The pragmatic attitude of the trial towards cluster
selection, cluster assignment, participant selection,
participant recruitment, informed consent and outcome
measurement supports generalisation to other
jurisdictions.
Postrandomisation selection bias is limited by the
use of existing, objective measures of eligibility.
The use of secondary data for baseline data collection
and follow-upmeasurement increases the
a ccuracy of the data collection and limits the loss
to follow-up compared with primary collection
methods.
It is not possible to conceal the treatment assignment,
which exposes half of the primary and secondary outcomes measures to placebo and observer-expectancy effects.
The jurisdictions included in the study used a convenience,
non-probability sampling approach in cluster selection, which may limit external validity.