Timings for saved lives and protection to health from the rapid COVID vaccine rollout in Sydney

In the previous post, we discussed how graphs of Australian COVID-19 hospital occupancy reveal dramatic improvement in health protection because of a more complete rollout of vaccination in Victoria by the start of the outbreak. 

The following graph shows the main point from that post:

Even though diagnosed daily cases were much greater in New South Wales, hospital occupancy was about 50 percent lower.

The Pundit has been continuing to explore what’s happing here, and detailed weekly NSW government reports showing numbers of new patients entering hospital and ICU beds each week enable better understanding of how health impact of vaccination changed during the NSW epidemic.

A useful way of doing this exploration is to plot cumulative disease events versus the cumulative number of COVID-19 cases. This effectively turns curves into straight lines and from these lines revealing rate factors are easily displayed as line slopes and calculated.

It’s a trick the Pundit learned a long time ago when he grew bacteria for a living.

Here’s such a plot using death and case numbers for the New South Wales epidemic obtained from weekly disease surveillance reports.

Before doing this plot, cumulative death counts have been shifted back 2 weeks to align with the infection number count that causes them.

What the plot reveals is that very early in the outbreak there is a dramatic change in slope of the plotted line.

The Pundit was delighted to make this discovery, which is most obviously revealed by this graphical approach.

The initial slope corresponds to a death rate of 1.4 lives lost per hundred cases, while the subsequent slope is 0.74 percent which appears unchanged for much of the subsequent part of the outbreak.

It implies that continued vaccine rollout, in addition to eventually halting new cases, saved several hundred lives among the cases of infection that got through health protection measures. To see this, just extend the initial slope upward in your imagination.

But there is more.

Hospitalisation rates changed

A similar graph can be done using cumulative numbers of hospitalisations instead of deaths, as follows.

In this graph the count of total hospital entries (vertical axis) has been shifted by one week to allow for the lag between illness onset and hospital entry.

The graph reveals a change of slope after after about 30,000 infections into the outbreak, to a lower value.

Doing some arithmetic on these slopes reveals that initially 19.3 people in 100 infection had to go to hospital but subsequently this dropped to 12.1 percent.

Rollout of vaccines during the outbreak cut chances of going to hospital in half (and at least partly explains changes in hospital occupancy during the outbreak).

ICU entry rates also changed.

Here is an analogous graph for cumulative ICU admissions versus case numbers:

Here cumulative ICU admissions are plotted on the vertical axis, adjusted for a 1 week lag.

Vaccine rollout in New South Wales was just in time.