Claims behind the RSA promotion of helmets

monument

Given that the 'Organ donors' without helmets thread was locked, here's the RSA's reply to my request as posted on that thread.

When I get the chance I'll be asking for further details, they have not answered my questions.

The RSA's response:

Mr. Cian Ginty

I refer to the request which you made under the European Communities (Access to Information on the Environment) Regulations 2007 (S.I. No. 133 of 2007) for access to records held by the Road Safety Authority. I attach the report as requested and includes relevant comment on the report below. The person in charge of compiling the report is Dr. Yaw Bimpeh, PhD (Statistics), MSc(Mathematics), BSc(Hon’s) Mathematics

The report is basically review of all cyclist road traffic collisions reported to the Garda Síochána, where details have been recorded and forwarded to the Road Safety Authority, involving fatalities and personal injury which occurred on public roads in Ireland over the period 1997 and 2009.

The relative risk, also called the risk ratio, is an important measure in epidemiological Studies. It is a useful index to measure the increased risk (if any) of incurring a particular death or disease if a certain factor is present.
Using this measure we found that over the the period 1997 and 2009, cyclists not wearing helmet had a risk of being killed on the road that was approximately two times the risk of those wearing helmet (Risk ratio: 2.1; 95% CI:0.87-5.0, P=0.064)

Helmet Wearing Status | Severity: Killed | Injured | Uninjured | Total | Risk | Relative Risk (Not in use/In use)
Crash helmet in use | 5 | 176 | 4 | 185 | 0.03 | -
Crash helmet not in use | 51 | 837 | 16 | 904 | 0.06 | 2.09


It is acknowledged that a large proportion of information on cyclist crash helmet usage in the collision database is missing. However, this information is missing at random therefore the analysis of the incomplete data is unbiased and statistically valid.

How to obtain the relative risk
Suppose that each subject in a study, at a particular time, is classified as positive or negative according to some risk factor, and as having or not having a certain disease under investigation. For any such categorization the population may be enumerated in a 2 _ 2 table below.

Factor | - | Disease | -
- | positive | negative | Total
positive | A | B | A+B
negative | C | D | C+D
Total | A+C | B+D | N=A+B+C+D

The entries A, B, C and D in the table are sizes of the four combinations of disease presence/ absence and factor presence/absence, and the number N at the lower right corner of the table is the total population size. The relative risk is given by
RR=[A*(C+D)] / [C*(A+B)]


References for analysis of incomplete data
Horton, Nicholas J., and Ken P. Kleinman. 2007. “Much Ado about Nothing: A Comparison of Missing Data Methods and Software to Fit Incomplete Data Regression Models.” The American Statistician. 61(1):79-90.
King, Gary, et al. 2001. “Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation.” American Political Science Review. 95(1): 49-69.
Little, Roderick J.A. 1988. “Missing-Data Adjustments in Large Surveys.” Journal of Business and Economic Statistics. 6(3):287-296.
Little, Roderick J.A., and Donald B. Rubin. 2002. Statistical Analysis with Missing Data, 2nd Edition. New York: Wiley.
Rubin, Donald B. 1987. Multiple Imputation for Nonresponse in Surveys. J. Wiley & Sons, New York.

I hope this information is what you were looking for.

Regards

Mick

Michael C Brosnan BA(Hons), MSc(Econ)
Research Manager
Research & Education
RSA

tomasrojo

However, this information is missing at random therefore the analysis of the incomplete data is unbiased and statistically valid.

It's not missing at random. The Gardaí decide to fill in the report field at the accident or decide not to. That's not very likely to be random.

Also, no mention of statistical significance. They include a 95% confidence interval, as in my post below. The CI crosses one, so that is taken in statistics as no evidence of an effect (though not, of course, as evidence of no effect). As I said before, their figure of 2.09 is just not statistically significant.

They really should drop this argument. It's the old story of garbage in and garbage out.

http://www.boards.ie/vbulletin/showpost.php?p=71566585&postcount=327

tomasrojo

Oddly, their confidence interval is not quite the same as the one I worked out. It's a standard enough formula. I wonder why theirs is different. (Not that the implication is different -- it still clearly crosses the value of no difference, 1.0)

Victor

However, this information is missing at random therefore the analysis of the incomplete data is unbiased and statistically valid.

