Revisiting Older Drivers’ Risks of At-fault Motor Vehicle Collisions in Japan

Background In Japan, older drivers have been encouraged to surrender their driving licenses for traffic safety, despite the potential adverse social and health outcomes of driving cessation. We reconsidered such policies and social pressure by comparing the risk of at-fault motor vehicle collisions (MVCs) across the age groups of drivers. Methods Using the national data of police-reported MVCs that occurred between 2016 and 2020, we examined the number of at-fault MVCs per licensed driver (MVC rate) and the number of fatally and non-fatally injured persons per at-fault MVC by the sex and age groups of at-fault drivers. Results The MVC rate of older drivers was higher than that of middle-aged drivers but lower than that of young drivers. The number of injured persons among the collided counterparts (collided car occupants, motorcyclists, bicyclists, and pedestrians) per MVC caused by older drivers was not greater than that by drivers in other age groups. In fatal MVCs caused by older drivers, drivers themselves or their passengers tend to be killed rather than their collided counterparts. Overall, the results were mostly consistent between male and female drivers. Conclusion The risk of at-fault MVCs increased with the advancing age of drivers after middle age; however, this risk among older drivers did not exceed that among young drivers, without posing a high risk of injuries to their collided counterparts.


INTRODUCTION
The driver population is getting older with population aging.As the number of older drivers increases, the proportion of their motor vehicle collisions (MVCs) also increases.In Japan, which has the oldest population in the world, people aged ≥65 years accounted for 29% of the total population and 24% of the driver population in 2021. 1,2From 2012 to 2021, the proportion of MVCs caused by drivers aged ≥65 years rose from 16% to 24%. 3 Consequently, their MVCs, especially fatal ones, have gained media attention, and public concern has been raised regarding the safety of older drivers.
Stringent licensing policies are imposed on older drivers because their driving abilities could be affected by functioning difficulties as they age.However, it is reported that older drivers adjust their driving by driving less or avoiding challenging situations, compensating for their declining abilities [4][5][6] and potentially attenuating their MVC risk.Nevertheless, older drivers in Japan are not only required to take driving lessons and cognitive screening tests at their license renewal but are further encouraged to surrender their driving licenses. 7][10][11] This is a critical issue in car-dependent societies, such as Japan, where 57% of people aged ≥60 years drive cars regularly. 12Additionally, it should be noted that driving cessation turns drivers to vulnerable road users, such as pedestrians and bicyclists.In Japan, the proportion of vulnerable road users in road traffic deaths is over 50%. 13This is another concern over older drivers' surrendering of their driving licenses.
5][16][17][18][19][20][21] Policy decisions should be based on a comparison between the risk of injury to other road users by older drivers and the risk of adverse events to older people with limited mobility.We previously examined the former risk, but our previous study had the limitations of the short temporal and limited geographical coverage of MVCs and used data from more than 10 years ago. 14Therefore, in the present study, we used recent nationwide and long-term data to examine the crash and injury profiles of MVCs caused by older drivers and drivers in other age groups to reconsider traffic safety policies and social pressure to promote driving cessation among older drivers.

Data
In this descriptive cross-sectional study, we obtained data on atfault MVCs from the Institute for Traffic Accident Research and Data Analysis, which compiles police-reported MVC data from the National Police Agency.The data included the number of atfault MVCs by collided counterparts (cars; motorcycles, including mopeds; bicycles; pedestrians; others, such as trains; and none [single crashes]), and the number of deaths, serious injuries, and minor injuries by at-fault drivers, their passengers, and collided counterparts.The data covered MVCs that occurred between January 2016 and December 2020 and were stratified by the sex and age group of at-fault drivers (18-19 years, 5-year intervals from 20 to 84 years, and ≥85 years).Only MVCs in which at-fault drivers drove cars were included because this was the interest of this study.We excluded at-fault MVCs collided with "others" from the analysis because the numbers of such MVCs caused by male and female drivers were as small as 373 and 158, respectively, for the entire study period.][24][25][26] In the police-reported MVC data, deaths are defined as deaths occurring within 24 hours of the crash, serious injuries as injuries diagnosed by physicians requiring treatment for at least 30 days, and otherwise minor injuries. 27The driver's responsibility for collisions was determined through police investigations at the scene of crashes. 27Drivers determined to be primarily responsible for crashes, including single-vehicle crashes, were referred to as at-fault drivers.

