Statistics

0
Research results must be judged in relation to the study's methodology.

In his latest blog, titled “Who are you going to trust?”, anti-assisted dying lobbyist Mr Paul Russell says:

“Polling noted today in the Australian shows a significant level of distrust in our political classes to get the issue of euthanasia and assisted dying right.”

He then goes on to quote some select statistics from said poll. In his blog, he mentions nothing about the sponsorship or conduct of the poll. After some searching, I found no other reference to said poll on his ‘HOPE’ website.

This is rather curious, because The Australian article he quotes, points out that the ‘poll’ was commissioned by him (his website is called ‘HOPE’).

Thus, Mr Russell tries to add credibility to his ‘poll results’ in his blog by citing only that it has been reported in a national newspaper. This ‘quote-someone-else-so-it-must-be-authoritative’ rhetorical strategy has been used before by opponents of assisted dying (see Box at end).

But as Mr Russell has himself promoted — happily republishing the opinion of the CEO of Christian Medical Fellowship (UK) — “opinion polls add up to very little.” That’s quite true… when they’re poorly designed and run, including the big no-no, ‘push-polling’, in which the researcher attempts to get the answer they want by crafting questions more likely to get it.

I searched hard for any reference to the methodology of said ‘poll’, but was unable to identify any despite a diligent search. Therefore, we don’t know what approach Mr Russell took: robust or otherwise.

Let’s assume for the sake of argument (and the absence of public evidence) that a poll of some kind was actually conducted. If it were a truly legitimate poll, you’d think that Mr Russell would be shouting about the study from his own rooftop (the ‘HOPE’ website). But so far, he hasn’t.

Mr Russell, while quoting statistics, has said absolutely nothing about the methodology — that I can find via a quite diligent search.

Results can only be interpreted in light of how the research was actually conducted, so quoting a 'study' while failing to publish its methodology in full is an absolute no-no. It only invites derision.

Rebecca Urban, for The Australian, quotes a number of ‘statistics’ from the ‘poll’ seemingly without question. But she’s hardly to blame: she’s skilled at journalism, not primary research.

So, for the benefit of Paul Russell, Rebecca Urban and all journalists reporting claimed statistics, here’s your minimum standard of conduct if the public are not to guffaw at the claims. All reported results must be in relation to properly disclosed methodology:

  • Who commissioned the research? (✔ Ms Urban reports who)
  • Who conducted — actually carried out — the research (e.g. a reputable research company)?
  • What precise population were respondents drawn from, how were they recruited, and screened in or out? What were the counts and percentage participation (approached/participated)?
  • What were the dates of the fieldwork?
  • What procedures were used to establish and maintain the authenticity of who was sampled (e.g. if an online poll, could people from anywhere technically participate in this Victorian poll)?
  • How was the questionnaire administered (e.g. paper self-complete, online, CATI)?
  • What was the script of stimuli administered to respondents? In other words, what prompts were given and what questions were asked: exact order and wording?
  • What results were obtained for each question (i.e. full rather than selective crosstabs)?
     

Until Mr Russell publishes in full how his ‘poll’ was conducted, the only honourable course of action for him to pursue is to withdraw the claimed results.

Until then, we can only see them as untrustworthy and a bit of a joke.

 

Rhetorical tactic — “Not” quoting yourself

This rhetorical tactic is also used by Mr Russell’s fellow Catholic, Prof. Margaret Somerville. For example, in her 2015 book Bird on an Ethics Wire, in relation to the supposed (but fanciful) fear of being euthanized in the Netherlands if adequate pain management is accepted, Somerville says in Chapter 4:

It has been alleged that Dutch physicians have interpreted patients’ consent to pain management as consent to euthanasia.38

If you’re like most people, you’d assume, given the effort of a citation (38), that an independent source had made the statement based on some evidence. Indeed, if you look at reference 38 you’ll see that the author is Lauren Vogel, and the source article is in the Canadian Medical Association Journal. All sounds like solid, legit stuff, doesn’t it?

