Article Text
PDF
Abstract
Objective The increasing frequency of bicycle-related injuries is due to the growing elderly population and their increasing physical activity. This study aimed to compare the protective effects of helmets on bicycle-related injuries in elderly individuals compared with those in younger adults.
Methods Data from the Korean emergency department–based Injury In-depth Surveillance database from eight emergency departments during 2011–2016 were retrospectively analysed. The subjects sustained injuries while riding bicycles. Cases with unknown clinical outcomes were excluded. Covariates included mechanism, place and time of injury. The primary outcome was traumatic brain injury (TBI) incidence, and the secondary outcomes were in-hospital mortality and severe trauma. The effects of helmets on these outcomes were analysed and differences in effects were determined using logistic regression analysis. Subsequently, the differences in the effects of helmets use between age groups were examined by using interaction analysis
Results Of 7181 adults, 1253 were aged >65 years. The injury incidents showed a bimodal pattern with peaks around ages 20 and 50 years. Meanwhile, the helmet-wearing rate showed a unimodal pattern with its peak at age 35–40 years; it decreased consistently with age. By multivariate analysis, helmet-wearing was associated with a reduced TBI incidence (OR 0.76; 95% CI 0.57 to 0.99) and severe trauma (OR 0.78; 95% CI 0.65 to 0.93). The effects of helmets increased in elderly individuals (TBI (p=0.022) and severe trauma (p=0.024)).
Conclusion The protective effects of helmets on bicycle-related injuries are greater for elderly individuals, thus reducing TBI incidence.
- injuries
- bicycling
- head protective devices
- aged
- protective factors
Statistics from Altmetric.com
Introduction
Elderly populations are growing rapidly, and the importance of quality of life in this age group is rising.1–4 Geriatric research contains recommendations of regular physical activity such as walking or cycling for healthy ageing.5–7 Accompanied by public health policies,5 6 8 these recommendations have resulted in an increased number of elderly cyclists and injuries.9 10
In January 2009, the Korean government announced ‘The National Bicycle Use Activating Plan’, which promotes bicycles as urban transportation to resolve its urban traffic problems and to reduce carbon emissions.11 Therefore, cycling-related legislation was revised, and cycling facilities such as bike paths and connexions to metro stations, numerous promotional events and education programmes11 have been quantitatively expanded as the number of cyclists who rides to school or to the workplace has increased. Consequently, the number of bicycle-related injuries increased from 8721 (4.1% of the total number of road traffic injuries) in 2007 to 17 366 (7.5%) in 2015. Moreover, the proportion of the number of deaths from bicycle-related injuries to the total number of deaths from road traffic injuries has increased.10 12 It is likely that the number of elderly individuals sustaining bicycle-related injuries has increased as the numbers of elderly individuals and cyclists increased.10 13 However, most public education programmes and interventions for helmet use have been targeted at young individuals or individuals attending educational institutions.
Helmet use is considered the most effective modality to prevent brain and facial injuries.14 15 Its effectiveness has been proven in several studies from laboratory level to population level.15 16 With increasing public awareness, the use of helmets is continuously increasing, and regulations are being developed worldwide.17
However, efforts to reduce bicycle-related injuries among elderly cyclists are limited thus far.18 19 Most public education programmes and interventions for helmets are targeted at young individuals.10 Considering that the susceptibility to complications and the long-term mortality of traumatic injury are greater in elderly populations, more vigorous efforts are required to prevent injuries in this population.20 21 Before initiating interventions that are specific to elderly individuals, the difference in the effects of helmets should be evaluated. The purpose of this study was to compare the protective effects of helmets on bicycle-related injuries in elderly individuals with those in younger adults.
Methods
Study design
This study was a retrospective analysis that used the emergency department–based Injury In-depth Surveillance (EDIIS) database of Korea. The EDIIS is a nationwide injury database that includes data of all injured patients admitted to emergency departments (EDs) across Korea. The requirements for informed consent were waived due to the retrospective design.
Data collection
Among 20 participating centres, eight provided data focused on patients who sustained road traffic injuries to the EDIIS database. The EDIIS database contains information on demographics, injury prevention and epidemiology, prehospital procedures, initial clinical findings at the ED, treatment in the ED, ED disposition and patient outcomes after admission. Primary information was acquired by physicians of each institution during their clinical practice and by trained coordinators of the EDIIS project who were assigned to each hospital. Coordinators collected the data from the standardised registry. The data of each ED were entered into a web-based database of the Korean Center for Disease Control and Prevention, and a quality improvement programme was conducted regularly.
Selection of participants
The study population included injured patients aged 20 years or older who sustained bicycle-related injuries in vehicles (riders and passengers) in the EDIIS database from January 2011 to December 2016. The final study population included injured patients who sustained injuries while riding bicycles. Cases with unknown helmet use or clinical outcomes were excluded. We divided the subjects into two groups by age: the young-adult group (aged 20 to 65 years) and the elderly group (older than 65 years).
