This Article Abstract Full Text (PDF) Services Download to citation manager PubMed PubMed Citation

Traffic-Related Fatalities Among Older Drivers and Passengers

Past and Future Trends

^a Lakehead Psychiatric Hospital, Thunder Bay, Ontario, Canada
^b Department of Psychology, Lakehead University, Thunder Bay, Ontario, Canada
^c Department of Clinical Epidemiology and Biostatistics, McMaster University, Ontario, Canada
^d Department of Health Studies and Gerontology, University of Waterloo, Ontario, Canada

Correspondence: Michel Bédard, PhD, Lakehead Psychiatric Hospital, 580 North Algoma Street, Thunder Bay, Ontario P7B 5G4 Canada. E-mail: mbedard{at}baynet.net.

Decision Editor: Laurence G. Branch, PhD

	Abstract

TOP Abstract Methods Results Discussion Appendix References

 
Purpose of the Study: This study was initiated to forecast the number of older drivers and passengers who may be fatally injured in traffic crashes in future years. Design and Methods: The study was based on data from the U.S. Fatality Analysis Reporting System covering the period from 1975 to 1998. Projections were based on least squares regression models. Results: About 35,000 drivers and passengers died in traffic crashes each year from 1975 to 1998. Older adults (65 and older) accounted for 10% of all fatalities in 1975, 17% in 1998, and a projected 27% by 2015, the same proportion predicted for drivers and passengers aged younger than 30. On the basis of these projections, the number of fatally injured women and men aged 65 and older will increase respectively by 373% and 271% between 1975 and 2015. Implications: If current trends continue, the number of fatalities among older drivers and passengers and those aged younger than 30, may be equivalent early in this century. These projections call for further research into conditions that may lead to crashes involving older drivers and for the development and implementation of initiatives to curb traffic-related fatalities among older adults.

Key Words: Fatalities • Older adults • Crashes • Trends • Projections

Traffic fatalities represent a serious public health issue, one usually attributed to young drivers. The common observation is that the magnitude of the problem posed by older drivers is relatively small compared with that of younger drivers, because there are fewer older drivers and they drive fewer miles (Brorsson 1989; McGwin and Brown 1999; Ryan, Legge, and Rosman 1998; Whitfield and Fife 1987). However, despite these assurances, safety analysts and planners are increasingly concerned by the demographic changes facing our society and the potential crash risk posed by older drivers (Janke 1991). One reason for this concern is that although fewer older drivers are involved in fatal crashes compared with other age groups, older drivers represent a crash risk equivalent to that of younger drivers when exposure is taken into account. Crash risk, when accounted for exposure, usually follows a -shaped curve, starting high for young drivers, declining and remaining steady for experienced drivers, and then rising with drivers aged 65 and older (Brorsson 1989; McGwin and Brown 1999; Ryan, Legge, and Rosman 1998; Stutts and Martell 1992; Williams and Carsten 1989; Zhang, Fraser, Clarke, and Mao 1998).

Predictions of the future involvement of older adults in crashes may provide some insight into this issue. Williams and Carsten courageously ventured a prediction (Williams and Carsten 1989). On the basis of their 1987 data, and census predictions, they reported that there would still be more than twice as many young drivers (younger than 30) involved in fatal crashes than older drivers (65 and older) by year 2030. This led them to conclude that crash issues related to older drivers "will remain relatively small well into the next century." (p. 327). However, past fatality trends paint a more alarming picture. Between 1980 and 1989, overall fatalities fell 20% from 19.8 per 100,000 drivers to 15.9. Despite this overall decrease, fatalities among drivers 65 and older increased 19% from 9.0 per 100,000 to 10.7. Whereas 2,323 drivers aged 65 and older died in 1980, accounting for only 8% of all 28,816 driver fatalities, 3,319 (13%) of all 26,389 drivers killed in 1989 were aged 65 and older (Barr 1991).

