Methods

Matched pair cohort study

Participants

Motorcycle crash driver/passenger pairs, identified by FARS (entire USA) from 1976‐1989, where both riders 14 years or older where one or both died. (N=8,816 pairs)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death within 30 days of crash

Notes

25% of eligible pairs excluded due to missing data on potential confounders or helmet use. Confounders measured incl: age, gender, seating position, 'police reported BAL' from FARS.
Study design assumed participant pairs matched for environmental factors including speed, road conditions etc.
FARS validity dependent upon police reporting ‐ differential misclassification of exposure and confounders unlikely
Provided a fatality risk ratio adjusted for age, gender and seating position (N=8816 pairs) and another adjusted for 'police reported BAL' (N=4265 pairs).
Authors note that when results stratified by year, effectiveness increases. Helmet effectiveness decreased in crashes involving collisions with other vehicles compared with non‐collision crashes and helmets appeared more effective in less severe crashes.

Methods

Retrospective cross‐sectional study

Participants

Police reported motorcycle crash victims where participants were able to be linked with medical record via probabilistic linkage in state of Wisconsin for 1991. (N=3009)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death and head injury as recorded in medical record

Notes

No potential confounders measured.
Approx 6% missing helmet status excluded. Also states approx 7% linking matches made by computer incorrect

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash victims presenting to one hospital in the USA from Jan 1, 1983 to May 31 1987. (N=367)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death, brain injury, neck injuries and maxillofacial injuries as defined from medical record

Notes

No potential confounders measured

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash victims over 15 years of age presenting to a level 1 trauma centre from July 1996 to Oct 2000. (N=216)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Mortality and head and/or neck injury AIS as recorded on trauma registry

Notes

Potential confounders measured but none adjusted for.
Raw numbers only given for mortality. Head, neck and facial injuries results recorded as average AIS compared between helmeted and unhelmeted riders

Methods

Prospective cross‐sectional study

Participants

Selection of motorcycle crashes identified from police and ambulance radio links and by hospital casualty officers over 4 month period from 1978‐1979. (N=45)

Interventions

Full‐face helmets compared with open‐face helmets.

Outcomes

Head injury and maxillofacial injury as recorded on medical records

Notes

Besides age, no potential confounders measured and none adjusted for.
11 deaths excluded. No indication of comparability of those selected for inclusion compared with general motorcycle riders in area.

Methods

Case‐control study.

Participants

Participants were motorcycle crash injured patients recruited from 7 hospitals in the area selected because they were more likely include patients with major trauma. Cases were those who had head injury (N=96) and controls were were non‐head injured participants (N=177).

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head trauma (and severity) as defined by medical records

Notes

Potential confounders measured but not adjusted for.
31% participants had unknown helmet status.
Quotes OR for death with helmet use as intervention factor but no CI given (OR 0.16)

Methods

Retrospective cross‐sectional study

Participants

Systematic sampling of motorcycle accident cases from Wisconsin state accident records from 1977 to 1979. (N=888)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury as classified on scene ‐ not verified by medical records

Notes

Potential confounders such as speed and 'manner of collision' measured and results stratifed by these factors. Found for all strata of speeds greater than 25mph, there was a significant difference in head injury incidence between helmeted and non‐helmeted.
3% missing helmet data.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle drivers involved in a crash identifed from one level 1 trauma centre trauma registry from 1995 to 2000. (N=311)

Interventions

Motorcycle helmet use compared with no helmet use.

Outcomes

Head injury and serious head head injury defined from ICD9 and AIS codes of medical record

Notes

Measured potential confounders such as age, gender, riding season, type and time of accident, drug screen, blood alcohol from trauma registry and adjusted for this in estimate of effect.
Only small loss of participants due to unknown helmet use.

Methods

Prospective cross‐sectional study

Participants

Motorcycle riders injured and admitted to any of the 4 hospital EDs in the region. (N=475)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury and serious head injury based on medical records.

Notes

Potential confounders measured but not adjusted for.
9% excluded due to unknown helmet use.

