Table of Contents
Next Section
Home

I. PROGRAM GOALS AND PROCEDURES

The singular objective of this work is to promote the safe mobility of older persons. Mobility is central to quality of life. There is a well-established link between restricted mobility among older persons and the onset or acceleration of diverse physical and mental health problems. Costs to society to provide care for seniors who lose their mobility also rise dramatically. To preserve independent functioning, to retain the dignity and self-esteem that result from providing for one's own mobility needs as long as it is possible to do so without unacceptable risk to oneself or to others-- these are the overriding goals in a U.S.DOT policy initiative, Safe Mobility for Life, that provides the framework for application of material in this Notebook.

In our society personal mobility, to an overwhelming degree, is tied to the ability to drive a car. People who drive automobiles can exercise the freedom to choose where to work, live, and recreate; their social needs and maintenance requirements can be self-fulfilled; and they can travel virtually at any time they desire. These attributes of a contemporary lifestyle, and the means most often used to attain them, are perceived to be among the basic rights of every adult.

As people age, however, their ability to safely drive a car may be compromised by a variety of functional impairments. The functional abilities at issue include vision, attention, perceptual skills, memory, decision making, reaction time and different aspects of physical fitness and performance. With increasing age, the occurrence of disease and pathology are more common and, even in their absence, declines in functional abilities are to be expected as a normal consequence of aging. There is an accumulating body of evidence to show that impairments in one or more areas of functional capability significantly increase a driver's risk of a crash. And because of their higher vulnerability, older persons involved in an automobile crash are more likely than their younger counterparts to be seriously injured or killed. The leading cause of accidental death for older persons is a car crash.

The changing demographics in our society underscore the consequences of age-related driving impairments as an emerging public health issue. The population over age 65 will grow by 60 percent in the next 20 years; during the decade from 2020 to 2030, the proportion of Americans over the age of 65 will increase to more than 1 in 5. The development of screening procedures for license renewal and regulatory control that are fair, accurate, and which can be administered cost-effectively is therefore a clear priority. This was the premise behind a NHTSA research project, "Model Driver Screening and Evaluation Program;" the information presented in this Notebook was generated through performance of that project.

Improved practices for assessing drivers' abilities and driving skills are overdue. Through the decade of the 1990s and beyond, as people age 85 and older have emerged as the fastest growing segment of our driving population, the driving task itself has become characterized by ever-growing traffic volumes and congestion, plus novel highway features and vehicle technologies that demand greater attention by the driver. Most seniors are as capable of driving safely as their younger counterparts, and when they become aware that they have a problem they typically act responsibly by limiting or modifying their driving habits. Still, some diminished functional capabilities are more difficult to detect or may be denied, and the margin for 'human error' in many driving situations has become vanishingly small. Thus the payback for more accurate prediction of who is at greatest risk of causing a crash is substantial--both for the individual and for society.

The Model Program's first priority has accordingly been to identify the most useful tools for evaluation of drivers' functional capabilities. For many reasons, it is anticipated that functional screening will not be confined to Departments of Motor Vehicles, and that the DMV may not even be the most important setting for early screening to occur. Providing tools for self-evaluation by older drivers, and for screenings in various health care and social service settings in the community, is strongly emphasized in the Notebook. A need for multiple tiers of evaluation activities is also emphasized, such that results of early screening for gross impairments lead to more comprehensive, diagnostic testing by appropriate professionals whenever warranted.

While identifying and assessing the ability of older people to remain safely mobile receives the largest share of attention in the Notebook, other goals are also defined. When it has been determined that an individual has one or more functional limitations that are likely to produce driving impairments, the Model Program supports remediation of the problem if possible, and the provision of mobility counseling to inform the individual about local alternative transportation options and how to access available services. More broadly, the Model Program also includes a public information and education component to help meet the assessment, remediation, and counseling goals by informing senior citizens and care givers about the link between functional decline and driving safety, and about resources that exist to help preserve or extend their mobility as they grow older.

The procedures described and reported on in the Notebook will give readers an understanding of the current state-of-the-knowledge in a given topic area, and will identify the principal sources of information and evidence for the included conclusions and recommendations. At the same time, the conclusions stated in this Notebook are preliminary and current knowledge may derive from research-in-progress. Where readers note significant omissions in material or material that is out of date with current practices it is requested that they bring such items to the attention of the authors. This reference document is, and should remain, a work-in-progress as jurisdictions throughout North America prioritize local issues relating to seniors' mobility needs, and implement the best solutions that are feasible at the time.

I.A. IDENTIFY OLDER PEOPLE WHO ARE AT HIGH RISK OF CRASHES


I.A.1. Epidemiology

(a) Dementia
(b) Cataracts
(c) Diabetes and Associated Conditions
(d) Glaucoma
(e) Foot Abnormalities
(f) Falls
(g) Cardiac (and Cardiopulmonary) Conditions
(h) Feet or Legs Cold on Exposure to Cold
(I) Bursitis
(j) Renal Disease
(k) Seizure Disorders
(l) Back Pain
(m) Overview: Comparative Risk Tables


The NHTSA/AAMVA (1980) document entitled, Functional Aspects of Driver Impairment: A Guide for State Medical Advisory Boards states that "... there is evidence that, as a group, individuals with certain types of medical impairment constitute a greater risk on the highway than does the population at large." However, while researchers have been trying for decades to determine the extent to which medical impairments lead to increased crash risk, none of the commonly studied medical conditions (e.g., diabetes, heart disease, stroke, Parkinson's disease) have been consistently associated with a high vehicle crash rate in older drivers (Hu, 1997). In fact, it is not the mere presence of the disease, but instead the functional limitations caused by the disease, that is key to predicting driving impairment. Unfortunately, as noted by Janke (1994), the degree of severity of the medical condition has not been typically considered in past research studies. Also, as people age, they are likely to develop multiple medical conditions, which makes it difficult to determine which specific condition was most impairing to the driving task. The information provided in this section of the Notebook presents findings from recent studies conducted by physicians, occupational therapists, epidemiologists, and other researchers who have sought to control for many of the extraneous variables that so often cloud the investigations of medical conditions and driving performance in older persons. From these data, the Notebook attempts to summarize the associations between age-related diminished functional abilities and crash risk in Section IA2. Section IC2(b)v (Test Procedures: comprehensive physical examination), and Section IC3(b)i (Rehabilitation Procedures: physician/occupational therapist review) provide more information about how physicians can identify at-risk older drivers and specific diagnoses, their effects on driving, and potential remediation.


IA1(a). Dementia


Summary:

Alzheimer's Disease (AD) is the most common cause of dementia, with a prevalence--based on correlation between autopsy data and the outcomes of strict clinical diagnostic procedures--estimated to be as high as 11.6 percent for those 65 and older and 47.8 percent for those over the age of 85 (Evans, Funkenstein, Albert, Scheer, Cook, Herbert, Hennekens, and Taylor, 1989). Drivers with dementia are less likely to report driving problems than cognitively unimpaired drivers, and their perception of their driving ability does not correspond either to that of their caregivers (as assessed by questionnaire) nor their actual driving performance (Cushman, 1992; Tallman, Tuokko, and Beattie, 1993). Thus, they are less likely to limit their exposure to high risk driving situations than are drivers who have diminished visual and physical capabilities, but intact cognitive capabilities. Throughout the first three years the crash rate for AD patients is only slightly higher than that for drivers of all ages in the United States, and remains well below that of young adults aged 16 to 24. Although the course of AD may vary considerably, study findings suggest that the increase in crash risk develops toward the end of the third year, and more than doubles in the fourth year (see Staplin, Lococo, McKnight, McKnight, and Odenheimer, in press, for a review of dementia and diminished driving skills).

A recent matched-pair, case-control study, with close (1-year) age matching was conducted in Sweden, using the Clinical Dementia Rating (CDR) scale to measure dementia severity. In this study, questionable dementia (CDR=0.5) and mild dementia (CDR=1) were found significantly more often in the case group (37 drivers age 65+ with license suspended due to crashes or moving violations) than in the matched control group (37 drivers age 65+ with no license suspensions in past 5 years). Dementia was found in 49 percent of the cases versus 11 percent of the controls. Comparison of the 23 case subjects with crashes and the 29 control subjects with no crashes in the past 5 years showed that the crashed drivers had more incidence of dementia/CDR>0 (p<.001), worse cube copying (p<.015), poorer 5-item recall (p<.003), a lower Mini-Mental Status Examination (MMSE) score (p<.019), and more EEG abnormalities. (see Johansson, Bronge, Lundberg, Persson, Seideman, and Viitanen, 1996; Johansson, 1997).

In a recent study to assess the reliability and stability of a standardized road test for healthy aging people and those with dementia of the Alzheimer type, a significant relationship between global rating on the road test and Clinical Dementia Rating (CDR) was found, such that most CDR-0 subjects (no dementia) were rated as "safe" drivers [78 percent (45/58) of CDR-0 subjects], compared to 67 percent (24/36) of CDR-0.5 subjects (very mild dementia) and 41 percent (12/29) of CDR-1 subjects (mild dementia)]. Only 3 percent of CDR-0 subjects were judged "unsafe," but 19 percent of CDR-0.5 and 41 percent of CDR-1 subjects were judged "unsafe." The remaining subjects in each CDR group were rated "marginal." (see Hunt, Murphy, Carr, Duchek, Buckles, and Morris, 1997a, and 1997b).

