Is it just me?

Just zis Guy said:

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

errrr, "making it up"? ;-/

Pete.
--
Peter Clinch Medical Physics IT Officer
Tel 44 1382 660111 ext. 33637 Univ. of Dundee, Ninewells Hospital
Fax 44 1382 640177 Dundee DD1 9SY Scotland UK
net [email hidden] http://www.dundee.ac.uk/~pjclinch/

Peter Clinch said:

Just zis Guy said:

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

errrr, "making it up"? ;-/

Yes and no.. Data is collected by overstretched staff who don't want to be
sued so they go for a more extreme diagnosis. Thus many of the serious head
injury are not so serious and thus the final data collected is faulty.

...d

you know? said:

I've been having a jolly argument with someone on rec.bicycles.misc[1],
during which he posted the following:

If you look at <http://www.smf.org/articles/report.html>, for
example, the authors (who include ones Krygowski and Kunich
regularly disparage), you'll see a statement that "Overall,
helmets were found to prevent 69 percent of head injuries, 65
percent of brain injuries, and 74 percent of severe brain
injuries" but these authors use the following definitions:

* Head injury: All injuries to the forehead, scalp, ears, skull
and brain, including superficial lacerations, abrasions and
bruises on the scalp, forehead and ears, as well as skull
fractures, concussion, cerebral contusions and lacerations and
all intracranial hemorrhages (subarachnoid, subdural, epidural
and intra-cerebral).
* Brain injury: A diagnosis of concussion or more serious
intracranial injury, excluding skull fractures without
accompanying brain injury.
* Severe brain injury: An intracranial injury or hemorrhage,
including all cerebral lacerations/contusions, and subarachnoid,
subdural and extradural hemorrhages.

This is reminiscent of the Thompson, Rivara and Thompson finding that
helmets prevent 85% of head injuries and 88% of brain injuries.

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

Perhaps the helmet reduces the deceleration to a point below which
severe brain injury occurs, but that's still in the region in which
less sever injuries occur.

In fact, thinking about it further, if you posit that helmets almost
always have a positive effect, and that the number of people in the
groups decreases with severity of injury (actually 4 groups if you
count death), it might be explainable. IE, if of 100 accidents, 60
result in head injury, 30 in brain injury, 15 in severe injury and
5 in death, if you move 2 people from each group into the group below
you have a 2/5 decrease in the death rate, but only a 2/30 decrease
in the brain injury rate.

Sod it, of course there's 5 groups, including the non-injured, but
I can't be bothered rewriting the above.

--
Trevor Barton

Just zis Guy said:

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

Ummm... the helmet converts a proportion of serious injuries to
non-serious ones?

--
Mark.

you know? said:

I've been having a jolly argument with someone on rec.bicycles.misc[1],
during which he posted the following:

If you look at <http://www.smf.org/articles/report.html>, for
example, the authors (who include ones Krygowski and Kunich
regularly disparage), you'll see a statement that "Overall,
helmets were found to prevent 69 percent of head injuries, 65
percent of brain injuries, and 74 percent of severe brain
injuries" but these authors use the following definitions:

* Head injury: All injuries to the forehead, scalp, ears, skull
and brain, including superficial lacerations, abrasions and
bruises on the scalp, forehead and ears, as well as skull
fractures, concussion, cerebral contusions and lacerations and
all intracranial hemorrhages (subarachnoid, subdural, epidural
and intra-cerebral).
* Brain injury: A diagnosis of concussion or more serious
intracranial injury, excluding skull fractures without
accompanying brain injury.
* Severe brain injury: An intracranial injury or hemorrhage,
including all cerebral lacerations/contusions, and subarachnoid,
subdural and extradural hemorrhages.

This is reminiscent of the Thompson, Rivara and Thompson finding that
helmets prevent 85% of head injuries and 88% of brain injuries.

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

They don't claim (as far as I can see) that the difference between 69, 65
and 74% is statistically significant or provide confidence intervals. I've
only skimmed through the paper so I probably shouldn't be posting...

