Is 40-year-old data is driving key health policy in 2024?

Opinion piece by Jordan Williams

Commentators, the Opposition and media went apoplectic when the new Government repealed Labour’s smoking laws.  Despite being years away from having any effect, these policies, somehow, apparently ‘save thousands.’  The worrying innovation of ‘prophecy-based policymaking’.  

But the same experts who insist upon ‘evidence’ have relied on some of the oldest we’ve seen.  Whether an oversight, or wilful blindness, the much repeated “4,000 to 5,000 who die from smoking every year” seemingly originates from 1981 Census via early to mid-1980s mortality data published in 1988. Policy in 2024 is seemingly driven by data when Sir Robert Muldoon was Prime Minister.

I've always suspected the figure was bogus and self perpetuating. Thanks to the help of an intern and a very good researcher (a retired historian), we've managed to track down, that it's so old, it's laughable (or worse) that it continues to be wheeled out.

How smoking death numbers became an advocacy tool:

Professor Janet Hoek recently co-authored: ‘The 5000 annual deaths caused by smoking also matter,’ published on “Public Health Communications Centre Aotearoa”.  This new podium of truth is hosted by Grant Robertson’s University of Otago Ōtākou Whakaihu Waka’s Department of Public Health.  For good measure, Professor Hoek added this on Newshub in February, when discussing nicotine: “Public health Professor Janet Hoek said she hadn't seen any evidence that caffeine is killing 5000 New Zealanders - 13 a day.”

Except nicotine doesn’t kill.  Smoking does. 

How could someone who is a professor in public health forget what Professor Michael Russell famously said in 1976: “people smoke for nicotine but they die from the tar”.  It makes Professor Hoek’s jab at Minister Costello a foot in mouth moment for Otago and the credibility of the ‘Public Health Communications Centre Aotearoa’.  Newshub’s piece, worryingly, showed politicians falling into the same trap as Hoek, despite the UK’s Royal Society for Public Health publishing nine years ago: Nicotine “no more harmful to health than caffeine”.  The NZ Drug Harm Ranking Study even ranked nicotine (in e-cigarettes) as having the same harms as Kava in adults. Conflating smoking with nicotine impedes quitting and if nicotine was as bad as Professor Hoek claimed on Newshub, why isn’t Otago calling for Nicorette and Habitrol to be pulled from supermarket shelves?

Are 5,000 New Zealanders truly dying from smoking each year?

This number, “5,000 New Zealanders - 13 a day” who die from smoking got our attention because it triggered déjà vu.

Being academics, the Public Health Communications Centre Aotearoa referenced a HealthNZ webpage on quitting.  Lazy, considering they’re meant to be experts and when a categorical number is cited.  This is where you go down a 40-year rabbit hole.  Academics cite the Ministry, Ministry officials cite academics, politicians cite both and back and forth.  This is how a ‘fact’ has become its own fact.

In 2021, former health Minister, Dr Ayesha Verrall in a speech previewing her now repealed laws said, “But smoking still kills 4,000-5000 people a year.”  The 2021 consultation document that led to that speech had “approximately 4,500” die a year citing the Global Burden of Disease Study 2017.  The Ministry’s 2020 report on that study actually said smoking “accounted for one-tenth (or 9.7 percent) of health loss in 2017...” 

Google reveals the “5,000” who apparently die each year from smoking, goes back to the dawn of search engines with variations like, “over 4,000”, “4,000 to 5,000” “4,500 to 5,000” and “up to 5,000”.  Otago in 2024 went a step further and just made it “5,000.”

Inconveniently for Otago, in 2010, some of Professor Hoek’s colleagues published a New Zealand Medical Journal article entitled, “If nobody smoked tobacco in New Zealand from 2020 onwards, what effect would this have on ethnic inequalities in life expectancy?” One author was Diana Sarfati and if that name is familiar, she is the current Director-General of Health. 

In it they wrote: “The public health case for concern and action on tobacco use in Aotearoa-New  Zealand is overwhelming; 4500 to 5000 deaths per year…”  This raises quality control questions at the Public Health Communications Centre Aotearoa, given 14-years on, an even larger claim is published despite being hosted by Otago’s Department of Public Health.  That 2010 article referenced a Ministry of Health Tobacco Trends 2008 study that while gone from the website, isn’t to us:

Extracted Ministry of Health graph (2009): Deaths peaked in 1980 below 5,000 but showing rapid declines (for men) while the trend for women is uncertain:

The Ministry of Health graph mirrors the 4,500 number for 1990 that was cited in the New Zealand Yearbook 1996.  It is however difficult to reconcile the “4,500 to 5,000” deaths claimed in the 2010 NZMJ paper, given what 2005 numbers point to. This begs another question. How could it be that a further 16-years later, the number of deaths claimed by the Public Health Communications Centre Aotearoa and Ministry of Health, in 2024, are at levels last seen since 1980 when smoking prevalence has collapsed?  

Zeroing in on the source:

In 2000, Labour’s then Minister, Annette King, issued a media release with, you guessed it, “new research indicates some 388 deaths a year are attributable to second-hand smoke. This is over and above the 4700 New Zealanders who die each year from smoking-related illnesses” (unreferenced). This seems well above what the graph above indicates.  Her National predecessor Wyatt Creech in 1999, the year before, put smoking deaths at 4,500 from a Drug Strategy but that also is unreferenced.  An MSD publication, authored by a Ministry of Health employee in 1998, put the number at smoking deaths 4,250 (also unreferenced).  In the space of three consecutive years are three different numbers that are tracking upwards when the graph shows deaths were going in the opposite direction.

