JAGS 49:431–435, 2001© 2001 by the American Geriatrics Society 0002-8614/01/$15.00

Driving Cessation: What Older Former Drivers Tell Us

Ann M. Dellinger, PhD, MPH,* Meena Sehgal, MPH,* David A. Sleet, PhD,* andElizabeth Barrett-Connor, MD

†

OBJECTIVES:

To understand why older drivers living ina community setting stop driving.

DESIGN:

A cross-sectional study within a longitudinalcohort.

SETTING:

A geographically defined community in south-ern California.

PARTICIPANTS:

1,950 respondents age 55 and older whoreported ever being licensed drivers.

MEASUREMENTS:

A mailed survey instrument of self-reported driving habits linked to prior demographic, health,and medical information.

RESULTS:

Of the 1,950 eligible respondents, 141 hadstopped driving within the previous 5 years. Among thosewho stopped, mean age was 85.5 years, 65.2% were fe-male, and the majority reported they were in very good(43.4%) or good (34.0%) health. Nearly two-thirds re-ported driving less than 50 miles per week prior to stop-ping and 12.1% reported a motor vehicle crash during theprevious 5 years. The most common reasons reported forstopping were medical (41.0%) and age-related (19.4%).In bivariate analyses, age and miles driven per week wereeach associated with cessation (

P

#

.001). Medical condi-tions, crashes in the previous 5 years, and gender did notreach statistical significance at the

P

#

.05 level. Logisticregression results found that the number of medical condi-tions was inversely associated with driving cessation.

CONCLUSION:

The relationship between medical condi-tions and driving is complex; while medical conditions werethe most common reason given for driving cessation, thosewho stopped had fewer medical conditions than currentdrivers. This suggests that a broader measure of generalhealth or functional ability may play a dominant role in de-cisions to stop driving.

J Am Geriatr Soc 49:431–435, 2001.Key words: driving cessation; older drivers; motor vehicle

T

he safety of older drivers has received increased atten-tion in recent years. The structure of our population

is changing so that the proportion of older persons is in-creasing. The population age 65 and over grew 11-foldfrom 1900 to 1994, while the population under 65 grewjust threefold. U.S. Census Bureau projections estimatethat by the year 2050 the population age 65 and overcould be as large as 80 million.

1

If just 75% of these per-sons are licensed to drive, we can expect 60 million olderlicensed drivers in 2050. Evidence suggests that the pro-portion of older persons licensed to drive is rising,

2

largelythe result of more older women driving.

3

Older drivers have among the highest motor vehiclecrash death rates per vehicle mile traveled.

4

Each year, ap-proximately 3,000 older drivers are killed and more than100,000 are nonfatally injured in traffic crashes. Fatal motorvehicle crash rates per 100 million vehicle miles traveled fol-low a U-shaped curve, with the highest rates among theyoungest and oldest drivers. In 1998, people 65 and olderrepresented 13% of the population, but constituted 17% ofall traffic fatalities.

5,6

These elevated crash statistics among older driverspersist even though many older drivers stop driving volun-tarily (i.e., without legal intervention). Because little isknown about the predictors of voluntary driving cessationamong older adults, this study was conducted to under-stand why active older drivers living in a community set-ting stop driving. A better understanding of the factorsthat influence older drivers to stop driving may help publichealth and medical practitioners advise those who need tostop driving or are considering it for safety reasons.

METHODS

Study Design

This was a cross-sectional study in a longitudinal cohort ofcommunity-dwelling adults in southern California. The co-hort was originally established in 1972–1974 as part of theLipid Research Clinics Prevalence Studies. Eighty-two percentof the adult residents age 50 to 79 living in a geographicallydefined community, Rancho Bernardo, were enrolled at thattime. The cohort was largely white and middle to upper-mid-dle class. Enrollees were recruited by telephone and surveyedin person for a variety of demographic and health status mea-sures.

