To me, this is often the biggest point. RCTs are held up as the gold standard, but there are multitudes of terrible RCTs that don't demonstrate causation at all. Meanwhile, there are plenty of high-quality epi studies that do. It's more complex than a yes/no RCT binaryhttps://twitter.com/ADAlthousePhD/status/1088056748693471237 …
Case in point: I'm currently working on an RCT, a pragmatic observational cohort, several retrospective case-controls, and a few clinical audits. All useful for different things!