You (MSB) used propensity-matching to demonstrate the...

You (MSB) used propensity-matching to demonstrate the remestemcel-L survival benefit in high-risk patients based on Baseline MAP scores.
https://investorsmedia.mesoblast.com/static-files/cc68683d-f5c5-4dc0-b718-499c11fd8ca

YES This was done at a later date to show that the data created in the chart corresponded to improved survival. FDA did acknowledge improved potency assays in CRL 2 so we still have the 54 available patients. JB seems to be saying that a large number were censored due to not having samples taken. I would find this to be unlikely in the case of 001, samples would have been collected on day one and reasonable regular afterwards. Indeed it would have been standard of care. ( take a look at the Magic Coinsortium data were they concluded that there MAP scores could be used as a tool to show responses and long term survival) I do agree with JB, some would not have had the second sample taken however a 7 - 14 or 21 day sample could be used as a substitute for the 28 day result adjusted to the probable value.

https://pubmed.ncbi.nlm.nih.gov/15494428/
Regarding confounding factors - sure, but MSB is trying to demonstrate that the product works independent of the donor that provided the bone marrow aspirate. How can you do that if a patient gets MSCs from potentially 2 (or more) different donors, especially since MSB have shown different levels that could have affect clinical outcome (INFgamma - see Aggarwal & Pittenger: )."

It's not hard given the amount of data that would have been collected over the years. Let's say the first dose has been shown to be the dominant contributor of day 28 results, perhaps 50% the second 20% down to the last which may have very little influence on the day 28 result. The rest is mathematical . It would actually be considered a study failure not to include the data. given the amount of data that would have been collected over the years. Let's say the first dose has been shown to be the dominant contributor of day 28 results, perhaps 50% the second 20% down to the last which may have very little influence on the day 28 result. The rest is mathematical . It would actually be considered a study failure not to include the data.

Unless you say that is such confounding factor that can be found in any study. If that is in fact the case, then we may have a big problem.
MSB realised that this could be an issue (frankly, one that could have been avoided) after being prompted by the FDA and that's how you ended up with 11 data points (single lots in a single patient).

Yet the chart clearly states 11 lots.
Just how important is this anyway? Be it 11 patients or 40 , the purpose of the chart is to summaries the data and show effect. What is more important is the validation of the results.