7

The cohort is seen intermittently for health and medi-cal assessment and mailed a short annual survey designed

From the *National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Atlanta, Georgia and

†

University of Cali-fornia, San Diego, School of Medicine, Department of Family and Preven-tive Medicine, San Diego, California.Address correspondence to Ann M. Dellinger, PhD, MPH, National Centerfor Injury Prevention and Control, Centers for Disease Control andPrevention, 4770 Buford Highway NE, Mailstop K-63, Atlanta, GA 30341.

432

DELLINGER ET AL

. APRIL 2001–VOL. 49, NO. 4 JAGS

to determine vital status and to address specific researchquestions. This study was part of a health survey mailed in1994. Approximately 69% responded by the end of 1995.

The present sample consisted of 1,950 respondentsage 55 and older who reported ever being licensed drivers.We asked about miles driven, “In an average week, abouthow many miles do/did you drive?” motor vehicle crashes,“In the last 5 years, have you had any automobile accidents,even minor ones, when you were driving?” and why they hadstopped driving, “If you have STOPPED driving, what is themain reason?” Respondents were asked to check one of thefollowing categories for stopping: licensing or license renewalproblems, costs of keeping an automobile, someone else canalways drive me, medical problems, changes due to aging, orother reason. We also matched information on driving tomedical data and health status measures collected during in-person interviews from 1992 and 1995. As an indication ofcomorbidity, we summed the number of serious self-reportedmedical conditions including angina, heart problems, highblood pressure, diabetes mellitus, arthritis, and osteoporosis.

Medical, health status, and comorbidity data were col-lected on all drivers, but analyses focused on the 141 driverswho reported they had stopped driving within the previous5 years. Our analysis excluded 92 persons who reportedstopping more than 5 years before the survey, four personswho reported their age at stopping as greater than their ageat the time of the survey, and six persons who reportedstopping within the previous 5 years but who also reportedbeing current drivers. Additionally, we excluded 29 personswho reported that they had never been licensed drivers.

Data Analysis

We used Statistical Analysis System (SAS) software to gen-erate all statistics.

8

Frequency of stopping was reported bydemographic characteristic. Reasons for stopping drivingwere grouped into six categories: (1) licensing or licenserenewal problems, (2) costs of keeping an automobile, (3)someone else can always drive me, (4) medical problems,(5) changes due to aging, and (6) other. The self-reportedlevel of general health was categorized as excellent, verygood, good, fair, or poor.

Pearson chi-square tests were used to: (1) assess theassociation between driving cessation in the previous 5years and present age, health, gender, miles previouslydriven in an average week, number of medical conditions,and number of motor vehicle crashes in the previous 5years; (2) compare those persons who stopped driving inthe previous 5 years to current drivers; and (3) comparemen and women who stopped driving.

In order to simultaneously control for multiple variablesof interest, we used a logistic regression modeling procedureto explore the association of several independent variables(age, gender, miles previously driven in an average week,number of crashes in previous 5 years, and number of medi-cal conditions) with the dependent variable (stopped driving).

RESULTS

All Drivers

Surveys were sent to 3,988 participants and 2,770 re-sponded for a response rate of 69%. There were 1,950 re-spondents who had ever been licensed drivers, all were age

55 and older, with a mean age of 73.8 years, 58.7% werefemale. Most (91.7%) reported that they were in excellent,very good, or good health. Only 8.3% reported fair orpoor health, but one-third reported one or more of the co-morbid conditions. Two hundred seventy-five (14.6%) re-spondents reported at least one motor vehicle crash in theprevious 5 years while they were driving (Table 1).

Drivers Who Stopped in the Previous 5 Years

Nearly two-thirds of the 141 persons who stopped drivingwithin the previous 5 years were female (65.2%), their meanage was 85.5 years. Age at time of stopping ranged from 65to 95 years. Just 2% stopped in their 60s, 18% in their 70s,63% in their 80s and 17% in their 90s. This represents 0.6%of those respondents in their 60s, 4.2% of those in their 70s,15.8% of those in their 80s, and 23.3% of those in their 90s.The majority reported very good (43.4%) or good (34.0%)health. Nearly two-thirds reported driving less than 50 milesper week prior to stopping, although 73% had none of theserious comorbid conditions. Of those who stopped driving,12.1% reported a motor vehicle crash in the previous 5 yearswhile driving (Table 1). Driving cessation did not vary bygender, and there were no statistically significant differences