In a study of healthy elderly controls (n=13; mean age=73.5; CDR score=0); subjects with very mild dementia (n=12 ; mean age=72.5; CDR score=0.5), and subjects with mild dementia (n=13; mean age=73.4; CDR score=1.0), the correlation between the pass/fail outcome on the road test and performance on the Logical Memory test was significant at the p<.0009 level. Five subjects--all in the CDR-1 stage-- "failed" the in-car on-road test. The Logical Memory subscale of the Wechsler Memory Scale assesses immediate or delayed recall of verbal ideas presented in two paragraphs, read aloud by the experimenter. (see Hunt, Morris, Edwards, and Wilson, 1993).

Most recently, Salzberg and Moffat (1998) evaluated the driving records of 46 older drivers who had psychiatric conditions (Alzheimer's, bipolar disorders, dementia, and confusion/memory loss) who were referred to the Washington State Special Examination Program (and passed), and 449 control group drivers. An additional 20 drivers with psychiatric conditions failed the special exam, and their licenses were canceled. This constituted 30 percent of the drivers with psychiatric conditions who underwent the special exam. This program is described in more detail in Section IA1(m) of the Notebook. A "special exam" includes an in-depth interview, and an extended or specialized on-road drive test, typically conducted near the driver's residence. The most common outcome of the "special exam" is to impose driving restrictions (time of day, area, equipment).

Crash and violation records of drivers with psychiatric conditions were compared with that of the control group, for a period of 1.75 years before the exam, and 3.25 years after the exam (a 5-year period). Crash and violation rates were calculated to describe the number of incidents per 100 subjects per year, since the pre- and post-observation periods differed in length. The crash and violation rates for the 46 drivers with psychiatric conditions who passed the "special exam" and the (entire) control group are presented below, for the pre-exam and post-exam period. For comparison purposes, in Washington State during 1996 there were 140,215 total collisions and 4,037,534 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers in a one-year period.

Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
Special-Exam Psychiatric Conditions (n=46) 12.4224 4.6823 23.6025 8.0268

Older drivers with psychiatric conditions who passed the "special exam" and received consequent driving restrictions showed a greatly reduced collision and violation rate. However, the rate reduction still resulted in a crash and violation risk that was approximately 4 times that of the control group of older drivers, who did not receive exams and consequent restrictions but also showed reductions in their crash and violation rates over the 5-year period. Of particular interest is that the post-exam collision rate of the psychiatric group (4.6823) was 1.35 times higher than the collision rate of the population of licensed drivers in the State (3.47). This point illustrates that restricting the driving privileges of drivers with psychiatric conditions brings their crash rate more in line (although still higher) with that of the general population of drivers, however, the rate is still much higher than that of a comparison group of older drivers without psychiatric conditions, who (probably) are practicing self-restriction.

Hunt (1994) describes the following situations in which demented drivers experience difficulty:

Familiar routes are no longer well remembered, and the demented individual may become lost while driving.

In an emergency, the driver may confuse the brake pedal with the gas pedal or press on both pedals simultaneously.

Driving situations that demand complex or rapid cognitive processing and problem solving may cause a demented driver to stop in the middle of traffic or otherwise fail to negotiate traffic safely. To an observer, there may seem to be no apparent reason to stop.

In making a left turn at an intersection, the driver may fail to yield the right-of-way or inappropriately attempt to proceed on a green light when the sign reads "left turn on arrow only."

Verbal commands or suggestions from a passenger (i.e., directions; reminders to check traffic before making a lane change) are not interpreted correctly or in time for the proper action to occur.

The American Psychiatric Association's Position Statement on the Role of Psychiatrists in Assessing Driving Ability was drafted by the Council on Aging, approved by the Assembly in November 1993, and by the Board of Trustees in December 1993 (Council on Aging, 1995). It states that: (1) a mental disorder per se does not imply impaired driving capacity; (2) persons suffering from mental disorders may experience symptoms that can interfere with their ability to drive; (3) usually, accurate assessment of the impact of symptoms on functional abilities is not possible in an office or hospital setting because such an assessment typically requires specialized equipment or actual driving observation which goes beyond the scope of ordinary psychiatric care; and (4) since psychiatrists do not have special expertise in assessing patients' ability to drive, they should not be expected to make these assessments in the course of clinical practice. However, the position statement specifies that psychiatrists do have a role to play in advising patients about the potential impact of their illness and treatments on driving ability, as follows: (1) when appropriate, psychiatrists should discuss with their patients symptoms of their mental disorders that may be serious enough to substantially impair their driving ability; (2) psychiatrists should warn their patients about the possible effects of prescribed psychotropic medications on alertness and coordination, and about the possibility that such medications could magnify the effects of alcohol; and (3) when clinically appropriate, medication with a low potential to impair ability should be chosen preferentially, depending on the patient's driving requirements and habits. Finally, the statement mentions that given the importance of maintaining confidentially in psychiatrist-patient relationships, psychiatrists should not be required to report information on a patient's driving ability to state departments of motor vehicles. However, a statute that allows, but does not require, reporting when there is clear-cut evidence of substantial driving impairment (e.g., a family's statement that a moderately demented patient has had several recent minor crashes) is socially desirable and can be clinically useful. The position is that ultimate responsibility for assessment of patients' driving ability should lie with the DMVs. Reports made in good faith, however, should be accompanied by immunity for psychiatrists from subsequent liability.

Conclusions/Preliminary Recommendations:

Diagnosis is not an adequate predictor of function, since there is great heterogeneity in the rate of progress as well as the cognitive strengths and weaknesses among patients with dementing disorders. Diagnosis could thus be important as a way to identify persons for tracking, with decisions on whether driver status should be terminated then based on functional assessments.

Mental status evaluations may be useful in identifying older drivers who are beginning to show evidence of cognitive decline, but on-road or off-road tests, especially those requiring the driver to follow sequential directions, are more likely to measure the skills required for driving. Cutoff scores (MMSE) must be considered as being relative, forming a small part of the basis of making decisions about driving, and secondary to a clinical evaluation; however, MMSE score 10, accompanied by a diagnosis of dementia, indicates a sufficiently low level of cognitive functioning to justify recommending immediate cessation of driving (Lundberg, Johansson, Ball, Bjerre, Blomqvist, Braekhus, Brouwer, Blysma, Carr, Englund, Friedland, Hakamies-Blomqvist, Klemetz, O'Neill, Odenheimer, Rizzo, Schelin, Seideman, Tallman, Viitanen, Waller, and Winblad, 1997).

It is important to note that MMSE scores are influenced by race and level of education, so some adjustment of cutoffs may be necessary.

Patients who have had AD for more than two years should have their driving ability closely monitored if they are to continue driving, as the overall risk to society during the first two years is well within the accepted range for other drivers. This is dependent upon whether AD is defined as early stage (CDR = 0.5) or later stage (CDR >1.0) however.

References:

Council on Aging (1995)

Evans, Funkenstein, Albert, Scheer, Cook, Herbert, Hennekens, and Taylor (1989)

Hunt (1994)

Staplin, Lococo, McKnight, McKnight, and Odenheimer (in press)

Excerpts from Annotated Research Compendium of Driver Assessment Techniques for Age-Related Functional Impairments (Hunt, Morris, Edwards, and Wilson, 1993; Tallman, Tuokko, and Beattie, 1993; Cushman, 1992; Odenheimer, Beaudet, Jette, Albert, Grande, and Minaker, 1994; Johansson, 1997; Lundberg, Johansson, Ball, Bjerre, et al., 1997; Keyl, Rebok, Bylsma, et al., manuscript under review; Duchek, Hunt, Ball, Buckles, and Morris, 1997; Rizzo and Dingus, 1996; Rizzo, Reinach, McGehee, and Dawson, 1997; Hunt, Murphy, Carr, Duchek, Buckles, and Morris, 1997a, and 1997b; Janke and Eberhard, 1998; Staplin, Gish, Decina, Lococo, and McKnight, 1998; DriveAble Testing, March 1997; Dobbs, 1997)


IA1(b). Cataracts


Summary:

Owsley, Stalvey, Wells, and Sloane (1999) conducted a study that included 279 drivers with cataract (mean age = 71) and 67 drivers with no cataract (mean age = 67). This on-going project is an intervention evaluation study to determine how improvement in vision impacts crashes and driving habits. Crash data from 5 years prior to enrollment and 3 years following enrollment were obtained from Alabama Dept. of Public Safety. Findings are as follows:

Subjects in the cataract group averaged 20/60 and 20/40 in the worst and best eye respectively, compared to the no cataract group who averaged 20/25 and 20/20 respectively. This difference was significant (p<.001).

Contrast sensitivity was significantly worse in both eyes for subjects with cataracts (p<.001). Age adjusted log CS for cataract group was 1.39 (best eye) and 1.19 (worst eye) compared to 1.61 (best eye) and 1.52 (worst eye) for no cataract group.

Cataract subjects detected fewer points in their visual field than the no cataract subjects.

Proportionately more cataract subjects preferred to have someone else drive when they traveled in a car, drove slower than the general traffic flow, and received advice that they limit or stop driving (self-reports on driving habits questionnaire).

Cataract was associated with reduced number of days driving per week and a reduced number of destinations. (Cataract drivers 2 times more likely to reduce driving).

Subjects with cataracts were (2 times) less likely to drive beyond their neighboring towns than subjects without cataracts.

Cataract was significantly associated with driving difficulty in the rain, driving alone, making left turns across traffic, driving on interstates, in high traffic, in rush hour, and at night (Cataract drivers 4 times more likely to report these difficulties).

After adjusting for driving exposure, the association between cataract and at-fault crash involvement was defined as relative risk equal to 2.48, (95% CI = 1.0-6.14).

