The Curse of GIGO

The purpose of computer modelling is to attempt to forecast the future. When real-world data and experimental research results conflict with computer modelling outputs, it tells us the inputs fed into the models were wrong and need changing. The curse of computer modelling is garbage in, garbage out (GIGO).

The inputs are the assumptions the modeller made. Modellers are not always honest about what they based those assumptions on. They can be even more coy about why they made the assumptions they did, and possible conflicts of interest. 

In highly contentious fields of inquiry, speculative models have been wielded as a weapon, with the public unaware they might as well be viewing science fiction. The purpose is not to inform policy decision-making, but to influence, scare, nudge and manipulate the unwary with something that sounds sciency.

So, what do we make of the following finding in this NZ Herald news story?

Unvaccinated people are nine times more likely to seed fresh Covid-19 outbreaks in New Zealand, finds new modelling that’s prompted a plea for those who haven’t had their shots to think before travelling this summer.

The giveaway as to the real purpose of the modelling and story is in the second part of the sentence. Many assumptions were made by researcher Dr Leighton Watson, but here are the key ones mentioned in the news story.

it was much more likely that an infected person who’d had both shots of the Pfizer vaccine would not spread Covid-19 (54 per cent chance of not seeding a new outbreak) compared to an infected, unvaccinated person (6 per cent).

Watson explained this was because the vaccine was assumed to be 50 per cent effective at preventing onward transmission as well as 70 per cent effective against infection.

“The vaccine was assumed to be 50% effective…” highlights the point I’m making.  How do these assumptions stack up against real-world data and empirical research evidence? Not too well is the short answer.

Like New Zealand, Gibraltar is a small island nation that has claimed, with some pride, that its residents are fully vaccinated. In fact, more than 100%, given vaccinated Spanish commute to the island for work. Despite this, it has experienced a spike in breakthrough cases. They can’t blame the unvaccinated because they don’t have any unvaccinated to blame.

What accounts for the spike in cases is explained in empirical research done by Oxford University, published in Nature Medicine. When it comes to Delta the researchers found that

Delta infections after two vaccine doses had similar peak levels of virus to those in unvaccinated people; with the Alpha variant, peak virus levels in those infected post-vaccination were much lower.

In another study on community transmission, researchers found that while vaccination could reduce the risk of infection and accelerate viral clearance,

fully vaccinated individuals with breakthrough infections have peak viral load similar to unvaccinated cases and can efficiently transmit infection in household settings, including to fully vaccinated contacts.

Research from Israel, whose population is highly vaccinated with Pfizer (like New Zealand), showed similar findings. Israeli epidemiologist, Professor Nadav Davidovitch of Ben-Gurion University, found that the vaccine was only 39% effective against infection. He made the following observation,

What we see is that the vaccine is less effective in preventing transmission, but it’s easy to overlook that it’s still very effective in preventing hospitalization and severe cases.

Other more recent research, below, throws into question that second part. Vaccination might initially prevent severe illness; however, Swedish research showed that this waned quickly and by months 7-9, those vaccinated with Pfizer were as protected as the unvaccinated. Efficacy was less than 50% 4-6 months after the last Pfizer shot was administered.

The Oxford researchers note that in the UK there were rising cases of Delta even amongst the vaccinated. This reflects the experience in highly vaccinated places like Gibraltar.

Of particular concern is the Delta variant (B.1.617.2), which has caused sharp rises in infections in many countries, including some with relatively high vaccination coverage, such as the United Kingdom (UK).

In the NZ Herald article, Ministry of Health Chief Medical Officer, Andrew Connolly, conveniently chimes in with the following:

If you add in the fact that most vaccinated people don’t get the disease, the modelling suggests a more than 80 per cent reduction in household transmission.

Not so fast Mr Connolly. Published on the 19th of November in the Lancet is a paper with data from the UK, Germany and Israel showing vaccinated people now represent the majority of symptomatic cases. They note research data from the UK showing

secondary attack rates among household contacts exposed to fully vaccinated index cases was similar to household contacts exposed to unvaccinated index cases (25% for vaccinated vs 23% for unvaccinated)

Similarly, the US CDC identifies “four of the top five counties with the highest percentage of fully vaccinated population (99.9 – 84.3%) as “high” transmission counties.”

The researcher includes a review of an Israeli study of a “nosocomial outbreak” involving 16 healthcare workers, 23 patients and two family members. The source of the outbreak was a fully vaccinated person. 14 of the fully vaccinated patients became severely ill or died. The authors conclude the following:

Many decision-makers assume that the vaccinated can be excluded as source of transmission. It appears to be grossly negligent to ignore the vaccinated population as a possible and relevant source of transmission when deciding about public health control measures.

So let’s repeat the key findings for Mr Connolly and Dr Watson. “Delta infections after two vaccine doses had similar peak levels of virus to those in unvaccinated people.” In short, vaccinated people are just as capable of spreading the virus as the unvaccinated. In many highly vaccinated countries, the number of infected vaccinated is now more than infected unvaccinated. Contrary to repeated messaging that the jab will stop you from getting severely ill, that doesn’t look so certain anymore either. Not for the most vulnerable.

Remember when we were told that it was just a few weeks of lockdown to flatten the curve, that the vaccines were almost 100% effective and 70% vaccination would give us herd immunity?

The real purpose of the NZ Herald story and the modelling may be to counter growing suspicion and COVID fatigue in the population. The government needs to scapegoat the unvaccinated because they’re lifting internal border restrictions. What’s happened in places like Gibraltar will happen here too. As political tactics go, scapegoating is an oldie but no less despicable. The government needs a ready narrative to sandbag ‘grossly negligent’ policy and explain rising cases.