Of the 177 versions of the ArriveCAN application that were released, we examined the ones that were considered major, so those that included substantial changes. We would expect the 25 versions to be well-documented, but for half of them, so 12, there was no documentation of user testing. Of the remaining 13 releases, only three were well documented, and the documentation was incomplete for the other 10.
We did not examine more releases, because we found that the best practices we expected would be followed when changes are made to an application's software and a new version is released weren't.