When adjusted for impaired health, the association between cataract and crash involvement was defined as relative risk = 2.49, (95% CI = 1.0-6.27).

Salzberg and Moffat (1998) evaluated the driving records of 45 older drivers with cataracts who were referred to the Washington State Special Examination Program (and passed), and 449 control group drivers. This program is described in more detail in Section IA1(m) of the Notebook. A "special exam" includes an in-depth interview, and an extended or specialized on-road drive test, typically conducted near the driver's residence. The most common outcome of the "special exam" is to impose driving restrictions (time of day, area, equipment).

Crash and violation records of drivers with cataracts were compared with that of the control group, for a period of 1.75 years before the exam, and 3.25 years after the exam (a 5-year period). Crash and violation rates were calculated to describe the number of incidents per 100 subjects per year, since the pre- and post-observation periods differed in length. The crash and violation rates for the 46 drivers with cataracts who passed the "special exam" and the (entire) control group are presented below, for the pre-exam and post-exam period. For comparison purposes, in Washington State during 1996 there were 140,215 total collisions and 4,037,534 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers in a one-year period.

Washington State Special Exam Program Analysis
Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
Special-Exam Cataracts (n=45) 5.0794 2.0513 15.2381 2.0513

Older drivers with cataracts had a pre-exam crash risk that was 1.33 times that of a control group of older drivers without medical conditions, and 1.46 times higher than the population of licensed drivers in Washington State. After taking and passing a special exam and receiving license restrictions, their risk dropped substantially, to a level below that of the general population, but still higher than that of the older drivers comprising the control group. The authors explain the drop in crash and violation rate shown by the control group as the result of lower driving exposure with increasing age, which is a trend that has been demonstrated in many studies employing older drivers. It is unknown to what degree the cataract group would self-restrict in the absence of the special exam and its formal license restrictions, however, the drop in violation rate for the cataract group as a function of having taken the exam was over twice the reduction shown for the control group. Thus, the special exam program (an on-road test in a driver's home area, plus the tailoring of license restrictions) showed a beneficial effect in reducing crash and violation risk for older drivers with cataracts.

Conclusions/Preliminary Recommendations:

Older drivers with a cataract experience a restriction in their driving mobility and a decrease in their safety on the road. Vision impairment from cataract is now largely reversible due to technological advances in surgical techniques and interocular lens design, with over 85 percent of cases reaching 20/40 acuity or better post-surgery. Cataract surgery is the most common surgical procedure performed on medicare beneficiaries representing 12 percent of the overall Medicare budget.

Owsley et al.'s in-progress study will determine whether improvement in vision following cataract surgery expands driving habits and improves safety. Cataracts are related to increased crash frequency; however, drivers with cataracts are candidates for remediation through eye surgery. Study findings may provide the basis for recommending earlier surgery to remove cataracts. Optometrists and ophthalmologists should counsel patients regarding the dangers associated with driving with cataracts, and suggest driving restrictions (e.g., at night/dusk, in reduced visibility conditions such as rain, fog, etc.) for their cataract patients. The findings from Washington State (Salzberg and Moffat, 1998) indicate that such licensing restrictions reduce the crash and violation risk of older drivers with cataracts to a level that is lower than that posed by the general population of licensed drivers.

References:

Owsley, Stalvey, Wells, and Sloane (1999)

Salzberg and Moffat (1998)


IA1(c). Diabetes and Associated Conditions


Summary:

Hu, Young, and Lu (1993) state that 26 out of 1,000 persons are diagnosed as having diabetes, based on the 1998 National Health Interview Survey, and that the prevalence rate increases with age. Diabetes Mellitus is the most prevalent metabolic disease that may have implications for driving (NHTSA, 1980). Hu et al. (1993) provide the following brief description of the disease. Diabetes Mellitus describes a variety of related medical conditions that affect the body's ability to produce appropriate levels of insulin. Insulin regulates blood sugar levels that provide nutrients to the brain; blood sugar levels that are too high (hyperglycemia) or too low (hypoglycemia) may lead to unconsciousness. Diabetes affects other parts of the body, including the circulatory system and vision. Diabetes in all age groups is associated with thickening of the arteries that can lead to faintness or unconsciousness. The longer a person has diabetes, the more likely that retinal damage (vision impairment) will occur. Approximately 60 percent of patients having diabetes for 15 years or more have some blood vessel damage in their eyes (American Academy of Ophthalmology, 1984). Diabetes Mellitus can be controlled by diet alone, by a combination of diet and oral medication, or by injection of insulin. NHTSA (1980) states that since the level of successfully controlling the disease varies, the following factors should be considered in determining whether a patient should be considered for driver licensing: (1) whether an individual is under regular medical supervision; (2) whether insulin is required; (3) whether the individual is in compliance with the prescribed medical/dietary regimen; (4) whether a warning is experienced before onset of any symptoms; and (5) whether the disease is under control.

A study by Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998) included 294 older drivers, ages 56-90 years at enrollment, drawn from the population of licensed drivers in Jefferson County over age 55. They were divided into three groups as follows: 33 percent had 0 crashes on record; 49 percent had 1 to 3 crashes over the prior 5-year period; and 18 percent had 4 or more crashes over the prior 5-year period. A significant, independent association with crash risk in 3-year follow-up was found for subjects with a diagnosis of diabetic retinopathy (5 times greater risk, 95% CI = 1.13 - 21.8).

Koepsell, Wolf, McCloskey, Buchner, Louie, Wagner, and Thompson (1994) conducted a case-control study of 234 older drivers (age 65+) who were injured in a crash during the previous 3-year period, and 446 older drivers who had no injury crashes during the same period. Injury risk was 2.6 times higher in older diabetic drivers, and higher for those treated with insulin (odds ratio = 5.8), or oral hypoglycemic agents (OR=3.1), or those having diabetes for more than 5 years (OR= 3.9), or those with both diabetes and coronary heart disease (OR=8.0).

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press) analyzed citation rates and crash rates (all crashes and at-fault crashes) for 10,069 drivers reporting diabetes mellitus and other metabolic conditions (including thyroid, parathyroid, pituitary) who had unrestricted licenses, and 358 drivers reporting diabetes and other metabolic conditions with restricted licenses [see Notebook section IA1(m) for further details regarding methodology]. Drivers with multiple medical conditions were excluded from these analyses, which significantly reduced the number of drivers with only diabetes, whose operating privileges were restricted in some way. Their crash and citation rates were compared to a control group of drivers (selected randomly from all licensed drivers without medical conditions), matched on age, gender, and county of residence. Accordingly, different control groups were established for restricted drivers and for unrestricted drivers with this medical condition.

Rates for drivers with diabetes (and other metabolic conditions) and their control groups per 10,000 license days for citations, for all crashes, and for at-fault crashes, are presented in the following table, by license status (not restricted and restricted). Also presented are the relative risk ratios (case rate/control rate).


Utah Rates and Relative Risk Ratios of Adverse Driving Events Per 10,000 Days of Driving
License Status Adverse Driving Event
Not Restricted Citation All Crashes At-Fault Crashes
Drivers with Diabetes 2.61 1.70 1.02
Matched Controls 2.52 1.20 0.64
Rate Ratio 1.04 1.41* 1.58*
Restricted Citation All Crashes At-Fault Crashes
Drivers with Diabetes 4.43 2.03 1.48
Matched Controls 3.16 1.42 0.82
Rate Ratio 1.40 1.43 1.79

* The rate for drivers with diabetes is significantly higher than the rate for their matched controls who have no reported medical conditions.

Drivers with diabetes (both restricted and unrestricted) had a higher risk of adverse driving events than control drivers without a medical condition. Drivers with diabetes whose operating privileges were restricted showed higher rates of adverse driving events than drivers with diabetes licensed without restrictions. This is noteworthy even though their rates were not statistically different than the rates of their control group. This may be the result of the small number of cases with restricted licenses (n=358) and the lower number of days of driving available to this group (54,199), as well as different population characteristics.

Salzberg and Moffat (1998) evaluated the driving records of 14 older drivers with diabetic retinopathy and 27 older drivers with diabetes mellitus who were referred to the Washington State Special Examination Program (and passed), and 449 control group drivers. This program is described in more detail in Section IA1(m) of the Notebook. A "special exam" includes an in-depth interview, and an extended or specialized on-road drive test, typically conducted near the driver's residence. The most common outcome of the "special exam" is to impose driving restrictions (time of day, area, equipment).

Crash and violation records of drivers with diabetic retinopathy and diabetes mellitus were compared with that of the control group, for a period of 1.75 years before the exam, and 3.25 years after the exam (a 5-year period). Crash and violation rates were calculated to describe the number of incidents per 100 subjects per year, since the pre- and post-observation periods differed in length. The crash and violation rates for the drivers with diabetes and related conditions who passed the "special exam" and the (entire) control group are presented below, for the pre-exam and post-exam period. For comparison purposes, in Washington State during 1996 there were 140,215 total collisions and 4,037,534 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers in a one-year period.


