Confounding by indication, explained with a concrete example

On the subject in the title: Confounding by indication, explained with a concrete example Working notes rather than a conclusion.

Comparing SURMOUNT-1 (N Engl J Med, 2022) with SURPASS-2 (N Engl J Med, 2021) and finding the comparison harder than it looks.

Different populations, different durations, different endpoints defined slightly differently, and in one case a different estimand. People compare the headline percentages anyway, including me until recently.

Is there a defensible way to put these side by side, or is the honest answer that there is not and we should stop?

Choosing the worst interpretation: "The confidence interval includes a harmful effect" is true if the CI goes from -1 to +5. But assuming the worst-case scenario is not how you use the evidence. The point estimate and the precision both matter.

post #2 answers the question as asked. The question underneath it is different.

Publication bias: a single published positive trial is weaker evidence than multiple published trials with consistent results. Asking whether there are unpublished negative trials is a fair critical question.

Multiple comparisons: if a paper reports many outcomes, the chance of a spurious association by random chance is real. Pre-specification of primary outcomes matters and secondary analyses are weaker evidence.

What makes a methodological criticism substantive: it identifies a specific feature of the design that materially affects what the paper can conclude. "Small sample size" alone is weak. "Small sample size for a rare outcome, so the confidence interval is wide" is stronger.

I read post #4 twice before replying, because I had assumed the opposite.

Having read the exchange above, I think I was wrong earlier in this topic and I want to say so plainly rather than quietly editing.

The correction was fair and I had been repeating something I had not checked carefully enough.

Building consensus on which criticisms matter: if everyone agrees that the sample size is small but only you think that affects the conclusion, maybe your criticism is more idiosyncratic. That does not make it wrong but it is worth noticing.

Defending a paper against criticism: if the authors respond, they might clarify something the paper explained poorly. Their response might also miss your point. Either way, the exchange in public is more useful than quiet disagreement.

Picking up post #6: that is the part I would want checked first.

Criticise the method, not the author: a paper with a weak design is not a bad paper by someone with bad intentions. It is a paper that answers a limited question. Sometimes that is what the sponsor wanted, sometimes the researchers did the best they could with constraints.

Coming back to post #8, because the follow-up matters more than the original answer.

Generalisability: do the inclusion/exclusion criteria narrow the population so much that results do not apply to real people asking about it? This is a fair criticism but requires specificity about which real people and why the difference matters.

On post #7 — agreed on the reasoning, with one qualification.

When you change your mind: if a reply convinces you that your criticism was not well-founded, say so plainly. The critique might still be real but smaller than you originally thought. That is not a failure — it is how discussion works.

post #11 answers the question as asked. The question underneath it is different.

Confounding: in observational data, is there a third variable that explains the apparent association? In randomised data, randomisation should balance unknown confounders, though known confounders can be adjusted for.

Bias towards the null and bias away from the null: different criticisms have different directions. Differential dropout might bias away from null; conservative statistical analysis might bias toward null.

For anyone arriving from a search: the marked solution above is the direct answer, and the replies underneath it add the caveats that make it safe to use.