Measures
We collected information on demographic factors (age, sex and medical history), injury-related factors (time and day of injury) and prehospital factors (emergency medical services use, helmet use, injury place, opponents of injury and the purpose of bicycle riding).
Clinical outcomes (traumatic brain injury (TBI), excess mortality ratio–adjusted ISS (EMR-ISS) and in-hospital mortality) were also included. TBI was defined based on the International Statistical Classification of Diseases and Related Health Problems 10th Revision, Clinical Modification (ICD-10-CM) codes F07.2, S02.0, S02.1, S02.3, S02.7, S02.8, S02.9, S06, S07.1, T90.2 and T90.5.
The EMR-ISS is a diagnosis-based injury severity scale converted from the ICD-10-CM. The EMR for specific ICD-10-CM codes was calculated based on the Korean National Health Insurance Database as 1 of 5 grades.22 The EMR-ISS ranged from 1 to 74. In this study, we classified the degree of severity into three categories: mild (1≤EMR-ISS≤9), moderate (9≤EMR-ISS≤24) and severe (EMR-ISS ≥25).
Primary and secondary outcomes
The primary outcome was the incidence of TBI and the secondary outcomes were severe trauma (EMR-ISS ≥25) and in-hospital mortality.
Injury pattern of deceased patients
To describe the injury pattern of deceased patients, we gathered all diagnosis of mortality cases, which are coded in the ICD-10-CM system.
Statistical analysis
For each age group, continuous and categorical variables were compared between helmet-wearing versus non–helmet-wearing groups by Wilcoxon rank-sum test and Fisher’s exact test, respectively. We used logistic regression analysis to identify risk factors of TBI and severe trauma. First, candidate risk factors were selected with p value <0.05 in univariate logistic regression analyses and included in the multivariate analysis. After analysing the effects of helmets on targeted outcomes, the differences in effects by helmet use between age groups were further examined by an interaction analysis. A logistic regression with interaction between age group and helmet-wearing was performed using an interaction term (age group×helmet-wearing) and TBI, EMR-ISS and mortality to calculate the statistical significance of effect size comparison. To describe the injury pattern of deceased patients, we calculated the ratio of certain affected body region versus the total number of diagnosis of mortality cases, which are coded in the ICD-10-CM system.
All statistical analyses were performed using R statistical software V.3.4.0 (R Foundation for Statistical Computing, Vienna, Austria), a language and environment for statistical computing.
Results
Of 75 084 injured patients, 7181 patients were eligible for study inclusion. We excluded patients younger than 20 years old (n=14 417), those who sustained injuries via other mechanisms (n=52 912), those who were passengers, (n=17) and those with unknown information on helmet use (n=558). Of 7181 adults, 1253 aged over 66 years were categorised as the elderly group (online supplementary figure 1).
Supplemental material
Figure 1 shows the incidence of bicycle-related injuries and the helmet-wearing rate by age. The injury incidence showed a bimodal pattern with peaks around ages 20 and 50 years. Meanwhile, the helmet-wearing rate showed a unimodal pattern with its peak at age 35–40 years, and the rate decreased consistently with age (figure 1).
Total number of any kind of bicycle injuries and the helmet-wearing rate by age from 2011 to 2016. The injury incidents showed a bimodal pattern with peaks around ages 20 and 50 years. The helmet-wearing rate showed a peak at age 35–40 years and declined with age.
The demographics and clinical outcomes are summarised in table 1. There were some consistent findings between the two groups. The non–helmet-wearing group had a significantly higher rate of injuries on weekdays, motor vehicles as crash opponents, daily activities as the purpose of bicycle riding and injuries on the roadway.
Demographic findings and clinical outcomes of bicycle-related injuries in adults
In young adults, the non–helmet-wearing group was significantly younger (43.4 years old vs 41.7 years old, p<0.001), more likely to ride from Monday to Friday (48.5% vs 61.3%, p<0.001) between 18:00 and 06:00 (4.7% vs 13.1%, p<0.001), more likely to ride on the roadway and less likely on bike paths (29.9% vs 34.9%, p<0.001), more likely to be riding for daily activities other than for commute or leisure (6.4% vs 14.2%, p<0.001) and more likely to have a motor vehicle as a crash opponent (15.3% vs 23.9%, p<0.001). In elderly adults, the non–helmet-wearing group was significantly older (70.8 years old vs 73.7 years old, p<0.001), more likely to ride from Monday to Friday (61.0% vs 74.4%, p=0.010), more likely to ride on the roadway and less likely on bike paths (51.2% vs 36.6%, p<0.001), more likely to be riding for daily activities other than for commute or leisure (4.9% vs 16.1%, p<0.001) and more likely to have a motor vehicle as a crash opponent (28.0% vs 44.4%, p<0.001). The time of injury was not significantly different between the two groups in elderly individuals.
