This is a continuation of a long topic, addressed by post number rather than by page. Start at post 1 · go to the accepted answer.
Individual participant data pooling is a much stronger design than aggregate pooling and is rare because it requires cooperation rather than a search.
Where I part company with post #91, and it is a narrow parting.
Where the pooling trials with different estimands discussion usually stalls is that nobody wants to say "I do not know" and everyone is willing to say "it varies". Those are the same sentence with different clothes on.
Answering the question post #91 raises rather than the one it answers.
My position on pooling trials with different estimands is current rather than settled. I have revised it once already and I expect to again, so treat it accordingly.
The arithmetic in post #95 is right; the assumption feeding it is the part to check.
The arithmetic on pooling trials with different estimands is the easy part and it is where the errors are, which is an uncomfortable combination. Show your working and someone will catch it.
Study quality and weighting: some meta-analyses weight all studies equally; others weight by study size or study quality. The choice affects the result and should be stated and justified.
If this contradicts something upthread, the upthread version may well be the better one.
Thank you — that answers what I came here to find out.
Taking post #102 at face value and following it one step further.
Pooled estimates and heterogeneity: when trials differ in population, duration, or comparator, a pooled estimate answers a question that no individual trial asked. High heterogeneity means effects genuinely differ across studies. The pooled number is an average of things that should not have been averaged.
The literature is thinner on this than the confidence in the thread implies.
Fixed-effects versus random-effects models: fixed-effects assumes all studies are estimating the same thing and variation is sampling error. Random-effects assumes studies are estimating effects from different distributions and allows between-study variance. Choice matters if heterogeneity is high.
Worth separating two things that post #104 runs together.
If you are new and reading this thread for the answer to pooling trials with different estimands: the answer is conditional, the conditions are in the third reply, and the rest of the thread is worth skipping.
This follows post #107 rather than contradicting it.
When trials differ in population, duration and comparator, the pooled estimate answers a question no individual trial asked. That is worth saying before the number is quoted.
Not disagreeing with anyone above, just adding the bit I keep having to look up.
I had written a reply contradicting post #107 and deleted it. Here is what survived.
The honest answer on pooling trials with different estimands is that it depends, and the useful part is the list of what it depends on. Four items, in rough order of how much they matter.
Most people get the first two right and then argue about the fourth.
This is the sort of exchange that makes the archive worth searching.
Two questions I would want answered before drawing anything from the pooling trials with different estimands data above: how were the cases selected, and what happened to the ones that dropped out.
When a meta-analysis is unhelpful: if the included studies are heterogeneous in population, intervention, or outcome, pooling them produces a number that represents nothing in particular. Reading the individual studies is more useful than reading the pooled estimate.
Adding the measurement that post #111 says would settle it.
The thing about pooling trials with different estimands that took me longest to accept is that a plausible mechanism is not evidence of an effect. It is a reason to look, not a result.
Post #112 describes the usual case. This is about the unusual one.
Posting my pooling trials with different estimands numbers with the method attached so they can be discounted properly. Uncontrolled, unblinded, and collected by someone who wanted a particular answer.
Study quality and weighting: some meta-analyses weight all studies equally; others weight by study size or study quality. The choice affects the result and should be stated and justified.
I would put a moderate confidence on that and no more.
Overlapping populations across included trials inflate the apparent sample size. It happens more than people expect where the same programme reports multiple papers.
On post #116 — agreed on the reasoning, with one qualification.
Fixed-effect and random-effects models answer different questions. The first assumes one true effect; the second assumes a distribution of them. Choosing between them is an assumption, not a technicality.
Building on post #118 rather than restating it.
Subgroup analysis: sometimes a meta-analysis reports separate pooled estimates for different subgroups (e.g., by baseline body mass index or by trial duration). Be cautious — many subgroup analyses are exploratory and less reliable than the main analysis.
I have kept the units in throughout, for the obvious reason.
Post #116 put the caveat in the right place and I want to underline it.
I have been on both sides of the pooling trials with different estimands argument in this category within eighteen months, which should tell you how strong the evidence for either side is.