Analysis
First, we calculated the number of at-fault MVCs per 100,000 licensed drivers per year (at-fault MVC rate) by collided counterparts according to the sex and age group of at-fault drivers, to examine the extent to which older drivers caused MVCs by their collided counterparts.Next, we calculated the number of injured persons (all injuries [fatal, serious, or minor injuries]), and fatally or seriously injured persons per 100 at-fault MVCs among at-fault drivers, their passengers, collided car occupants, motorcyclists, bicyclists, and pedestrians by the sex and age group of at-fault drivers, to examine how serious, on average, the consequences of MVCs caused by older drivers were.Finally, we calculated the proportion of at-fault drivers, their passengers, collided car occupants, motorcyclists, bicyclists, and pedestrians among all deaths in fatal MVCs by the sex and age group of atfault drivers, to examine who was killed in fatal MVCs caused by older drivers.
The at-fault MVC rate was higher among males than among females, irrespective of their age (Table 1).Among both sexes, the at-fault MVC rate tended to increase as the age of the at-fault drivers advanced after middle age.However, among males, the at-fault MVC rate was the highest among drivers aged 18-19 years (1,811), followed by drivers aged 20-24 years (1,034) and 25-29 years (762), whereas it was 499, 548, 595, and 661 among drivers aged 70-74, 75-79, 80-84, and ≥85 years, respectively.Among females, the at-fault MVC rate was the highest among drivers aged 18-19 years (964), and it was almost the same among drivers aged 20-24 years (572) and drivers aged ≥85 years (571), followed by drivers aged 80-84 years (530).Regarding collided counterparts, the at-fault MVC rate with cars was higher than that with other counterparts in all age groups of at-fault drivers for both sexes.
Across the age groups of at-fault drivers, the at-fault MVC rate with cars and motorcycles was much higher among young drivers than among other drivers, especially among males.However, the at-fault MVC rate with bicycles and pedestrians was relatively higher among older drivers than among other drivers.The single crash rate was high toward both the upper and lower ends of the age groups of at-fault drivers.
The number of injured persons per 100 MVCs was much higher among collided car occupants (73.5 and 73.0 per 100 MVCs caused by male and female drivers, respectively) than bicyclists (16.6 and 17.0), motorcyclists (11.4 and 10.6), pedestrians (10.3 and 10.4), at-fault drivers (3.3 and 5.7), and at-fault drivers' passengers (2.7 and 1.9) (Table 2).Across the age groups of atfault drivers, the number of injured persons among collided car occupants and motorcyclists in MVCs caused by older drivers was mostly lower than the average, whereas the number of injured persons among at-fault drivers, passengers, bicyclists, and pedestrians in MVCs caused by older drivers was mostly higher than the average.
The number of fatally or seriously injured persons per 100 MVCs was the highest among pedestrians (2.1 and 2.1 per 100 MVCs caused by male and female drivers, respectively), followed by motorcyclists (1.6 and 1.5), bicyclists (1.4 and 1.4), collided car occupants (1.4 and 1.3), at-fault drivers (0.8 and 0.7), and at-fault drivers' passengers (0.5 and 0.3).Typically, there were no distinct differences in the number of fatally or seriously injured persons per MVC among collided car occupants and motorcyclists irrespective of the age group and sex of at-fault drivers, and the number of fatally or seriously injured persons among bicyclists and pedestrians in MVCs caused by older drivers was almost equivalent to that in MVCs caused by middle-aged drivers.However, the number of fatally or seriously injured persons among at-fault drivers and their passengers in MVCs caused by older drivers tended to be greater as the drivers advanced in age.
The proportion of at-fault driver deaths among all deaths in fatal MVCs caused by older drivers was much higher than that in fatal MVCs caused by other drivers; older drivers tended to be killed in their fatal MVCs (Table 3).This proportion was 31% in MVCs caused by all male drivers, whereas it was 43%, 53%, 60%, and 67% in MVCs caused by male drivers aged 70-74, 75-79, 80-84, and ≥85 years, respectively.The proportion of at-fault drivers' passenger deaths among all deaths was higher in MVCs caused by young and older drivers than in those caused by other drivers, especially among males.In contrast, the proportion of pedestrian deaths among all deaths in fatal MVCs caused by older drivers was smaller than that in fatal MVCs caused by other drivers, as was the proportion of deaths of collided car occupants, motorcyclists, and bicyclists among all deaths in fatal MVCs caused by older drivers.The proportion of pedestrian deaths among all deaths was 38% in fatal MVCs caused by all male drivers, whereas it was 32%, 24%, 14%, and 11% in fatal MVCs Older Drivers' Risks of At-fault Motor Vehicle Collisions caused by male drivers aged 70-74, 75-79, 80-84, and ≥85 years, respectively.Generally, similar trends were observed among female drivers.