This increase cannot be explained by demographics only. One possible explanation is that of cyclical and irregular fluctuations as is often observed with market-based data. Another possibility is an increase in exposure among older drivers. Data from North Carolina for the period from 1974 to 1988 illustrate this point (Stutts and Martell 1992). Whereas the proportion of residents aged 75 and older increased by 39%, the proportion of licensed drivers aged 75 and older increased by 129%, and the proportion of crashes involving this age group increased by 75%. Such changes in licensing rates are apparent in other jurisdictions as well. Overall in the United States, the proportion of licensed adults aged 65 and older passed from 62% in 1980 to 70% in 1989 (Barr 1991). Other researchers reported an increase from 61% in 1983 to 75% in 1990 (Massie, Campbell, and Williams 1995). In the United States, the proportion of adults aged 70 and older with valid driver's license is expected to more than double by 2020 (Eberhard 1996), and in Australia, the number of license holders among elderly adults is expected to double within the next 20 years (Darzins and Hull 1999). Distance traveled is also changing. From 1980 to 1989 the 65-and-older population increased their yearly distance traveled by 31% (Barr 1991). Comparing data for 1983 and 1990, drivers aged 75 and older increased their distance traveled by 90%, the largest percentage increase in travel by any age group (Massie et al. 1995).

Similar issues apply to comparisons of men's and women's involvement in fatal crashes. Men consistently comprise a larger proportion of individuals involved in, and fatally injured in, crashes, an observation attributed to men's risk-taking behavior and women's lower exposure (Chipman, MacGregor, Smiley, and Lee-Gosselin 1993; Evans 1988a; Li, Baker, Langlois, and Kelen 1998; Massie et al. 1995; Massie, Green, and Campbell 1997; Stutts and Martell 1992; Whitfield and Fife 1987). However, older women's exposure is also changing. In North Carolina the number of licensed women aged 65 and older increased at four times the rate of men of the same age (Stutts and Martell 1992). Similarly, whereas less than 7% of all women and men aged 70 or older today hold a driver's license, this proportion will climb to 18% for women but only 13% for men by year 2020 (Eberhard 1996). Finally, women, who accounted for 33% of the distance traveled in the United States in 1983, accounted for 37% of the distance traveled in 1990 (Massie et al. 1995).

The data on exposure highlight an increasing salient phenomenon. Although the demographic changes alone will create pressures on crash issues for older adults, these changes will be compounded by the increasing independence and mobility of coming older generations. Today's older adults are more likely to retire earlier, with more disposable income, and are more likely to maintain an active lifestyle (Fildes 1997). The coming generation of "baby boomers" will be more mobile than the current generation of older adults (Weinand 1996). Hence, available predictions of older adults' involvement in fatal accidents based only on demographics likely represent underestimates; updated predictions are required for the better planning of safety initiatives.

From a safety-planning and policy perspective, the problems related to young drivers (e.g., aggressive driving, driving while impaired) are very different than those related to older drivers (e.g., health problems, attention or perception; Lilley, Arie, and Chilvers 1995). However, because young drivers account for the majority of fatalities, most safety initiatives target their needs, not those of older drivers. Yet, given that over 5,000 vehicle occupants aged 65 and older die each year, reducing this number by even a small percentage would translate into a large number of lives saved; fatalities are preventable events. Furthermore, safety initiatives may reduce the financial burden related to crashes. Traffic-related costs average about 2.5% of the gross national product of industrialized countries (Elvik 2000), and 3.2% of U.S. health care spending (Miller, Lestina, and Spicer 1998). Altogether, traffic-related costs in the United States alone exceed $50 billion per year (Miller, Lestina, Galbraith et al. 1997).

