Comparison of vaccine efficacy must be based on good clinical data.

The conclusion by Nur Asheila et al that “the mortality rate of those who received inactivated vaccine was higher than the recipients of the BNT162b2 and ChAdOx1 vaccines” cannot be supported by their data because patients receiving the 3 types of vaccine may not be comparable. The authors in fact noted that in their sample,“vaccine allocation was not entirely random”. They can correct this limitation by applying statistical tests on the baseline characteristics of the 3 patient groups, especially the number and severity of co-morbidities, to see if indeed comparison across the 3 groups is valid. There is good evidence that vaccine efficacy drops with time. Therefore their conclusion is further weakened because we do not know the time interval from vaccination to death in their study population. For any comparison of vaccine efficacy to be valid, we must be sure to compare similar time intervals after full vaccination. Finally their logic defying result in the inactivated vaccine group, showing fully vaccinated patients to have higher mortality than those partially vaccinated, suggests that sample size is too small to correct for important mortality affecting bias in this population group. After all, in actual community use, total Covid-19 deaths per million is lower in Laos (59 deaths per million) which uses inactivated vaccines for a majority of its population compared to richer regional neighbour Singapore (154 deaths per million) with its mRNA based program.

Declaration of interests

None.