Table 1. Characteristics of the Sample

Characteristic

“EverDrivers”

n

5

1,950

“DrivingNow”

n

5

1,686*

“StoppedLast 5 Years”

n

5

141

%

Age

†

55–64 24.2 27.6 0.965–74 28.5 31.4 5.975–84 32.5 32.5 33.1

$

85 14.8 8.5 61.0Gender

Male 41.0 42.8 34.8Female 58.7 57.2 65.2

Self-reported healthExcellent 22.3 23.9 5.7Very good 43.1 43.6 43.4Good 26.3 25.3 34.0Fair 7.3 6.4 15.0Poor 1.0 0.8 1.9

Miles driven per week

†

,

50 34.7 31.2 64.950–100 33.8 35.2 21.1

.

100 31.5 33.6 14.0Selected medical

conditions

‡

0 65.7 65.1 73.11 22.0 22.8 17.72 8.8 8.8 5.73

1

3.5 3.3 3.5Motor vehicle crash

last 5 years 14.6 15.2 12.1

*

“Ever drivers” less those who stopped at any time in the past, those who reportedcurrently driving and stopping, and those who gave other inconsistent answers.

†

Statistically significant at the

P

,

0.05 level.

‡

Sum of the number of medical conditions reported including angina, heart prob-lems, high blood pressure, diabetes mellitus, arthritis, and osteoporosis.

JAGS APRIL 2001–VOL. 49, NO. 4

DRIVING CESSATION: WHAT OLDER FORMER DRIVERS TELL US

433

between men and women who stopped driving with respectto number of miles driven before stopping, number of medi-cal conditions, or the number of crashes in the previous 5years while driving. Nevertheless, there were differences intheir reasons for stopping (

P

#

.04); compared with men,women more frequently reported licensing problems, costs ofkeeping an automobile, and that someone else could drivethem.

Respondents who stopped driving were asked tocheck one of six categories for the

main

reason theystopped driving; 139 respondents indicated at least onereason. Medical problems (41.0%) and changes due to ag-ing (19.4%) were the most common categories checked,followed by licensing or license renewal problems (12.2%)and other reasons (12.1%) (Table 2). The survey alsoasked the respondent to specify (i.e., write in) the mainmedical, age-related, or other reason for stopping if theyhad checked one of these boxes. Although respondentswere asked to choose one main reason for stopping, 13(9%) of 139 participants who responded to this questionwrote in more than one reason or diagnosis. Thus, Table 2gives the percentage of respondents who checked off eachcategory, and Table 3 gives their specified responses.

Of the 57 persons checking medical problems as theirmain reason for stopping, 52 wrote in 60 different reasonsfor stopping. The most common reason was vision (n

5

20) followed by cardiovascular conditions (9), Parkinson’sdisease (6), arthritis (4), slow reactions or slow driving (3),accidents (2), and 16 persons wrote in other reasons suchas sciatica, leg numbness, or osteoporosis (Table 3). Of the27 persons checking changes due to aging as their mainreason for stopping, the most common reasons given werevision (7), slow reactions or slow driving (3), and time tostop or not safe (3). For the 17 persons who checked li-censing or license renewal problems as their main reasonfor stopping there was no area to write in a specific re-sponse (see Table 2 for question format). Nevertheless,eight persons wrote in ten responses. A similar result oc-curred in the someone else can always drive me categorywhere four persons gave six reasons for stopping (Table 3).