Washington State Special Exam Program Analysis
Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
Special Exam Diabetic Retinopathy (n=14) 12.2449 .0000 8.1633 2.1978
Special Exam Diabetes Mellitus (n=27) 6.3492 1.1396 8.4656 2.2792

Older drivers with diabetic retinopathy had a pre-exam crash risk that was 3.2 times that of a control group of older drivers without medical conditions, and 3.5 times higher than the population of licensed drivers in Washington State. The pre-exam crash risk for drivers with diabetes mellitus was 1.67 times higher than the control group of older drivers. After taking and passing a special exam and receiving license restrictions, their risk dropped below that of the control group. The authors explain the drop in crash and violation rate shown by the control group as the result of lower driving exposure with increasing age, which is a trend that has been demonstrated in many studies employing older drivers. Since the drop in crash and violation rates was greater for drivers with diabetes and related conditions than that demonstrated by the control group of older drivers over the 5-year period, it may be concluded that the Special Exam Program (on-road driving exam and license restrictions) was effective in reducing crash risk without eliminating mobility for these drivers. What is not known is the actual driving exposure of these groups of drivers and the severity of disease in the exam group. Thus, the drop in rates for the special exam group could have resulted from being too sick to drive for some period of time during the study.

Finally, in the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that older drivers who reported having diabetes were slightly more likely to be involved in a crash (OR=1.34). For female subjects only (n=163), the odds ratio was 2.13. Subjects ranged in age from 68 to 89 (mean age = 75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997).

Conclusions/Preliminary Recommendations:

Diller et al. (in press) and Salzberg and Moffat (1998) found that drivers licensed with diabetes and other metabolic conditions have a higher rate of crashes than the general population of drivers. In the Owsley et al. study, the association between crash rate and diabetic retinopathy was independent of visual functional problems, since these variables were addressed separately in the modeling. The authors state that this implies that features of eye conditions unrelated to the visual functions assessed in the study (Letter Acuity - ETDRS chart; Contrast Sensitivity - Pelli-Robson chart; Stereoacuity - TNO Test; Disability Glare - MCT-8000 (VisTech); Visual Field Sensitivity) may be associated with crash involvement; factors such as medication usage and other systemic and functional complications. The authors also state that diabetic retinopathy is relatively common in the elderly and is treatable (ophthalmologic laser surgery to seal or photocoagulate the leaking blood vessels or a surgical procedure called a vitrectomy, which is the removal of the blood-filled vitreous from the eye and replacement with a clear artificial solution). If elevated crash rate is independent of visual function, diabetes (not diabetic retinopathy) may actually be responsible for the elevation in crash rate. Physicians and ophthalmologists should counsel their patients with diabetes regarding the importance of complying with treatment recommendations (diet and medications) for maintaining safe driving, and recommend driving restrictions/cessation on an individual basis, depending on the extent and severity of the symptoms.

Regarding the effectiveness of restricting the licenses of drivers with diabetes, results are mixed. This is because actual exposure data have not been available. Diller et al. (in press) attempted to control for the effects of exposure, but only used available days (as opposed to actual miles driven). The reduction in crash and violation rates shown in the Salzberg and Moffat (1998) study are noteworthy; however, caution needs to be taken in generalizing the results. Older drivers with medical conditions may either choose to restrict their driving because they know that they are at an increased crash risk, or they may not feel well enough to drive as often as healthy older drivers. Lower exposure leads to a lower crash risk. Also, the sample size of drivers with diabetes in this study was small. But the crash and violation rate reductions reported above indicate that restricting the driving privileges has promise in improving safety while maintaining mobility.

References:

American Academy of Ophthalmology (1984)

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press)

Hu, Young, and Lu (1993)

Koepsell, Wolf, McCloskey, Buchner, Louie, Wagner, and Thompson (1994)

National Center for Health Statistics(1989)

NHTSA (1980)

Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998)

Salzberg and Moffat (1998)


IA1(d). Glaucoma


Summary:

Glaucoma is one of the leading causes of blindness in the U.S., affecting 2 out of every 100 persons over age 35 (American Academy of Ophthalmology, 1983).

A study by Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998) included 294 older drivers, ages 56-90 years at enrollment, drawn from the population of licensed drivers in Jefferson County over age 55. They were divided into three groups as follows:

33% had 0 crashes on record.

49% had 1 to 3 crashes over the prior 5-year period.

18% had 4 or more crashes over the prior 5-year period.

A significant, independent association with crash risk in a 3-year follow-up was found for subjects with a diagnosis of glaucoma: (Relative Risk =5.20, 95% Confidence Interval = 1.19-22.72). The relationship for glaucoma and crashes was stronger for males (RR=9.81) than for females (RR=5.14).

The association between crash rate and glaucoma was independent of visual functional problems, since these variables were addressed separately in the modeling. The authors state that this implies that features of eye conditions unrelated to the visual functions assessed in the study may be associated with crash involvement (such as medication usage and other systemic and functional complications).

In another study of 193 older drivers between age 55-87 (mean = 71 years), identified through Alabama Department of Public Safety Files, 78 drivers (cases) had at least 1 crash in the prior 5-year period that resulted in an injury to anyone in the involved vehicles, and 115 drivers (controls) had no crashes in the same 5-year period. Glaucoma was independently associated with crash risk in the multivariate analyses: cases were 3.6 times more likely to report glaucoma than were controls (Owsley, McGwin, and Ball, 1998).

In a panel data analysis of remaining eligible drivers in 1993 (507 female drivers and 375 male drivers) who participated in the Iowa 65+ Rural Health Study from 1981-1993, none of the commonly studied medical conditions (e.g., diabetes, heart disease, stroke, Parkinson's Disease) were associated with crashes. The only medical condition that increased crash risk in older drivers was glaucoma. And, the association between glaucoma and highway crashes was evident only among older male drivers (odds ratio = 1.7) (Hu, Trumble, Foley, Eberhard, and Wallace, 1998).

Stewart, Moore, Marks, May, and Hale (1993) found no association between glaucoma and increased crash risk, in a sample of 1,431 older drivers. Both independent and dependent variables, however, were comprised of self-reports (of medical conditions and crashes, respectively).

Conclusions/Preliminary Recommendations:

Glaucoma is relatively common in the elderly and is associated with an increased crash risk. In multiple studies, the risk of an older driver being involved in a crash is 1.7 to 5.2 times higher if glaucoma is present. Two studies showed that the risk appears to be higher for males than for females. The American Optometric Association (AOA) recommends that people ages 10 to 40 see an optometrist every 2 to 3 years; people ages 41-60 every two years; and people age 61+ every year. Individuals age 61+ have an increasing risk for the development of cataracts, glaucoma, and macular degeneration and other sight threatening or visually disabling eye conditions as well as systematic health conditions. The American Academy of Ophthalmology recommends that persons over age 35 be checked for glaucoma every 2 or 3 years. Glaucoma is treatable (eye drops, pills to decrease pressure either by assisting outflow of fluid from the eye or by decreasing the amount of fluid entering the eye, or surgery to perform a new drainage canal).

References:

American Academy of Ophthalmology (1983): Glaucoma

Hu, Trumble, Foley, Eberhard, and Wallace (1998)

Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998)

Owsley, McGwin, and Ball (1998)

Stewart, Moore, Marks, May, and Hale (1993)

 

IA1(e). Foot Abnormalities


Summary:

Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994) studied 283 community-dwelling individuals age 72 to 92 (mean age = 77.8) from the Project Safety cohort living in New Haven, CT who drove between 1990 and 1991. Fifty-seven percent of the sample were males.

The number of the following foot abnormalities was noted in addition to the ability to stand on toes and heels: toenail irregularities, calluses, bunions, and toe deformities such as hammer toes. Analyses were conducted contrasting driving outcomes for patients with 0 to 2 foot abnormalities versus 3 to 8 foot abnormalities.

The outcome variable was self-reported involvement in automobile crashes, moving violations, or being stopped by police in the year following administration of the test battery.

Persons with 3 or more foot abnormalities were more likely to have adverse events (23 percent had adverse events) compared to persons with 0-2 foot abnormalities (10 percent had adverse events). The difference was significant at p<0.01 level (relative risk = 2.0, 95% CI=1.0 to 3.8).

A multivariate analysis adjusting for driving frequency and housing type found the following factors to be associated with the occurrence of adverse events: poor design copying on the MMSE (relative risk=2.3, 95% CI=1.5 to 5.0), fewer blocks walked--0 versus > 1 (relative risk=2.3, 95% CI=1.3 to 4.0) and more foot abnormalities--3 to 8 versus 0 to 2 (relative risk=1.9, 95% CI=1.1 to 3.3).

Combining these 3 factors to assess their ability to predict adverse driving events showed that if no factors were present, 6 percent of drivers had adverse events; if 1 factor was present, 12 percent had events; if 2 factors were present, 26 percent had events; and if all 3 factors were present, 47 percent had events.


Conclusions/Preliminary Recommendations:

There is a significant relationship between foot abnormalities in the elderly and increased crash risk. The association between foot abnormalities and crashes is logical, because such abnormalities may affect the ability to maneuver between the brake and accelerator. Physicians should take notice of foot abnormalities in older patients and include driving history-taking and counseling as part of routine exams.

References:

Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994)


IA1(f). Falls


Summary:

Sims, Owsley, Allman, Ball, and Smoot (1998) conducted a study to explore associations between a history of at-fault vehicle crashing in older subjects (between 1985-1991) and several medical and functional variables collected on them in 1991. Seven questionnaires and 10 physical examination/ performance measures were employed to assess medical and functional domains. Lists of drivers and number of crashes for each driver were made available by the AL Dept. of Public Safety.

Subjects included 174 drivers ages 55-90 (mean age 71.1), residing in Jefferson County, AL. Case drivers has at least 1 state-recorded at-fault crash in the 6 years preceding the assessment (n=99).

Controls had no state-recorded at-fault crashes in the prior 6 years (n=75).