I strongly suspect it isn't missing at random - a more thorough investigation is going to be done (a) for more severe accidents (b) helmet wearing is more common amongst higher income groups, who are also more likely to seek medical attention and one presume, report an inceident to the Garda.

tomasrojo

I've heard that wealthier people are less likely to be involved in serious collisions in the first place.

But anyway, even if this assertion of random gaps in the data were true (which I don't concede for a minute), their claim is still not supportable by the statistical test they're doing. I see they've included a p-value there, which is another way of looking at the statistical significance. In science and medicine, a p-value of 0.05 or lower is acceptable if you wish to claim an effect (95% confidence being the lowest admissible confidence level); any more than 0.05 and you can't claim an effect. Their value is 0.064. So, again, not statistically significant.

morana

tomasrojo wrote: »

I've heard that wealthier people are less likely to be involved in serious collisions in the first place.

But anyway, even if this assertion of random gaps in the data were true (which I don't concede for a minute), their claim is still not supportable by the statistical test they're doing. I see they've included a p-value there, which is another way of looking at the statistical significance. In science and medicine, a p-value of 0.05 or lower is acceptable if you wish to claim an effect (95% confidence being the lowest admissible confidence level); any more than 0.05 and you can't claim an effect. Their value is 0.064. So, again, not statistically significant.

so what are you saying without all the statisical mumbo jumbo?

Victor

Based on the data, while there might be a tendency for helmets to have an effect, the RSA haven't proven their case.

tomasrojo

Victor wrote: »

Based on the data, while there might be a tendency for helmets to have an effect, the RSA haven't proven their case.

Pretty much.

If you were trying to publish this analysis in a peer-reviewed scientific or medical journal, you wouldn't be allowed to make the claim that the RSA are making. You might get away with saying something like "though the relative risk is not statistically significant, there are indications that helmets provide a protective effect." However, the RSA are being very categorical, claiming a strong protective effect, and making this claim with great certainty. Neither the claim of strength nor the great certainty is warranted.

tomasrojo

The other issue, in plain English, is that the Gardaí omitted to record whether the cyclist was wearing a helmet in the vast majority of records. So you have to wonder what difference those missing data would make. The RSA is assuming claiming none. But as Christine Keeler Mandy Rice-Davies might put it, they would say that, wouldn't they.

Cookie_Monster

185 incidents out of how many cyclists wearing a helmet or how many KM per year?
904 out of how many?

that 185 could be out of 10,000 people
the 904 could be out of 500,000 people

which would suggest it's more dangerous to wear one (obviously figures would be closer in reality)

so until that is answered it's kind of meaningless either way.

Accident per KM cycled or by total cycling pop helmeted / un-helmeted is the only way to remove the massive uncertainty IMO

tomasrojo

Cookie_Monster wrote: »

185 incidents out of how many cyclists wearing a helmet or how many KM per year?
904 out of how many?

that 185 could be out of 10,000 people
the 904 could be out of 500,000 people

which would suggest it's more dangerous to wear one (obviously figures would be closer in reality)

so until that is answered it's kind of meaningless either way.

Accident per KM cycled or by total cycling pop helmeted / un-helmeted is the only way to remove the massive uncertainty IMO

The strict version of the claim they're making is that helmeted cyclists have about half the risk of dying in the event of a collision . If their data weren't so full of omissions and if the relative risk were statistically significant, they could make this claim; you don't need kilometres travelled or general rates of wearing to make the strict version of the claim, since we have a Garda report for most serious collisions. Pity they didn't bother filling the forms out fully.

tomasrojo

Using this measure we found that over the the period 1997 and 2009, cyclists not wearing helmet had a risk of being killed on the road that was approximately two times the risk of those wearing helmet

This claim, on the other hand, is not supportable by the test done by the RSA. For this claim you would need to know a lot more about rates of helmet-wearing, km travelled, and so on. Or at least have some idea of how likely helmet-wearing cyclists are to be in a collision compared to non-helmet-wearing.

tomasrojo

Just to sum up, the RSA is being really sloppy in about three ways here.

• They are pushing their statistics as meaning "likelihood of being killed on the road", which is not what they calculated.
  