DISCUSSION
The at-fault MVC rate increased with the advancing age of drivers after middle age, but the rate among older drivers was still lower than that among young drivers.Moreover, the total number of injured persons per MVC caused by older drivers was not excessively higher than that caused by drivers in other age groups.In the fatal MVCs of older drivers, the drivers themselves and their passengers tended to be killed more than their counterparts.This was not the case for drivers in other age groups.Overall, the results were mostly consistent between male and female drivers.
It should be noted that older drivers cause more collisions with bicyclists and pedestrians than other drivers.This is possibly attributable, in part, to older drivers' lowered ability to perceive and react to road hazards, 28,29 which affects their ability to avoid conflicts with bicyclists and pedestrians.Bicyclists and pedestrians can suddenly appear and cross the road, especially in residential areas where the roads are narrow without separate sidewalks and bicycle lanes.
Nevertheless, in terms of the total number of injured persons per at-fault MVC, older drivers did not pose a high risk of injuries to those involved in their at-fault MVCs.This implies that the consequences of MVCs caused by older drivers are, on average, not more serious than those caused by other drivers.However, the risk of injuries to older drivers and their passengers was notably high in their MVCs.This is likely because older drivers are physically fragile 17,18 and often drive with older passengers. 30ased on our findings, continued efforts are needed to reduce older drivers' MVCs, especially with bicyclists and pedestrians; however, we cannot draw a conclusion that older drivers pose excessive injury risks to other road users, which would support the claims that they should be restricted from driving.We believe that, if it is possible to identify older drivers who are at high-risk of causing MVCs, they should be dissuaded from driving, but generally, older drivers should be assisted in driving safely for at least two reasons.
The first is the risk of traffic injuries when turning to vulnerable road users.In 2021, 58% of road traffic deaths occurred among people aged ≥65 years, 49% of road traffic deaths occurred among pedestrians and bicyclists, and 75% of pedestrian and bicyclist deaths occurred among people aged ≥65 years. 31The proportion of older adults and vulnerable road users in traffic deaths in Japan is two to three times as high as that in other high-income countries. 13[10][11] This study has some limitations.First, using aggregated data, we could not identify serious MVCs in terms of the number of injured persons in one MVC (ie, high-profile MVCs resulting in multiple victims).Thus, we only showed the average number of injured persons per MVC.We are unsure whether older drivers caused high-profile MVCs.Second, using the number of at-fault MVCs per licensed driver, we might have underestimated older drivers' risk of at-fault MVCs relative to their driving exposure (per distance driven) because the distance driven by older drivers is less than that driven by drivers in other age groups.][17][18][19][20][21] In conclusion, the risk of at-fault MVCs increased as at-fault drivers advanced in age after middle age.However, the risk of atfault MVCs among older drivers did not exceed that among young drivers.Moreover, older drivers did not pose a high risk of injuries to their collided counterparts; however, they posed a high risk to themselves and their passengers in their at-fault MVCs.

Table 1 .
Number of licensed drivers and their at-fault motor vehicle crashes (MVCs) and the at-fault MVC rate (per 100,000 licensed drivers per year) by collided counterparts and the sex and age group of at-fault drivers from 2016 to 2020 a Sum of the number of licensed drivers or at-fault MVCs in each year from 2016 to 2020.Ichikawa M, et al.J Epidemiol 2024;34(6):295-300 j 297

Table 2 .
Number of injured persons (all injuries [fatal, serious, or minor injuries] and fatal or serious injuries) per 100 at-fault motor vehicle crashes among at-fault drivers, their passengers, collided car occupants, motorcyclists, bicyclists, and pedestrians by the sex and age group of at-fault drivers from 2016 to 2020

Table 3 .
Number and proportion of at-fault drivers, their passengers, collided car occupants, motorcyclists, bicyclists, and pedestrians among all deaths in fatal motor vehicle crashes by the sex and age group of at-fault drivers from 2016 to 2020 J Epidemiol 2024;34(6):295-300 j 299