Accordingly, the aim of this study was to forecast the future involvement of older adults, in comparison with younger and middle-aged adults, in fatal crashes, and that of older women in comparison with older men. To achieve this aim we used fatality data from previous years, and projected these trends into the future. Using data from previous years, other researchers have shown that older adults' fatalities increased at a faster rate than explained by demographics alone, probably because of changes in the number of licensed older adults and changes in distance traveled. Estimating fatalities from these data should provide a more accurate picture than that projected from census data alone. We hypothesized that older adults would make up an increasing proportion of all fatalities. Furthermore, because larger proportions of older women, compared with older men, are becoming licensed and using automobiles, we hypothesized that fatalities among women 65 and older would increase at a faster rate than that of men aged 65 and older.

	Methods

TOP Abstract Methods Results Discussion Appendix References

 
Data
This study relied on data contained in the U.S. Fatality Analysis Reporting System (FARS). For every U.S. traffic fatality, information about the crash situation, drivers and passengers, and the vehicles involved is collected and added to the FARS database. Data from 1975 to 1998 (inclusive) were available for the analyses. The total number of fatally injured individuals was calculated for each year. This number included all individuals involved, including pedestrians. Subsequently, analyses and projections focused on vehicle drivers and passengers only to facilitate the interpretation of the findings. Further analyses were conducted with age and gender stratified. Age was categorized as younger than 30, 30 to 64, and 65 and older.

Statistical Analyses
Statistical analyses included counts and percentages, measures of association (chi-square), and time-series projections based on yearly fatality counts (to avoid the cyclical variations observed with seasonality) from 1975 to 1998 using a least squares regression model to provide the best fit to the actual data (Hamburg 1977). The least squares regression method was used to fit the actual number of fatalities recorded from 1975 to 1998 (Snedecor and Cochran 1989; Spiegel 1992). One regression model providing the best prediction of actual yearly fatality data was produced for each stratum studied (younger than 30, 30–64, all adults 65 and older, men 65 and older, and women 65 and older). The regression equations then were used to predict future fatalities to year 2015 (trend), a period of 17 years in addition to the available 24 years of data.

Because the relationship between the number of fatalities and the year was possibly nonlinear, different curve estimations were used to determine the best fitting model. For each stratum the model best fitting the data was determined through a combination of the highest multiple correlation coefficient (r value) and visual fit of the data (George and Mallery 2000). Plotting the actual data against the predicted values of several models provides an initial judgment of the validity of the models tested (Snedecor and Cochran 1989). Furthermore, an additional criterion to select the best model was its parsimony as determined by its degrees of freedom. Specifically, in the presence of similar r values and visual fit, the model with the most degrees of freedom was selected (Dowdy and Wearden 1991).

	Results

TOP Abstract Methods Results Discussion Appendix References

 
Fatalities From 1975 to 1998
For every year from 1975 to 1998, with the exception of 1992, more than 40,000 individuals were fatally injured in traffic-related incidents, totaling more than one million lives. When only drivers and passengers were considered, an average of more than 35,000 persons died each year. The number of driver and passenger fatalities remained relatively constant across years (see Total column, Table 1 ). However, the characteristics of individuals accounting for these fatalities changed considerably over time.

We used the fatality data presented in Table 1 to generate regression models (trend) to fit fatality data for each group. Logistic models best fit the data (see the A). The logistic model for the younger group, F(22) = 99.67, p < .001, r = .82, projected 9,701 fatalities by 2015, representing 27% of all fatalities. For drivers and passengers aged 30 to 64 the model, F(22) = 17.19, p < .001, r = .44, projected 16,125 fatalities (46% of all fatalities). For the older group the model, F(22) = 281.48, p < .001, r = .93, projected 9,569 fatalities by 2015, the equivalent of 27% of all fatalities (see Fig. 1).

View larger version (15K): [in this window] [in a new window]   Figure 1. Number of fatalities (data) for years 1975 to 1998 and projected trends to year 2015 (trend) by age group.

View larger version (13K): [in this window] [in a new window]   Figure 2. Number of fatalities (data) for years 1975 to 1998 and projected trends to year 2015 (trend) by gender for individuals aged 65+.

