Ian Macleod – Discovery Health faces scrutiny over data transparency in actuarial conference

At the 2023 Actuarial Society of South Africa Convention, a surprising turn occurred when Roseanne Harris, a health policy actuary, expressed professional disheartenment after a presentation on Covid data. Criticising colleagues who question lockdowns, Harris claimed superior access to real-world data. However, her emotional remarks lacked evidence, drawing scepticism. Actuary Nick Hudson challenged her to share Discovery Health’s Covid data for scrutiny. The clash revealed a divide between emotional assertions and scientific rigour, emphasizing the importance of transparent data in challenging times. Here is a report on the event from Ian Macleod, a consulting economist and feature writer based in Cape Town.

Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox at 5:30am weekdays. Register here.


Ian Macleod: Discovery Health and apologetic actuaries: Put up, or shut up

By Ian Macleod

“I guess I feel quite sad listening to this,” opened the comment. That’s not the sort of gushy emotion one expects from a Q&A session at an actuarial conference. Especially after a 45-minute presentation covering the hard data behind death, illness and professional failure in unvarnished detail. But this is how an extraordinary comment began at the 2023 Actuarial Society of South Africa (ASSA) Convention in Sandton this October. 

Actuary and chairman of PANDA, Nick Hudson, had just completed his presentation, Actuarial and statistical problems around the Covid phenomenon, at the Sandton Convention Centre. After addressing the first few questions that followed his talk, one audience member, who identified herself as Roseanne Harris, piped up with “more of a comment than a question”. 

Her LinkedIn profile confirms she is a health policy actuary at Discovery Health and adjunct professor at the School of Statistics and Actuarial Science at Wits University. According to the ASSA website, she has also served as president of that association and as president of the International Actuarial Association. She’s a hotshot. 

After Harris’s unusual opening sentence, she went on to say she feels “professionally disheartened. While some may have been locked away in their ivory towers, trawling through secondary data and evidence, trying to desperately support an ill-conceived initial stance, I’ve been fortunate enough to be working with dedicated teams who have approached this problem with integrity and rigour and scientific curiosity, to unpack the evidence and contribute to saving lives and saving the economy.” 

There are some odd claims in there. For one thing, Discovery happens to occupy the shiniest ivory tower in all of Sandton. That’s fine. Some great work gets done in sparkling office buildings. But the statement is hypocritical at least. Especially for a company that coerced people to take a hastily approved drug with demonstrated severe side effects, including heart damage, to earn the privilege of entering their glassy castle at the top of Sandton Drive. 

The claim is also not accurate. Most people I know who make scientific and ethical arguments against lockdowns and coerced medical treatment do so far away from ivory towers. Most of us do so with the understanding that the establishment is against us. We do it for no pay, and at significant risk to our own careers. As a rule of thumb, ivory towers are unwelcoming to independent thought that challenges the official narrative. 

Stranger still is Harris’s suggestion that secondary data is somehow inferior. Her fellow actuaries and Wits academics surely use plenty of secondary data. It is different to primary data, and should be treated that way. It is not, as Harris’s statement suggests, inherently inferior. 

Harris also didn’t explain her disdain for those who use secondary data. She failed to identify the primary data that she uses. She failed to point out where anyone else has inappropriately used secondary data in a circumstance where primary data is required. 

This would prove to be the formula for the rest of her comment: sweeping claims with nothing to back them up.  

“I work for Discovery Health,” Harris continued. “And for us, Covid was real. Our team was dealing with real-world primary data. People who were sick, who died. People who lost loved ones. There were sixteen-and-a-half thousand confirmed Covid deaths in our population. We recorded a case fatality of 2.5% and hospital mortality of 15.5%. That is one death in every six-and-a-half admissions. Just under 82 000 of our members were admitted to hospital during Covid.”

This would have been Harris’s time to challenge, for example, Hudson’s claim that rapid introduction of untested new hospital protocols contributed to death rates. But we got nothing of the sort. Just the assertion. 

She went on, “I assure you that there was not a conspiracy of doctors admitting those members. There is no evidence in our data to support the conclusion that vaccines did more harm than good. Quite the opposite in fact. And transmission reduction was a valuable side effect of that programme that was intended to protect people from severe illness and death. These are real numbers, and I can’t help but feel the need to apologise to the many health professionals and everyone who has worked tirelessly and selflessly during Covid. And to everyone who was affected themselves or their loved ones. When we hear such hurtful misinformation.” 

Is that all? 

Cue some awkward applause. Some participants clearly endorsed Harris’s message. There seem to be a number of apologetic actuaries who endorse emotion over evidence. 

Others may have been scratching their heads as to the value of an emotional broadside without any evidence at a gathering of number crunchers. 

But that was it. Harris was done. Accusations without evidence and an emotional display. Straw manning her opponents as propagators of conspiracy theories. Rubbishing someone’s data without offering her own. Hardly actuarial. Hardly professorial. 

