Expert Witness Journal Issue 67 June/July 2026 - Flipbook - Page 85
Dealing 昀椀rst with the M-NIHL method, the Court
noted that this method had been superseded, two
years after publication, by the rM-NIHL method.
The revised method was necessitated by issues
which had been identi昀椀ed with the speci昀椀city of the
original method. Although he had said in evidence
that the rM-NIHL method should be preferred, Prof
Moore also suggested that, if the revised method
was not met, it would be appropriate to go back to
the original method to see if it yielded a positive
diagnosis. As to that, the Court, having set out in
detail the identi昀椀ed shortcomings in the M-NIHL
method, concluded:
(i) Pure Tone Audiometry (PTA) conducted in
compliance with the protocol and standards set by
the British Society of Audiology (BSA) is the gold
standard of audiometric testing and should be used
in medico-legal cases as the best evidence whenever
it is available.
(ii) However, screening audiograms, including
military screening audiograms, are ordinarily
suitable for screening and triage. Furthermore,
especially when they are part of a consistent pattern,
they may also be used, as part of the exercise of clinical
judgment by clinicians advising in medico-legal cases,
for diagnostic and quanti昀椀cation purposes.
350. I unhesitatingly reject the suggestion that, all
that notwithstanding, it might in certain cases be
reasonable, if a case did not receive a diagnosis
under rM-NIHL, to go back to the original method
on the basis that it has a positive predictive value
or ‘PPV’ greater than 0.5 (which, Prof. Moore said,
indicates the balance of probabilities was met.). It
was de昀椀ciencies in the original method, notably its
tendency to produce false positives, that prompted
the revised method and it would be nonsensical, if
not improper, to seek to advance a case on the basis
of what has been shown to be a de昀椀cient test, merely
because the better test did not produce the desired
outcome.
(iii) It may be possible to obtain from screening
audiograms an understanding of the hearing
pro昀椀le and trends of an individual (in other words
his auditory capabilities and the changes in those
capabilities) where there is a consistent pattern in an
audiometric series. That may not be possible where
the screening audiograms show signi昀椀cant and
unexplained variation.
(iv) Audiometry evidence (including BSA-compliant
PTA) should, wherever possible, be considered in
context and as a whole, rather than in isolation.
(v) The arrangements operated by the MoD are
designed to give a good degree of oversight over the
hearing capacity of members of the armed forces. The
majority of such audiograms are conducted properly
and in good faith. This is particularly so once a
referral for medical assessment is made. However,
there may, on occasions, be systemic or operational
inadequacies, or even deliberate falsehood, and the
Court should be alive to that possibility.
The MLP(18) method was addressed in similarly
robust terms by the Court.
MLP(18) is a machine learning AI methodology
which was trained by Prof Moore using a database of
audiograms obtained by Claimants making claims
for noise-induced hearing loss and a control database
of Germans not known to have had signi昀椀cant
noise exposure. It is fair to say that, during cross
examination, Prof Moore conceded that it was not
possible to tell why MLP(18) had given a diagnosis in
a particular case:
Method for Diagnosis of Military Noiseinduced hearing loss
The Court had before it four methodologies for the
diagnosis of noise-induced hearing loss in a military
context, three of which were authored principally
by Professor Brian Moore and put forward by the
Claimants, namely:
Q. It gives a result, doesn’t it, that you can’t say why it
has made a particular diagnosis in a given case? The
output is still the black box?
•
•
•
Q. So, for instance, an ENT surgeon who is using it
couldn’t – wouldn’t be in a position to interrogate or
challenge the result that comes out the other end?
A. That is true. That is true.
The M-NIHL 2020 method;
The rM-NIHL 2022 method;
The MLP(18) method.
The 昀椀nal methodology considered was the method
proposed by Coles, Lutman and Bu昀케n (CLB) which
has been in conventional use in cases of industrial
noise exposure since the Nottinghamshire and
Derbyshire Deafness Litigation. This was the
method relied upon by the Defendant.
A. Well, no – they wouldn’t – the result will come out
positive or negative, they won’t know why…
Having carefully set out the competing methods,
the Court concluded that the M-NIHL method and
MLP(18) method could be rejected in short order.
A. That’s correct.
EXPERT WITNESS JOURNAL
Q. …with the individual case of Mr Craggs, for
instance – his Lordship would not be in a position to
understand why that number has come out the other
end, would he?
83
JUNE 2026