Methods

Retrospective cross‐sectional study

Participants

Injured motorcycle riders who were treated at participating Level 1 & 2 trauma centres across the USA from 1982‐1988 and identified on the trauma registry. (N=1066)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Mortality, head injury and/or spinal cord injury as recorded on medical records (trauma registry)

Notes

Potential confounders measured but not adjusted for. 87% of selected participants had missing helmet data and were excluded from analysis.
Found average severity of head/brain/spinal injury significantly less for helmeted versus unhelmeted riders.

Methods

Retrospective cross‐sectional study

Participants

Patients (ages 16 to 24 years) that sustained a head injury following a motorcycle or moped accident. Selection of participants not described. (N=192)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Mortality and head injury severity

Notes

Potential confounders measured but not adjusted for. No description of method of selection of participants and all from a subgroup of those alread with a head injury.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash presenting to hospital ED in 1990. (N=2498)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death (or survival) up to 4 months after discharge from hospital.
Head injury (as per AIS score)

Notes

Measured confounders but none adjusted for.
<20% missing data due to either unknown helmet use or injury status.

Methods

Matched pair cohort study.

Participants

Motorcycle crash driver/passenger pairs identified by FARS (entire USA) during 1975‐1986 where both riders were 16 years or older and one or both riders died.
Pairs had to be matched for age (driver and passenger ages within 3 years of one another) and only males included (N=4714 fatalities)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death within 30 days of crash

Notes

Study design matched for age, gender (by excluding females as too few all female pairs). This resulted in loss of 42% fatality data.
Authors found driver seating position had greater risk of fatality.

Methods

Prospective cross‐sectional study

Participants

Consecutive motorcycle drivers presenting as trauma victims to one hospital from 1972‐1974. (N=124)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death because of head trauma.
Head trauma ‐ defined as a "contusion".
Facial injuries.

Notes

Potential confounders not measured nor controlled for.
No loss to follow up or missing information data provided.
Only outcome of 'death because of head trauma' used because inadequate definitions given for other outcomes.

Methods

Case‐control study

Participants

Cases and controls identified from traffic accident reports of motorcycle crashes investigated by Division of Motor Vehicles during Jan1, 1989 to Dec 31, 1990 in El Paso County (Colorado, USA) ie: all motorcycle crashes where there was personal injury or property damage. Cases were those who crashed and sustained a traumatic brain injury or skull fracture identified thought the Colorado dept of health severe head injury surveillance system (based on death certificates, discharge ICD‐9 codes, text diagnoses). (N=71)
Controls were those who crashed and did not sustain a head injury (ie: were not identified by the head injury surveillance system) (N=417)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury: traumatic brain injury or skull fracture as defined by ICD‐9 codes or comparable medical record diagnoses

Notes

Confounders such as DUI, age, passenger status, crash time and type, motorcycle speed, citatation for various motorcycle offences measured and adjusted for.
Misclassification of minor head injury cases (ie: superficial lacerations or concussions) as controls is possible and if helmets are protective, this will result in underestimate of effect.

Methods

Prospective cross‐sectional study

Participants

Used participants from Hurt 1981 study. See description of this study. (N=644)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Fatality, head and neck injury

Notes

Uses econometric model to take account of confounders such as age, alcohol consumption, rider on‐road experience and speed in predicting effect of motorcycle helmets on outcomes. Model used has been criticised.
Excluded some data (28%) due to missing values and for some models assigned a mean value to missing data.

Methods

Retrospective cross‐sectional study

Participants

Identified initially from Motor Vehicle Accident files for motorcycle crashes occuring in 4 counties during Sept 1, 1986 to Dec 31, 1987 where at least one crash victim was transported to hospital. (N=742)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury as defined by a medical record

Notes

Potential confounders measured but not adjusted for.
21% loss of participants due to missing injury data or unknown helmet data.

Methods

Retrospective cross‐sectional study

Participants

Non‐fatally injured motorcyclists admitted to any of 28 non‐randomly selected hospitals across 10 Californian counties from1991 to 1993. (N=4895)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Facial injuries defined from medical records.