However, Ms Vogel is a journalist, not a Dutch medic or a researcher, and what she reports in relation to Somerville’s claim is merely a quote of what someone said. And who is that someone? Why, it’s Margaret Somerville — what a coincidence!

Somerville could have just said “I’ve argued this before…”, but instead gives a seemingly robust reference to a source that has the appearance of independence and scholarship. Yet obviously she knows that the source is merely herself saying so.

Let’s be clear: something is not true just because someone alleges it. Even if they allege it twice or more. And happen cite themselves via someone else in the process.


Share This Post:
0
The Age reports a 'gloves off' campaign of misinformation

Both the Herald Sun and The Age reported last week that religious anti-assisted dying crusaders are running a 'gloves off' campaign in Victoria.

Religious forces are gathering once again to attempt to thwart the views of the great majority of Victorians in favour of assisted dying law reform.

Matt Johnston in the Herald Sun quoted Paul Russell, a long-term figure in Catholic circles, and Greek Orthodox Bishop Ezekiel, in statements against assisted dying.

Farrah Thomazin in The Age quoted religious stalwarts Margaret Tighe of Right To Life, and the Australian Christian Lobby, in further statements against assisted dying.

The crux of the story is that 'pollsters' claim to have run a survey in Victoria. They refuse to be identified. They refuse to publish their methodology. And they refuse to publish all their results. Enough said.

They cherry-pick an item from their supposed poll to claim that 33% of Victorians who oppose assisted dying will change their vote against a supporting politician at the next election. They neglect to mention that only a tiny minority of Victorians actually oppose assisted dying. Their analysis is astonishingly superficial, even assuming they ran a proper, robust poll and didn't manufacture the numbers themselves.

They then use this tidbit of 'data' to put the fear of electoral defeat into politicians who will soon to face an assisted dying Bill in the Victorian Parliament.

What rubbish. Assisted dying (AD) opponents seem to be utterly shameless in misrepresenting and distorting cherry-picked data to push their religious agenda — which they pretend isn't religious.

The real situation in respect of AD is the exact opposite of their claims as I show in a proper, robust analysis of legitimate data, demonstrating that:

  • A massive 78.9% of Victorians support AD, with only a tiny 8.1% opposed. Strong supporters outnumber strong opponents by more than ten to one.
  • Significantly more supporters of AD believe that law reform is personally important, than opponents believe the status quo (no law) is personally important.
  • At a general election, far more Victorian voters will punish Members who oppose the AD Bill than will punish Members who support it (3.5 to 1 overall, 2.4 to 1 for the Liberal/National Coalition and 6.6 to 1 for Labor).
  • The co-sponsors of Victoria’s 2008 AD Bill were returned with greatly increased majorities (including relative to their party’s overall performance) despite campaigns against them by anti-AD crusaders.

 

You can read the full analysis here.

So that Victorian politicians are not misled, I have forwarded my report to the Victorian Government's Cabinet and other selected members of Parliament.

The only way in which this campaign could be called 'gloves-off' is that opponents, lurking around with their shadowy misinformation, don't want to get bullshit on their mittens. Hands seem to be much easier to wash. And hide.


Share This Post:
0
Dr Megan Best and the AMA profoundly misrepresent what this research says.

In my report about Dr Megan Best misreporting research findings about assisted dying practice in Belgium, I said that I would ask the Australian Medical Association (AMA) to publish a correction statement. I did just that, posting a courteous notice explaining the misinformation and requesting the correction, as a comment to their online article. What happened next was supremely revealing about how the conservative medical establishment opposes your right to choose assisted dying. I expose six reprehensible aspects of Dr Megan Best’s ‘expert’ opinion and MJA InSight’s response to my report.

 

Response 1—Try to bury the correction request

How did the AMA respond? Well, first of all, they simply deleted my post from the Medical Journal of Australia article (in MJA InSight).1

Subsequently, AMA member Dr Rosemary Jones posted my advice and request for correction. They dared not delete the post of a member, because Dr Jones’ post remains and MJA InSight have now responded… in a fashion.

