Letters
Emergency medicine
Volume 198 - Issue 8

The effectiveness of helmets in reducing head injuries and hospital treatment costs: a multicentre study

Authors:  Michael M Dinh, Kate Curtis and Rebecca Ivers

Med J Aust 2013; 198 (8): 415-417. || doi: 10.5694/mja12.11580
Published online: 20 May 2013
Trauma registry data on patients admitted to Sydney hospitals following bicycle or motorcycle accidents reveal the protective effect of wearing a helmet.
Download PDF (150 KB)
Share: Share on Blue Sky Share on Facebook Share on LinkedIn Share by email

To the Editor: Debate continues regarding the health benefits and consequences of helmet use in pedal cyclists.1 Australia is one of few countries in the world with mandatory helmet laws for both pedal cyclists and motorcyclists. To place the protective effect of helmets in pedal cyclists into perspective, we report on the relationship between helmet use and head injury severity in a retrospective cohort of both pedal cyclists and motorcyclists.

Trauma registry data on such patients admitted to seven tertiary level hospitals in Sydney, New South Wales (Liverpool, St George, Royal Prince Alfred, Westmead, Royal North Shore, St Vincent’s and Prince of Wales hospitals) between July 2008 and June 2009 were obtained. Patients were included if they were aged 15 years or over with an incident occurring on a public road. The Abbreviated Injury Scale and Injury Severity Score were used to classify body regions and severity of injury, respectively. Helmet use, incident and other injury details were routinely collected by trained data and case managers from standard ambulance and trauma clinical case notes. Inhospital costs were calculated using standardised cost weights (NSW Program and Product Data Collection, 2008–09). Primary outcomes were any head injury and severe head injury (Abbreviated Injury Scale severity score ≥ 3), including significant intracranial haemorrhages, and diffuse axonal injury. Logistic regression was used to determine odds ratios for head injury and severe head injury, adjusting for age (as a continuous variable) and location of incident (based on incident postcode) as a-priori confounders based on previous work.2

There were 398 cases identified. Of these, 50 patients (13%) had missing helmet information, leaving 348 cases analysed. Baseline characteristics stratified by helmet use are shown in the Box. For any head injury associated with helmet non-use, the adjusted odds ratio was 5.6 (95% CI, 2.1–14.9; P < 0.001) for pedal cyclists and 2.2 (95% CI, 0.9–5.0; P = 0.06) for motorcyclists, compared with helmeted patients in each group. For severe head injury associated with helmet non-use, the adjusted odds ratio was 5.5 (95% CI, 1.5–20.6; P = 0.01) for pedal cyclists and 3.5 (95% CI, 1.3–8.9; P = 0.01) for motorcyclists, compared with helmeted patients in each group. For the 50 patients with severe head injury, inhospital costs (AUD) were around three times higher in non-helmeted patients (median, $72 000; interquartile range, $33 000–$140 000) compared with helmeted patients (median, $24 000; interquartile range, $15 000–$60 000) (P = 0.02).

The protective effect of helmet use with respect to head injury prevention therefore appears to be greater in pedal cyclists compared with motorcyclists. There was no association observed between helmet use and diffuse axonal injury. Limitations to our study include the small number of patients with severe head injury, and the inability to control for other incident factors such as speed, collision details and intoxication. The use of hospital data biases observations towards patients with more severe injuries. Nevertheless, the results add to the growing weight of observational data supporting the use of helmets, 3-5 which should therefore be considered at least as protective for pedal cyclists as they are for motorcyclists.

Characteristics of cyclists and motorcyclists with head injuries, by helmet use

Cyclist (n = 110)

Motorcyclist (n = 238)

Characteristic

Helmet (n = 70)

Non-helmet (n = 40)

P

Helmet (n = 206)

Non-helmet (n = 32)

P

Demographic

Median age (IQR), years

41 (29–53)

35 (23–44)

0.02

31 (24–43)

25 (21–38)

0.06

Male

64 (91%)

35 (88%)

0.51

194 (94%)

31 (97%)

0.53

Incident details

After hours*

25 (36%)

14 (35%)

0.94

65 (32%)

14 (44%)

0.17

Location†

Inner Sydney

21 (30%)

11 (28%)

40 (19%)

13 (41%)

Suburban Sydney

26 (37%)

20 (50%)

76 (37%)

11 (34%)

Regional/rural

23 (33%)

9 (23%)

0.37

91 (44%)

8 (25%)

0.02

Injury severity

Median ISS (IQR)

9 (5–14)

9 (5–21)

1.0

9 (5–17)

15 (5–25)

0.15

Multiregion injury (%)

46 (66%)

29 (73%)

0.46

153 (74%)

21 (66%)

0.32

ICU

11 (16%)

7 (18%)

0.81

47 (23%)

12 (38%)

0.07

Outcomes

Head injury

27 (39%)

30 (75%)

< 0.001

68 (33%)

14 (44%)

0.23

Severe head injury

6 (9%)

9 (23%)

0.04

26 (13%)

9 (28%)

0.02

Diffuse axonal injury

0

0

na

5 (2%)

3 (9%)

0.08

Rehabilitation‡

3 (4%)

6 (15%)

0.07

35 (17%)

4 (13%)

0.53

Median cost (IQR), AU$1000

6.5 (2.8–10.7)

5.6 (2.5–15.2)

0.91

7.7 (3.0–20.7)

11.4 (4.4–41.0)

0.05

ICU = intensive care unit admission required. IQR = interquartile range. ISS = Injury Severity Score. na = not applicable. * Recorded incident times between 19:00 and 07:00 hours. † Postcode of location of incident was used to classify incident locations in the inner Sydney (within 10 km of central business district), suburban Sydney (bounded by Hornsby to the north, Royal National Park to the south and Penrith to the west), and regional and rural regions of New South Wales. ‡ Discharge from hospital to a rehabilitation facility.

Authors

  • Michael M Dinh1
  • Kate Curtis2
  • Rebecca Ivers3
  • 1. Sydney Medical School, University of Sydney, Sydney, NSW.
  • 2. St George Hospital, Sydney, NSW.
  • 3. Injury Division, The George Institute for Global Health, Sydney, NSW.

Correspondence: michael.dinh@sswahs.nsw.gov.au

Competing interests

Kate Curtis was supported by a Sydney Nursing School, University of Sydney Postdoctoral Fellowship. Funding support was provided by the NSW Institute of Injury and Trauma Management and St George Honda Trauma and Critical Care Research Program. Rebecca Ivers was supported by a National Health and Medical Research Council Career Development Fellowship. The study sponsors had no involvement in the study design, collection, analysis and interpretation of data, or the writing of the manuscript.

Acknowledgements

We acknowledge the work of Mary Lam from the University of Sydney and Kendall Bein from Royal Prince Alfred Hospital for cost and data analysis. We also thank the trauma data managers and trauma coordinators from the major trauma centres who submitted data for the original study.

References

View this article on Wiley Online Library

More like this

View more