« Story time in vaccine trial reporting | Main | Notes on the Astrazeneca-Oxford vaccine trial results 3: it's raining numbers »


Feed You can follow this conversation by subscribing to the comment feed for this post.


This was all a bit of a mess. Deciding on what happens in Phase 1/2 trials and then what happens in Phase 3 is a bit of an art rather than science. The Phase 2 are rarely powered enough to make a definite decision on dosing etc for Phase 3. After all that is why they are Phase 2. They made life difficult also by not really doing a real Phase 2.

I think they panicked. They actually had designed their Phase 2/3 to detect a VE of 70%. They must have decided that a single dose was not going to be sufficient. It is possible that they compared the antibody response to that of another vaccine, and realised they didn't have as good a vaccine. So they added a second dose. While it is messy, it doesn't invalidate the trial, as randomisation still holds.

The different countries isn't a problem. Their modelling included an effect for study for the rate of infections. Somewhere in the protocol that also allowed for checking for a treatment by study interaction which I don't think they found. Some statisticians would argue for a random effects model. Similarly there are opinions that they should have ignored the LD/SD and SD/SD groups, as differences between them are likely to be spurious.

I would like to see a survival analysis approach to analysis. This would allow for setting up a time-dependent effect of treatment to see how the effect of treatment differs for the first and second dose.


Ken: Ordinarily, different countries are not a problem as evidenced by multi-site trials (like Pfizer). But when these are separate trials, done at different times, not controlling for demographics, and even having varying treatments (dose intervals) and placebos, that's a lot to stomach. I think they could have helped their case by releasing analyses to support the assertion that these trials can be pooled, as well as various other claims throughout. It's certainly possible but hard to tell without seeing supporting evidence.

The comments to this entry are closed.

Get new posts by email:
Kaiser Fung. Business analytics and data visualization expert. Author and Speaker.
Visit my website. Follow my Twitter. See my articles at Daily Beast, 538, HBR, Wired.

See my Youtube and Flickr.


  • only in Big Data
Numbers Rule Your World:
Amazon - Barnes&Noble

Amazon - Barnes&Noble

Junk Charts Blog

Link to junkcharts

Graphics design by Amanda Lee

Next Events

Jan: 10 NYPL Data Science Careers Talk, New York, NY

Past Events

Aug: 15 NYPL Analytics Resume Review Workshop, New York, NY

Apr: 2 Data Visualization Seminar, Pasadena, CA

Mar: 30 ASA DataFest, New York, NY

See more here

Principal Analytics Prep

Link to Principal Analytics Prep