Of the 139 persons who specified any reason for stop-ping, 5.0% (n

5

7) indicated accidents and an additional8.6% (n

5

12) indicated they might be unsafe drivers.Where drivers who stopped within the previous 5

years were compared with current drivers (Table 1), thosewho stopped were more likely to be older (

P

#

.001) andin poorer health (

P

#

.003) than those driving at the timeof the survey. There were no differences in the number ofmotor vehicle crashes (

P

#

.33), or the number of medicalconditions (

P

#

.21).We performed Pearson chi-square tests to measure the

association between demographic characteristics and driv-ing cessation. Age and miles driven per week were eachstatistically significant at the

P

#

.001 level. Medical con-ditions, crashes in the previous 5 years, and gender did notreach statistical significance at the

P

#

.05 level.A logistic regression model containing number of

medical conditions, crashes in the previous 5 years, gen-der, age, and miles driven per week was used to assess theindependent associations of each variable with driving ces-sation. Results were similar to the bivariate analyses inthat age (

P

#

.0001) and miles driven per week (

P

#

.0001) were statistically significant, gender (

P

#

.55) andcrashes (

P

#

.90) were not. The one difference in the logis-tic regression results was that the number of medical con-ditions became significantly associated with driving cessa-tion (

P

#

.02) but this relationship was in an unexpecteddirection—the more medical conditions, the less likely aperson was to have stopped driving.

DISCUSSION

We found that respondents who had stopped driving inthe previous 5 years tended to be older and female, al-though gender differences did not reach statistical signifi-cance. Medical problems were the most commonly reportedreason for driving cessation, and those who stopped drivingwere twice as likely to report fair or poor health. Neverthe-less, those who stopped had fewer medical conditions thanthose who continued to drive and logistic regression re-sults found that the number of medical conditions was in-versely related to cessation. These seemingly inconsistentresults may be explained by a broader or more general no-tion of health. In this case, people who stopped drivinghad fewer medical conditions but lower levels of self-reported health than current drivers. This paradox may il-lustrate the less-than-perfect match between medical diag-noses and general functioning and suggests that for theseolder drivers, making the decision to stop driving wasmore likely to be based on an individual assessment of ca-pabilities than on a medical diagnosis. The factors thatcomprise this individual assessment are not known, but,given the wide variety of reasons participants gave fordriving cessation, are likely to be multifaceted.

Other studies that have explored the reasons olderpersons stop driving had differing results, supporting thesuggestion that the decision to stop driving is complex,and likely influenced by several factors. Marottoli et al.looked at possible predictors of driving cessation andfound that increasing age, unemployment, low income,neurological disease, cataracts, decreased physical activity,and functional disability were all associated with cessa-tion.

9

In a study of how social factors affect driving deci-sions, Chipman et al. found that gender and marital statuswere more influential than the presence of chronic dis-eases.

10

In contrast, Hakamies-Blomqvist and Wahlstromfound that deteriorating health was the most frequent rea-son given by older men for stopping driving.

11

Campbell etal. also studied the prevalence of medical conditions and

Table 2. Self-Reported Reasons for Driving Cessation asAsked on Survey n

5

141

*

If you have STOPPED driving, what is the mainreason? (Check only one) Number (%)

†

Medical problems (main one?) ____________ 57 (41.0)Changes due to aging (main one?) _________ 27 (19.4)Licensing or license renewal problems 17 (12.2)Other reason (specify) ___________________ 17 (12.1)Someone else can always drive me 15 (10.8)Costs of keeping an automobile 6 (4.3)

*

Total adds to 139 because two respondents did not give a reason for stopping.

†

Percentages do not sum to 100 due to rounding.

434

DELLINGER ET AL

. APRIL 2001–VOL. 49, NO. 4 JAGS

their relationship to driving cessation. They reported thatage, activity limitation, syncope, macular degeneration,stroke (among males), Parkinson’s disease (among females),and retinal hemorrhaging (among females) predicted cessa-tion.

12

Persson conducted focus groups in which 56 olderdrivers identified why they had stopped driving. The mostcommon reasons were advice from a doctor, increased ner-vousness behind the wheel, trouble seeing pedestrians andcars, and medical conditions.

13

Most of those former driversfelt they had stopped driving at the right time and that theywere the best ones to make the decision to stop.

In the present study, 25% of respondents specified vi-sion (e.g., macular degeneration, cataracts, glaucoma, blind-ness) as a reason for cessation. Some characterized it as amedical condition, some as a change related to aging, andsome as an “other” reason.

