{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigtkvdlzq7fko67hivrr5ws6yf7wb2zjhrm3taihb3do6gbdzgkam",
    "uri": "at://did:plc:wwyqal4cnqhuwyacdj7rqq3n/app.bsky.feed.post/3mibw3cochx42"
  },
  "path": "/t/thinking-clearly-about-association-studies-risk-factors-and-causal-salad-included/28679#post_8",
  "publishedAt": "2026-03-29T21:25:09.000Z",
  "site": "https://discourse.datamethods.org",
  "tags": [
    "@Pavlos_Msaouel",
    "both preclinically",
    "clinically",
    "this work",
    "Target Trials",
    "immortal time bias",
    "collider selection bias",
    "here",
    "commit the Table 1 fallacy",
    "omit key visual elements of survival plots",
    "incompletely report toxicity data",
    "more frequent unsupported claims of differential treatment effects",
    "this fascinating history of adaptive Bayesian clinical trials",
    "this Judea Pearl festschrift book chapter",
    "@Sander"
  ],
  "textContent": "ESMD:\n\n> @Pavlos_Msaouel has been using DAGs elegantly to _inform design_ _of his oncology RCTs_. The DAGs in his papers seem, clinically-speaking, eminently sensible. I’d be interested to hear how he develops them.\n\nThat is very kind of you. Most of our work in the lab and clinic involves experimental designs, which are well suited for causal diagrams both preclinically and clinically. When we do observational research it is usually connected with experimental models (e.g., this work evaluating high-intensity exercise as a risk factor for renal medullary carcinoma in the presence of sickle cell trait).\n\nIt is typically much harder to encode the causal assumptions of purely observational studies. But in our practice I have found the concept of Target Trials to be helpful particularly because it emphasizes the importance of carefully specifying the “time zero” of the interventions. Doing so helps notice and potentially reduce common design biases such as immortal time bias and collider selection bias. A recent high-profile example of collider selection bias here. These design biases can influence causal inferences at least as much or more than other more commonly cited biases such as confounding.\n\nESMD:\n\n> First, I suspect that the evidence underpinning _his_ DAGs is of a much higher volume and rigour, as necessitated by the high-stakes nature of the drug development gauntlet (?) Second, it occurred to me that _his_ incentives, and the incentives of _drug sponsors_ , are very different from those of academic observational researchers.\n\nThere is indeed some misalignment between the “publish or perish” academic culture and truth-seeking. This is supported, e.g., by our (observational ) data showing that academic oncology RCTs are more likely than industry ones to commit the Table 1 fallacy, omit key visual elements of survival plots, and incompletely report toxicity data. On the other hand, industry incentives can also be misaligned resulting in more frequent unsupported claims of differential treatment effects, and I recently read this fascinating history of adaptive Bayesian clinical trials (certainly of interest to datamethods contributors) discussing how the industry sponsor (and the academic investigators) chose to not report the ROAR trial using the estimates from the Bayesian hierarchical modeling of the primary endpoint and erroneously decided to report the frequentist estimates – despite the FDA prompting them to prioritize the Bayesian hierarchical analyses.\n\nAlso related to this thread, this Judea Pearl festschrift book chapter by @Sander superbly argues that every real data analysis rests on a causal model of the data-generating process, even for purely descriptive goals.",
  "title": "Thinking Clearly about Association Studies (Risk Factors and Causal Salad included)"
}