{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreidbp4efsoonoxo5aggkfkhnmtev4uacprxb26pqwctruqxjxhn6mq",
    "uri": "at://did:plc:wwyqal4cnqhuwyacdj7rqq3n/app.bsky.feed.post/3momw3j4gvjj2"
  },
  "path": "/t/randomized-non-comparative-trials-an-oxymoron/20863?page=4#post_77",
  "publishedAt": "2026-06-19T07:21:20.000Z",
  "site": "https://discourse.datamethods.org",
  "textContent": "Great points, excellently put! From my experience the second point is key, and under-appreciated. I find myself saying a lot that there is always selection bias in the patients that end up in a trial - this is obvious, we have eligibility criteria so they can’t be a random selection - and it’s different for every trial (even if the eligibility criteria are the same). So doing anything apart from comparing with a concurrent group that is randomly allocated (or matched in another way) is weakening any inferences, potentially to the point that they are meaningless. That’s not to say historical data can’t be useful, but we need to be careful and honest about their use.",
  "title": "Randomized non-comparative trials: an oxymoron?"
}