At the univariate level, crash-involvement was significantly associated with falling in the prior two years (p=0.004). All non-collinear variables that were significant at the univariate level were entered into logistic regression models; these included falling, reduction of 40 percent or more in the useful field of view, and not taking a beta-blocking drug. The logistic regression model indicated that having fallen in the prior two years was related to crash involvement with an odds ratio of 2.6 (CI=1.1-6.1, p=0.025).

Note: In another study by Owsley, McGwin, and Ball (1998), subjects with crashes were 3.6 times more likely to report a diagnosis of glaucoma compared to controls. These authors cited Glynn et al. (1991). Although medication information was not collected, Glynn et al. (1991) reported that the use of topical eye medications in elderly patients with glaucoma increased their risk of falling (an adverse mobility outcome).

Koepsell, Wolf, McCloskey, Buchner, Louie, Wagner, and Thompson (1994) conducted a case-control study of 234 older drivers (age 65+) who were injured in a crash during the previous 3-year period, and 446 older drivers who had no injury crashes during the same period. Injury risk was 1.4 times higher in older drivers who had fallen in the previous year. The authors caution that this association could have arisen by chance.

In the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that self-reported falls in the past two years was related to crashing (Odds Ratio for all subjects=1.53; OR for females only=1.38; OR for males only=1.61). Subjects ranged in age from 68 to 89 (mean age=75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997).

Conclusions/Preliminary Recommendations:

Crash involvement in the elderly is significantly related to having fallen in the past two years. Professionals conducting geriatric assessments should include a question about falling as part of history-taking, and DMVs should include a question about falling on license renewal applications for tracking of associations between falling and automobile crashes.

References:

Glynn, Seddon, Krug, Sahagian, Chiavelli, and Campion (1991)

Koepsell, Wolf, McCloskey, Buchner, Louie, Wagner, and Thompson (1994)

Owsley, McGwin, and Ball (1998)

Sims, Owsley, Allman, Ball, and Smoot (1998)

 

IA1(g). Cardiac (and Cardiopulmonary) Condition


Summary:

Stewart, Moore, Marks, May, and Hale (1993) studied 1,431 participants in the Florida Geriatric Research Program (Dunedin, FL), for whom 8 years of longitudinal data were available (1975-1987). Subjects included 874 females (mean age = 77.8 years, s.d. = 4.6) and 596 males (mean age = 78.6 years, s.d.=4.5).

The dependent variable was self-reported crashes. Independent variables included self-reported information on 31 diseases, 26 symptoms, 34 clinical and laboratory values, number of drugs reported, number of symptoms reported, number of diseases reported. Subjects completed a questionnaire containing 180 questions and a form listing prescribed and nonprescribed medications used on a regular basis. Biochemical profile includes hemogram, red cell indices, and SMAC-23. Clinical assessment includes electrocardiogram and carotid auscultation, plus MMSE and Beck Depression Inventory at 8th visit.

The correlation between irregular heartbeat (palpitations) and crashes is significant (p=0.0017, Odds ratio = 1.83, 95% CI = 1.25-2.68). No other cardiovascular symptoms or diseases investigated in the present study were predictive of crashes.

No other signs or symptoms were of significance in crashes (paroxysmal nocturnal dyspnea, temporary loss of limb, dizziness/spinning, lightheadedness, syncope, tinnitus, dysphagia, amaurosis fugax, pain in abdomen, swollen feet/ankles, headache, paresthesia, diarrhea, recurrent cough, hematuria, incontinence (urine), aphasia, dysphonia, dyspnea, orthopnea, nocturia, claudication, dysuria, memory loss, feel awkward, effort angina, angina with tension, hemoptysis, constipation, thin bowel movements, blood in stools, melena, swollen joints, ache/painful joints, urinary hesitancy, and carotid bruits).

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press) analyzed citation rates and crash rates (all crashes and at-fault crashes) for 18,990 drivers with cardiovascular conditions (including heart disease, rhythm disturbances, or history of myocardial infarctions, heart surgery, or hypertension) who had unrestricted licenses, and 160 drivers with cardiovascular conditions with restricted licenses [see Notebook section IA1(m) for further details regarding methodology]. Drivers with multiple medical conditions were excluded from these analyses, which significantly reduced the number of drivers with only cardiovascular conditions, whose operating privileges were restricted in some way. Their crash and citation rates were compared to a control group of drivers (selected randomly from all licensed drivers without medical conditions), matched on age, gender, and county of residence. Accordingly, different control groups were established for restricted drivers and for unrestricted drivers with this medical condition.

Rates for drivers with cardiovascular conditions and their control groups per 10,000 license days for citations, for all crashes, and for at-fault crashes, are presented in the following table, by license status (not restricted and restricted). Also presented are the relative risk ratios (case rate/control rate).

The data indicate that unrestricted drivers with cardiovascular disease have significantly higher crash rates (all crashes and at-fault crashes) than their matched controls without a medical condition. Drivers with cardiovascular disease whose driving privileges are restricted, also have a higher rate of adverse events than their matched control group, although the differences are not statistically significant. The higher rate may be explained by the small sample size (n= 160) and resulting number of eligible licensed driving days (22,290). In the time period under analysis (1992-1996), these restricted drivers experienced only 7 citations, 3 crashes, and 2 at-fault crashes.


Utah Rates and Relative Risk Ratios of Adverse Driving Events Per 10,000 Days of Driving
License Status Adverse Driving Event
Not Restricted Citation All Crashes At-Fault Crashes
Drivers with Cardiovascular Conditions 1.23 1.04 0.55
Matched Controls 1.60 0.91 0.47
Rate Ratio 0.77** 1.14* 1.15*
Restricted Citation All Crashes At-Fault Crashes
Drivers with Cardiovascular Conditions 3.14 1.35 0.90
Matched Controls 2.0 0.83 0.52
Rate Ratio 1.57 1.61 1.72

* The rate for drivers with cardiovascular conditions is significantly higher than the rate for their matched controls who have no reported medical conditions.

** Differences in rates between medical conditions and control groups are statistically significant, with higher rates for control group.

Salzberg and Moffat (1998) evaluated the driving records of 47 older drivers with cardiovascular conditions who were referred to the Washington State Special Examination Program (and passed), and 449 control group drivers. This program is described in more detail in Section IA1(m) of the Notebook. A "special exam" includes an in-depth interview, and an extended or specialized on-road drive test, typically conducted near the driver's residence. The most common outcome of the "special exam" is to impose driving restrictions (time of day, area, equipment).

Crash and violation records of drivers with cardiovascular conditions were compared with that of the control group, for a period of 1.75 years before the exam, and 3.25 years after the exam (a 5-year period). Crash and violation rates were calculated to describe the number of incidents per 100 subjects per year, since the pre- and post-observation periods differed in length. The crash and violation rates for the drivers with cardiovascular conditions who passed the "special exam" and the (entire) control group are presented below, for the pre-exam and post-exam period. For comparison purposes, in Washington State during 1996 there were 140,215 total collisions and 4,037,534 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers in a one-year period.


Washington State Special Exam Program Analysis
Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
Special Exam Cardiovascular Conditions (n=47) 7.2948 1.9640 20.6687 2.6187

Older drivers with cardiovascular conditions had a crash rate almost twice as high as that of the control group of older drivers prior to taking the special exam and receiving driving restrictions, and a violation rate over 2.5 times higher than control group drivers, during the pre-exam period. After undergoing the special exam process, their crash and violation rates fell significantly, to almost the level of that shown by the control group, which is less than the crash rate of the population of licensed drivers in the State of Washington. Thus, it appears that appropriate license restrictions (e.g., driving only within a specific radius of residence, daylight driving only, driving only between the hours of 10 a.m. to 3 p.m., no freeway driving, and/or driving within city limits only) are effective in reducing the risk posed by older drivers, without unduly restricting their mobility.

In Janke's (1994) review of cardiovascular conditions and driving, it is concluded that increased societal risk due to the driving of patients (in personal vehicles) with cardiovascular disease has not been shown. There is evidence that cardiac patients cut down on their mileage considerably and reduce long-distance driving, driving in bad weather, driving alone, driving after dark, and driving in heavy traffic (Waller, 1981, 1987; Potvin, Guibert, Philibert, and Loiselle, 1990, Potvin, Guibert, and Loiselle, 1993: in Janke, 1994). Potvin, Guibert and Loiselle, 1993 (in Janke, 1994) note methodological problems in the studies they review, including low occurrence of crashes, difficulty in defining a suitable comparison group, classification difficulties (e.g., healthy controls may develop a cardiovascular condition in the course of the study, unknown to the experimenter), and uncontrolled variations in exposure to crash risk.

Diller et al. (in press) also analyzed citation rates and crash rates (all crashes and at-fault crashes) for 2,615 drivers with pulmonary conditions (including pulmonary disease or symptoms, impaired function, or severe respiratory difficulties) who had unrestricted licenses, and 244 drivers with pulmonary conditions and whose licenses were restricted. Drivers with multiple medical conditions were excluded from these analyses, which significantly reduced the number of drivers with only pulmonary conditions, whose operating privileges were restricted in some way. Their crash and citation rates were compared to a control group of drivers (selected randomly from all licensed drivers without medical conditions), matched on age, gender, and county of residence. As mentioned earlier, different control groups were established for restricted drivers and for unrestricted drivers with this medical condition.

Rates for drivers with pulmonary conditions and their control groups per 10,000 license days for citations, for all crashes, and for at-fault crashes, are presented in the following table, by license status (not restricted and restricted). Also presented are the relative risk ratios (case rate/control rate).