• They are making strong claims of a dataset that is substantially incomplete, and making an unwarranted assumption that the missing data would have no effect
  
• They are making strong claims from a statistical test that actually shows that we cannot conclude that rates of fatality for helmet-wearers are different from non-helmet-wearers

.

tomasrojo

Finally, just for a bit of fun, you can do these relative risk calculations to other Garda collision data.

For example, you can apply this test to Table 34 of the RSA's 2009 Road Collision Fact Book, and you find that unhelmeted motorcyclists are at less risk of death in the event of a collision than helmeted motorcyclists.

6/33 helmeted versus 12/230 unhelmeted
Relative risk: 3.48
95% Confidence Interval: 1.40 -> 8.65

This time the relative risk is statistically significant. So wearing a motorcycle helmet increases your chance of dying in a collision. By RSA logic.

Cookie_Monster

tomasrojo wrote: »

The strict version of the claim they're making is that helmeted cyclists have about half the risk of dying in the event of a collision

Indeed but what if those wearing a helmet are twice or three as likely to be in a collision too due to the perception of better protection, reduced motorist clearances and greater risk taking as was discussed in the last thread?
That would make the initial statistic less relevant if that were the case. As does leaving out other data...

Thomas_S_Hunterson

That email pretty much epitomises the danger of letting non-statisticians do statistics.

See also.

morana

Thomas_S_Hunterson wrote: »

That email pretty much epitomises the danger of letting non-statisticians do statistics.

See also.

I'll have a shot at it so !!!:D

Irish_Elect_Eng

Thomas_S_Hunterson wrote: »

That email pretty much epitomises the danger of letting non-statisticians do statistics.

See also.

On the other side of the coin, this thread epitomizes the danger of relying on statistics when common sense is all that is required to point out the obvious.

A lack of clear data should not be a barrier to adopting safer practices.

However I do agree that the phrasing of the RSA is "statistically" not supported to a 95% confidence limit based on Irish Data, but it is incorrect to state that lesser confidence limits are not valid, depending on the usage of the data. Looking at the statistics in isolation is also a process fraught with the probability of error. Without expert interpretation statistical data analysis is pretty meaningless in many cases. There is always the classic strong link between the sales of ice cream and shark attack fatalities.

There is some interesting information on this site:

http://www.bhsi.org/stats.htm

But perhaps how the British phrase it in their report is more appropriate, but the message is still valid, and the use of stats in an awareness campaign woulf just confuse the important message.

Statistics from the UK Department of Transport

The Potential for Cycle Helmets to Prevent Injury:
A Review of the Evidence
TRL Report PPR 446 - Findings

• Assuming that cycle helmets are a good fit and worn correctly, they should be effective at reducing the risk of head injury, in particular cranium fracture, scalp injury and intracranial (brain) injury.
• Cycle helmets would be expected to be effective in a range of accident conditions, particularly:
• the most common accidents that do not involve a collision with another vehicle, often simple falls or tumbles over the handlebars; and also
• when the mechanism of injury involves another vehicle glancing the cyclist or tipping them over causing their head to strike the ground.
• A specialist biomechanical assessment of over 100 police forensic cyclist fatality reports predicted that between 10 and 16% could have been prevented if they had worn an appropriate cycle helmet.
• Of the on-road serious cyclist casualties admitted to hospital in England (HES database):
• 10% suffered injuries of a type and to a part of the head that a cycle helmet may have mitigated or prevented; and a further
• 20% suffered ‘open wounds to the head’, some of which are likely to have been to a part of the head that a cycle helmet may have mitigated or prevented.
• Cycle helmets would be expected to be particularly effective for children, because:
• the European Standard (EN 1078) impact tests and requirements are the same for adult and child cycle helmets – both use a 1.5 m drop height test; and so
• given that younger children are shorter than older children and adults, their head height would be within the drop height used in impact tests, so a greater proportion of single-vehicle accidents are likely to be covered by the Standard for children.
• No evidence was found for an increased risk of rotational head injury with a helmet compared to without a helmet.
• In the literature reviewed, there is a difference between hospital-based studies, which tend to show a significant protective effect from cycle helmets, and population studies, which tend to show a lower, or no, effect. Some of the reasons behind this were due to:
• the lack of appropriateness of the control groups used; and
• limitations in the available data, such as knowledge of helmet use and type of head injury.

tomasrojo

The TRL report is a curious one in that they found no statistically reliable evidence that cycle helmets had proved effective in reducing head injury, and then proceeded to speculate what that effect might be.