The relationship between driving and vision has beenthe topic of several other studies. Rabbitt et al. found thatvision was the most commonly reported health problem in

a group of more than 2,000 drivers and former drivers inEngland,

14

and Johnson and Keltner’s study of 10,000 vol-unteers found that drivers with binocular visual field losshad crash and conviction rates twice those of drivers withnormal fields of vision.

15

These results suggest that visionmay play a key role in how often and how far people driveand may demonstrably affect their level of crash risk.

One logical extension of this hypothesis is that visiontests could discriminate between drivers who reduce theirdriving and those who do not, and between crash-free andcrash-involved drivers. Hennessey studied how well visualtests (tests included the following: Pelli-Robson Low Con-trast Acuity Test, Smith-Kettlewell Low-Luminance Care,Berkeley Glare Tester, Modified Synemed Perimeter, Vi-sual Attention Analyzer/Useful Field of View) could pre-dict self-restriction and crashes. He found that in 3,699applicants for driver’s license renewal, neither vision testscores nor age accounted for more than 7% of the varia-tion in types of self-restriction. In addition, the five visiontests were not significantly associated with crashes that oc-curred within the 3 previous years for all ages combined.In the 70 and older age group, however, the Useful Fieldof View and standard visual field tests were significantlycorrelated with previous crashes, although they explainedonly a small amount of the variance in crash involve-ment.

16

This may be partly explained by the nature of thevisual tests involved. Tests that measure higher-level visualfunctioning, for example, visual processing speed, mayprove to be better predictors of crashes than tests thatmeasure acuity alone. Another factor influencing this rela-tionship is driver self-regulation. There is evidence to showthat drivers with visual impairment (e.g., cataract, age-related macular degeneration) may modify their drivingbehavior,

17

yet at the same time individuals are sometimesunaware of their visual problems.

18

If some types of visualimpairment are less likely to be recognized, for example,decreased visual processing speed, persons would not beinclined to modify their driving behavior. We can conceiveof several groups of drivers: (1) those with recognized vi-sual impairment who have modified their driving, (2) thosewith recognized visual impairment who have not modifiedtheir driving, (3) those with unrecognized visual impair-ment who have modified their driving (for other reasons),and (4) those with unrecognized visual impairment whohave not modified their driving. Visual testing would mostlikely affect persons in group four above who would thenbe willing to modify their driving. Given this scenario, it isnot surprising that while visual impairment is a factor indriving ability and crash risk, it is not, by itself, a good pre-dictor of most crashes or of who will stop driving.

This study found no clear indication that crashes causedrespondents to stop driving, only 5% gave this reason. Ofthe respondents who had stopped driving, 12.1% stated thatthey had had a motor vehicle crash while driving within theprevious 5 years. Stewart et al., in their longitudinal study onaging, reported similar results—9.1% of women and 11.2%of men had crashed in the previous 5 years.

19

This study has several limitations. First, data wereself-reported. Respondents may have interpreted the ques-tion, “In the last 5 years, have you had any automobile ac-cidents, even minor ones, when you were driving?” as re-ferring only to crashes reported to the police. In that case,

Table 3. Specified Reasons for Driving Cessation as Reportedon Survey

Category Reason (Number)

Medical problems (n

5

57) Vision (20)Cardiovascular (9)Parkinson’s Disease (6)Arthritis (4)Slow reactions/slow driving (3)Accidents (2)Other* (16)Nothing specified (5)

Changes due to aging (n

5

27) Vision (7)Slow reactions/slow driving (3)Time to stop/not safe (3)Cardiovascular (1)Other* (4)Nothing specified (9)

Licensing problems (n

5

17) Vision (4)Cardiovascular (2)Time to stop/not safe (1)Other* (3)Nothing specified (9)

Other reasons (n

5

17) Accidents (4)Moving (3)Vision (2)Other* (7)Nothing specified (1)

Someone else can alwaysdrive me (n

5

15) Vision (2)Arthritis (1)Slow reactions/slow driving (1)Other* (2)Nothing specified (11)

Costs (n

5

6) Accidents (1)Other* (1)Nothing specified (4)

*

Other reasons included anemia, breathlessness, bad back, broken shoulder, pros-tate cancer, sold car, poor circulation, not confident, depressed, diabetic neuropa-thy, don’t like, hip fracture, no license renewal, leg numbness, memory loss, os-teoporosis, sciatica, seizures, stress, surgery, telangiectasis, and weakness.