Delving into Hansard we close upon the source and demonstrate how numbers and sources become juxtaposed.  During a 1991 debate in the House, the late Jim Anderton said: “The Department of Health stated clearly that every year nearly 5000 New Zealanders die from cigarette-smoking and its attendant results. The fact that, by the last day of 1988, 4920 people had died from cigarette-smoking that year.”

Except that is not what the research says given the Department of Health published it in 1988 but by 1991, the data was already six-years old.  It also shows how numbers become absolutes in the retelling.

Helen Clark in 1990, when introducing the Smoke-free Environment Bill, said: “Every year more than 4000 New Zealanders are dying from diseases directly attributed to their smoking of tobacco, and 273 of them do not smoke at all but are dying from other people's smoke”. 

The source for all of this is most likely to be 1988 research in the New Zealand Medical Journal and also undertaken by the Department of Health using the same methodology as the Journal. 

Then Health Minister, David Caygill, in a 1988 debate on smoking said: “…nearly 20 percent of all deaths in this country are due to cigarette smoking (at least 4000 every year). This does not include deaths due to passive smoking; or from smoking; pipes and cigars…I am also forwarding to the Member a summary sheet from “The Big Kill’ as well as a New Zealand Medical Journal article 'The Cost of Cigarette Smoking; in New Zealand.’”

The 1988 New Zealand Medical Journal article Mr Caygill referenced is based on the 1981 Census using mortality data to 1984 and states: “In all 4137 deaths per year are attributable to smoking.” 

The Big Kill , a departmental publication, uses the 1981 Census and mortality data to 1985 and  states: “4920 die each year from smoking” based on a population of 3,260,470 in which “Cigarette smoking caused 19.4 % or 1 in 5 of all adult deaths between 1981 and 1985.”  “In 1985 cigarette smoking causes 16470 hospital admissions of 18 days each.  An extra 827.7 hospital beds are needed for the 302093 bed days used each year because of smoking. They cost the taxpayer $99914239 each year.”

These two papers are seemingly the basis for “4000 to 5000” deaths each year from smoking.  It is not sound or credible to use this number in 2024.  The 1988 research methodology needs to be repeated using the most up to date mortality data, albeit, not to the electorate level that The Big Kill did.

Conclusion

All ministers and policymakers must pay close attention to the sources of ‘evidence’ used in policy making.  Since 1981, the number of New Zealanders who die annually per 100,000 has fallen 51% to 355.8 in 2021.  Despite Covid-19, 2021 is lower than deaths per 100,000 recorded in 2019.  The “4,000 to 5,000” claim is not credible and despite searching for it, cannot be sustained in 2024.  Professor Hoek’s article on the Public Health Communications Centre Aotearoa puts it at the top end of all estimates; “5,000 New Zealanders - 13 a day.”  That becomes an own goal.  If true, it means the entire smoking control programme over the past four decades has failed.  More likely is that smoking deaths have declined rapidly reflecting falls in smoking and vastly improved medical treatment.  A smaller number does not serve the interests of advocates and lobbyists who have secured millions in public research funds and built careers in this area.  To paraphrase Deming, ‘In God we trust. All others must bring data and the original sources of that data.” 

So what to be done?

The motivation for this piece, which has been on the cutting board for a long time, is to highlight how even the most 'well researched' areas of public policy become highjacked by good intentions zeallots and fictional 'truths'.  If I was in procurement of or needed to rely on those advising on public policy, I'd be asking hard questions about why there is such little intellectual discipline on such a key claim. I'd be ensuring public health research money for smoking is better directed to those of the highest quality. ASH NZ, the NZ Health Survey and NZ Drug Harm Ranking Study come to mind.  I am looking forward to getting responses from those agencies and NFPs that continue to make the claim on websites and publications without acknowledgement or caution that the data is some 40-years old.

But putting the irony aside about all of those who scream 'evidence based', we actually do need the evidence!  Maybe the Health Research Council could direct some Health Research Funding to contract independent research of the same quality, which generated the core 1988 research. Create a modern datum inclusive of say 2022 mortality and repeated at five-yearly intervals?

Ministers Reti, Seymour and Costello must also ask the Ministry of Health and Health NZ to undertake a quality review audit of health data and model assumptions to ensure old data is not affecting advice and service delivery.  Other ministers should similarly ask their departments for a quality check assurance of data and model assumptions, I doubt this is isolated to one department or policy area.

Disclosure: As described on our About Us page and public disclosures, part of our funding mix (less than 3% of our total revenue) is from industry membership/support. Of that, a subset are related to the nicotine and tobacco industries. While I held the final pen on this piece, I did run it past (and sought feedback and comments on whether anything was wrong) from someone within industry.  I would hope that the piece encourages discourse rather than result in predictable and lazy criticisms about funding. I am confident the facts in this piece are such to support the premise: that the emperor has, at minimum, very old fashioned cloths when it comes to the key claim by the moral panic who argue for ever more tax on tobacco (and now vapes). No one in the Government, as far as I know, knew about this piece or work prior to publication.