You seem to be saying that the RSA can say whatever they like, as long as they're promoting helmets?

I wouldn't set much store by the BHSI. Your link there includes this:

Helmet use has been estimated to reduce head injury risk by 85 percent.

Anyone who includes that statistic hasn't been keeping up to date with the research, or else is a zealot that doesn't care that it isn't true.

In fact, Randy Swart (the one man behind that one-man operation) had this to say about using discredited statistics:

We are aware of the second study, but by the time it appeared the 85% figure was so deeply ingrained in the injury prevention community that a change will not be helpful. I regard any similar numbers published in those or other studies as approximations anyway. Over the years observation of who gets head-injured and who does not seems to support the 85% number. So we have left it up that way. We do have references to the later study for those who need more.

http://www.chapmancentral.co.uk/wiki/Bicycle_Helmet_Safety_Institute

Ryder

+1 on the above.
Not sure why helmets are such an issue for people........all of the studies are observational and suffer from many biases including reporting bias. It may never be possible to prove that helmets reduce the severity of head injuries because of the variety of confounding variables, but they are a simple cheap measure which will reduce the severity of injury if you sustain head trauma in most cases. This will reduce health costs, is a reasonable public health measure and so while im not totally supportive, I dont think its unreasonable to make them compulsary.

Arguements involving pedestrians etc etc make little sense

Cookie_Monster

Ryder wrote: »

Not sure why helmets are such an issue for people........

the nanny state H&S mentality mainly and the impression they give that cycling is dangerous and requires protective gear when it blatantly does not.

I dont think its unreasonable to make them compulsary.

but it's been shown that compulsory helmet usage has a negative impact on overall public health (Aus and NZ) as significantly less people cycle and therefore get less exercise and increase all the negative health issues relating to lack of exercise and overweightness, compared to a tiny minority who may be spared some head trauma from a crash...

tomasrojo

Irish_Elect_Eng wrote: »

However I do agree that the phrasing of the RSA is "statistically" not supported to a 95% confidence limit based on Irish Data, but it is incorrect to state that lesser confidence limits are not valid, depending on the usage of the data.

It's an interesting point, but the RSA itself has used the 95% confidence interval because that's the lowest confidence interval that is standardly used, and has been the standard in research since 1925 or so. Since they used the relative risk because it is "an important measure in epidemiological studies", they should also use the standard method by which the relative risk is judged:

http://clinicalevidence.bmj.com/ceweb/resources/glossary.jsp

If the 95% confidence interval for a relative risk (RR) or an odds ratio (OR) crosses 1, then this is taken as no evidence of an effect.

Some other words on using confidence intervals lower than 95%:

http://www.childrensmercy.org/stats/ask/why95.asp

[...] most researchers will react strongly and negatively to any change in the confidence level. When you change the confidence level, it appears to some that you are manipulating your data to get a pre-ordained conclusion.

If you want to justify a different confidence level, you need to present that justification in the protocol, before any data is collected.

I remember when I worked as a researcher that papers that used unorthodox confidence intervals or that claimed that the outcome was nearly significant always smacked of special pleading.


Is this getting a bit far away from the subject of cycling now?

monument

International pro helmet research is nearly as flawed and sometimes more flawed than the RSA's claims. Nearly all of it is on thin ground and is highly disputed. As tomasrojo points out, some of the wildest claims have been retracted by their authors.

Ryder wrote: »

Arguements involving pedestrians etc etc make little sense

Actually, using that logic that cyclists should wear helmets, the arguments about pedestrians make perfect sense. You apparently don't even need stats or proof... Watch this...

Irish_Elect_Eng wrote: »

On the other side of the coin, this thread epitomizes the danger of relying on statistics when common sense is all that is required to point out the obvious.

A lack of clear data should not be a barrier to adopting safer practices.