JAGS APRIL 2001–VOL. 49, NO. 4

DRIVING CESSATION: WHAT OLDER FORMER DRIVERS TELL US

435

the number of crashes would have been underestimated.Also, 5 years may have been a long period of recall forsome respondents. Those who stopped driving nearer thesurvey time may have had better recall of the reasons theystopped or whether they had any crashes during that timeperiod. Moreover, those who stopped nearer the surveyhad more opportunity for a crash, since they were drivingfor a larger part of the previous 5 years. However, sincethose who stopped also drove very few miles before stop-ping, this is unlikely to be a large source of bias. Althoughpersons who had stopped driving more than 5 years priorto the survey were not included in the analyses, bothgroups reported medical problems and age-related reasonsfor cessation in similar proportions. Respondents were ed-ucated, Caucasian residents of a southern California sub-urb. Results may not apply to low-income, racial minori-ties, or rural or inner city seniors.

Some respondents were not able to select just onemain reason for cessation and listed several, varied rea-sons. It appears that as aging progresses and skills deterio-rate, the complex task of driving becomes more difficultand the decision to stop may be motivated by several fac-tors. The decision to stop driving may be a gradual processthat is not explained by one factor alone. The cessationprocess may occur in stages so that it is not recognized as asingle decision to quit, but as a gradual progression of self-imposed restrictions that culminate in cessation. If there isa

natural history

of cessation, or a

cessation continuum

among older drivers, then the reported reasons for stoppingwould vary depending on when the person was questioned.

Some reasons given for stopping were unusual (e.g.,anemia, telangiectasia) and difficult to interpret becausethey do not appear to be related to driving skills. Personaland perceptual reasons for stopping included not confi-dent, don’t like, not safe, and slow reactions. There were115 instances in which respondents wrote down one ormore reasons for stopping, yet none of these instances in-cluded advice from family, friends, or physicians, suggestingthat a subjective assessment of one’s own driving ability andskills contributes to the decision to stop driving. We shouldlearn more from drivers about their motivations to stop andthe process of self-regulation, to provide a more-reasonedapproach to reducing crash rates among older drivers.

Physicians may be in a unique position to aid in thisprocess by providing accurate and timely patient feedback.In our study of mostly healthy community-dwelling volun-teers, respondents who quit driving were twice as likely toreport fair or poor health as current drivers. This suggeststhat these older drivers may have regular contact with themedical system. In addition, there is some evidence thatpatients take advice from physicians about driving very se-riously, while advice from their families may be less influ-ential.

14

The American Medical Association has adoptedH-140.925 Impaired Drivers and Their Physicians, whichstates that physicians have an ethical responsibility to as-sess physical or mental impairments that might affect driv-ing. If necessary, physicians should use their best judge-ment when deciding when to report a patient to theDepartment of Motor Vehicles.

20

Currently, there are nonationally accepted guidelines for physicians to use incounseling older drivers; however, the medical advisoryboards of several states have information to help physi-

cians deal with these issues.

21,22

From a public health per-spective, these programs need rigorous scientific evalua-tion before they can be recommended on a national level.

Better understanding of the natural history of driving ces-sation is needed to determine whether there are critical stagesin this process when advice or counseling would be most ben-eficial to older adults. This information can then be used totailor prevention programs to maximize their effectiveness.

REFERENCES

1. U.S. Bureau of the Census. 65

1

in the United States, Statistical Brief (No. SB/95-8). Washington, DC: Bureau of the Census, U.S. Department of Com-merce, 1995.

2. U.S. Department of Transportation. Highway Statistics Summary to 1995(No. 050-001-00323-6). Washington, DC: Office of Highway InformationManagement, Federal Highway Administration, U.S. Department of Trans-portation, 1997.

