Guardian article - worth a read
Guardian article - worth a read
Suppose that this room is a lift. The support breaks and down we go with ever-increasing velocity.
Let us pass the time by performing physical experiments...
--- Arthur Eddington (creator of the Eddington Number).
Let us pass the time by performing physical experiments...
--- Arthur Eddington (creator of the Eddington Number).
Re: Guardian article - worth a read
I can't quite understand what they've done but it appears to be the usual mistake of the wrong denominator in their rates calculation. The part of the paper that describes it is:
I assume this means they have calculated the number of admissions per head of population because of the bit I've underlined (not quite sure what the person years is unless its allowing for the fact that not every person was resident in the province or territory for a complete year). This is a measure that has been used in other studies BUT it fails to take into account any change in the number of people cycling as a result of the legislation.
If it is as per my assumption then its not worth a read and is another of those poor quality pieces of research that tells you nothing useful about helmets other than by their measure you can't see any effect of helmet laws. The key figure that illustrates this is:
Firstly, we calculated the annual rates of admissions to hospital for cycling related head injuries per 100 000 person years in all provinces and in the territories. Population counts for the denominator were obtained from CANSIM (Canadian Socioeconomic Information Management System).
I assume this means they have calculated the number of admissions per head of population because of the bit I've underlined (not quite sure what the person years is unless its allowing for the fact that not every person was resident in the province or territory for a complete year). This is a measure that has been used in other studies BUT it fails to take into account any change in the number of people cycling as a result of the legislation.
If it is as per my assumption then its not worth a read and is another of those poor quality pieces of research that tells you nothing useful about helmets other than by their measure you can't see any effect of helmet laws. The key figure that illustrates this is:
Re: Guardian article - worth a read
The report states without a hint of criticism ...
... citing Thompson, Rivara and Thompson.
The report uses two measures: cycling head injuries per 100,000 of population, and cycling head injuries per injured cyclist. They briefly mention the possibility that reduced cycling rates might account for lower head injuries per population.
From graphs of cycling head injuries per 100,000 of population plotted against years, showing the year of legislation, we see the legislaton had no obvious effect, and the report says this.
Similarly, there was no significant change in the "cycling head injuries per injured cyclist" measure pre- and post-law. According to the graphs, in almost all provinces where this measure reduced for children, it increased for adults. And vice versa. I don't suppose this is statistically significant either.
<rant>Cycling head injuries per 100,000 of population is an appalling measure. Any researcher who uses it should be made to cycle a thousand times around the Bow roundabout (with helmet, if they wish). It is an appalling measure because a country that totally banned cycling would achieve an excellent score: zero cycling head injuries per 100,000 of population.</rant>
Helmets reduce the risk of injuries to the brain by up to 88%, the head by up to 85%, and the face by up to 65%.
... citing Thompson, Rivara and Thompson.
The report uses two measures: cycling head injuries per 100,000 of population, and cycling head injuries per injured cyclist. They briefly mention the possibility that reduced cycling rates might account for lower head injuries per population.
From graphs of cycling head injuries per 100,000 of population plotted against years, showing the year of legislation, we see the legislaton had no obvious effect, and the report says this.
Similarly, there was no significant change in the "cycling head injuries per injured cyclist" measure pre- and post-law. According to the graphs, in almost all provinces where this measure reduced for children, it increased for adults. And vice versa. I don't suppose this is statistically significant either.
<rant>Cycling head injuries per 100,000 of population is an appalling measure. Any researcher who uses it should be made to cycle a thousand times around the Bow roundabout (with helmet, if they wish). It is an appalling measure because a country that totally banned cycling would achieve an excellent score: zero cycling head injuries per 100,000 of population.</rant>
Re: Guardian article - worth a read
One of the most interesting indicators of the effect of helmet wearing on cycling injuries comes from Canada - Ontario to be precise. It show just what effect on head injuries in injured cyclists is from large changes in helmet wearing rates.
- Photo Apr 14, 6 17 37.gif (9.21 KiB) Viewed 1576 times
Re: Guardian article - worth a read
These results repeat the findings in New Zealand and Australia where time trend analysis of cyclists populations does not support the case controlled studies of helmet safety.
You may be able to show that helmets have a protective effect, but that's not translating into improved safety when you look at injury trends in populations after helmet compulsion.
The Canadian linear trend graph is not dissimilar to the New Zealand equivalent:
You may be able to show that helmets have a protective effect, but that's not translating into improved safety when you look at injury trends in populations after helmet compulsion.
The Canadian linear trend graph is not dissimilar to the New Zealand equivalent:
For a successful technology, reality must take precedence over public relations, for nature cannot be fooled - Richard Feynman
Re: Guardian article - worth a read
Geriatrix wrote:These results repeat the findings in New Zealand and Australia where time trend analysis of cyclists populations does not support the case controlled studies of helmet safety.
You may be able to show that helmets have a protective effect, but that's not translating into improved safety when you look at injury trends in populations after helmet compulsion.