Response 2—Plead it’s not the AMA

First up, Cate Swannell, editor of MJA InSight as well as “MJA news and online,” wrote that MJA InSight is:

“…editorially independent from the AMA and the AMA does not influence our content and editorial decisions in any way, shape or form … Thanks.” — Cate Swannell

It would be fair to take Ms Swannell’s word for it that the AMA doesn’t make a habit of sticking its nose into individual items at the Medical Journal of Australia (MJA). But consider this: MJA, of which MJA InSight is a part, is published by a wholly owned subsidiary of the AMA. That is, even in the absence of direct influence, one might reasonably expect that the AMA has broader, long-term influence over “content and editorial decisions.” After all, a core purpose of establishing a wholly-owned subsidiary is to help promote owner’s interests (which at present are officially and trenchantly opposed to the legalisation of assisted dying).

Those interests are further intertwined at DoctorPortal which is co-owned and operated by the AMA and its entity that publishes MJA. The portal includes the ‘MJA Bookshop’ where doctors must log in using their AMA credentials to obtain discounts.

MJA InSight is in real and practical terms a publication of the AMA.

Refresher: The original allegations

Before we cover the further AMA responses, let’s remind ourselves of the ‘expert opinion’ the AMA obtained from Dr Megan Best about Belgian assisted dying practice as published in MJA InSight:

“I am concerned by the reduction in referral to palliative care doctors and specialists in the euthanasia approval process, as GPs are less likely to know whether or not the suffering can be alleviated – the keystone of the act,” she said.

“This is a weakening of the due process of the act and suggests that the ‘safeguards’ are seen more as a barrier to be overcome than an opportunity to improve life to the extent that euthanasia is no longer necessary.” — Dr Megan Best

Response 3—Confirm the ‘cherry-picking’ misrepresentation

Here’s how Sarah Colyer, author of the article in which Dr Best’s nonsense was quoted, responded to Dr Rosemary Jones’ objections:

“…there are two findings related to the involvement of palliative care doctors in the study. First, the study found palliative care physicians were less likely to be the legally-required ‘second doctor’ to sign off on the decision for euthanasia, as more GPs took on the role.” — Sarah Colyer

What a profound fudge. The paper does not say that. There is a notional mathematical decrease in palliative care (PC) ‘sign-offs,’ but with a p value of 0.3 it is not in the slightest bit statistically significant. In other words, it is not possible to validly assert that it became “less likely.” And this is the reason that the researchers correctly didn’t draw such an inference in their discussion and conclusions.

This is extraordinary. In its defence, this AMA publication confirms and defends that Dr Best had cherry-picked a supposed trend with no statistical significance and which seems to (but doesn’t) support opposition to assisted dying law reform, while utterly failing to mention a highly significant trend (p = 0.001) in the opposite direction, present in the very same table of data (Table 3 of the cited research paper).2

While the original MJA article reported on both data series, it failed to identify that one was in no way statistically significant while the other was highly so.

Response 4—Execute an impressive flip-flop

“Second, and as Sarah Edelman is quoted saying, 39% of patients had contact with palliative care teams in the lead up to euthanasia (but these doctors were not necessarily involved in sign-off).” — Sarah Colyer

Both Dr Megan Best, and Sarah Collyer in defence, express concern about what they see as the insufficient ‘signing off’ or approval of assisted dying by Belgian PC specialists. Yet Palliative Care Australia explicitly states that:

“The practice of palliative care does not include euthanasia or physician assisted suicide. Palliative care does not intend to hasten or postpone death.” — Palliative Care Australia3

So on the one hand Australia's national PC body demands that PC have nothing whatever to do with assisted dying, while an Australian PC specialist (who is opposed to assisted dying) demands that it must: not just ‘incidental’ involvement to ensure the patient is well-informed as to what PC may offer, but to directly insert itself in the assisted dying approval process.