	Discussion

TOP Abstract Methods Results Discussion Appendix References

These projections, however, assume that current trends will continue unabated. Events such as retraining, engineering advances, higher mobility, and lower or higher speed limits may result in lesser or higher fatalities than projected. Further, the impact of the baby-boomers generation on these projections remains difficult to assess. The baby boomers may be healthier than previous older generations (Allaire, LaValley, Evans, et al. 1999; Reynolds, Crimmins, and Saito 1998). However, this may result in a better ability to sustain the traumatic effects of crashes, yet at the same time encourage travel and increase exposure.

In addition, it is not possible to extrapolate these findings to individual states or other countries; these projections provide overall trends for the United States only. Unique settings and older adult characteristics may affect future fatalities. For example, the changes in mobility observed in North Carolina may be occurring at different rates in other jurisdictions. Although projections could be established on a state-by-state basis, the underlying contributions of other sociodemographic characteristics (e.g., retirement income, health) cannot be determined from information contained in FARS. Another caveat of this study is the focus on fatal crashes only. It is unclear if older adults' involvement in nonfatal crashes will also increase in similar proportions.

Notwithstanding these concerns, the projections presented here highlight the need for safety initiatives directed specifically at older adults. However, given that most past initiatives were directed at young drivers, the selection of initiatives appears limited. Imposing blanket restrictions on older drivers would not serve the goal of maximizing the independence of older adults while ensuring public safety and their own. Whereas the abilities of some older adults will decline below acceptable thresholds, that of many others will remain above these thresholds for their entire lives (Bedard, Molloy, Guyatt, Stones, and Strang 1997).

Achieving a balance between public safety and older adults' own safety and independence can be realized by understanding why crashes and fatalities occur within older age groups. Three types of information must be obtained to reach this goal. First is the identification of situations in which older drivers may be at greater risk of initiating fatal crashes. Compared with young adults, older adults are overinvolved in crashes during the daytime (Mortimer and Fell 1989; Ryan, Legge, and Rosman 1998; Zhang, Fraser, Clarke, and Mao 1998), on weekdays (Zhang, et al. 1998), at intersections and merging situations (McGwin and Brown 1999; Pruesser, Williams, Ferguson, Ulmer, and Weinstein 1998; Ryan et al. 1998; Viano, Culver, Evans, Frick, and Scott 1990; Zhang et al. 1998), and during good weather (Zhang et al. 1998). Although useful, these data may result from higher exposure, higher crash initiation, or a combination of both. If researchers are to reduce crash risks in older adults, the crucial information needed is whether older adults are at greater risk of initiating crashes in some situations than in others. This information would allow researchers to devise age-specific interventions to reduce older adult involvement in crashes (Zhang et al. 1998). Furthermore, because many women have less driving experience than men (i.e., obtained their licenses later in life, drive lesser distance) their needs may be different than men's; they may possibly benefit from gender-specific interventions. Interventions aimed at older drivers, such as retraining, may yield substantial reductions in crashes while being cost-effective.

The second type of information required is the identification of older drivers who may be at increased risk of initiating crashes. Whereas lack of experience, aggressive driving, and driving while impaired are related to crash initiation in younger drivers, reduced driving abilities and health-related impairments are more likely agents behind the increased risk of crashes among older drivers (Lilley et al. 1995; Retchin and Anapolle 1993; Reuben, Stilliman, and Traines 1988). Enhanced screening protocols and strategies, such as graduated licenses and copilots (Bedard, Molloy, and Lever 1998; Shua-Haim and Gross 1996), may reduce the risk of crash initiation among drivers whose abilities are declining. Larger scale testing of these options and others is desirable.

