Re: Covid-19: Vaccine candidate may be more than 90% effective, interim results indicate
I read the analysis of NNTV provided on 13th November by Dr Allan S. Cunningham. Although the numbers appear correct, the applicability of the analysis appears to rely on the permanently living under lockdown conditions.
The basis for the argument that 90.7 % efficacy provides little value is based on the attack rate for the placebo group of 0.0043 and for the vaccinated group 0.0004. This represents a factor of 10.75 or 90.7 % effectiveness. Dr Cunningham then subtracts the two percentages to claim that the reduction in attack rate is just 0.0004 reducing the absolute attack rate to 0.0039. From this he concludes that it would be necessary to vaccinate 1/0.0039 = 256 to avoid one infection and claims this means the vaccine adds almost no value. I find this conclusion very troubling.
Absolute attack rates during lockdown are probably around 0.005 and do not represent normal living conditions. The absolute impact fo a virus when in lockdown will scale with the attack rate. Without mitigation, the attack rate would be similar or higher than other airborne viruses like the flu. If we pick 50 % as an attack rate in an unguarded society and assume a 2% mortality rate then in a population of 330 million, the expected deaths would be 3.3 million. If that population had all been vaccinated using a 90.7 % effective vaccine then the death figure would drop to 306,900.
If we now repeat the trick of subtracting attack percentages, the figures are now 50 % and 4.65 %, with the difference being 45.35 %. I.e. one infection saved for every two people vaccinated. That is a far more realistic analysis of the impact of a vaccine with 90.7 % efficacy than the 1 in 256 calculated by Dr Cunningham which is an accurate figure for a lockdown but not how we want to live the rest of our lives.
Competing interests: No competing interests