They say their sample size would need to be 53x bigger in order to
determine whether the difference between 81% and 72% injuries prevented
by Snell/Ansi helmets (worn by 54% v. 28% of riders respectively) is
significant or not, so my guess is that the data _can't_ say whether
helmets are more effective in preventing serious or non-serious brain
injuries or not.

I think I could make up semi-plausible scenarios for either case.

AC

Quoted message said:

The authors must have thought this through, surely, and come up with some
kind of rationalisation?

I think it's inherently implausible, far more likely to be a confounding
factor due to the self-selecting nature of the study groups, but maybe Im
missing something.

[1] Bill "Evasion" Zaumen, whose major debating technique is to hurl
playground insults and then sulk when they are returned in kind as an excuse
for not addressing substantive points which demolish his assertions :-)

Guy

On 29/7/04 3:19 pm, in article [email hidden], "Just zis Guy,

you know? said:

I've been having a jolly argument with someone on rec.bicycles.misc[1],
during which he posted the following:

I gave up rec.bicycles.misc many years ago. Such a waste of time and space.
It seems the same faces are still engaging in the same arguments..

...d

Mark Tranchant said:

Ummm... the helmet converts a proportion of serious injuries to
non-serious ones?

But the non-serious figure is smaller than the serious figure, so that
implies that there is differential effect, reducing with reducing severity
of injury. What I can't understand is that every other piece of PPE is
expected to have a much greater effect in preventing trivial injuries than
in preventing serious ones, but the helmet promoters do not even think to
question the analysis when it shows that helmets, alone among PPE, are more
effective in preventing serious injuries than trivial ones.

I haven't found any reference to this in any of the studies, but then again
I haven't gone back and re-read them looking for it.

Guy
--
May contain traces of irony. Contents liable to settle after posting.
http://www.chapmancentral.co.uk

88% of helmet statistics are made up, 65% of them at Washington
University

On 29/7/04 4:21 pm, in article [email hidden], "Just zis Guy,

you know? said:

Mark Tranchant said:

Ummm... the helmet converts a proportion of serious injuries to
non-serious ones?

But the non-serious figure is smaller than the serious figure,

ITYMTP:

If a serious becomes a non serious, and a nonserious becomes no injury then
there will be the same number of non-serious but less serious, ie there will
be a greater reduction in serious than nonserious.

Of course there will also be a reduction in deaths, leading to serious
instead.

but AFAICT the model doesn't support the data.

...d

In article said:

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

A helmet which effectively cushions severe accelerations, but is lined with
sandpaper, causing minor injuries? Oh wait, you said "credible".

David Martin said:

On 29/7/04 4:21 pm, in article [email hidden], "Just zis Guy,

you know? said:

Mark Tranchant said:

Ummm... the helmet converts a proportion of serious injuries to
non-serious ones?

But the non-serious figure is smaller than the serious figure,

ITYMTP:

If a serious becomes a non serious, and a nonserious becomes no injury then
there will be the same number of non-serious but less serious, ie there will
be a greater reduction in serious than nonserious.

But if there are far more non-serious injuries than serious (as there are)
then the serious injuries that are converted to non-serious injuries won't
be enough to significantly influence the non-serious injury statistics.

Head hurts...

AC

Quoted message said:

Of course there will also be a reduction in deaths, leading to serious
instead.

but AFAICT the model doesn't support the data.

..d

you know? said:

I've been having a jolly argument with someone on rec.bicycles.misc[1],
during which he posted the following:

If you look at <http://www.smf.org/articles/report.html>, for
example, the authors (who include ones Krygowski and Kunich
regularly disparage), you'll see a statement that "Overall,
helmets were found to prevent 69 percent of head injuries, 65
percent of brain injuries, and 74 percent of severe brain
injuries" but these authors use the following definitions:

* Head injury: All injuries to the forehead, scalp, ears, skull
and brain, including superficial lacerations, abrasions and
bruises on the scalp, forehead and ears, as well as skull
fractures, concussion, cerebral contusions and lacerations and
all intracranial hemorrhages (subarachnoid, subdural, epidural
and intra-cerebral).
* Brain injury: A diagnosis of concussion or more serious
intracranial injury, excluding skull fractures without
accompanying brain injury.
* Severe brain injury: An intracranial injury or hemorrhage,
including all cerebral lacerations/contusions, and subarachnoid,
subdural and extradural hemorrhages.