The non–helmet-wearing group had a significantly higher rate of TBI (young adults, 5.9% vs 7.9%, p=0.024; elderly individuals, 4.9% vs 14.5%, p=0.012), severe EMR-ISS (young adults, 13.4% vs 16.6%, p=0.006; elderly individuals, 17.1% vs 34.8%, p=0.001) and in-hospital mortality (young adults, 0.4% vs 0.8%, p=0.248; elderly individuals, 0.0% vs 6.0%, p=0.012) in both age groups.
Logistic regression analysis
Univariate analysis found primary and secondary outcomes to be significantly related to variables such as age, helmet-wearing, sex, remarkable medical history, date of injury and injury place. Motor vehicles as crash opponents were not significantly associated with the primary and secondary outcomes (table 2).
Logistic regression analysis of the primary and secondary outcomes
Regarding the primary outcome of TBI, there was a significant association between age and TBI (OR 1.46; 95% CI 1.18 to 1.79), and helmet-wearing and TBI (OR 0.72; 95% CI 0.55 to 0.93). Regarding the secondary outcomes, there were also significant associations with age (OR 1.82; 95% CI 1.57 to 2.12) and helmet-wearing (OR 0.80; 95% CI 0.67 to 0.96) (table 2).
Comparison of the effect of helmets on outcome by age groups
We compared the effect size of helmet-wearing on each outcome. Regarding the incidence of TBI, there was a significant interaction between age group and helmet-wearing (p=0.022). Regarding the incidence of severe trauma (EMR-ISS ≥25), we also found a significant interaction (p=0.024). The effect of helmet-wearing increased in the elderly group (figure 2). However, the p value for the interaction term (age group×helmet-wearing) and in-hospital mortality was not statistically significant (figure 2C).
Effect differences of helmet in bicycle-related injuries. Logistic regression of the primary and secondary outcomes using interaction terminology (age×helmet-wearing). (A) Interaction plot of age group and TBI. The p value for interaction term (age group×helmet-wearing) and TBI rate was 0.022. (B) Interaction plot of age group and severe EMR-ISS. The p value for interaction term (age group×helmet-wearing) and the severe EMR-ISS rate was 0.024. (C) Interaction plot of age group and mortality rate. The p value for interaction term (age group×helmet-wearing) and mortality was not statistically significant. EMR-ISS, excess mortality ratio–based ISS; TBI, traumatic brain injury.
Injury pattern of deceased patients
The most common diagnosis was injuries to the head in both groups. The proportion of injuries to the head was higher in the helmet-wearing group (47.7%) than in the non–helmet-wearing group (38.6%). The ratio of minor head trauma to severe head trauma showed a slight difference between the two groups (helmet-wearing group, 6:3; non–helmet-wearing group, 34:39) (table 3).
Injury pattern for deceased patients
Discussion
Bicycles have been widely encouraged for environmental and health purposes since the Korean government announced the National Bicycle Use Activating Plan in January 2009.6 9 As the population and the physical activities of elderly individuals increase, the frequency of injuries has also increased.6 7 The numbers of cyclists and bicycle-related injuries in Korea have steadily increased; moreover, the proportion of the number of deaths from bicycle-related injuries to the total number of deaths from road traffic injuries increased.12 13
However, studies on the characteristics and the incidence of bicycle-related injuries in elderly cyclists are limited. Previous studies on elderly individuals have focused on the incidence and trends of bicycle-related injuries.10 21 Studies have revealed that the incidence of bicycle-related injuries and the helmet-use rate increased, but to the lowest degree in the elderly group, and a lack of helmet use was significantly associated with serious outcomes.10 15 23–25 However, to the authors’ knowledge, no study has analysed the effect of helmet-wearing between the two age groups.
Our study showed that non–helmet-wearing groups were more likely to ride a bicycle on roadways for daily activities during weekdays and consequently were more likely to collide with motor vehicles in both age groups. Considering that most young adults are involved in collisions without opponents, non–helmet-wearing elderly adults have a high probability of colliding with a motor vehicle. In the elderly adult group, non–helmet-wearing individuals were older than the helmet-wearing individuals, a pattern which is opposite to that seen in the young adult group.
