Cumulative excess deaths in New Zealand in the COVID-19 era: biases from ignoring changes in population growth rates: reply

A prominent narrative about New Zealand's experience in the COVID-19 era is that we were almost unique in having negative excess mortality -fewer deaths than would be expected in the absence of the pandemic.Gibson (2024) gives three examples of this claim from leading figures in public health, plus examples from journalists and politicians.Underpinning these claims was a statistical series from Our World in Data (OWID) where the expected number of deaths in the COVID-19 era simply extrapolates a trend from annual death totals for 2015 to 2019. 1 A problem with this series for New Zealand is that while total deaths rose by 2.1% per annum from 2015 to 19 the death rate was unchanged; rising deaths reflected population growth of over 1.9% per annum.Population growth fell to just 0.3% per annum after border closures in March 2020, so expected deaths thereafter using this extrapolation are too high.Amongst the 36 (mainly OECD) countries in the Short-Term Mortality Fluctuations (STMF) database (Nmeth, Jdanov, & Shkolnikov, 2021) with required data, New Zealand's population growth rate change was 2nd largest; making extrapolations that ignore changes in population growth rates especially inappropriate for us.After adjusting for population growth rates, New Zealand's cumulative excess deaths appeared positive rather than negative, and rather than being the 2nd lowest (of the 36 STMF countries) we were only 7th lowest (Gibson, 2024).In his comment on this article, David Hood argues that analyses should use age-standardized data and when he does this for New Zealand he finds negative excess mortality, of -2%.The basis of Hood's argument seems to be that migrants are predominantly young people so while border closures reduced New Zealand's population growth by a lot, the reduction in expected mortality should not be as large because migrants who did not come would have been young people with low mortality risk (yet when what should be the same age-group were inward migrants in 2011-19 total deaths were rising at about the same rate that population rose).In response, it should first be noted that the examples from local public health authorities, politicians and journalists critiqued by Gibson (2024) did not use age-standardized data, so Hood's comment is somewhat tangential, and can be thought of as opening up a new front in debates about excess deaths in New Zealand.Moreover, other local studies that used death rates (so needed no critique about their ignoring changes in population growth rates) also did not use age-standardized data (Kung, Doppen, Black, Hills, & Kearns, 2021, 2023) -notably, these studies are in The Lancet, one of the world's leading medical journals.To provide more evidence on this I used the Scopus database to search for journal articles with ('excess deaths' OR 'excess mortality') AND 'New Zealand' AND 'COVID-19' in titles or abstracts.Of the 12 records, two referred to different versions of the same paper and one referred to Hood's comment, leaving a sample of ten papers.Details for these papers are in Appendix A.