Interpretation of our results and prioritisation at the national level might also need to take into account the variable strength of evidence supporting the causal connection for each risk-outcome pair. In GBD 2016, we have continued to use the World Cancer Research Fund criteria of convincing or probable evidence to select risk-outcome pairs for inclusion. Some aspects of these definitions are subjective. Not all researchers would agree on the interpretation of the available evidence as fulfilling these criteria. For example, there are six studies on non-exclusive breastfeeding and LRI; there are two studies on discontinued breastfeeding and diarrhoeal diseases. We have sought to quantify the number of studies of different kinds that are available to support these judgements in table 1, but not all studies support causality to the same extent. Randomised trials, if well conducted, provide the strongest evidence of causality, because they are likely not affected by confounding. But even randomised trials can have biases when there are missing observations, as is often the case. Randomised trials are also not feasible in many cases, or if feasible, not
