{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreibgxgr2vidkwpgjsxjvudkfvsiqs4f53o642whxe4eoeh3mowrdam",
    "uri": "at://did:plc:wwyqal4cnqhuwyacdj7rqq3n/app.bsky.feed.post/3merkriaaf5q2"
  },
  "path": "/t/statin-related-side-effects-the-recent-lancet-publication-is-biased-toward-false-negatives/28635#post_5",
  "publishedAt": "2026-02-11T06:45:00.000Z",
  "site": "https://discourse.datamethods.org",
  "tags": [
    "Women’s Health Initiative",
    "earliest warnings"
  ],
  "textContent": "Dear Frank,\n\nI am not familiar with the history of the statin literature, so I will refrain from commenting on that specifically. That said, I consider it highly implausible that a single paper could be regarded as conclusive, starting with the well-known problems of transportability. The effort I am asking for is simply this: do not canonize the literature, because 1) even under the best of circumstances, it is extraordinarily difficult to guarantee its robustness - the history of medicine makes this clear with cases such as the Women’s Health Initiative - and 2) the current landscape is driven by high-level interests that range - as I argued in my response to Erin - from the pursuit of funding to the endorsement of political and ideological positions.\n\nWe are developing increasingly sophisticated methods that rest on extremely strong and difficult-to-handle assumptions (e.g., g-methods), in a context where the scientific community has proven capable of adopting and institutionalizing rituals - such as p < 0.05 - that had far less justification (virtually none, I would argue). More than 100 years after the earliest warnings about the distinction between “statistical significance” and “practical significance,” we still find prestigious journals that fail to differentiate between the two concepts.\n\nFor these reasons, I continue to regard my strong concerns about the entire process that generates data and results as well founded, and I believe it is a deontological duty of methodologists - especially those far more skilled than I am - to maintain the highest level of vigilance.",
  "title": "Statin-related side effects: the recent Lancet publication is biased toward false-negatives"
}