Tue Mar 22

Most COVID patients in NZ's Omicron outbreak are vaccinated, but that's no reason to doubt vaccine benefits

Written by: Thomas Lumley, Professor of Biostatistics, University of Auckland

New Zealand’s Omicron wave may be peaking, but we’ll continue to record thousands of new cases ^[1] each day and most people who test positive or are hospitalised with COVID will have been vaccinated.

This is exactly what we should expect and it’s no reason to doubt vaccine effectiveness.

The principal reason why a lot of COVID cases are vaccinated is because most New Zealanders are now vaccinated. As of today, about 94% of people 12 years and older have had two or more vaccine doses, and even if their risk of catching COVID is significantly lower than for an unvaccinated person, they vastly outnumber those who aren’t.

In the week ending March 13, about 93% of the 118,000 confirmed cases 12 years and older were in people with two or more doses. But such crude proportions of cases aren’t all that good an indicator of vaccine effectiveness.

Last year, during the Delta outbreak, the proportions were misleading in the other direction. The rate of cases in people who were unvaccinated was about 20 times that in vaccinated people.

Unfortunately, some commentators talked about that ratio as if it was all a real benefit of vaccination. It wasn’t ^[2].

The outbreak in Auckland was nearly under control and was spreading among unvaccinated people partly because they had less resistance to infection, but also because they were more likely to come into contact with infected people. Social clustering leads to disease clustering.

What case numbers can tell us

For Delta, two doses of the vaccine produced very good immunity ^[4], especially in the short term. The vaccine is less effective for Omicron ^[5]; two doses give only partial immunity even in the short term, and the effectiveness wears off over time.

About 60% of people 12 years and older have had a booster dose, and in the week ending March 13, only 42% of cases were in people who had been boosted. We can see that boosters help.

Counting cases remains important, because even a non-hospitalised case of COVID can be unpleasant, and because we don’t know how likely a mild case is to lead to long COVID and months or years of disability. We can’t draw strong conclusions from numbers of cases, though.

Many cases, probably most cases, are not being diagnosed at the moment. Unvaccinated people will be less likely to get tested, especially in mild cases of the disease, either because of poor access to the health system or because they don’t think COVID is important. We can’t really tell how much bias this introduces into the numbers.

Hospitalisations and deaths are much more reliably counted than cases. Results from clinical trials and careful population studies of COVID vaccines consistently show the vaccines to be more effective in preventing more serious disease, especially with the new variants. There are plausible biological explanations for this, based on different parts of our immune response.

Antibodies against the COVID virus ^[6] seem to be affected more by differences between strains than T-cells are; antibodies are probably more important for preventing initial infection and less important for fighting serious disease.

More benefit in protecting from serious disease

When we look at hospitalisations and deaths, the difference between vaccinated and unvaccinated people is much more dramatic. In the week ending March 13, 65% of people over 12 hospitalised were vaccinated, compared to 94% in the population; 32% had a booster dose, compared to 60% in the population. The 5% of unvaccinated people over 12 contributed 20% of hospitalisations.

The number of deaths is, fortunately, too small for the Ministry of Health to publish detailed weekly breakdowns, but vaccinated people are a minority over the period since August.

The relatively small number of deaths in New Zealand’s Omicron wave also shows the effectiveness of the vaccine. Hong Kong had largely eliminated COVID until Omicron; they are now getting a large outbreak similar to New Zealand’s, but only in the number of cases. Over the past week, Hong Kong averaged 280 deaths per day, in a population less than twice that of New Zealand.

The vaccination rate in Hong Kong is much lower ^[7]. About 71% are fully vaccinated and only 30% have had a booster. Among elderly people, who are at much greater risk from COVID, the vaccination rate is especially lower, with two-thirds of people over 80 and more than a third of those aged 70-80 having been unvaccinated when Omicron hit.

Towards fair comparisons

Comparing across whole populations this way gives some indication of the vaccine benefit, but it is very imprecise. We don’t choose randomly who gets the vaccine and who doesn’t.

In New Zealand, for example, essentially everyone over 75 has been vaccinated. Since people over 75 are much more likely to need hospital care than younger people, the higher vaccination rate in people over 75 makes the vaccine look less effective than it really is.

Statisticians call this “confounding by indication”. Auckland has always had more exposure to new outbreaks and had higher vaccination rates than the rest of the country; this again tends to make the vaccine look less effective that it really is.

More reliable comparisons require either random allocation of vaccine to people, as in the clinical trials performed before the vaccines were approved, or careful statistical matching of vaccinated and unvaccinated groups to get a fair comparison.

Omicron is too recent to have useful clinical trial data, but peer-reviewed statistical analyses of individual case data from the United Kingdom ^[9], the United States ^[10], and South Africa ^[11] all agree the vaccines are beneficial.

There’s some evidence vaccination also reduces the risk and severity of long COVID ^[12], the most likely bad outcome for healthy people. But there obviously hasn’t been time to do this sort of comparison specifically for the Omicron variant.

Overall, the most reliable comparisons between vaccinated and unvaccinated people have consistently shown a benefit of vaccination. The effectiveness of the vaccines does wear off over time, and the effectiveness is lower against Omicron than it was against Delta or the original COVID strain, but it still improves your chances of avoiding infection, keeping out of hospital and making a full recovery.