It is known that risk-taking behaviour is highest among younger people.26 Moreover, Gaudet et al reported that the helmet-wearing rate was 26% and that it increased with age up to 50%.27 It is interesting that our study revealed contradictory results regarding the helmet-wearing rate; compared with previous studies, more than 20% of individuals between the age of 30 and 50 years and less than 5% of those 75 years and older were found to wear helmets. Several studies have reported high rates of helmet-wearing in France (26%),28 Seattle (52.5%)29 and Australia (89%)30; however, Asian countries such as Taiwan,31 Singapore32 and Lao People’s Democratic Republic33 reveal low rate of helmet-wearing. Furthermore, the third Global Status Report on road safety showed that low-income and middle-income countries have twice the fatality rate of high-income countries and 90% of global road traffic deaths.34 Vulnerable road users such as pedestrians, cyclists and motorcyclist make up half of these fatalities.34
Since the National Bicycle Use Activating Plan, most public education and intervention programmes for helmet use have been targeted at young individuals and those who attend educational institutions.11 Changing cycling behaviour in the 10 years since the introduction of programmes for cyclists may be difficult for those who have been riding since before this education programme, unless targeted by specific campaigns. Moreover, elderly risk-takers may not recognise riding on the roadway as an involuntary risk since they would have been riding on the roadway for a long time. The disparity observed in the helmet-wearing rate between young and elderly adults may be because of the rapid transition to a bicycle-friendly country.
Helmet-wearing may not have done to prevent catastrophic injuries to thorax, abdomen and pelvis.35 However, helmet-wearing is considered to be one of the most effective modalities to reduce TBI, severe TBI and facial injuries.13 Previous studies found that the protective effect of helmets was 3-fold to 8-fold higher in the helmet-wearing group, which indicates that unmeasured confounders cannot explain the differences in the risk of injury between helmet-wearing and non–helmet-wearing groups.36–38 Our study revealed a similar result after controlling for potential confounding variables.
Bicycle incidents are caused by various factors. Identifying and addressing these factors may help to reduce bicycle-related injuries, including improvements in road infrastructure, socioeconomic factors, as well as stricter legislation and enforcement.39 The greatest impact on road safety would be achieved through adopting the Safe System from WHO, which involves managing vehicles, road infrastructure, driving speeds, road users and the interactions between these components.34 Adopting stronger legislation and enforcement is a critical in changing road user behaviour.34 As the elderly population grows rapidly worldwide, more focus should be placed on their injury prevention. To improve road safety for users in this targeted population, assessments to identify the minority of elderly riders, educational efforts to encourage safer riding habits, including helmet-wearing but not only through self-regulation, and adopting stricter legislation for helmet-wearing are key.
Limitations
First, the registry only included patients who presented to EDs; this could have influenced the study population by eliminating patients with very minor injuries. Patients with bicycle-related incidents but without any injuries (possible prevented by helmets) were also not included in the study, which could have led to underestimating the rate of helmet-wearing. However, we have compared the helmet-wearing rate by groups, not by the number of cases, which would decrease the bias. To describe the injury pattern of deceased patients, we gathered all diagnosis of mortality cases. However, it was limited to provide precise information of what actually kills cyclists in this study due to lack of information of the registry we used. Because several important variables to calculate disability-adjusted life year are missed, it was impossible to quantify the burden of disease from bicycle-related injuries.
Second, we could not use mortality as an outcome. Although mortality is primarily used as a primary outcome in many injury-related studies, there were no mortalities in the elderly helmet-wearing group, which made it impossible to implement a statistical model to compare with the other groups. We could only show the difference in size by plain graph without statistical significance.
Third, the EMR-ISS, which was a secondary outcome, is not a commonly used index. Its development was published in a peer-reviewed journal.22 Although the index was used in multiple studies in Korea, its uses in other countries are not reported yet.
Conclusion
This study revealed two major results. The first is that the helmet-wearing rate continuously decreases after 40–45 years of age. The second is that the effect of helmet-wearing increases with age. Its effect was determined both on the incidence of TBI and severe trauma (EMR-ISS ≥25). This implies that more vigorous education is required for elderly individuals because of the low usage rate and high effect size. In this study, we found that the protective effects of helmets on bicycle-related injuries were greater for elderly individuals.
What is already known on the subject
Helmet use is considered the most effective modality to prevent brain and facial injuries.
However, most public education and interventions for helmets are targeted towards young individuals and efforts to reduce bicycle-related injuries among elderly cyclists are limited thus far.
Considering that the susceptibility to complications and the long-term mortality of traumatic injury is greater in elderly populations, more vigorous efforts are needed.
What this study adds
The helmet-wearing rate continuously decreases after 40–45 years of age.
The effect of helmet-wearing both on the incidence of TBI and severe trauma increases with age.
This implies that more vigorous education is required for elderly individuals because of the low usage rate and high effect size.
References
Footnotes
Collaborators Won Chul Cha.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The study was approved by the Korean Center for Disease Control and Prevention (KCDC) and the institutional review board of Samsung Medical Center (IRB file no. 2017-07-152).
Provenance and peer review Not commissioned; externally peer reviewed.