, we evaluated evidence from case-control studies as well (appendix 1 p 344). Table 1 summarises the evidence using multiple dimensions, which supports our assessment that each included risk-outcome pair meets the criteria of convincing or probable evidence (appendix 1 p 10 contains a justification of the criteria presented to support causality). In this summary of evidence, we have focused on randomised controlled trials and prospective observational studies, along with supporting evidence, like dose–response relationships and biologically plausible mechanisms.Table 1Descriptive cataloguing of