Surly it's a common sense and obvious safety practice that people walking should wear helmets? It would protect against a wide range of things from people falling on their own to people getting hit by cars and cyclists to assaults. Surly you want to adopt safer practices? It's common sense and the obvious safe thing to do.

What seems like "safer practices", "common sense", and "the obvious" often are not actually safer in the real world, and in some cases it takes a long time for this to be accepted:

• Pedestrian barriers seem like they are a common sense safety device, but have since been shown to have major negative affects (the likes of increasing speeding and trapping peds and cyclists).
  
• It seems obvious to many people that cycle lanes are always safer than the road or a normal traffic lane, but that's not always the case.
  
• Helmets use and promotion -- without much proven benefits -- contribute to making cycling look less safe, make it less convenient, and generally less attractive. We know that safety is an issue for cyclists and non cyclists and we know the most proven way to increase safety is to make cycling more attractive to more people to cycle more often (safety in numbers -- unlike helmets is proven over and over again and is not highly disputed; when you have more cyclists or peds, you have less injuries and deaths), thus an unneeded focus on safety gear hinders the promotion of cycling and safer cycling.

Let me guess, you want stats and this logic is somehow not as valid as your common sense? BTW if you want I can point to data on safety in numbers and the perception of danger being a barrier to cycling.

Irish_Elect_Eng wrote: »

...Without expert interpretation statistical data analysis is pretty meaningless in many cases. There is always the classic strong link between the sales of ice cream and shark attack fatalities.

Even looking at one set of stats can be problematic for a number of reasons. Helmet use is a good example:

• There's a low chance of being in a collision or a crash on your own.
  
• If that happens, there's a lower chance of getting killed or seriously injured.
  
• If those happen, there's a lower chance of head or brain injuries to be a factor.
  
• And even when head or brain injuries are a factor, helmets can still be useless because force goes well beyond their limits and/or other injuries would have killed the person even where head injuries are recorded.

Irish_Elect_Eng wrote: »

There is some interesting information on this site:

http://www.bhsi.org/stats.htm

This one is a little more interesting: http://cyclehelmets.org/

Ryder wrote: »

Not sure why helmets are such an issue for people........all of the studies are observational and suffer from many biases including reporting bias. It may never be possible to prove that helmets reduce the severity of head injuries because of the variety of confounding variables, but they are a simple cheap measure which will reduce the severity of injury if you sustain head trauma in most cases. This will reduce health costs, is a reasonable public health measure and so while im not totally supportive, I dont think its unreasonable to make them compulsary.

Arguements involving pedestrians etc etc make little sense

Seriously: For such a cheap measure why not make helmets compulsory for motorists and all people on streets?

Not only is helmet promotion (or compulsion) not a reasonable public health measure, it is fearmongering which stands in the way of one of the most promising public health measures around -- large numbers of people cycling.

A Professor of Public Health at University of Sydney recently wrote about "Ditching bike helmets laws better for health".

Cycling is proven to help with, source:

• have lower rates of: coronary heart disease, high blood pressure, stroke, diabetes, colon and breast cancer, and depression;
• have increased functionality, lower risk of falling and of hip or vertebral fractures;
• are more likely to maintain their weight and prevent weight gain
• have reduced levels of depression and better cognitive function

It also could help further with:

• Improving health and reducing health costs for the above outlined reasons
• Reducing air and noise pollution, thus further improving population health
• Making our towns and cities more inviting and productive

The big question is: Do you spend money promoting helmets or promoting cycling?

Capt'n Midnight

Have we had any incidents of children being killed / injured (usually strangulation) as a result of wearing a bicycle helmet when not on a bike.

It's happened in other countries and perhaps childrens helmets should carry a warning that they are not to be worn off a bike / always worn with adult supervision.



Are there any stats on hardshell helmets at cycling speeds ?
Do they record the speed motocyclists are travelling at when in collision with motorists / inaminate objects ? ( yes you would have to take into account the weight of a motorcycle as it will slow a car down more than a bicycle )

The point here is that cycling helmets are NOT designed to signifigantly reduce the energy levels involved in a collision above 20Kmhr - it would be interesting to see if hardshell helmets offer more protection at speeds below 20Kmhr

blorg

Capt'n Midnight wrote: »

Have we had any incidents of children being killed / injured (usually strangulation) as a result of wearing a bicycle helmet when not on a bike.