The irony is face-punchingly obvious. Harris could have picked on any of the allegedly “conspiratorial” evidence and broken it down. If it was so weak, she could surely have demolished Hudson’s presentation with ease. She could have applied her primary evidence – or any other evidence – to any of the data points that Hudson presented. But she did not. 

Harris was playing the role of cheerleader, not scientist. Cheerleaders are okay, as long as they’re identified as such. Then we can ignore them. But when masquerading as scientists, occupying influential roles, relying on secret data, and contributing to decisions that result in things like coerced medicine and universal lockdowns, blurring the line between cheerleader and scientist can be ruinous. 

As he always does, Hudson has made his presentation available for anyone to challenge. Here it is: https://pandauncut.substack.com/p/actuarial-and-statistical-problems. You’ll notice the detailed referencing for all sources. 

For current purposes I make no attempt to analyse the substance of this data-rich, often esoteric and complex presentation – fascinating and important, though that it is. For the record, I find most of PANDA’s work rather convincing. But for now, I address only the fundamentals of how good science is done – and how bad science is done. 

Based on the cornerstones of scientific integrity, we can comfortably say that Hudson and PANDA have behaved like scientists, and our Discovery actuary and Wits academic has not. In fact, Harris’s behaviour is an affront to science. 

Prove it

In this realm, you show your working. You publish, and then expect other scientists to challenge you. You relish it. It’s a thrill. Good scientific debate can be spicy. That’s often a good way to find better answers, provided you “play the man, not the ball”. 

But Discovery has rocked up at the game, pretended there was a ball, and claimed they scored a goal with it. They’ve promised they have footage of their stunning shot through the posts, but they won’t let anyone see it. That’s why Harris is forced to trumpet “these are real numbers” and “assure” us she’s right. 

To that, the genuine scientist’s response is a simple one: prove it. 

That is easily done. Just show us your data (anonymised to ensure patient confidentiality). For a highly successful company that proclaims itself to be “science-based”, this is a minimum requirement. Not an optional extra. You can hoard your evidence, or you can be a science-based organisation. You cannot do both. 

I have requested their anonymised data via official channels. They turned me down. This seems to be the standard response. 

Hudson confirms that PANDA has repeatedly offered to analyse the anonymised data supporting Discovery’s claims. And been denied each time. In Hudson’s words, this is a “posture that is incompatible with sound scientific protocol. Furthermore, Discovery has admitted to not disclosing their all-cause mortality analysis, ostensibly for fear that the public ‘would not understand it’.” 

In preparation of this piece, I asked Nick to confirm that he and PANDA remain committed to analysing Discovery’s data. He answered in the affirmative. 

The contrast between Discovery and Hudson is stark. The video of Nick’s entire presentation is available online. Every claim is referenced, and the presentation slide deck is freely available on the PANDA website (see link above).

The message that the scientific community must send to Discovery and Harris is this: Put up or shut up. 

I am left with several questions: 

For Discovery Health:

  • Will you share your anonymised data on Covid? 
  • If not, why not? 
  • If not, will you stop making claims based on data that nobody is able to verify? 

For Wits University and Discovery Health:

  • Do you endorse your staff/faculty publicly making evidence-free personal attacks on people sincerely presenting scientific arguments they disagree with? 
  • If not, do you intend to address Harris’s utterances at the ASSA Conference? 

For Roseanne Harris:

  • Will you identify each individual argument that Hudson made in his presentation and challenge it, using appropriate scientific arguments? 
  • Will you identify the arguments you believe are conspiratorial? 
  • If not, will you apologise to the ASSA members whose time you wasted with an emotional diatribe, and retract your accusations of dishonest science? That is, will you put up or shut up? 

For completeness – and contrast – here is the response Hudson gave to Harris:

“Needless to say, this is not the first time I’ve heard such a speech. Prepared, never responding to any of the analysis performed. Just stop [and] apply your minds to the data. That is what actuaries are meant to do how do. 

“How do we explain that in multiple regions around the planet, 2020 ended with very high seropositivity and no excess mortality? How can you explain that if we have a deadly virus on the loose and not an epidemic of catastrophically deteriorating standards of care? It’s very hard. Germany ends up with 10% or 20%, upwards in some areas, and zero excess mortality at the end of 2020. Zero. And there are entire swathes of the planet that had exactly this outcome. 

“And I take incredible exception to the idea that somehow a volunteer organisation of academics who are forced to work in private because the cancel culture led by organisations such as Discovery is so intense that they cannot do honest work in their own organisations. Those people are working for no pay and tirelessly and putting themselves at great risk. 

“There is no question in my mind… and this is not about a conspiracy theory. What an immature approach to take towards a person who sits there presenting statistics. It’s not a conspiracy theory. It’s an alternate view, and we can engage. I’d love to engage on your data and methodologies, but you don’t share it.” 

At that point the tension got too much for the MC, who called the session to an end.  

*Harris makes her comment at the 45-minute mark in the video. Hudson’s response begins at 47m35s. 

Read also:

Visited 5,594 times, 14 visit(s) today