I despair, I really do.

[Tony Raven]

Nick Maclaren wrote on 29/04/2007 11:14 +0100:
> |>
> |> I don't think that is correct. Rogers in the US, with statistics on 8
> |> million cyclists found that "the bicycle-related fatality rate is
> |> positively and significantly correlated with increased helmet use".
>
> That doesn't mean quite what you seem to think that it means.

Come on then Mr Cryptic. Explain what it does mean.

>
> What evidence there is, is that the physical factors related to helmet
> wearing are nil to within the experimental accuracy, at least for all
> normal forms of cycling. And the evidence for even a psychological
> effect is VERY weak, except via the mechanism of discouraging cycling
> (which puts up the risk).
>
> Or that is what I found when I last looked.
>

Maybe you should look again. Perhaps you should look at the work by
John Adams, Mayer Hilman, Ian Walker and Brent Hagel (yes Brent Hagel
the mandatory helmet promoter)
Mok D, Gore G, Hagel B, Mok E, Magdalinos H, Pless B., Risk compensation
in children’s activities: A pilot study; Paediatr Child Health.
2004;9(5):327-330.

--
Tony

"The most savage controversies are those about matters as to which there
is no good evidence either way."
- Bertrand Russell

 

[Peter Clinch]

Tony Raven wrote:

> The Australian particularly and NZ longditudinal data on the number of
> head injuries, the number of people cycling and the percentage wearing
> of helmets across the introduction of mandatory helmets is more than
> enough to demonstrate that the numbers wearing helmets increased
> dramatically and the numbers cycling fell dramatically such that the
> head injury rate increased signficantly.

I don't really see that. Or otherwise, in the infamous "pick the
year of the mandation" graphs you could easily tell when helmets
were introduced. It works just the same the other way about, with
no obvious increase as well as no obvious decrease.

> And that is just on the basic
> data before you start factoring in other factors such as trends in other
> groups such as pedestrian head injuries. The work of Hewson in the UK
> shows similar results.

From my readings (admittedly a while ago) Hewson seemed to find
nothing much changed with increasing wearing rates. Things like
"but we know boys take more risks" is pretty speculative: can you
show a much higher injury rate for boys than girls before lots of
girls started wearing lids?

> So a correlation between helmet wearing and head injuries has been more
> than demonstrated in epidemiological studies. What has not been shown
> is whether it is because of physical factors related to the helmet
> wearing or psychological factors related to motorist and helmet wearing
> cyclist behaviour.

I can't see anything obvious in Robinson's BMJ summary of what has
actually happened on the ground saying folk are clearly getting
hurt more in lids. For Victoria she specifically says the rate of
decline of head injuries did not change. To me that reads as no
change, not "decreased".

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 protected] http://www.dundee.ac.uk/~pjclinch/

 

[Tony Raven]

Peter Clinch wrote on 29/04/2007 12:02 +0100:
>
> I don't really see that. Or otherwise, in the infamous "pick the year
> of the mandation" graphs you could easily tell when helmets were
> introduced. It works just the same the other way about, with no obvious
> increase as well as no obvious decrease.
>

Not so. If the numbers cycling had stayed constant and helmet wearing
had doubled and you couldn't see the effect in those graphs you could
say helmets did nothing. The fact that at the same time the number of
cyclist dropped significantly and you can't see an effect in the graphs
(which are total numbers, not rates) strongly indicates that the effect
of the helmet step change was offset by there being fewer cyclists.

>
> For Victoria she specifically says the rate of decline of
> head injuries did not change. To me that reads as no change, not
> "decreased".
>

Yes but the numbers cycling in Victoria dropped about 40% (while the
helmet wearing rate went from 31% to 75%) with as you note, no change in
head injuries. Which means the head injury _rate_ increased by ~40%


--
Tony

"The most savage controversies are those about matters as to which there
is no good evidence either way."
- Bertrand Russell

 

[Peter Clinch]

Tony Raven wrote:

> Not so. If the numbers cycling had stayed constant and helmet wearing
> had doubled and you couldn't see the effect in those graphs you could
> say helmets did nothing. The fact that at the same time the number of
> cyclist dropped significantly and you can't see an effect in the graphs
> (which are total numbers, not rates) strongly indicates that the effect
> of the helmet step change was offset by there being fewer cyclists.

Perhaps I'm looking at different graphs... The one I generally
think of is the one Hagel says of Robinson's Fig. 2 that the effect
can "clearly be seen". No such effect can be seen, any way.

