RE: RE: DSMB infoFrom what I read, the results are blinded even to the DSMB and unblinded only if there is evidence of severe adverse effects, such as death, in a percentage of the tests subjects, then the test subjects are unblinded so the lead investigator and statician can determine if there are adverse effects are from the drug.
Here is a quote:
In multi-center studies, performance monitoring should be an ongoing activity performed by the study investigator and statistician, and reviewed by the DSMB. Performance data are reviewed in a blinded fashion, often in aggregate by site, and thus do not raise unblinding or bias issues. The investigator and statistician also perform ongoing safety review of the data, and safety reports are reviewed at regularly scheduled DSMB meetings.
Full DSMB protocal:
https://www.niddk.nih.gov/patient/Clinical_Trials/PlnwthDSMB.pdf