Notes

Potential confounders measured but not adjusted for. 15% of particpants excluded due to unknown helmet status.

Methods

Retrospective cross‐sectional study

Participants

Included by linking databases including death certificates, hospital data, highway patrol motor vehicle crash report over 1977 to 1980 for one US state. (N=2874)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head/neck/facial injury and deaths defined from medical records and death certificates.

Notes

Potential confounders measured but not adjusted for.
Unknown proportion of participants lost in linkage process, 11% unknown helmet status.

Methods

Prospective cross‐sectional study

Participants

Non‐random selection of reported motorcycle crash victims that investigators were notified of by emergency services and able to investigate on‐scene. (N=878)

Interventions

Motorcycle helmet use compared with no helmet use.
Different helmet types.

Outcomes

Head and neck injuries in relation to helmet or no helmet use and type of helmet use.

Notes

Potential confounders measured but not adjusted for.

Methods

Retrospective cross‐sectional study

Participants

All injured motorcycle crash victims admitted to a regional level 1 trauma centre over 4 years. (N=331)

Interventions

Motorcycle helmet use compared with no helmet use.

Outcomes

Injuries including skull fracture, facial fracture and cervical spine injury as recorded in medical records

Notes

Potential confounders measured but not adjusted for.
No mention of any lost data or participants. Incidence of skull fracture found to be significantly less in those wearing helmets (p<0.01)

Methods

Retrospective cross‐sectional study

Participants

All drivers of motorcycles involved in police reported crashes in 7 US states that were able to be linked to injury databases (EMS, hospital). (N=10353)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury and death as confirmed through linkage with medical records and death certificates.

Notes

Potential confounders measured but not adjusted for.
Unclear as to lost data through non‐linkage. Also 38% unknown helmet use in NY state data and one state, Utah, excluded due to inability to distinguish between helmeted and non‐helmeted riders.
No raw data or confidence intervals provided with estimates of effect.
Also provided information on seat belt effectiveness.

Methods

Prospective cross‐sectional study

Participants

Motorcycle rdiers involved in a crash presenting less than 24 hours after the crash to one of 8 hospitals in 4 counties. Engine size must be 150cc or greater and have known helmet status. (N=398)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death and injuries including head and/or neck injury, facial injury and neck injury as recorded from medical records.

Notes

Potential confounders measured but not adjusted for with injury as outcome (confounders controlled for in outcome of overall injury severity).

Methods

Retrospective cross‐sectional study

Participants

All motorcycle and moped riders killed in traffic accident identified through autopsy reports from 1977‐1983. (N=132)

Interventions

Motorcycle helmet use compared with no helmet use.
Full‐face and open‐face helmet types.

Outcomes

Head injuries and neck injuries as defined on autopsy report

Notes

Potential confounders not measured.
Authors stated that autopsies are conducted on all deaths in traffic accidents in Sweden and therefore likely to have included all deaths in region.

Methods

Retrospective cross‐sectional study

Participants

All motorcycle riders who crashed and required medical treatment as identified from police reports, death certificates, hospital records in county. Non‐county residents and females excluded (N=626)

Interventions

Motorcycle helmet use compared with no helmet use.

Outcomes

Serious and non‐serious head injury. No clear indication of definition of head injury although serious head injury defined as that resulting in death, hospitalisation, boney fracture and requiring continuous medical care beyond 2 visits

Notes

Potential selection bias as only 628 male drivers responded to questionnaire of 1273 injured persons. Furthermore, only 268 of the 628 male drivers had speed and helmet use data for the stratified analysis.

Methods

Retrospective cross‐sectional study

Participants

Drivers from fatal or severe injury motorcycle crashes reported to police in LA county from July 1988 to Oct 1989 where drivers records could be linked to coroner or hospital records.
(N=477)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head Injury from medical records.
Fatality from medical record or coroner

Notes

Potential confounders measured but no adjusted for.
60% data missing due principally to non‐linkage of reported crashes.

Methods

Retrospective cross‐sectional study

Participants

Non‐fatally injured motorcycle crash victims presenting to 18 non‐randomly selected hospitals in 10 California counties over a period Jan 1, 1991 to Dec 31, 1993.