> Yes but the numbers cycling in Victoria dropped about 40% (while the
> helmet wearing rate went from 31% to 75%) with as you note, no change in
> head injuries. Which means the head injury _rate_ increased by ~40%

I'm seeing these as rates.

To quote the summary points of Robinson's piece:
"Before and after data show enforced helmet laws discourage cycling
but produce no obvious response in percentage of head injuries".

To quote John Franklin's assessment of Hewson's work which you say
similarly shows clear disbenefit, "no association could be found
between the differing patterns of helmet wearing rates and casualty
rates for adults and children".
That's no association, /not/ a clear disbenefit. He goes on:
"research for the Scottish Executive found that wearing a helmet
made no significant difference to outcome in the case of the more
serious head injuries measured by need of follow-up or hospital
admission." Again, no change in outcome is not a worsened outcome.

And that's another chap with no reason to pull his punches if
unambiguous data is there, but he says there's nothing there.

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 protected] http://www.dundee.ac.uk/~pjclinch/

 

[Tony Raven]

Peter Clinch wrote on 29/04/2007 13:10 +0100:
>
> Perhaps I'm looking at different graphs... The one I generally think of
> is the one Hagel says of Robinson's Fig. 2 that the effect can "clearly
> be seen". No such effect can be seen, any way.
>

Penny drops, we are. I thought you were referring to the graphs of
total number of head injuries (I'll dig out the reference if you need it
later). The error I think you are making is the Robinson graphs
(actually Hendrie D et al graphs) you are looking at are the percentage
of cyclist injuries that were head injuries. What they do not tell you
and which the graphs I was referring to do, is whether the total number
of head injuries went up or down, just their proportion.

> To quote the summary points of Robinson's piece:
> "Before and after data show enforced helmet laws discourage cycling but
> produce no obvious response in percentage of head injuries".
>

Correct which is an indicator that the effect may be more pyschological
than physical i.e. helmeted cyclists are having more of the same type of
accident so that the relative proportion involving head injuries does
not change.

> To quote John Franklin's assessment of Hewson's work which you say
> similarly shows clear disbenefit, "no association could be found
> between the differing patterns of helmet wearing rates and casualty
> rates for adults and children".

Hewson's work is worth reading first hand and he is very circumspect
with what he claims. His conclusions assume that girls and boys
experience the same risk profile when you would conclude no benefit. It
is highly likely that this in not the case in which case there would be
disbenefit.

> That's no association, /not/ a clear disbenefit. He goes on: "research
> for the Scottish Executive found that wearing a helmet made no
> significant difference to outcome in the case of the more serious head
> injuries measured by need of follow-up or hospital admission." Again,
> no change in outcome is not a worsened outcome.
>
> And that's another chap with no reason to pull his punches if
> unambiguous data is there, but he says there's nothing there.

I think most people including myself err on the side of saying they do
no good as safe and clearly defendable ground. The disbenefit argument
is much harder to get people to believe, unnecessary to making the case
and not as clear cut in the evidence but I actually believe after
looking at lots of it that there is strong indications that it is the
case. YMMV.


--
Tony

"The most savage controversies are those about matters as to which there
is no good evidence either way."
- Bertrand Russell

 

[Nick Maclaren]

In article <[email protected]>,
Tony Raven <[email protected]> writes:
|> > |>
|> > |> I don't think that is correct. Rogers in the US, with statistics on 8
|> > |> million cyclists found that "the bicycle-related fatality rate is
|> > |> positively and significantly correlated with increased helmet use".
|> >
|> > That doesn't mean quite what you seem to think that it means.
|>
|> Come on then Mr Cryptic. Explain what it does mean.

That statement implies nothing whatsoever about causality, direct or
indirect. And you were assuming some indirect causality.

|> Maybe you should look again. Perhaps you should look at the work by
|> John Adams, Mayer Hilman, Ian Walker and Brent Hagel (yes Brent Hagel
|> the mandatory helmet promoter)
|> Mok D, Gore G, Hagel B, Mok E, Magdalinos H, Pless B., Risk compensation
|> in childrens activities: A pilot study; Paediatr Child Health.
|> 2004;9(5):327-330.

Yuck. I don't see how I can get hold of it without subscribing to an
(expensive) electronic database. Cambridge University does not subscribe
to anything that seems to have it. Do you have any ideas?


Regards,
Nick Maclaren.