The relative risk ratios for all events between drivers with pulmonary conditions who drove with unrestricted licenses and their matched controls are significantly different (at alpha = 0.05). This finding suggests that drivers who have pulmonary conditions and unrestricted driving privileges have a higher risk of crash events than drivers in the general population who do not report medical conditions. For citations, pulmonary conditions appear to have a protective effect, possibly due to self-restriction (a factor which was not taken into account in the data collection), or to other differing population characteristics. However, the differences between drivers with medical conditions who were restricted in their driving privileges and their corresponding control groups were not significantly different.

Utah Rates and Relative Risk Ratios of Adverse Driving Events Per 10,000 Days of Driving
License Status Adverse Driving Event
Not Restricted Citation All Crashes At-Fault Crashes
Drivers with Pulmonary Conditions 2.24 1.52 0.85
Matched Controls 2.54 1.22 0.63
Rate Ratio 0.88** 1.25* 1.35*
Restricted Citation All Crashes At-Fault Crashes
Drivers with Pulmonary Conditions 0.69 1.04 1.04
Matched Controls 1.39 1.11 0.64
Rate Ratio 0.50 0.93 1.63

* The rate for drivers with pulmonary conditions is significantly higher than the rate for their matched controls who have no reported medical conditions.

** The rate for drivers with pulmonary conditions is significantly lower than the rate for their matched controls who have no reported medical conditions.

Conclusions and Preliminary Recommendations:

The correlation between irregular heartbeat and crashes in the elderly was significant in a study that used self-reporting both of crash occurrence and of medical conditions. Another study found that drivers with cardiovascular conditions and drivers with pulmonary conditions who drive without restrictions on their licenses have a significantly higher citation and crash risk than drivers without these medical conditions. Restricted drivers in Utah have either a 3-month interval for review (cardiovascular conditions) or a 6-month interval for review (pulmonary conditions), and generally have the following restrictions placed on their driving privileges: speed limitations (profile level 6); speed and area limitations (level 7); speed, area, and time of day (level 8); and speed, area, time of day, and must be accompanied by licensed passenger (levels 9-10). Thus, it appears that restricting the driving privileges of persons with cardiovascular conditions and those with pulmonary conditions reduces citation and crash risk to the level of risk posed by the general population without these medical conditions. One limitation to the methodology in the Diller et al. study was that no actual measure of exposure was collected; therefore, it is unknown to what degree the restricted drivers (whose impairments were more severe than unrestricted drivers) reduced their own risk by lowering their exposure.

Larsen et al, 1994 (in Janke, 1994) recommended that doctors should advise their arrhythmia patients not to drive for 7 months after discharge from the hospital.

References:

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press)

Janke (1994)

Stewart, Moore, Marks, May, and Hale (1993)

 

IA1(h). Feet or Legs Cold on Exposure to Cold


Summary:

Stewart, Moore, Marks, May, and Hale (1993) studied 1,431 participants in the Florida Geriatric Research Program (Dunedin, FL), for whom 8 years of longitudinal data were available (1975-1987). Subjects included 874 females (mean age = 77.8 years, s.d. = 4.6) and 596 males (mean age = 78.6 years, s.d.=4.5).

The dependent variable was self-reported crashes. Independent variables included self-reported information on 31 diseases, 26 symptoms, 34 clinical and laboratory values, number of drugs reported, number of symptoms reported, number of diseases reported.

Subjects completed a questionnaire containing 180 questions and a form listing prescribed and nonprescribed medications used on a regular basis. Biochemical profile includes hemogram, red cell indices, and SMAC-23. Clinical assessment includes electrocardiogram and carotid auscultation, plus MMSE and Beck Depression Inventory at 8th visit.

The correlation between feet or legs cold upon exposure to cold and traffic crashes is significant (p=.0074, odds ratio = 1.82, 95% confidence interval = 1.17 - 2.82).

Conclusions/Preliminary Recommendations:

One study has found that older drivers who indicate that their feet/legs feel cold upon exposure to cold are at increased crash risk. Professionals conducting geriatric assessments should include a question about these symptoms as part of history-taking, and important data may be obtained if DMVs included a similar question on license renewal applications for tracking of associations between feet/legs becoming cold upon exposure to cold and automobile crashes.

References:

Stewart, Moore, Marks, May, and Hale (1993)

 

IA1(i). Bursitis


Summary:

Stewart, Moore, Marks, May, and Hale (1993) studied 1,431 participants in the Florida Geriatric Research Program (Dunedin, FL), for whom 8 years of longitudinal data were available (1975-1987). Subjects included 874 females (mean age = 77.8 years, s.d. = 4.6) and 596 males (mean age = 78.6 years, s.d.=4.5).

The dependent variable was self-reported crashes. Independent variables included self-reported information on 31 diseases, 26 symptoms, 34 clinical and laboratory values, number of drugs reported, number of symptoms reported, number of diseases reported. Subjects completed a questionnaire containing 180 questions and a form listing prescribed and nonprescribed medications used on a regular basis. Biochemical profile includes hemogram, red cell indices, and SMAC-23. Clinical assessment includes electrocardiogram and carotid auscultation, plus MMSE and Beck Depression Inventory at 8th visit.

Bursitis is an inflammation of a bursa, especially of the shoulder or elbow. Bursae are closed synovial spaces located at the site of friction between skin, ligaments, tendons, muscles and bones; the most common site of bursitis is in the shoulder. Bursitis may cause severe pain and limitation of mobility. The correlation between bursitis and traffic crashes was significant (p=.0005, odds ratio = 2.18, 95% confidence interval = 1.41 - 3.38).

In the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that self-reported bursitis was related to crashing for females only (Odds Ratio = 1.57). Subjects ranged in age from 68 to 89 (mean age = 75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997). Only 13 of the 146 females who responded to this health question reported having bursitis.

Conclusions/Preliminary Recommendations:

Older drivers with bursitis are at increased crash risk. Professionals conducting geriatric assessments should include a question about bursitis as part of history-taking (and note its presence during the assessment), and DMVs should include a similar question on license renewal applications for tracking of associations between bursitis and automobile crashes.

References:

Stewart, Moore, Marks, May, and Hale (1993)

 

IA1(j). Renal Disease
Protein in Urine


Summary:

Stewart, Moore, Marks, May, and Hale (1993) studied 1,431 participants in the Florida Geriatric Research Program (Dunedin, FL), for whom 8 years of longitudinal data were available (1975-1987). Subjects included 874 females (mean age = 77.8 years, s.d. = 4.6) and 596 males (mean age = 78.6 years, s.d.=4.5).

The dependent variable was self-reported crashes. Independent variables included self-reported information on 31 diseases, 26 symptoms, 34 clinical and laboratory values, number of drugs reported, number of symptoms reported, number of diseases reported.

Subjects completed a questionnaire containing 180 questions and a form listing prescribed and nonprescribed medications used on a regular basis. Biochemical profile includes hemogram, red cell indices, and SMAC-23. Clinical assessment includes electrocardiogram and carotid auscultation, plus MMSE and Beck Depression Inventory at 8th visit.

The correlation between protein in the urine and traffic crashes was significant (p=.0021, odds ratio = 1.84, 95% confidence interval = 1.25 - 2.72).

Conclusions/Preliminary Recommendations:

Increased urinary excretion of protein is a common sign of renal disease, and is significantly related to older driver crashes. Urinalysis should be a part of a physical examination for older persons.

References:

Stewart, Moore, Marks, May, and Hale (1993)


IA1(k). Seizure Disorders


Summary:

Hu, Young, and Lu (1983) state that epilepsy may cause sudden loss of consciousness, muscular convulsions or spasms, or it may only cause a slight temporary change in a person's conscious awareness. They report that although the actual number of Americans who have epilepsy is unknown, the National Center for Health Statistics (NCHS, 1989) estimated a rate of 3.8 in every 1,000 persons.

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press) analyzed citation rates and crash rates (all crashes and at-fault crashes) for 2,620 drivers with epilepsy who had unrestricted licenses and 775 drivers with epilepsy with restricted licenses [see Notebook section IA1(m) for further details regarding methodology]. These groups were not mutually exclusive, as during the study period, a number of drivers may have fluctuated between restricted and nonrestricted licensing privileges. Their crash and citation rates were compared to a control group of drivers (selected randomly from all licensed drivers without medical conditions), matched on age, gender, and county of residence. Accordingly, different control groups were established for restricted drivers and for unrestricted drivers with this medical condition.

This category of medical condition (epilepsy and other episodic conditions) is defined as follows in Utah's Guidelines and Standards for Health Care Professionals), included as Appendix A in Diller et al.: "Epilepsy includes any recurrent loss of consciousness or conscious control arising from intermittent changes in brain function. Because of the similarity of consequences, other disorders affecting consciousness or control such as syncope, cataplexy, narcolepsy, hypoglycemia, episodic vertigo interfering with function, etc., have been included in this section, to be considered in a similar fashion."

Rates for drivers with epilepsy and their control groups per 10,000 license days for citations, for all crashes, and for at-fault crashes, are presented in the following table, by license status (not restricted and restricted). Also presented are the relative risk ratios (case rate/control rate).


Utah Rates and Relative Risk Ratios of Adverse Driving Events Per 10,000 Days of Driving
License Status Adverse Driving Event
Not Restricted Citation All Crashes At-Fault Crashes
Drivers with Epilepsy 4.06 2.69 1.76
Matched Controls 3.96 1.49 0.84
Rate Ratio 1.03 1.81* 2.11*
Restricted Citation All Crashes At-Fault Crashes
Drivers with Epilepsy 4.13 2.67 2.40
Matched Controls 3.94 1.73 0.97
Rate Ratio 1.05 1.55* 2.47*

* The rate for drivers with epilepsy is significantly higher than the rate for their matched controls who have no reported medical conditions.