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury and severe head injury as recorded on medical records

Notes

No potential confounders measured.
<20% participants excluded due to missing helmet or injury data.

Methods

Prospective cross‐sectional study

Participants

injured motorcycle riders following a crash aged 14‐35 years presenting to 2 selected hospitals in Italy during Jan to June 1999. (N=736).

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head Injury based on data collected by investigators or those recruited by investigators.

Notes

Potential confounders measured by none adjusted for.
No apparent missing data.

Methods

Cohort study

Participants

Junior college students from 3 randomly selected colleges in a rural and urban area of Taiwan.
Participants followed for 18 months from Nov 1994 to June 1996. (N=1889 crashes)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head, neck and facial injury reported by participants on a questionnaire and supplemented by school records.

Notes

Potential confounders measured and although states a multivariate analysis conducted, this is not shown and attempts to contact authors have been unsuccessful.
Average response rate to questionnaire 92%.
20% participants lost to follow up due to graduation of one year. Participants could be included more than once in this study as investigators collected relevant injury data for each crash sustained by the participant and there were more crashes (N=1889) than individual riders involved (N=1284) and therefore despite having raw numbers, no RR were extrapolated.
Reliability of questionnaire responses assessed through re‐test of 150 randomly selected questionnaires.

Methods

Retrospective cross‐sectional study

Participants

Injured motorcycle riders presenting to one hospital in Texas during Feb 1985 to Jan 1986 (N=88)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury as recorded on trauma registry

Notes

No potential confounders measured.
Only reported a difference in average nervous system score between helmeted and nonhelmeted riders. No estimate of statistical significance provided.
45% participants excluded due to unknown helmet use status.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle accident victims presenting to a US trauma centre from July 1, 1978 to Nov 30, 1979. (N=263)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death within first week following admission to hospital.
Major head injury from medical records

Notes

Potential confounder not measured.
15% participants with unknown helmet use.

Methods

Retrospective cross‐sectional study

Participants

Victims of motorcycle crashes requiring transport according to county triage criteria to one trauma centre (N=213)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury as recorded on medical record

Notes

Potential confouders measured but not adjusted for.
5% participants unknown helmet use.
Found significant head injuries accounted for 9% of injuries in helmeted patients compared with 37% in unhelmeted.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash victims seen a one level 1 trauma centre over 45 months (N=347)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head and/or neck injury and neck injury alone as described in medical records.
Death as recorded from medicla record.

Notes

No potential confounders measured.
28% of participants had unknown helmet status.

Methods

Matched pair cohort study

Participants

Motorcycle crash driver/passenger pairs, identified by FARS (entire USA) during 1980‐1998, where riders were 16 years or older and one or both riders in the pair died. (N=9,222 pairs)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death within 30 days of crash

Notes

Study design matches for motorcycle characteristics such as type, speed and environmental factors.
20% pairs excluded due to missing data. Those with missing helmet data had similar age and gender distribution as those with helmet data. Confounders measured and adjusted for included gender, age, rider position.

Methods

Prospective cross‐sectional study

Participants

Motorcyclists who died in a crash in the Adelaide (Australia) metropolitan area between 1983‐1991 (N=159)

Interventions

Full face motorcycle helmet compared with open‐faced motorcycle helmet

Outcomes

Cervical spine injury verifed by autopsy examination (i.e. only in motorcyclists who died)

Notes

Authors comment on subgroups with head impact cases and helmet retention.
Study measured confounders such as age, head impact crash type, BAL but did not find any significant predictor of cervical spine injury and therefore did not control for these in final OR. Study base includes all crashes in the area but selects from this a subset of all those who died. Presents evidence to suggest there is no systematic difference between those motorcycle riders who live or die and the type of helmet worn. 8% missing autopsy data

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash victims admitted to a level 1 trauma centre between Jan 1, 1985 to Jan 1 1990. (N=425)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death, head injury and neck injury as recorded in medical record