What an impressive ‘must not, yet must’ flip-flop.

Response 5—Confirm bias about consultations

Sarah Colyer goes on to say:

“Although this latter figure is an increase from 33%, the article makes clear that this is more about palliative care doctors being ‘notified’ than necessarily having any professional input into the decision-making process.” — Sarah Collyer

If my point 3 above exposed a ‘profound’ fudge, this is a spectacular one. Indeed it’s utter fiction. Nowhere in the cited research report do the data or authors suggest in any way that PC teams were merely ‘notified.’ They were consulted. The word is used explicitly in no fewer than seventeen times specifically in relation to PC. What is a professional ‘consultation’?: “a meeting or discussion to obtain advice.”

To suggest that these are mere ‘notifications’ is absurd: why would a physician even bother to formally ‘notify’ other relevant carers if he or she had no intention of the notification having any meaningful consequence?

The only outcome of this ‘notification’ fiction is to make Belgian physicians seem careless or disinterested—a shabby and biased approach to describing overseas colleagues who engage in lawful conduct of which the commentators (at least, Dr Best) personally disapprove.

Response 6—Demonstrate fundamental ignorance about research

Remember that Dr Best complained about her perception of a low assisted dying PC consult rate as obstructing the “opportunity to improve life to the extent that euthanasia is no longer necessary”. Sarah Colyer defends and amplifies the ‘low consult rate’ message of the Belgian research:

“Although Dr Edelman and the study authors said this [increase in palliative care consults] was a positive development, Dr Best and another respected palliative care doctor who provided background for the article said it was still overall a very low rate of palliative care involvement.” — Sarah Colyer

These remarks reveal a deep ignorance about research evidence. The reported data the 'experts' criticised was only in respect of PC involvement after a request for assisted dying. It did not include any PC consults that occurred in the patient’s care trajectory prior to the assisted dying request. (Indeed, the assisted dying request may in some cases have developed as a consequence of palliative futility.) The research authors are even explicit about this:

“The reporting form does not record whether patients have previously received palliative care.” — Dierickx, Deliens, Cohen et al2

Contrary to Best and Colyer’s indefensible assumption, empirical research evidence indicates that around 74% of assisted dying cases in 2013 at least in Flanders, the Dutch-speaking larger half of Belgium, involved specialist PC services.4 The Dierckx et al2 data is also in respect only of PC-primary-specialist consults. Many Belgian physicians, especially but not only many oncologists, themselves hold PC sub-specialty endorsements, yet only their primary organ or disease specialty is recorded on the report forms.

The 74% PC consult figure of course by definition excludes those who made a request but who subsequently rescinded it (i.e. not captured in the data because there was no assisted death). The patient may have rescinded their request as a consequence of PC consultation. Other published scientific research indicates that in 2013 around 18% of Belgian assisted dying requesters revoked their request.5

Indeed, more than half (59%) of Belgian patients making a request in 2013 died before the consulting team had reached a decision, and 7.5% of requests were denied by physicians because the patient wasn’t terminally ill even though that is not a requirement of the law.5 The wider evidence suggests that Belgian doctors remain conservative and cautious rather than pursuing poor practice as Dr Best and Ms Colyer invalidly attempt to conclude.

What is “very low”?

Just for good measure, let’s compare a Belgian assisted dying 74% PC consult rate with the general PC rate in Australia (where assisted dying is illegal). There’s precious little data around and Palliative Care Australia didn’t respond to two requests. [Addendum 5th Oct: Palliative Care Australia responded but advised that to the best of their knowledge the specific statistics were not available.] The closest relevant official data I could find with hours of intensive research, reports that in 2011-12, 39.5% of hospital inpatients who died received PC consults.6 Of course, PC consults can only be relevant in respect of non-sudden deaths because there’s no opportunity to make such decisions in a relatively sudden death. About two thirds of all deaths are not ‘sudden,’ so let’s adjust up the original figure to reflect that, resulting in a hospital non-sudden-death PC consult rate of around 60%.

