Analysis And Comment

No clear evidence from countries that have enforced the wearing of helmets

BMJ 2006; 332 doi: (Published 23 March 2006) Cite this as: BMJ 2006;332:722

Re: Objective observation of helmet use is essential

Reply to Keatinge “Objective observation of helmet use is essential”
(1) and two email communications to us on the same topic. Your questions
and/or statements are listed below with our responses.

YOUR STATEMENTS “Cyclists without head injuries will report rates of
helmet wearing much higher than their true rate. Cyclists with recent head
injuries, as in the case-control studies, will not be able to deceive
themselves or their interviewers, and will report much lower, truer rates
of helmet wearing. This bias is enough to account for all the positive
findings in case-control studies:” (1) AND “The promising results of the
case-control studies could trivially be accounted for by the hypothesis
that most cyclists exaggerate their helmet use, but head-injured ones do
not because after a demonstrable injury to the head they can’t fool
themselves in that particular way. (Robinson’s results do not rely on self
-reporting.)” 22 June email AND “in one case-control study based in
Seattle, 7% of the cyclists with head injuries reported wearing helmets,
as compared with 24% of the emergency room controls and 23% of community
cyclists who had had an accident(2) However, less than 6% of cyclists on
the Seattle streets at the time were actually observed to wear helmets (3)
Either helmets are a serious cause of accidents, or self-reports of helmet
use are not valid” (1)

OUR ANSWERS: Your claim that there is differential reporting of
helmet use by head injured (cases) and non head injured (controls)
cyclists in published case-control studies is mere speculation. There is
no evidence that controls, or for that matter cases, report helmet use
incorrectly. There is good evidence that self-reported helmet use is
accurate. Remember, all these cyclists had an injury that caused them to
seek medical care. This is an important event which people remember
accurately. Additionally, there were no helmet laws in Seattle at the
time. Observations of helmet use in the general cycling population after
helmet legislation (4) or before and after a helmet promotion campaign (3)
provide no information as to whether any of these cyclists were wearing
helmets when they crashed and went to the hospital. The studies of
observed helmet use in the work cited above provide information on the
prevalence of helmet use in the general cycling population. The case
control studies provide the prevalence of helmet use in cyclists who seek
ED treatment following a crash, a much different circumstance.

YOUR STATEMENT: “Robinson presents the best available evidence
derived from objective assessment of helmet wearing. Her demonstration
that cycle helmet laws do not work is likely to remain the definitive
answer.” (1)

OUR ANSWER: The studies that Dorothy Robinson cites in her recent
BMJ article (1) are time series studies or ecological designs which do not
have appropriate comparison groups. We have pointed this out in our
discussions with you, Dorothy Robinson, Bill Curnow and Mayer Hillman over
the past ten years. Please re-read our replies to your criticisms of our
Cochrane review. These are available at the end of the review (5) or on
the web at the following address: www.cochrane- Additionally, Robinson selects
portions of the studies which support her arguments and ignores the
portions which indicate a reduction in head injuries following
legislation. (6) She also omits 2 studies from North America which
indicate legislation reduces head or brain injury. (7,8)

YOUR QUESTION: “I would be interested to know how this data
collection was done in this study, and in particular how helmet use was
ascertained. Was it done by observation, by questionnaire in hospital, or
by later follow-up?" (Email of 29 May, 2006 to inquire about our Cycle
Helmet studies: NEJM, 1989 and JAMA, 1996.(2,9)

OUR ANSWER: In reply, both studies were prospective case control
studies so the data were collected in an ongoing fashion as the bicycle
crashes occurred and the riders sought treatment in the emergency
departments (ED’s) at the participating hospitals. Hospital ED’s were
visited at least once a week to identify bicycle riders who were injured.
Helmet use was determined by self-report using mailed questionnaires sent
to each identified cyclist following the ED visit. People who did not
respond promptly were interviewed by telephone using the same
questionnaire. Helmet use was ascertained from these completed

YOUR STATEMENTS: “I would be interested in any information you can
give on the degree of independence of these two sources of data on helmet
wearing, and on any records of whether emergency room staff actually saw
the helmets themselves.” AND “The hypothesis that most cyclists exaggerate
their helmet use, but the head-injured ones do not because after a
demonstrable injury to the head they can’t fool themselves in that
particular way.” (22 June Email )