Notes

Potential confounders measured. Gives an estimate effectiveness of helmets for mortality and head injury weighted by a non‐head Injury severity score.
14% participants have no helmet data

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash victims identified from 28 hospital databases across 4 US states. (N=1056)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head Injury according to AIS from medical records.
Mortality from medical records and some on‐the‐scene ambulance and police data.
Cervical spine injury as recorded in medical records

Notes

Potential confounders measured by none adjusted for.
<20% participants excluded due to lack of helmet use data.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle crash victims identified via a state public health trauma registry including all level 1 & 2 trauma centres in state from July 1, 1991 to Dec 31, 1992. (N=819)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head Injury according to AIS

Notes

Potential confounders measured but none adjusted for.
26% of those identified had missing helmet status but investigators noted no significant difference in demographics of those with missing helmet status.

Methods

Retrospective cross‐sectional study

Participants

Identified by traffic police as any motorcycle riders involved in a motor vehicle accident where at least one person was killed or injured in 1985 and 1994 in Greece.

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death rather than injury according to traffic police statistical department

Notes

Masured age and gender and states final estimate of effect is adjusted for confounders but does not state what these are. Attempts to contact authors to clarify this have been unsuccessful.
Authors state that approximately 20% of information was missing due to incomplete returns.

Methods

Prospective cross‐sectional study

Participants

Injury motorcycle riders directly transferred to one of two selected hospital emergency departments where the accident occurred in the Hadyai municipality. from April to Sept 1997 (N=581)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head and/or neck injury according to AIS

Notes

Potential confounders measured but not adjusted for

Methods

Retrospective cross‐sectional study

Participants

All fatally injured motorcyclist, moped, motorscooter and minibike riders as identified by California FARS during 1987‐1988 and able to be linked with California MCOD and SMD files with known helmet status. (N=1025)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury in those who died defined by ICD‐9 codes 800‐803, 850‐854 inclusive.

Notes

Reports adjusted OR for odds of head injury with helmet use adjusted for gender, seating position, cycle damage and crash type.
Authors report limitation of high proportion (40%) of deaths have unspecified injuries thereby potential misclassification of those with head injury. Authors recalculated OR re‐classifying those with unspecified injuries as non‐head injuries and found that OR still showed helmets protective against death.
Confounders such as speed, BAL not considered.

Methods

Retrospective cross‐sectional study.

Participants

Motorcycle drivers only who crashed in Washington state in 1989 as identified by State patrol records and linked to hospital and death records

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head Injury defined by ICD‐9 codes and then mapped to AIS scores.
Death defined by death certificate.
Facial injury defined as AIS>0.

Notes

Reports adjusted RR for risk of death with helmet use (Rivara 2003) and adjusted OR for odds of head injury with helmet use.
Confounders measured included age, gender, locality of crash, environmental conditions. 23% participants missing from head injury data because of non‐linkage.

Methods

Retrospective cross‐sectional study

Participants

All motorcycle riders involved in a crash hospitalised in any of 8 level 1 or 2 trauma centres in state during Oct 1, 1987 to Jan 1, 1991. (N=460)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury defined by AIS from medical record.

Notes

Potential confounders measured but none adjusted for.
48% of participants excluded due to unknown helmet use.

Methods

Retrospective cross‐sectional study

Participants

Dead motorcycle crash victims identified from police and coroner reports in one county from July 1, 1988 to Oct 31, 1989. (N=164)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head and/or neck injury, facial fracture or neck injury as recorded in medical record or autopsy

Notes

Measured other injuries as potential confounders and stratified findings according to those with equally severe non‐head injuries.

Methods

Retrospective cross‐sectional study

Participants

All motorcycle drivers involved in a crash that was reported to police and transported to hospital in Maryland USA during July 1987 to June 1988. (N=721)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury defined by medical records

Notes

Potential confounders measured but none adjusted for.
25% participants had missing data and were excluded.

Methods

Retrospective cross‐sectional study.

Participants

Traffic accidents reported by police in Fukuoka Prefecture (Japan) in 1990 catagorised into motorcycle crashed and motorcar accidents. (N=1077)

Interventions

Motorcycle helmet use compared with no helmet use.
Seat belt use compared with no seat belt use.