3. Spain D. Societal trends: The aging baby boom and women’s increased inde-pendence (No. DTFH61-97-P-00314). Washington, DC: 1, 1997.

4. Cerrelli E. Crash data and rates for age-sex groups of drivers, 1996. Wash-ington, DC: National Highway Traffic Safety Administration, 1998.

5. U.S. Department of Transportation. Traffic Safety Facts 1998 (Publicationno. DOT HS 808 983). Washington, DC: National Center for Statistics andAnalysis, National Highway Traffic Safety Administration, 1999.

6. National Highway Traffic Safety Administration. General Estimates System(GES), Data for 1998. Washington, DC: National Center for Statistics andAnalysis, National Highway Traffic Safety Administration, 1998.

7. Criqui MH, Barrett-Connor E, Austin M. Differences between respondentsand non-respondents in a population-based cardiovascular disease study. AmJ Epidemiol 1978;108:367–372.

8. SAS Institute. Statistical Analysis System (SAS) Software (Version 6.11).Cary, NC: SAS Institute, 1998.

9. Marottoli RA, Ostfeld AM, Merrill SS et al. Driving cessation and changes inmileage driven among elderly individuals. J Gerontol B Psychol Sci Soc Sci1993;48:S255–S260.

10. Chipman ML, Payne J, McDonough P. To drive or not to drive: The influ-ence of social factors on the decisions of elderly drivers. Accid Anal Prev1998;30:299–304.

11. Hakamies-Blomqvist L, Wahlstrom B. Why do older drivers give up driving?Accid Anal Prev 1998;30:305–312.

12. Campbell MK, Bush TL, Hale WE. Medical conditions associated with driv-ing cessation in community-dwelling, ambulatory elders. J Gerontol B Psy-chol Sci Soc Sci 1993;48:S230–S234.

13. Persson D. The elderly driver: Deciding when to stop. Gerontologist 1993;33:88–91.

14. Rabbitt P, Carmichael A, Jones S et al. When and why older drivers give updriving. Hampshire, United Kingdom: AA Foundation for Road Safety Re-search, 1996.

15. Johnson CA, Keltner JL. Incidence of visual field loss in 20,000 eyes and itsrelationship to driving performance. Arch Ophthalmol 1983;101:371–375.

16. Hennessy DF. Vision Testing of Renewal Applicants: Crashes Predicted whenCompensation for Impairment Is Inadequate (RSS-95-152). Sacramento, CA:California Department of Motor Vehicles, 1995.

17. Ball K, Owsley C, Stalvey B et al. Driving avoidance and functional impair-ment in older drivers. Accid Anal Prev 1998;30:313–322.

18. Ball K, Owsley C. The Useful Field of View Test: A new technique for evalu-ating age-related declines in visual function. J Am Optom Assoc 1992;64:71–79.

19. Stewart RB, Moore MT, Marks RG et al. Driving cessation and accidents inthe elderly: An analysis of symptoms, diseases, cognitive dysfunction andmedications. Washington, DC: AAA Foundation for Traffic Safety, 1993.

20. Council on Ethical and Judicial Affairs, American Medical Association. Reportof the Council on Ethical and Judicial Affairs, American Medical Association(CEJA Report 4-A-99). Chicago, IL: American Medical Association, 2000.

21. Medical Advisory Board. Functional Ability Profiles Governing Physical,Emotional and Mental Competence to Operate a Motor Vehicle, FAP-II. Au-gusta, ME: Bureau of Motor Vehicles, Department of the Secretary of State,State of Maine, 1995.

22. Anapolle J, Gnazzo JA. Survey of licensing procedures for older, physicallyand/or mentally impaired drivers (Publication no. 17271). Boston, MA: De-partment of Elder Affairs, Registry of Motor Vehicles, Commonwealth ofMassachusetts, 1992.

23. Petrucelli E, Malinowski M. Status of Medical Review in Driver Licensing:Policies, Programs and Standards (Publication no. DOT HS 807 892).Springfield, VA: National Highway Traffic Safety Administration, 1992.