It's happened in other countries and perhaps childrens helmets should carry a warning that they are not to be worn off a bike / always worn with adult supervision.

Most, maybe all helmets I have had came with this specific warning (never to use off the bike.) At least one had a diagram of a kid playing in a tree with an X through it.

amacachi

Capt'n Midnight wrote: »

Have we had any incidents of children being killed / injured (usually strangulation) as a result of wearing a bicycle helmet when not on a bike.

Around here we had a kid who fell off some goalposts on a football pitch and died but I've yet to hear of one dying while driving a car. I'm pretty sure that's a 100% correlation right there.

Anyone not wearing a helmet on the road is a tool (and more power to them, if I start cycling again I'll be one of them) but whoever keeps giving the RSA funds to tell people to wear helmets, to indicate while driving, to not walk on motorways and to not ignore wrong-way signs is a bigger tool.

Irish_Elect_Eng

tomasrojo wrote: »

Is this getting a bit far away from the subject of cycling now?

True, but you make well reasoned points on the use of statistics which are interesting in themselves.

I work for a company that had traditionally used a 95/90, 30 sample methodology, but recently to your point changes to a 95/95, 59 sample methodology, just to avoid even the possibility of our data or results being questioned.

But one last point on the cycling, I would think that the impact of helmet usage on the on the number of people cycling would be impacted by two key factors.

(1) National characteristics, that is to say ho compliant are the populace in a particular country, some countries follow "rules" better than others, therefore the impact of introducing rules would vary widely by country.

(2) Entry Requirement, once a rule is in place and becomes accepted as the norm, the impact that rule has on a sport or activity declines, as it is now an entry requirement for that activity. Amateur boxers wear helmets, hurlers wear helmets, soccer plater wear shin-guards, martial artists wear various protections. Now I know that these a not legal requirements,just requirements of the sports bodies, but you do not hear people say, I won't let my son play x because he will have to wear a helmet.

What do you think?

ThisRegard

Thomas_S_Hunterson wrote: »

That email pretty much epitomises the danger of letting non-statisticians do statistics.

See also.

But he feels his BA and M.Sc brings weight to his opinion.

Thisregard. B.Sc Hons

blorg

Irish_Elect_Eng wrote: »

(2) Entry Requirement, once a rule is in place and becomes accepted as the norm, the impact that rule has on a sport or activity declines, as it is now an entry requirement for that activity. Amateur boxers wear helmets, hurlers wear helmets, soccer plater wear shin-guards, martial artists wear various protections. Now I know that these a not legal requirements,just requirements of the sports bodies, but you do not hear people say, I won't let my son play x because he will have to wear a helmet.

A helmet is already mandatory for cycling as a sport, and has been, particularly at amateur level, for a long time. No one has a problem with this, but many do not think it should be a mandatory requirement for popping down the shops to get a pint of milk. The current situation in cycling would be similar to motorsport incidentally, helmets are mandatory there too but not for driving a car in general.

opinion guy

Thomas_S_Hunterson wrote: »

That email pretty much epitomises the danger of letting non-statisticians do statistics.

See also.

Doesn't the email say the info comes from a guy with a Masters and Phd in statistics ?

Thomas_S_Hunterson

opinion guy wrote: »

Doesn't the email say the info comes from a guy with a Masters and Phd in statistics ?

The report may be produced by a statistician but what the RSA are doing is plucking random statistics from the report and misrepresenting them to suit their own agenda.

I think it's unlikely that Dr. Yaw Bimpeh, PhD draws the same conclusions from the report that the RSA does.

The RSA are going about it all wrong. It appears they had their conclusion before even doing any research: That cyclists should wear helmets. Now they're data-mining looking (unsuccessfully) for anything which supports this conclusion.

MrCreosote

Not sure if this has been up. A good review on the Cochrane website:
http://www2.cochrane.org/reviews/en/ab001855.html is the abstract. Full text and some interesting replies on the "library" link.

Basically they say the benefit from helmets is so large for head injury, it's extremely unlikely to be outweighed by changes in riding or driving behaviour

tomasrojo

A recent paper by Elvik is rather critical of the Cochrane Review.

