Many thanks to Dr Charles Beardall for this week’s note. It’s a novel one. The paper was called “Red and processed meat intakes and cardiovascular disease and type 2 diabetes mellitus: An umbrella systematic review and assessment of causal relations using Bradford Hill’s criteria” and it was by Hill et al (Ref 1).
We’ve looked at umbrella reviews before. They are supposed to be the best of the best of evidence. They look for systematic reviews and meta-analyses, which pool together all data and they pool these together. This process has a couple of initial flaws. First, all pool-together articles are only as good as the original studies. If the original studies are poor, then poor has been pooled with poor. Second, systematic reviews and meta-analyses should find the same studies. Exclusion and inclusion criteria may differ, but the same studies will appear in different pool-together articles. This means that a number of studies will be duplicated. In meta-analysis, the larger/longer studies carry more weight. If these are duplicated (as they likely will be, since they are the major studies), then they carry even more weight in an umbrella review, likely to the point of disproportion.
The novel aspect of this review was that most researchers present risk ratios as conclusions and then infer that association can mean causation. Bradford Hill established famous criteria for assessing if association is likely to mean causation (Ref 2). This study examined the findings against the Bradford Hill criteria to see if causation might be likely. I’ll use the full name, Bradford Hill, throughout this note so that we don’t confuse Bradford Hill with the lead author of this week’s paper – Hill.