This is reminiscent of the Thompson, Rivara and Thompson finding that
helmets prevent 85% of head injuries and 88% of brain injuries.

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet,

You're implying that it seems implausible that helmets prevent 69% of head
injuries if <50% of head injuries are to the areas covered by a helmet?

The authors seem to be of the opinion that helmets can protect against
a large proportion of facial injuries in areas that are not covered by
helmets. It would be nice to know how this is supposed to work - I agree
it seems surprising.

Quoted message said:

can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

Could you have answered half of your question? Here head injuries include
facial injuries, and the study shows that helmets have no protective
effect for injuries to the lower face (e.g. jaw) - which seems plausible.

If helmets are designed to protect against brain injuries (major)
rather than facial injuries (minor) then this might explain why they are
seemingly equally good at preventing both - if minor injuries are
ordinarily easier to prevent.

As for severe / non severe brain injuries - I vaguely recall something
about helmets not 'crushing' enough to absorb minor impacts. If they do
crush enough to turn major impacts into minor impacts, then this could
explain this finding.

Quoted message said:

The authors must have thought this through, surely, and come up with some
kind of rationalisation?

All my arguments above are hand-waving - I wouldn't want to argue them in
a paper.

AC

Quoted message said:

I think it's inherently implausible, far more likely to be a confounding
factor due to the self-selecting nature of the study groups, but maybe Im
missing something.

[1] Bill "Evasion" Zaumen, whose major debating technique is to hurl
playground insults and then sulk when they are returned in kind as an excuse
for not addressing substantive points which demolish his assertions :-)

Guy

Quoted message said:

Leaving aside the fact that Australian data suggests that under half
of head injuries are to the areas covered by a helmet, can anybody
think of a credible mechanism by which helmets could be more effective
against serious brain injuries than against non-serious brain injuries
or all head injuries, including trivial ones?

Read/watch Yes Minister - it has an entire episode devoted to the subject of rubbishing
studies & reports.

Brain Injuries Vs head injuries should be easy to resolve. The helmet prevents
lacerations etc. while not necessarily preventing concussion & other brain damage. The
hard shell protects against the head scraping along the ground quite well, but isn't so
good at preventing the brain bouncing about the inside of the skull. Also, their
definition of head injuries also includes injuries to the brain. Are two boxes allowed
to be ticked on the forms?

Also, unless the population sample is 100% of the population then all sorts of errors can
creep in. But I'm guessing you know more about that side than me :-) Still, as long as
they didn't use a self-selecting sample they'll hopefully be able to minimise this.

Quoted message said:

I think it's inherently implausible, far more likely to be a
confounding factor due to the self-selecting nature of the study
groups, but maybe Im missing something.

Ah.

anonymous coward said:

Head hurts...

Not wearing your helmet?

d.

Just zis Guy said:

Leaving aside the fact that Australian data suggests that under half of head
injuries are to the areas covered by a helmet, can anybody think of a
credible mechanism by which helmets could be more effective against serious
brain injuries than against non-serious brain injuries or all head injuries,
including trivial ones?

There's a very high chance that a severely crashed head is going to have
some minor head injuries (only meaning facial cuts and bruises) even
with a helmet. Of course in this case the helmet will also probably have
actually prevented lots of minor injuries as well but they are not
counted as individual events within a single crash, are they?

The really really minor bangs that a helmet might completely protect
against will probably not have been reported in any case, helmet or not.

James
--
If I have seen further than others, it is
by treading on the toes of giants.
http://www.ne.jp/asahi/julesandjames/home/

On Thu, 29 Jul 2004 17:37:43 +0100, anonymous coward
<[email hidden]> wrote in message
<[email hidden]>:

Quoted message said:

But if there are far more non-serious injuries than serious (as there are)
then the serious injuries that are converted to non-serious injuries won't
be enough to significantly influence the non-serious injury statistics.