Publication bias and time-trend bias in meta-analysis of bicycle helmet efficacy: A re-analysis of Attewell, Glase and McFadden, 2001
Elvik R. Accident Analysis & Prevention, in press. 2011.

One problem is that the review is dominated by the work of Thompson, Rivara and Thompson (the 85% people that compulsionsist love to quote). The authors of the review are also -- you guessed it -- Thompson, Rivara and Thompson. No post-compulsion studies from Australia were included, for example.

Elvik said:

The Department for Transport in Great Britain issued a report in 2002 entitled: Bicycle helmets -- a review of their effectiveness: a critical review of the literature. The report includes an assessment of the quality of 16 studies that have evaluated the effects of bicycle helmets. Studies were rated as good, reasonable or weak. Of the seven studies Thompson et al. (2009) included in the Cochrane review, one was rated as good, two as good/reasonable, three as reasonable and one as reasonable/weak. Three of the eight studies Thompson et al. (2009) omitted were also rated by the Department for Transport (2002). One was rated as good, one as good/reasonable and one as reasonable. Thus, if the rating developed by the Department for Transport (2002) is applied, it is by no means obvious that all the seven studies that were included by Thompson et al. (2009) ought to have been included. Nor is it clear that all the omitted studies were of lower quality than the studies included.

The really important one Thompson et al. omitted was Dorothy Robinson's analysis of head injury rates pre- and post-compulsion.

EDIT: For stats geeks: they also misused odds ratios in their analysis, using it instead of relative risk, which is what the RSA, correctly, has used.. They've been making this mistake since 1989. I suspect it's not a mistake, since odds ratios are bigger than relative risk and therefore sound more impressive.

Civilian_Target

They're some interesting statistics.

There's some strong statements coming from the RSA in what is clearly shown above to be a weak correlation, even when the integretity of the data is not questioned.

This inspired me to do a little google, and I think I came across the most compelling reason for not wearing a helmet that I've seen to date:

Cycle Helmets: The Case for and Against, Mayer Hillman, 1993 wrote:

An appropriate solution to the problem of serious accidents to cyclists requires an understanding of the circumstances in which accidents occur. Cyclists rarely ride into motor vehicles. It is motor vehicles driven without sufficient care which are the source of most of the danger and which pose the threat to the life and limb of cyclists. Calling on cyclists to increase their safety by wearing a helmet shifts responsibility away from the drivers, the agents of accidents, on to cyclists who are nearly always the victims. Were cycle helmets to be made compulsory, it would reinforce public perceptions of the bicycle as a dangerous form of transport and encourage the view that cyclists are responsible for their own injury.

http://members.pcug.org.au/~psvansch/crag/psi.htm

In short - my interpretation of all this is that wearing a helmet isn't for the benefit of cyclists, it's for the benefit of drivers.

MrCreosote

tomasrojo wrote: »

One problem is that the review is dominated by the work of Thompson, Rivara and Thompson (the 85% people that compulsionsist love to quote). The authors of the review are also -- you guessed it -- Thompson, Rivara and Thompson. No post-compulsion studies from Australia were included, for example.

There was a fair bit of self referencing alright.

The big problem I can see is taking a fairly narrowly defined area of research like "do helmets reduce the risk of head injury in event of a crash" and using this to justify compulsory wearing without looking at all the other social effects, like reducing cycling overall.

Lived in a country with compulsory helmets, and although you get used to it quickly, the short distance rides are almost nonexistent. And nothing like Dublin Bikes would work there.

tomasrojo

Mayer Hillman is a very interesting man. I first came across him in this article:
http://www.guardian.co.uk/theguardian/2002/nov/02/weekend7.weekend2

Clip this article. Photocopy it, send it to a friend, file it. In 10 years' time, if the person it's about is right (and doubt doesn't figure in his lexicon), you'll be amazed that the views it expresses ever seemed outlandish or unfeasible. What sounds now like wild ecotopian fantasy will have turned into an unexceptionable statute governing daily life.

Well, I guess we've one year left before we see widespread carbon rationing (article published 2002).

But it's a good article, and it shows what interesting ideas he has.

This bit resonated strongly with me.