Drivers with epilepsy (both those with restrictions and those without restrictions) have a higher risk of crashing than their matched control groups.

Conclusions/Preliminary Recommendations:

The data analysis conducted by Diller et al. (in press) indicates that licensed drivers with epilepsy/episodic conditions (both those who have restricted operating privileges and those without license restrictions) are at a significantly higher risk of a crash than the general population of drivers.

The American Academy of Neurology, American Epilepsy Society, and Epilepsy Foundation of America (1994) have drafted consensus statements on driver licensing and epilepsy, based on a Consensus Workshop held in 1991. These groups agree that a seizure-free interval should be stated, and that 3 months is preferred, starting from the date of the seizure. Both favorable and unfavorable modifiers could alter the interval. The groups also agree that "restricted licenses may be appropriate under certain circumstances in which such restrictions will allow driving with an acceptable risk of seizure occurrence." They further state that physician and/or medical advisory board input should be obtained for individualized determination of the terms of each restricted license. There is unanimous agreement among the groups that physicians should not be required to report their patients to the DMV; they should, however, advise patients about the medical risks involved, about DMV requirements, about self-reporting obligations, and should tell the patient the physician's own recommendation about driving. The patient should be responsible to self-report the condition initially to the DMV and to report recurrent seizures. However, the group stated that if the physician believes the patient has not self-reported and is endangering the public by driving, the physician should have the right to report the patient, with immunity. The participants of the Consensus Workshop determined that medical criteria for licensing are best handled in the form of medical guidelines or regulations. Sample statutory language is provided in the document; many are based on Wisconsin Statutes.

References:

American Academy of Neurology, American Epilepsy Society, and Epilepsy Foundation of America (1994)

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press)

Hu, Young, and Lu (1993)

National Center for Health Statistics (1989)

IA1(l). Back Pain


Summary:

Hu, Trumble, Foley, Eberhard, and Wallace (1998) conducted a panel data analysis of the remaining eligible drivers in 1993 (507 female drivers and 375 male drivers) who participated in the Iowa 65+ Rural Health Study from 1981-1993. The study included all noninstitutionalized individuals in two counties age 65+. The resulting sample was 6,553 female person-years and 5,414 male person-years. The survey data were obtained from in-home and telephone interviews, and included demographic attributes, onset of medical conditions, symptoms and ailments, functional status, physical functioning, physical activities, vision, drug use, cognitive abilities, and annual miles driven. The survey data were linked to crash files maintained by the Iowa DMV. The association between crash risk and persistent back pain was significant for combined gender (6,553 female person-years and 5,414 male person-years). The risk ratios (RR) are as follows, for the specified mileage levels: RR=1.25 for 3,000; 6,000, and 12,000 miles driven annually; RR = 1.54 for 9,000 and 18,000 annual miles.

Foley, Wallace, and Eberhard (1995) interviewed 1,791 of the Rural Health Study participants in 1989. Between the period of 1985 and 1989, 206 drivers were involved in 245 state-recorded crashes. They found that a large proportion of drivers with existing back pain or an episode of back pain in the previous year (42%) had a significantly increased risk of crashing. Interestingly, none of the other disease histories obtained were related to crashing (heart disease, cancer, stroke, hypertension, diabetes, asthma, arthritis, osteoporosis, and emphysema). The crash involvement rate (number of drivers involved in crashes per 1,000 estimated person-years of driving) for older drivers with back pain in the past 12 months was 33. The relative risk was 1.5, with a confidence interval ranging from 1.2 to 2.0.

Conclusions/Preliminary Recommendations:

Foley et al. (1995) stated that the association of back pain with crash risk corroborates concern over the impact of musculoskeletal dysfunction on driving. Reasons for this association may include decreased motor function in driving tasks because of pain or underlying neurologic deficit in the lower extremities, as well as a dysfunction resulting from more generalized arthritic conditions.

Since the presence of self-reported arthritis did not correlate with crashes it seems reasonable to conclude that license renewal forms should specifically cite symptoms, as opposed to diagnoses alone, to query drivers about health risks that may be related to crashes.

References:

Hu, Trumble, Foley, Eberhard, and Wallace (1998)

Foley, Wallace, and Eberhard (1995)


IA1(m). Overview: Comparative Risk Table


Summary:

Tables permitting comparison of the risk associated with each of the conditions addressed in this section are presented below. First, the results of a recent and ongoing analysis of Utah's medical conditions database are presented, summarized in the form of relative risk values for all of the included conditions compared to matched control groups of drivers (Diller, Cook, Leonard, Reading, Dean, and Vernon, in press). These values are presented on page 36; a synopsis of the methodology used to derive these values is also given. Next, data from Washington State are presented in a table on page 37, along with a description of the study methodology (Salzberg and Moffat, 1998). Immediately following the relative risk table for the Utah and Washington data is another table labeled "Risk Ratios for Identified Medical Conditions." This table, shown on page 38, extracts risk ratios and odds ratios for crash involvement for various conditions as they could be extracted from the studies cited in each area, i.e., Notebook sections IA1(a) - IA1(l).

In Utah, driver license applicants must complete a general questionnaire designed to identify medical conditions related to physical, mental, and emotional health. Applicants who report a medical condition are placed into at least one of 12 functional ability categories (diabetes mellitus and other metabolic conditions; cardiovascular; pulmonary; neurologic; epilepsy and other episodic conditions; learning/memory/communications; psychiatric or emotional conditions; alcohol and other drugs; visual acuity; musculoskeletal abnormalities/chronic medical debilities; functional motor ability; and hearing) and further by functional ability level (1-12) within the functional category. Passenger vehicle drivers in functional ability profile levels 1-5 may drive without restrictions (speed, area, time of day, licensed passenger). Although severity of impairments increase with increases in assigned functional profile level, drivers in levels 4 and 5 are deemed safe to drive without license restriction, but may be required for reexam/medical review at intervals shorter than the standard renewal period, depending on their functional (medical) category. Drivers assigned to functional ability profile level 6 have a speed restriction placed on their licenses. A profile level of 7 indicates that the driving risk posed by the functional impairment justifies a speed and area limitation. A profile level of 8 indicates a speed, area, and time of day limitation. Drivers in profile level 9 must be accompanied by a licensed driver, and may have speed, area, and/or time of day limitations as recommended by their health care professional. Levels 10 and 11 are associated with special driving limitations recommended by health care providers or the Director of Licensing. A person assigned to level 12 may not drive until ability improves and functional ability can be assigned at a lower level.

Utah's Guidelines and Standards for Health Care Professionals (provided as Appendix A to Diller et al., in press) contains descriptions of basic concepts, definitions, and ground rules for each functional ability category. A brief description of conditions, symptoms, impairments, etc., that are subsumed under each category is presented next.

Diabetes Mellitus and Other Metabolic Conditions: Disturbances in the function of the endocrine glands cause many symptoms from generalized asthenia, muscle weakness, and spasm or tetany to sudden episodes of dizziness or unconsciousness. This category includes diabetes mellitus, parathyroid disorders, thyroid disorders, and hypoglycemia.

Cardiovascular: Cardiovascular disease may affect a driver's ability in a variety of ways, and therefore profile guidelines and standards are provided for four of the most common circumstances: general heart disease; rhythm; after myocardial infarction or cardiac surgery; and hypertension. The 12 profile levels are determined by the history and severity of these four circumstances. General heart disease, for example, is divided into four classes based on the functional classification of the American Heart Association, with Class I containing patients with heart disease but with no limitations of physical ability (ordinary physical activity causes no undue dyspnea, anginal pain, fatigue, or palpitation) and Class IV containing patients with inability to carry on any physical activity without discomfort (symptoms of cardiac insufficiency or of the anginal syndrome may be present, even at rest, and are intensified by activity).

Pulmonary: Although impaired pulmonary function is seldom the cause of sudden death, it may seriously affect operators of vehicles in the following ways: (1) sudden severe coughing while driving may result in a crash; (2) cough syncope may occur while driving; (3) impaired cerebral oxygenation caused by impaired pulmonary function may result in mental confusion and/or impaired judgment. In assessing the severity of pulmonary impairment, effort is made to limit the tests to those found in most medical offices, although occasionally, more sophisticated studies may be needed (e.g., arterial blood gases, maximal voluntary ventilation, etc.). The basic function tests (FVC and FEV) are the principal guidelines and standards currently recommended.

Neurologic: A wide variety of neurologic conditions may affect driving safety, that includes (but is not limited to) strokes; head injuries; Cerebral Palsy; Multiple Sclerosis; Parkinson's Disease; progressive conditions such as muscular atrophies and dystrophies; myasthenia gravis; and other spinal cord and brain diseases. The common element in all of these is the disturbance of sensory, motor, or coordinating functions sufficient to effect driving.

Epilepsy and Other Episodic Conditions: Epilepsy includes any recurrent loss of consciousness or conscious control arising from intermittent changes in brain function. Because of the similarity of consequences, other disorders affecting consciousness or control such as syncope, cataplexy, narcolepsy, hypoglycemia, episodic vertigo interfering with function, etc., have been included in this section, to be considered in a similar fashion.

Learning, Memory, and Communication: This broad category includes retardation; learning problems related to general intelligence; impairments relating to the recovery of head injuries; closed head injuries (resulting in diffuse cognitive deficits such as impaired judgment, impulsiveness, distractibility, impaired attention, neglect, slowed reaction time, or impaired cognitive endurance); Alzheimer's Disease; aphasia, and inadequate language skills.