 

[Peter Clinch]

Tony Raven wrote:

> Hewson's work is worth reading first hand and he is very circumspect
> with what he claims. His conclusions assume that girls and boys
> experience the same risk profile when you would conclude no benefit. It
> is highly likely that this in not the case in which case there would be
> disbenefit.

I did read Hewson 1st hand. As I asked before when you made the
above assumption, it's a not unreasonable assumption as a first
thought but you *need* to back it up with numbers if you're going
to say anything solid about it. I've not seen numbers showing
unhelmeted boys collect more serious injuries than unhelmeted
girls, and I find it a bit /too/ coincidental that the Bad Effect
of a helmet on a girl would just happen to be ~ exactly the same as
the Bad Effect from a boy being a bit madder. We should have
learned not to jump to conclusions about what will show up in the
data, or we'd have just assumed lids help in the first place and
left it at that.

> I think most people including myself err on the side of saying they do
> no good as safe and clearly defendable ground.

Indeed.

> The disbenefit argument
> is much harder to get people to believe, unnecessary to making the case
> and not as clear cut in the evidence but I actually believe after
> looking at lots of it that there is strong indications that it is the
> case. YMMV.

I'm willing to believe it /might/ be, but I prefer the solid ground
if I'm going out and making claims. That Robinson is willing to
come out and clearly say laws are counter productive on the basis
of falling cycling levels, but not in terms of actual injuries,
strikes me as quite telling. The points you flag up make a case
for more investigation, but that's not the same as close the case
as clear disbenefit.

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 protected] http://www.dundee.ac.uk/~pjclinch/

 

[Tony Raven]

Peter Clinch wrote on 29/04/2007 17:01 +0100:
>
> I'm willing to believe it /might/ be, but I prefer the solid ground if
> I'm going out and making claims. That Robinson is willing to come out
> and clearly say laws are counter productive on the basis of falling
> cycling levels, but not in terms of actual injuries, strikes me as quite
> telling. The points you flag up make a case for more investigation, but
> that's not the same as close the case as clear disbenefit.
>

It's probably a matter of personal interpretation. The Australian and
NZ data can clearly be interpreted as it is as a greater injury rate
correlated with increased helmet use. Rodgers study in the US shows a
small but significant correlation. Papers like Hewson and Hagel lend
supporting evidence but as you say there are assumptions that would have
to be made. Hence my saying throughout that there are strong
indications and not there is proof. And as Nick points out it is
correlation not necessarily causation although again the longditudinal
Aus and NZ data points against that.

As to why the figures match so closely I would suggest they do not
necessarily match because of the data variability and trending but it is
close enough not to be distinguishable. John Adam's use of the word
risk _homeostasis_ may provide the rest of the answer.

--
Tony

"The most savage controversies are those about matters as to which there
is no good evidence either way."
- Bertrand Russell

 

[Simon Brooke]

in message <[email protected]>, Tony Raven
('[email protected]') wrote:

> Peter Clinch wrote on 29/04/2007 09:37 +0100:
>>
>> AFAICT the error bars on the total effect are at /least/ as big as any
>> effect good or bad you can really draw out from the data we have.
>> Bearing that in mind, I think saying they do nothing much (that is 0 +/-
>> a wee bit) is a much safer proclamation than "slightly more deaths"
>> (that is "a small though definite disbenefit"). A plausible mechanism
>> on its own isn't enough for a safe conclusion.
>>
>
> I don't think that is correct. Rogers in the US, with statistics on 8
> million cyclists found that "the bicycle-related fatality rate is
> positively and significantly correlated with increased helmet use".

Yes, but, increases in helmet use are also statistically correlated with
decline in cycling (by any measure - distance or trips or participants).
So it's not clear that it's the helmets themselves that cause the excess
deaths. They may do, but it isn't proven.

> The Australian particularly and NZ longditudinal data on the number of
> head injuries, the number of people cycling and the percentage wearing
> of helmets across the introduction of mandatory helmets is more than
> enough to demonstrate that the numbers wearing helmets increased
> dramatically and the numbers cycling fell dramatically such that the
> head injury rate increased signficantly. And that is just on the basic
> data before you start factoring in other factors such as trends in other
> groups such as pedestrian head injuries. The work of Hewson in the UK
> shows similar results.
>
> So a correlation between helmet wearing and head injuries has been more
> than demonstrated in epidemiological studies. What has not been shown
> is whether it is because of physical factors related to the helmet
> wearing or psychological factors related to motorist and helmet wearing
> cyclist behaviour.