Outcomes

Death within 24 hours of accident compared to no injury for both motorcyclists and motorcar occupants.
Only outcome for motorcyclists examined in review.

Notes

OR stratified by gender and only given for male riders. Compared population who died with those with no injuries.
OR adjusted for age and alcohol use.
Speed found to be an effect‐modifier, therefore at speeds between 30‐50km/h helmets have protective effect against death but at speeds >50km/hr the protective effect is not significant.
Authors also note that at the time of the study, Japan had no on‐scene emergency management of injuries which may affect generalisability of results.

Methods

Prospective cross‐sectional study.

Participants

Injured motorcycle riders seen by author in one hospital during May to Dec 1985. (N=302)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head Injury measured by author according to AIS

Notes

Unclear description of methodology including selection of participants and blinding of assessor.

Methods

Case‐control study.

Participants

Motorcycle riders receiving care for crash injuries in the ED of one of 16 hospitals in Taipei (Taiwan) from August 1 to Oct 15 1990.
Cases were those recieving care for head injuries. (N=562)
ED Controls were randomly selected individuals seeking care for injuries other than head injuries. (N=789)
Street Controls: Were photographs of uninjured, non‐crash motorcycle riders matched for time and place of daytime cases. (N=1094)

Interventions

Motorcycle helmet use compared with no helmet use.
Full‐face motorcycle helmets compared to no helmet.
Non‐full face helmets (full helmet or partial coverage helmet) compared to no helmet

Outcomes

Head injury defined as brain injury, cerebral concussion, skull fracture, clinically proven unconsciousness, amnesia or neurologic sequalae on a re‐visit to the ED. Soft‐tissue/scalp injuries are not included
Head injury severity as measured by GCS scores.

Notes

Reported comparative estimates for ED and street controls.
One ED excluded because suspected bias in selection of participants (5% excluded).
Confounders including gender, age, rider position, motorcycle type, weather, place of accident measured.
Quasi‐random sampling of participants and unable to guarantee completeness of sample but odds of helmet use for street controls found to be similar in ED cases and controls.

Methods

Prospective cross‐sectional study

Participants

A consecutive sample of motorcycle and moped accident victims admitted to university hospitals in one town from May 1, 1992 to April 30, 1994. (N=221)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury, head injury severity and cervical spine injury as recorded from a medical record

Notes

Stratified results according to non‐head injuries (defined as a surrogate for kinetic energy) and found the ratio of head and facial injuries per a patient was more than double in non‐helmeted patients compared with helmeted.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle accident victims as identified by routine crash data in Sydney during a three month period. (N=1552)

Interventions

Full face motorcycle helmets compared with jet‐style motorcycle helmets

Outcomes

Head injury, facial injury and neck injury from police reports and supplemented by medical records

Notes

Older study may mean different helmet standards and manufacturing practices mean comparisions not generalisable.
No confounders measured.
Methodology brief and not always clear.

Methods

Retrospective cross‐sectional study

Participants

Motorcycle accident victims transferred to a major trauma centre on helicopter ambulane (Lifestar) over a 5 year period. (N=80)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Cervical spine injury and fatality from medical records

Notes

Potential confounders measured but none adjusted for

Methods

Prospective cross‐sectional study

Participants

Used participants from Hurt 1981 study. See description of this study.

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Head injury

Notes

Statistical model controlled for alcohol and speed of rider in estimating predicted effect of helmets. Found that helmets lead to 42% increase in riders with no head injury

Methods

Matched pair cohort study

Participants

Motorcycle crash driver/passenger pairs, identified by FARS (entire USA) from 1982‐1987, where both riders 14 years or older where one or both died. (N=5292 riders)

Interventions

Motorcycle helmet use compared with no helmet use

Outcomes

Death within 30 days of crash

Notes

Confounders including rider gender and age not measured nor adjusted for.
Helmet effectiveness 29% (no CI given). Effectiveness stratified by passenger (30%) and driver (27%)