Making sense of the NZ covid death statistics
Today, one report headlines “more than 1k deaths” and points out that 1017 people died with Covid.16 minutes earlier, the colleague reports “close to 450 Kiwis died from it”. That report goes on to say 447 had it as an underlying cause of death (COD), for 231 it was contributing to the COD, and 161 had a COD unrelated to Covid. 136 still need to be classified.
So, based on the the earlier published report, (447 + 231) 678 people died with Covid, and 136 still need to classified, how can we have a figure of 1017 floating around? Even if you added the unclassified (678 + 136 = 814), there still is a difference of 203, or 20%!
The reports don’t go as far as clarifying over what time frame those deaths were accumulated. From Worldometer we find out that the bulk of NZ cases have started in Feb 2022, and slightly delayed, the deaths went up from March 2022. Based on those numbers, what is referred to as the Infection Fatality Rate (IFR) is 814 / 1,070,000 = 0.076%. That is an IFR in the ball-park of an influenza outbreak. If we took the 447, then the IFR is 0.04%
Another aspect those articles don’t seem to address is whether or not those deaths lead to over-mortality. The graph from the NZ Health portal, taken from 2013 to 2022, but note, it doesn’t include NZ winter 2022 yet. It shows a distinct pattern of higher death rates in winter, lower over summer. Note, how that up and down is missing in winter 2020. The curve is flattened here, aka less in winter, more over summer. This is partially due to lock downs and delayed funerals, as the data entry comes from funeral dates. Based on timing, the bulk of the 814 deaths would fall in a time frame of 12 weeks, about 67 deaths/week (dpw). The NZ Portal over the same time averages around 700 dpw. Thus, the Covid deaths are around 9.5% of all deaths, and specifically, the 447 classified with Covid, 37 dpw, or 5% of total deaths for that period.
Is it actually over-mortality? You be the judge! Look at the graph and do your own visual analysis with respect to the average number. Keep in mind those numbers are not adjusted for population increase, aka, the population now is higher than in 2013, which is reflected in a slight increase of average over time (see blue line on the graph).
Again, the chart does not yet include winter 2022, but it will be an interesting review in 6 months time.