Zigzaggerly.

--
[email protected] (Simon Brooke) http://www.jasmine.org.uk/~simon/

;; Good grief, I can remember when England won the Ashes.

 

[Simon Brooke]

in message <[email protected]>, Tony Raven
('[email protected]') wrote:

> Penny drops, we are.  I thought you were referring to the graphs of
> total number of head injuries (I'll dig out the reference if you need it
> later).  The error I think you are making is the Robinson graphs
> (actually Hendrie D et al graphs) you are looking at are the percentage
> of cyclist injuries that were head injuries.  What they do not tell you
> and which the graphs I was referring to do, is whether the total number
> of head injuries went up or down, just their proportion.

Could you post these graphs (with citation) on the Web somewhere? Or are
they already available?

--
[email protected] (Simon Brooke) http://www.jasmine.org.uk/~simon/

X-no-archive: No, I'm not *that* naive.

 

[Peter Clinch]

Simon Brooke wrote:

> Could you post these graphs (with citation) on the Web somewhere? Or are
> they already available?

The ones in Robinson are online at bmj.com, along with the whole
paper, responses and the "response" from Hagel et al.

http://www.bmj.com/cgi/content/full/332/7543/722-a

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 protected] http://www.dundee.ac.uk/~pjclinch/

 

[Peter Clinch]

Tony Raven wrote:

> John Adam's use of the word
> risk _homeostasis_ may provide the rest of the answer.

Risk homeostasis is, ISTM, how someone keeps their level of risk
roughly the same, *not* how one group adjusts their level of risk
to match another group who you suggest are more prone to taking
risks.

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 protected] http://www.dundee.ac.uk/~pjclinch/

 

[Tony Raven]

Simon Brooke wrote on 29/04/2007 18:32 +0100:
> in message <[email protected]>, Tony Raven
> ('[email protected]') wrote:
>
>> Penny drops, we are. I thought you were referring to the graphs of
>> total number of head injuries (I'll dig out the reference if you need it
>> later). The error I think you are making is the Robinson graphs
>> (actually Hendrie D et al graphs) you are looking at are the percentage
>> of cyclist injuries that were head injuries. What they do not tell you
>> and which the graphs I was referring to do, is whether the total number
>> of head injuries went up or down, just their proportion.
>
> Could you post these graphs (with citation) on the Web somewhere? Or are
> they already available?
>

I'll have to dig out the graphs and citations but
http://members.pcug.org.au/~psvansch/crag/actlaw.htm gives an example of
the evidence I am referring to.

--
Tony

"The most savage controversies are those about matters as to which there
is no good evidence either way."
- Bertrand Russell

 

[Dave Larrington]

In news:[email protected],
John B <[email protected]> tweaked the Babbage-Engine to tell us:

[a tale told by an idiot, full of sound and fury, signifying nothing..."]

I rather suspect that if the lad had been shot and then set ablaze while
riding his bike, the parents - who are certainly worthy of this particular
fate - would /still/ be campaigning for an MHL.

Gagh!

--
Dave Larrington
<http://www.legslarry.beerdrinkers.co.uk>
Here, take these cheese-shaped stilts. You'll know when to use
them.

 

[Tony Raven]

Dave Larrington wrote on 30/04/2007 08:55 +0100:
>
> I rather suspect that if the lad had been shot and then set ablaze while
> riding his bike, the parents - who are certainly worthy of this particular
> fate - would /still/ be campaigning for an MHL.
>
> Gagh!
>

Perhaps we should start a campaign in that case for helmets to be made
from fire retardant polystyrene. After all that stuff is real nasty in
a fire ;-)

--
Tony

"The most savage controversies are those about matters as to which there
is no good evidence either way."
- Bertrand Russell

 

[John B]

Dave Larrington wrote:

> In news:[email protected],
> John B <[email protected]> tweaked the Babbage-Engine to tell us:
>
> [a tale told by an idiot, full of sound and fury, signifying nothing..."]
>
> I rather suspect that if the lad had been shot and then set ablaze while
> riding his bike, the parents - who are certainly worthy of this particular
> fate - would /still/ be campaigning for an MHL.

My youngest daughter came into the bikeshed (aka kitchen) yesterday with a
bike helmet strapped to her stomach.
With a wide grin she proclaimed she now can't fall off her bike and cut her
knee .

Letter has gone to the rag in question.

John B