Half (50%) of all deaths in Australia occur in hospitals, 38% in residential and community (i.e. other institutionalised) care, and just 12% elsewhere (mostly at home).6 Let’s say for the sake of argument (and lack of data) that the PC consult rate for other institutions is similar to hospitals. Let’s also assume that of (two thirds non-sudden) deaths at home, all of them received palliative care consults. That raises the Australian PC consult rate to around 68% of all non-sudden deaths. Maybe the real figure is somewhat higher or lower.

But it is wholly incorrect to conclude, given Belgium’s Flemish PC consult rate for assisted dying at 74%, that the Belgian palliative care consult rate is “very low” as the two ‘expert’ PC commentators wrongly assert and Sarah Colyer reports on the basis of data about a different measure.

Conclusion

The AMA widely promotes evidence-based decision making, but it has comprehensively offended its standard in regard to this issue. The AMA’s response to this further critique will give the Australian public a strong indication as to its intent, ethics and credibility. Of course, I'll let you know how they respond.

As for Dr Megan Best offering such ill-considered and evidence-defying nonsense as ‘expert opinion,’ I would hope that she permanently retires from offering commentary.

Now at least you can see the kinds of rubbish arguments that the conservative medical establishment uses to justify opposing your right to consider an assisted death.

 

Inexpert opinion via the AMA’s MJA

Through its MJA InSight article and response, the AMA has published and defended misinformation against assisted dying, a choice that the overwhelming majority of Australians want on the table, but which the AMA explicitly opposes.

  1. First up, the AMA tried to bury my critique of its incoherent and incorrect ‘expert opinion’ by deleting the critique.
  2. Then it tried to argue that its published misinformation was nothing to do with the AMA.
  3. It confirmed that it had reported a cherry-picked non-significant ‘trend’ from a research paper’s data that was, coincidentally, consistent with AMA opposition to assisted dying, while omitting to mention a highly statistically significant trend—in the same table of data—that contradicted the AMA and commentators’ stance.
  4. Its article and critique response argued that PC must be involved in approving assisted dying requests at the same time that Palliative Care Australia insists that assisted dying is nothing to do with PC.
  5. It fictionally asserted that the Belgian research paper suggested PC specialists were merely ‘notified,’ while the research paper said no such thing and referred repeatedly to ‘consultation.’
  6. It ignorantly claimed that the Belgian PC consult rate was “very low” (when it isn’t) on the basis of what happened only after an assisted dying request, ignoring that PC consults can and do occur before an assisted dying request.

 

References

  1. Colyer, S 2016, Belgian euthanasia model splits Aussie experts, MJA InSight, viewed 26 Sep 2016, https://www.doctorportal.com.au/mjainsight/2016/37/belgian-euthanasia-model-splits-aussie-experts/.
  2. Dierickx, S, Deliens, L, Cohen, J & Chambaere, K 2016, 'Euthanasia in Belgium: Trends in reported cases between 2003 and 2013', CMAJ, September 12, 2016.
  3. Palliative Care Australia 2016, Euthanasia and physician assisted suicide: Position statement, viewed 1 Sep 2016, http://palliativecare.org.au/download/2448/.
  4. Chambaere, K, Vander Stichele, R, Mortier, F, Cohen, J & Deliens, L 2015, 'Recent trends in euthanasia and other end-of-life practices in Belgium', N Engl J Med, 372(12), pp. 1179-1181.
  5. Dierickx, S, Deliens, L, Cohen, J & Chambaere, K 2015, 'Comparison of the expression and granting of requests for euthanasia in Belgium in 2007 vs 2013', JAMA Internal Medicine, 175(10), pp. 1703-1706.
  6. Swerissen, H & Duckett, S 2014, Dying well, Grattan Institute, Carlton, Victoria, pp. 37.

 

Share This Post:
Subscribe to RSS - Statistics