The Canadian linear trend graph is not dissimilar to the New Zealand equivalent:
Both this graph, and the one in the previous post, fall into the same trap: what is the scaling for head injuries? The original Wiki article doesn't say, either (I didn't follow as far as the source). If it is a 'percentage', percentage of what? If one takes this graph literally, have over 40% of cyclists suffered head injuries in a single year? Or is it 40% of the whole population?
The data, even if they demonstrate their failure to support the efficacy of helmets, are meaningless as displayed.
Last edited by 661-Pete on 16 May 2013, 1:36pm, edited 1 time in total.
Suppose that this room is a lift. The support breaks and down we go with ever-increasing velocity.
Let us pass the time by performing physical experiments...
--- Arthur Eddington (creator of the Eddington Number).
Let us pass the time by performing physical experiments...
--- Arthur Eddington (creator of the Eddington Number).
Re: Guardian article - worth a read
661-Pete wrote:Both this graph, and the one in the previous post, fall into the same trap: what is the scaling for head injuries?
The NZ graph measures head injuries as a percentage of limb injuries (Robinson 2001). The assumption being that helmets have no impact on limb injuries, so if helmets do have a safety benefit there should be a downward step in the head injuries when helmet compulsion is introduced.
For a successful technology, reality must take precedence over public relations, for nature cannot be fooled - Richard Feynman
Re: Guardian article - worth a read
Geriatrix wrote:661-Pete wrote:Both this graph, and the one in the previous post, fall into the same trap: what is the scaling for head injuries?
The NZ graph measures head injuries measured as a percentage of limb injuries (Robinson 2001). The assumption being that helmets have no impact on limb injuries, so if helmets do have a safety benefit there should be a downward step in the head injuries when helmet compulsion is introduced.
That explains a little. Though it ought to be 'percentage of all cycling-related injuries'. Presumably all recorded ones that is: ones that required attention of a doctor, nurse, ambulance or paramedic.
Suppose that this room is a lift. The support breaks and down we go with ever-increasing velocity.
Let us pass the time by performing physical experiments...
--- Arthur Eddington (creator of the Eddington Number).
Let us pass the time by performing physical experiments...
--- Arthur Eddington (creator of the Eddington Number).
Re: Guardian article - worth a read
It takes the assumption that helmets do not prevent serious limb injuries, and in fact have no affect on limb injuries (which is conservative - there are theories that suggest the opposite)
- As helmet wearing does not affect limb injuries then these injuries form our baseline, and a built in adjustment for other factors.
- IF helmets are effective on a population basis then we should see a change in the proportion of "serious" head/limb/both injuries amongst cyclists. (Serious is here defined as - bad enough to go into hospital, where we can count.)
Therefore the complete absence of any perceivable change in the overall trend of injury distribution provides excellent evidence that there is no significant benefit to wearing a helmet. I suspect that if data was collected about neck injuries then a similar pattern would present (although a smaller %age scale might be needed).
- As helmet wearing does not affect limb injuries then these injuries form our baseline, and a built in adjustment for other factors.
- IF helmets are effective on a population basis then we should see a change in the proportion of "serious" head/limb/both injuries amongst cyclists. (Serious is here defined as - bad enough to go into hospital, where we can count.)
Therefore the complete absence of any perceivable change in the overall trend of injury distribution provides excellent evidence that there is no significant benefit to wearing a helmet. I suspect that if data was collected about neck injuries then a similar pattern would present (although a smaller %age scale might be needed).
A shortcut has to be a challenge, otherwise it would just be the way. No situation is so dire that panic cannot make it worse.
There are two kinds of people in this world: those can extrapolate from incomplete data.
There are two kinds of people in this world: those can extrapolate from incomplete data.
Re: Guardian article - worth a read
[XAP]Bob wrote:It takes the assumption that helmets do not prevent serious limb injuries, and in fact have no affect on limb injuries (which is conservative - there are theories that suggest the opposite)
Yes my statement is badly worded. It should have read "helmets offer no protective value to limbs in the event of a crash".
For a successful technology, reality must take precedence over public relations, for nature cannot be fooled - Richard Feynman
Re: Guardian article - worth a read
Geriatrix wrote:Yes my statement is badly worded. It should have read "helmets offer no protective value to limbs in the event of a crash".
Where have you been all this time to not know that the Thompson Rivara & Thompson data shows that helmets prevent 85% of head injuries and 75% of leg injuries?
Re: Guardian article - worth a read
Geriatrix wrote:The Canadian linear trend graph is not dissimilar to the New Zealand equivalent
The nice thing about the Canadian graph is it shows a downward trend in head injuries when helmet use is both increasing and decreasing (due to a lack of enforcement of their helmet law). Too often people look at the New Zealand graph and say head injuries after are lower than head injuries before ergo helmets work, neglecting that it is just a background trend independent of helmet wearing.
Re: Guardian article - worth a read
Ben Goldacre' and David Spiegelhalter's comments on the Canadian report.
With a disproportionate emphasis on politics.
Standing over all this methodological complexity is a layer of politics, culture, and psychology.
With a disproportionate emphasis on politics.
For a successful technology, reality must take precedence over public relations, for nature cannot be fooled - Richard Feynman