Serious are only about 2% of non-serious hospital treated, I think.

But if there is an exponential decay from trivial to fatal (which is
how I've always pictured the injuries being distributed), and helmets
move the entire distribution to the left by some meaningful amount,
there would be a huge number of trivial injuries eliminated and a much
smaller proportion of serious ones (by whch time the curve is more
nearly flat).

Guy
--
May contain traces of irony. Contents liable to settle after posting.
http://www.chapmancentral.co.uk

88% of helmet statistics are made up, 65% of them at Washington University

Quoted message said:

If you look at <http://www.smf.org/articles/report.html>, for
example, the authors (who include ones Krygowski and Kunich
regularly disparage), you'll see a statement that "Overall,
helmets were found to prevent 69 percent of head injuries, 65
percent of brain injuries, and 74 percent of severe brain
injuries" but these authors use the following definitions:

Its a very difficult statistic, as how do you know that the remaining
"prevented" injuries would have happened?

Also, its expressed as a percent. The real picture could be that out of
hundreds and hundreds of cycle accidents, the "brain injury" count was
reduced from 3 to 1, so with those numbers perhaps helmets "prevented 66%
of brain injuries", or was it just statistical noise?

Also note that the bands include higher bands, so a serious brain injury
(compression perhaps) would count towards the statistics for "Head Injury",
"Brain Injury" and "Severe Brain Injury" all at once. Knock yourself out
for more than 2 minutes, and you've had 3 accidents by these books!

Also, and quite obviously, the thresholds are very low. Isn't concussion "I
felt dizzy for a bit"? or maybe knocked out for seconds/a minute or two? I
suppose technically it is a brain injury. I've had various "minor
lacerations" doing various things - you'd think I'd have to wear a helmet
all the time! I'd expect a helmet, or any hat like thing for that matter,
to be quite effective against minor lacerations, unless it causes them
itself.

Such high numbers as these do look too good to be true. Someone's trying to
overimpress with a point that they want to make, which is a shame, because
it means that the statistics are completely useless for anyone trying to
make an informed choice. At least they do define the statistics, rather
than just quote the numbers blindly, though it only takes the second
readers to start misquoting or incompletely quoting.

- Richard

(For dodgy statistics, it's probably comparable with Surf's "Cleans 99
stains". Is that 99 out of 100, or 99 out of a few million? What about the
special conditions that effectively say you carry a portable washing
machine with you? I expect most people read the implicit percent sign and
associate a warm feeling of goodness. Its a bit like those people that see
"60% Fat Free" and ignore the complex mathematical bit, but Surf always
ensure the percent sign is never there, so the ASA rejected my complaint
about that advert)

--
_/_/_/ _/_/_/ _/_/_/ Richard dot Corfield at ntlworld dot com
_/ _/ _/ _/
_/_/ _/ _/ Time is a one way street,
_/ _/ _/_/ _/_/_/ Except in the Twilight Zone.

Just zis Guy said:

But if there is an exponential decay from trivial to fatal (which is
how I've always pictured the injuries being distributed)

*Bzzt!* Assumption!

Any data for this? Seems like common sense, but then so do helmets... ;-)

--
Mark.

Mark Tranchant said:

Quoted message said:

But if there is an exponential decay from trivial to fatal (which is
how I've always pictured the injuries being distributed)

Quoted message said:

*Bzzt!* Assumption!
Any data for this? Seems like common sense, but then so do helmets...

All injuries ca. 100,000
Hospitalisations ca. 2000
Known Serious injuries ca. 500
Fatal ca. 25

Guy
--
May contain traces of irony. Contents liable to settle after posting.
http://www.chapmancentral.co.uk

88% of helmet statistics are made up, 65% of them at Washington
University

James Annan said:

The really really minor bangs that a helmet might completely protect
against will probably not have been reported in any case, helmet or
not.

Surely that's precisely what the study methodology must have been designed
to measure - injuries which never happened?

Guy
--
May contain traces of irony. Contents liable to settle after posting.
http://www.chapmancentral.co.uk

88% of helmet statistics are made up, 65% of them at Washington
University