By Helen Petousis Harris 15/01/2021

Happy New Year! No, experts are not concerned that “one of New Zealand’s COVID-19 vaccines will fail to protect the country” Here is why.

But first I wish to issue an expletive about this journalism (First in Australia and then in NZ). It exhibits utter failure to actually truly consult the experts, other than a passing token quote. I call false journalistic balance.

Grump out of the way, here is the story.

A virology society in Australia issued a public statement about the AstraZeneca COVID-19 vaccine (AZ). They compared it to the Pfizer RNA vaccine and said it would not be good enough to achieve ‘herd immunity’ and Australia should halt deployment of it. They stated that the efficacy of the AstraZeneca vaccine was 62% and that the efficacy of the Pfizer vaccine was 95%. Oh boy are they missing some pertinent information. I can only assume they are making these statements based on newspaper reports and not the actual data. Here are the reasons we should ignore this posturing.

  1. We do not have the same amount of data yet for the AZ vaccine as we do for the Pfizer vaccine.

So far there is data on about 20,000 participants in the AZ trials but just 11,636 included in the efficacy presented to date (efficacy trial here) and about 42,000 included for efficacy assessment so far in the Pfizer trials (efficacy trial here). As you can see, these are quite different! That means, right off the bat, that there will be less precision in the estimates from the AZ studies because there is way less data so far. This is evident when we look at the results. There was a total of  8895 persons getting two standard doses and 2741 persons received a low doses followed by a standard dose. The groups receiving the different dosing differed, including the fact that only younger people received the low dose.

  • If you pool all of the 11,000 together the efficacy against symptomatic COVID is 70.4% however there is a 95% chance the true rate is between 54.8% and 80.6%
  • If you only look at the high dose recipients the efficacy is 62.1% with a 95% chance the true rate is between 30.7% and 81.5%
  • If you only look at the low dose recipients the efficacy is 90.0% with a 95% chance that the true rate is between 67.4% and 97%

The reason I have presented these intervals is because when there is an overlap it basically means there is no significant difference between the groups. As you can see they all overlap.

In contrast, the data for the Pfizer vaccine indicates a 95% efficacy for the same outcome with the interval of 90.3 to 97.6%. The interval is much tighter because there is more data.

  1. The trials for the two different vaccines were not conducted in exactly the same way or in the same places

The AZ assessment pooled data from slightly different trials which were similar enough to allow for this. Importantly, the reason there are those different groups (high and low dose) in the AZ study is because AZ had some challenges with their manufacturing and their delivery of vaccine during the trials. This resulted in some people receiving a lower amount of active ingredient in their first dose, also some people received their second dose later.

The Pfizer trials were pretty tight about the dosing interval and participants received their second dose 21 days after the first. However, some participants in the AZ trials had gaps of two to three months between doses and the bigger gap resulted in higher efficacy (we come to that below). This is consistent with vaccinology 101, a decent space between doses tends to result in a superior immune response.  Right now we need some additional data to help guide this.

  1. How did the AZ vaccine really perform?

Severe Disease

Protecting against severe disease is a pretty important outcome, this keeps people out of hospital. The AZ and Pfizer vaccines had pretty much the level of protection against severe disease. Among 12,021 AZ trial participants there were two vaccinated people hospitalised with COVID-19 and 16 in the unvaccinated group. This equated to an efficacy of 88% (Not bad). Over in the Pfizer camp, there was one severe case in the vaccine group and 9 in the unvaccinated group equating to an efficacy of …88.9%.  However, the confidence intervals are wide on both as the numbers are small given the short follow-up time. We need more data to be more confident.

Asymptomatic infection

Unlike the Pfizer study, the AZ trial conducted swabbing of early participants each week and there was some indication of a protective effect which differed between low and high dose recipients. 59% in the low dose and 4% in the high dose. However the intervals widely overlapped so really more data is required.

  1. The importance of dosing interval

There  is a central dogma in vaccinology that the ideal vaccine schedule for non-live vaccines is a priming dose (or two) followed by a booster at least 4 months after the first dose.

When the AZ participants were analysed by how far apart they received their doses a different picture emerges, one that aligns nicely with the immune response. For the subgroup with dosing interval 8-11 weeks, the efficacy was 72.85%, for the subgroup with dosing interval > 11 weeks, it was 81.90%. This level of detail is provided in the Regulatory Agency public assessment report, a document that the critics might well have done to consult before airing their criticisms so widely.

Key messages

  • There is still a lot to be discovered about the AZ vaccine and new data is becoming available fast. Bottom line is that it is going to keep a lot of people out of hospital and save a lot of lives
  • As of writing we do not know how well these vaccines will perform against transmission of the virus but there are hints that they will do so to at least some extent
  • There is a lot of data that must be waded through and assessed in order to make recommendations for vaccine deployment, the bigger picture is critical.


See opinion piece by Prof Peter McIntyre in the Sydney Morning Herald on this topic. Peter is an internationally renown vaccine expert.