IARC on Glyphosate –
what to do when a mistake is made?
The Governing Council of the International Agency for
Research on Cancer (IARC[1])
is meeting[2]
today and tomorrow.
Not listed on the published agenda is glyphosate, but much
of the conversation will be about the hotly disputed decision[3]
that glyphosate is ‘probably carcinogenic to humans’ (Group 2A). Was it the
right finding, why was so much of the animal experimentation evidence deemed
unsuitable for consideration, is it ethical to make pronouncements of any sort
if there is no published evidence of how often IARC decisions are wrong? How
should scientific expert opinion be held to account? Who underwrites the effect
of mistakes? Is it ethical not to take responsibility for mistakes?
Holding IARC to account
It seems obvious in hindsight that institutions of all kinds
whether commercial or public should publish an account of how accurate their
published findings are[4].
In time, false positives, false negatives, true positives and true negatives
can be counted and simple statistics can be generated[5].
Has anyone done this for IARC’s Monograph series?
One of the problems in producing such a validation test is
what to do about chemicals where no pronouncement is made. If chemical x has
not been pronounced upon is this a false negative or a true negative?
Fortunately, the Positive Predictive Value[6]
(PPV) is not influenced by this problem and is the statistic which is
most needed.
Why is PPV important to the risk management industry?
When IARC say “is carcinogenic” or “is probably
carcinogenic” some regional authorities[7]
respond with requirements to publish warnings and this inevitably has an effect
on liability risk. Sometimes the IARC view leads to an automatic response[8],
sometimes the IARC view is further judged before action is triggered. I argue
here that PPV values for “is carcinogenic” and “is probably carcinogenic”
should in principle influence the form of the response. Some proposals are made
below.
For the purposes of liability risk management, if the
positive predictive value is below 50% then on average all pronouncements are
more likely wrong than right. I propose that in effect the public authority
position should be that implied common law responses be ruled out
automatically. If a warning issued by IARC or anyone else is ‘probably wrong’
then a warning label would probably misinform the public or other intended
audience. This cannot be in the public interest. If ruled out, generators of
risk and their insurers would continue to be free to apply their own assessment
and be prepared to justify it. How well is IARC doing? What are their PPVs? Surely
this is a matter that should be on today’s agenda.
For individual substances, if one authoritative body
disagrees with IARC then common law responses should in the same way be ruled
out automatically. One vs one cannot
pass the balance of probabilities test; a warning would be as likely wrong as
it is right and the public is not informed. WHO should say that ‘reasonable
people armed with the same evidence can come to opposite conclusions’. When
this happens, the IARC finding should be labelled as “pending review”.
In case of glyphosate, IARC is disagreed with by just about every
independent authority! The very least that IARC should do, in the public
interest, is to suspend the finding until a definitive view is
forthcoming. What will the Governing Council do today? And, do regional
authorities have the regulatory capacity to suspend regulatory action if IARC
don’t first suspend a finding? Would it be unethical for IARC to respond to independent
expert review?
Of course, liability risk management consultants should face
the same sort of scrutiny when they publically identify liability issues. But
this rarely happens[9].
If the PPV is less than 50% one logical view is that it is
automatically unreasonable to make an accounting provision such as a reserve. The
PRA[10]
would indicate their displeasure. Even if a deeper evaluation reveals a given
item to be a true positive, the question for liability risk managers is, how
many “predictions” can you afford to analyse in order to find one good one?
Dr Andrew Auty
[1]
The International
Agency for Research on Cancer was established in May, 1965, through a
resolution of the XVIIIth World Health Assembly, as an extension of
the World Health Organization (WHO).
[2] http://governance.iarc.fr/GC/GC61/index.php
[3] https://www.iarc.fr/featured-news/media-centre-iarc-news-glyphosate/
(2015)
[4] Recent audit scandals for example have drawn attention
to the need for the “big four” to be held to account for their false findings.
How often do they mislead shareholders? How are they held to account for their
errors?
[5] https://www.medcalc.org/calc/diagnostic_test.php
[6] PPV is the ratio of true positives to the sum of true
positives and false positives.
[7] Perhaps the best known is Proposition 65 in
California. https://oehha.ca.gov/proposition-65
[8] https://oehha.ca.gov/proposition-65/crnr/regarding-certain-iarc-international-agency-research-cancer-2b-chemicals
[9] Readers may like to know that Re: Liability (Oxford)
Ltd has developed a methodology to make such an evaluation and has applied it
to a well-publicised market report which contained more than fifteen liability
“predictions”. A full report on this is available for a fee. Contact: andrew@reliabilityoxford.co.uk
.
[10] In the UK, financial services firms are regulated by
the Prudential Regulation Authority (PRA) https://www.bankofengland.co.uk/prudential-regulation