Children's lives have been evolving in a way that mirrors the characteristics of the lives of criminals in prison. They, too, have a roof over their heads, regular meals and entertainment provided for them, but they are not free to go out. But children are not criminals.

He was in the news a bit a few months ago, when the British coalition government were looking into double summer time. He did a lot of work promoting the idea to save electricity.

He also is the source for the much-quoted statistic of the benefits of cycling dwarfing the risks by a ratio of 20:1. The BMA asked him to do some calculations, I think.

Iwannahurl

MrCreosote wrote: »

There was a fair bit of self referencing alright.

The big problem I can see is taking a fairly narrowly defined area of research like "do helmets reduce the risk of head injury in event of a crash" and using this to justify compulsory wearing without looking at all the other social effects, like reducing cycling overall.

Lived in a country with compulsory helmets, and although you get used to it quickly, the short distance rides are almost nonexistent. And nothing like Dublin Bikes would work there.

That thought occurred to me when I saw the Velib bikes operating in Paris. They were extremely popular, widely used and treated as a very practical and easy mode of transport. There is probably no way you could make such a scheme work so well with compulsory helmets.

Which means the marvellous Dublin bike scheme has probably put paid to any prospect of compulsory helmets in this neck of the woods. Yippee.

Capt'n Midnight

Civilian_Target wrote: »

In short - my interpretation of all this is that wearing a helmet isn't for the benefit of cyclists, it's for the benefit of drivers.

Motorists MUST be made to wear helmets - the benefits are measurable , there is no danger of increased rotational injury by being caught on the tarmac, and since the helmet is not visible other road users won't reduce the clearnace in the way they do with cyclists.

http://www.copenhagenize.com/2009/10/australian-helmet-science-for-motorists.html

But more impressive were the estimates of introducing protective headwear for car occupants. The authors of the report estimated that the annual reduction in harm would be in the order of $380 million.

That works out as a benefit of $10 a person per year across the whole population , not just motorists. has the RSA produced the potential savings that getting cyclist to comply would bring ?


When motorist helmet wearing is enforced then we can look at other road users including pedestrians.



Alternatively we could replace drivers airbags with harpoons. This would reduce the number of collisions considerably. It would reduce the cost to the health service since less individuals would need long term care or expensive operations.

Victor

Ultimately, what is need is for people to realise that protective equipment only protects you from certain risks, but exposes you to new risks.

with the advances in body armour, Iraq and Afghanistan have shown that taking on American soldiers in a gun fight is a losing battle. Therefore many attacks are now made with small and medium bombs.

Relative to previous wars, deaths from gunfire have dropped, but deaths and multiple amputations from bombs are way up.

Capt'n Midnight wrote: »

Alternatively we could replace drivers airbags with harpoons. This would reduce the number of collisions considerably. It would reduce the cost to the health service since less individuals would need long term care or expensive operations.

But those people would be dead.

CramCycle

Victor wrote: »

But those people would be dead.

There is no statistical proof :eek:

drunk of my ass and loving this thread, crashed today (while sober), I can honestly say if I was wearing a helmet, I would be dead

n=>3

, luckily all i have is a ripped knee and elbow :P

Capt'n Midnight

Victor wrote: »

Ultimately, what is need is for people to realise that protective equipment only protects you from certain risks, but exposes you to new risks.

with the advances in body armour, Iraq and Afghanistan have shown that taking on American soldiers in a gun fight is a losing battle. Therefore many attacks are now made with small and medium bombs.

Relative to previous wars, deaths from gunfire have dropped, but deaths and multiple amputations from bombs are way up.

risk compensation is the name of the game

most devices that improve driver safety have resulted in a signifigant number of motorists taking more risks
if you switch to a car or bikes with poorer brakes you will start braking earlier / enter blind corners at a slower speed

But those people would be dead.

only if they continue to drive as they currently do, once harpoons are made mandatory in all vehicles I'd predict the number of collisions would go down

also in these recessionary times dead people are a lot cheaper than long term care

tomasrojo

John Adams' formulation is: potential safety benefits are consumed as performance benefits.

It's used in this document anyway.
http://john-adams.co.uk/wp-content/uploads/2007/10/risk%20and%20freedom.pdf