The third type of information needed regards the greater susceptibility of older adults to fatal injuries (Evans 1988b). Although the aging human body may be less capable of sustaining trauma than younger bodies (Brorsson 1989; Evans 1988b; Viano et al. 1990), it is possible that vehicle characteristics (e.g., car design, seatbelts, air bags) provide a differential protective benefit to young and older adults and to men and women (Viano et al. 1990). A better understanding of the relationship between vehicle characteristics and fatal injuries may allow the development of strategies to enhance the probability of survival of older adults involved in crashes. Engineering advances targeted specifically to older adults may result in additional protection (Viano et al. 1990).

The results of this study attest to a growing public health issue faced by all aging and increasingly mobile societies. Older adults will represent a sizable proportion of future traffic fatalities early in this century unless researchers step up the development and implementation of safety initiatives targeting their specific needs.

Practice Concepts

The Forum

Book Reviews

	Acknowledgments

 
We thank Dr. Lori Chambers and two anonymous reviewers for their insightful comments on an earlier version of this article.

Received for publication April 13, 2001. Accepted for publication July 9, 2001.

	Appendix

TOP Abstract Methods Results Discussion Appendix References

 
The logistic model is defined as

where y is the predicted number of yearly fatalities, b0 and b1 are regression coefficients (derived from fatality data for years 1975 to 1998), and t is the year, ranging from 1 (1975) to 24 (1998).

The quadratic model is defined as

where y is the predicted number of yearly fatalities, b0, b1, and b2 are regression coefficients (derived from fatality data for years 1975 to 1998), and t is the year, ranging from 1 (1975) to 24 (1998).

	References

TOP Abstract Methods Results Discussion Appendix References

 