OUR ANSWER: We also abstracted the medical record of those cyclists
in the study and tabulated information about helmet use as recorded by ED
personnel. The medical record was written when the patient was treated in
the ED. This ascertainment is independent of helmet use reported
subsequently on the research questionnaire. This permitted independent
corroboration of reported helmet use. We compared the report of helmet use
in the medical record to the report of helmet use on the questionnaire and
assessed the agreement between the ED record and the questionnaire report
of helmet use. Agreement was almost 100% for both cases and controls. (9)

In many cases the attending physician or emergency staff saw the

In addition, independent laboratory analysis of 500 of the helmets
which were damaged in the crashes indicated that cyclists reported helmet
type correctly. This is a more complex memory task then reporting whether
a helmet was worn. (10) In conclusion, the multiple methods used to
ascertain helmet use provide solid evidence for the validity of self
reported helmet use.

SUMMARY : The question here for you and other doubters at this point
is; Are you really trying to pursue the science assessing cycle helmet
efficacy and population effectiveness or is this some sort of crusade on
your part to fit the science to your beliefs? Over the years my colleagues
and I have explained case control studies and other epidemiologic methods
of conducting studies and analyzing statistics. (See our Cochrane review,
our comments replying to numerous criticisms, Peter Cummings’ recent
article in Accident Analysis and Prevention,” Misconceptions regarding
case-control studies of bicycle helmets and head injury”.(5,11) You could
also read the article by Brent Hagel et al “Arguments against helmet
legislation are flawed” which is published the same March issue of BMJ as
Dorothy Robinson’s article. The rapid responses from anti-helmet writers
are simply offering the same old arguments, opinions and speculations that
have been written by the same people for years. If you truly wish to move
the science forward in this area, we respectfully request that you all
take some formal courses in the use of Epidemiological methods, non-
experimental design and biostatistics. After that you should review the
whole body of peer-reviewed published information with an open mind and
see where your review leads you. We predict it will lead you to the
conclusion that we have reached, that cycle helmets prevent head and brain
injuries and increased wearing rates result in decreased head and brain
injuries in cyclists. (5,6)

1. Keatinge,RM. Objective observation of helmet use is essential. BMJ 14
May 2006.

2. Thompson RS, Rivara FP,Thompson DC. A case-control study of the
effectiveness of bicycle safety helmets. NEJM 1989;320:1361-7.

3. DiGuiseppi CG, Rivara FP, Koepsell T, Polissar L. Bicycle helmet
use by children. Evaluation of a community-wide helmet campaign JAMA

4. Robinson DL. No clear evidence from countries that have enforced
the wearing of helmets. BMJ 2006;332:722-5.

5. Thompson DC, Rivara FP, Thompson RS. Helmets for preventing head
and facial injuries in bicyclists. Cochrane Database Syst Rev 2000;(2):

6. Hagel B, Macpherson A, Rivara FP, Pless B. Arguments against
helmet legislation are flawed. BMJ 2006:332:725-6.

7. Lee BH, Schofer JL, Koppelman FS. Bicycle safety helmet
legislation and bicycle-related non-fatal injuries in California. Accid
Anal Prev 2005;37:93-102

8. Macpherson AK, To TM, Macarthur C, Chipman ML, Wright JG, Parkin
PC. Impact of mandatory helmet legislation on bicycle-related head
injuries in children: a population-based study. Pediatrics 2002;110:e60.

9. Thompson DC, Rivara FP, Thompson RS. Effectiveness of bicycle
safety helmets in preventing head injuries: A case-control study. JAMA

10. Ching RP, Thompson DC, Thompson RS, Thomas DJ, Chilcott WC,
Rivara FP. Acc Anal Prev 1997;29:555-562.

11. Cummings P, Rivara FP, Thompson DC, Thompson RS. Misconceptions
regarding case-control studies of bicycle helmets and head injury. Accid
Anal Prev 2006;38:636-643.

Competing interests:
None declared

Competing interests: No competing interests

08 July 2006
Diane Thompson
Epidemiologist (retired)
Robert S. Thompson
Harborview Injury Prevention & Research Center, Univ of Washington, Seattle, WA 98104 USA