• Allaire S. H., LaValley M. P., Evans S. R., O'Connor G. T., Kelly-Hayes M., Meenan R. F., Levy D., Felson D. T., 1999. Evidence for decline in disability and improved health among persons aged 55 to 70 years: The Framingham Heart Study. American Journal of Public Health 89:1678-1683. [Abstract/Free Full Text]
• Barr R. A., 1991. Recent changes in driving among older adults. Human Factors 33:597-600. [Medline]
• Bédard M., Molloy D. W., Guyatt G. H., Stones M. J., Strang D. S., 1997. Competency to drive in cognitively impaired older adults. Annals of the Royal College of Physicians and Surgeons of Canada 30:346-352.
• Bédard M., Molloy D. W., Lever J. A., 1998. Factors associated with motor vehicle crashes in cognitively impaired older adults. Alzheimer Disease and Associated Disorders 12:135-139. [Medline]
• Brorsson B., 1989. The risk of accidents among older drivers. Scandinavian Journal of Social Medicine 17:253-256. [Medline]
• Chipman M. L., MacGregor C. G., Smiley A. M., Lee-Gosselin M., 1993. The role of exposure in comparisons of crash risk among different drivers and driving environments. Accident Analysis and Prevention 25:207-211.
• Darzins P., Hull M., 1999. Older road users. Australian Family Physician 28:663-667. [Medline]
• Dowdy S., Wearden S., 1991. Statistics for research Wiley, New York.
• Eberhard J., 1996. Safe mobility for senior citizens. Journal of the International Association of Traffic and Safety Sciences 20:29-37.
• Elvik R., 2000. How much do road accidents cost the national economy?. Accident Analysis and Prevention 32:849-851.
• Evans L., 1988a. Older driver involvement in fatal and severe traffic crashes. Journal of Gerontology: Social Sciences 43:S186-S193.
• Evans L., 1988b. Risk of fatality from physical trauma versus sex and age. The Journal of Trauma 28:368-378. [Medline]
• Fildes, B. (1997). Safety of older drivers: Strategy for future research and action initiatives. [Monash University Accident Research Centre, Report No. 118]. Clayton, Australia: Monash University.
• George D., Mallery P., 2000. SPSS for Windows step by step: A simple guide and reference 9.0 update Allyn and Bacon, Boston.
• Hamburg M., 1977. Statistical analysis for decision making Harcourt Brace Jovanovich, New York.
• Janke M. K., 1991. Accidents, mileage, and the exaggeration of risk. Accident Analysis and Prevention 23:183-188.
• Li G., Baker S. P., Langlois J. A., Kelen G. D., 1998. Are female drivers safer? An application of the decomposition method. Epidemiology 9:379-384. [Medline]
• Lilley J. M., Arie T., Chilvers E. D., 1995. Accidents involving older people: A review of the literature. Age and Ageing 24:346-365. [Free Full Text]
• Massie D. L., Campbell K. L., Williams A. F., 1995. Traffic accident involvement rates by driver age and gender. Accident Analysis and Prevention 27:73-87.
• Massie D. L., Green P. E., Campbell K. L., 1997. Crash involvement rates by driver gender and the role of average annual mileage. Accident Analysis and Prevention 29:675-685.
• McGwin G. J., Brown D. B., 1999. Characteristics of traffic crashes among young, middle-aged, and older drivers. Accident Analysis and Prevention 31:181-198.
• Miller T., Lestina D., Galbraith M., Schlax T., Mabery P., Deering R., 1997. United States passenger-vehicle crashes by crash geometry: Direct costs and other losses. Accident Analysis and Prevention 29:343-352.
• Miller T. R., Lestina D. C., Spicer R. S., 1998. Highway crash costs in the United States by driver age, blood alcohol level, victim age, and restraint use. Accident Analysis and Prevention 30:137-150.
• Mortimer R. G., Fell J. C., 1989. Older drivers: Their night fatal crash involvement and risk. Accident Analysis and Prevention 21:273-282.
• Pruesser D. F., Williams A. F., Ferguson S. A., Ulmer R. G., Weinstein H. B., 1998. Fatal crash risk for older drivers at intersections. Accident Analysis and Prevention 30:151-159.
• Retchin S. M., Anapolle J., 1993. An overview of the older driver. Clinics in Geriatric Medicine 9:279-296. [Medline]
• Reuben D. B., Stilliman R. A., Traines M., 1988. The aging driver—Medicine, policy and ethics. Journal of the American Geriatrics Society 36:1135-1142. [Medline]
• Reynolds S. L., Crimmins E. M., Saito Y., 1998. Cohort differences in disability and disease presence. The Gerontologist 38:578-590. [Abstract]
• Ryan G. A., Legge M., Rosman D., 1998. Age related changes in drivers' crash risk and crash type. Accident Analysis and Prevention 30:379-387.
• Shua-Haim J. R., Gross J. S., 1996. The "Co-pilot" driver syndrome. Journal of the American Geriatrics Society 44:815-817. [Medline]
• Snedecor G. W., Cochran W. G., 1989. Statistical methods Iowa State University Press, Ames, IA.
• Spiegel M. R., 1992. Theory and problems of statistics McGraw-Hill, London.
• Stutts J. C., Martell C., 1992. Older driver population and crash involvement trends, 1974–1988. Accident Analysis and Prevention 24:317-327.
• Viano D. C., Culver C. C., Evans L., Frick M., Scott R., 1990. Involvement of older drivers in multivehicle side-impact crashes. Accident Analysis and Prevention 22:177-188.
• Weinand M., 1996. Safety measures for elderly drivers: The situation in Germany. Journal of the International Association of Traffic and Safety Sciences 20:67-75.
• Whitfield R. A., Fife D., 1987. Changing patterns in motor vehicle crash mortality: 1940–1980. Accident Analysis and Prevention 19:261-269.
• Williams A. F., Carsten O., 1989. Driver age and crash involvement. American Journal of Public Health 79:326-327. [Abstract/Free Full Text]
• Zhang J., Fraser S., Clarke K., Mao Y., 1998. Age-specific patterns of factors related to fatal motor vehicle traffic crashes: Focus on young and elderly drivers. Public Health 112:289-295. [Medline]

This Article Abstract Full Text (PDF) Services Download to citation manager PubMed PubMed Citation