Psychiatric or Emotional Conditions: Psychiatric history and medications determine the functional levels under this category. There are a variety of behavioral conditions, extremes of mood, and impairments in thinking associated with psychiatric disorders which may correlate with accident proneness or driver risk. These include: inattentiveness which may accompany even minor disturbances; impulsivity, explosive anger, and impaired social judgment characteristic of personality disorders, especially antisocial personality; and suicidality, perceptual distortions, psychomotor retardation or frank irrationality in addition to the previously described symptoms which are common features of major psychiatric illnesses such as schizophrenia, major depressive disorder, bipolar (manic depressive) disorder, and organic brain syndromes.


Alcohol and Other Drugs: This category includes chronic use of alcohol; use of mood altering and hallucinogenic drugs (amphetamines, LSD, antihistamines, barbiturates, benzodiazepines, and anti-psychotics such as phenothiazine, haloperidol, and sleeping pills of all types); marijuana; and excessive or inappropriate use of drugs for the purpose of intoxication or stimulation (including prescription, nonprescription, legal, and illegal drugs). Users of alcohol and other drugs are well known for their tendency to under-report amounts used, and there is wide individual variation in the effects of such substances; therefore, the only valid basis for evaluating a person's probable safety as a driver is careful appraisal of the person's history including, but not limited to, the past effect on driving.

Visual Acuity: Guidelines for placing drivers in functional ability categories are based on acuity and visual fields. Correction must be less than 10 diopters to qualify for profile level 1 (20/25 vision in each eye; monocular visual fields 120 in each eye; binocular visual fields 70 to the right and to the left in the horizontal meridian). Other eye conditions that require special consideration, but which have no set standards, include: color vision; dark adaptation; heterophoria; stereopsis; monocular vision; refractive states; telescopic lenses; and chronic and recurrent disease.

Musculoskeletal Abnormality or Chronic Medical Debility: Includes chronic conditions not listed elsewhere, including osteoporosis, HIV, amputations, congenital abnormalities (unless compensatory devices are used as outlined in the Functional Motor Ability Category), that according to medical judgment may be of primary importance in determining limitations on driving.

Functional Motor Ability: Evaluations of this ability consist of an appraisal of an individual's ability to operate a vehicle with reference to muscular strength, coordination, range of motion of joints, spinal movement and stability, amputations or the absence of body parts, and/or other abnormalities affecting motor skill. The health care professional should indicate in their best judgment a provisional profile level without and with compensating devices. This will help the driver examiner who tests the applicant in the vehicle using compensatory devices, and makes the final determination of the functional motor ability profile.

Hearing: No hearing requirements have been formulated for drivers of private vehicles. For Meniere's Disease, see Episodic Disorders.

Recently, Diller et al. (in press) evaluated the medical conditions program by comparing the crash and citation rates per eligible licensed days for restricted and unrestricted drivers who had single medical conditions, by functional ability category (levels 3-5 vs 6-11) to the rates of control drivers (drivers licensed without a medical condition) matched on age group, gender, and county of residence. The relative risk ratios are shown in the table presented on page 36.

Salzberg and Moffat (1998) evaluated the Washington State Department of Licensing's Special Examination Program. A "special exam" includes an in-depth interview, and an extended or specialized on-road drive test, typically conducted near the driver's residence. The requirements of the "special" on-road exam are dependant of the Licensing Service Representative's (LSRs) assessment of the driver during the interview. The "special" drive test may be limited to specific roads or routes (e.g., form home to the doctor's office). Drivers come to the "special exam" program by being referred to the Department by law enforcement, physicians, family, or by LSRs who observe an impairment or disability when the driver comes in for license renewal. These drivers must undergo and pass a drive test (a "re-exam") and possibly a knowledge test. Drivers who fail the "re-exam" or those with medical/vision certificates who do not meet Department of Licensing standards have their license canceled. However, they may request a "special exam" that more completely assesses their driving ability. The most common outcome of the "special exam" is to impose driving restrictions, such as time of day (e.g., 10 a.m. to 3 p.m., daylight only); area (e.g., within an x-mile radius of residence, within city limits only, no freeway driving); and equipment (e.g., corrective lenses, hand controls, outside vehicle mirrors, power steering, power brakes). In some cases, drivers who retain their licenses must submit periodic medical or visual reports.

The study included 380 older drivers who were required to undergo a "special examination" (and passed) in 1994, and 449 control drivers matched on age, gender, and city of residence. Sixty-nine drivers failed the "special exam" and are not included in this analysis, because they have no post-exam driving exposure (97 percent had their licenses canceled and 3 percent voluntarily surrendered their licenses). Control group drivers averaged 75.6 years of age, and drivers who passed the "special exam" averaged 75.2 years of age. Documents retrieved to describe the medical conditions and driving performance of the subjects included medical certificates, vision certificates, driver license status and restrictions, and traffic violations and convictions. The most common reasons that drivers were given "special exams" were because of failing a re-exam (36 percent), a vision certificate being filed with the Department of Licensing (30 percent), or a medical certificate being filed (15 percent). Law enforcement accounted for 4 percent of the referrals, physicians for 6 percent, Licensing Service Representative for 7 percent, and family/friend/self for 3 percent of the referrals.

The following visual and medical conditions were represented among the "special exam" group: cataracts, diabetic retinopathy, macular degeneration, diabetes mellitus, cardiovascular conditions, neurological conditions, psychiatric conditions, and stroke/cerebral vascular conditions. The primary medical condition for a subject listed on the Department of Licensing record is the condition that was associated with a particular subject for this study.

Crash and violation records of drivers who underwent the "special exam" were compared with that of the control group, for a period of 1.75 years before the exam, and 3.25 years after the exam (a 5-year period). Since control group drivers did not undergo a special exam (by definition), an arbitrary date that was the same as the date for the matched exam group drivers was chosen to measure driving performance. Crash and violation rates were calculated to describe the number of incidents per 100 subjects per year, since the pre- and post-observation periods differed in length. For comparison purposes, in Washington State during 1996 there were 140,215 total collisions and 4,037,534 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers in a one-year period. Driving records for the "special exam" (passing) and control groups are shown below.


Washington State Special Exam Program Analysis
Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
Passed Special-Exam (n=380) 7.0677 3.2389 13.3835 5.2632

Control group drivers (who did not receive exams and consequent restrictions) showed reductions in their crash and violation rates over the 5-year period. The authors explain this phenomenon by noting that a normal trend exists among older drivers that as they age, they tend to reduce their driving, or to stop altogether. Although the exam group drivers also showed a reduction in crash and violation rate after passing the exam and receiving restrictions, their rates were significantly higher than the control group during the post-exam period. Comparing the post-exam collision rates of the "special exam" drivers (3.24 per 100 licensed drivers) with collision rates for the entire licensed population in Washington State over a 1-year period (3.47 per 100 licensed drivers) shows that drivers who pass special exams and receive driving restrictions are no larger a threat to the public than the population of drivers across all age groups. Rates for control- and exam-group drivers by medical condition are presented on page 37.

A discussion of the results found for drivers with neurological conditions and stroke/cerebral vascular conditions is provided here, instead of in a separate sub-section for several reasons. First, the sample sizes are rather small (~20). Second, the type of neurological condition is not specified (e.g., epilepsy, cerebral palsy, muscular dystrophy, poliomyelitis, multiple sclerosis, Parkinson's Disease, myasthenia gravis, tumors of the brain, etc.). Third, there is not a separate section under epidemiology in this Notebook that deals with strokes, because, depending on what area of the brain is affected, a stroke could have minimal, moderate, or severe effects that are either temporary or permanent. Also, a stroke may affect any of the following capabilities needed for driving: vision, perception, physical functionality, reaction time, and cognitive skills needed for decision making and judgment.

What is interesting about the drivers with neurological and stroke/cerebral vascular conditions in Salzberg and Moffat's study is that their post-exam crash and violation rates remained among the highest of all exam-group drivers with medical conditions, and these rates were well above (2.6 to 3.8 times) the post-exam rates of the control group of older drivers. In addition, drivers with strokes/cerebral vascular conditions had a post-exam crash rate that was 1.27 times that of the population of licensed drivers in the State of Washington. Therefore, restricting the licenses of drivers with these medical conditions was not sufficient to reduce their crash risk to the level posed by drivers across all age groups, nor did these drivers appear to reduce their exposure to the level of their age peers in the control group, who showed a reduction in crash risk over the 5-year period without any intervention. It is possible that drivers with these medical conditions are unaware of the risks they pose while driving, and demonstrate poor judgment and impulse control leading to adverse driving events. The number of incidents per year per 100 drivers in each group is presented below.

Washington State Special Exam Program Analysis
Group Pre-Exam ollision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
Special Exam Neurological Conditions (n=20) 8.5714 3.0769 17.1429 7.6923
Special Exam Stroke/Cerebral Vascular Conditions (n=21) 5.4422 4.3956 8.1633 7.3260

Note: For comparison purposes, in Washington State during 1996, there were 140,215 total collisions and 4,037,543 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers during this one-year period.

Conclusions/Preliminary Recommendations:

Study findings from Diller et al. are currently in press; it is expected that a report will be submitted to NHTSA later in 1999. Based on the preliminary findings cited in this Notebook, the Utah report is expected to present evidence to indicate which medical conditions are associated with a higher rate of at-fault crashes and citations for licensed drivers who have full privileges than for the general population of drivers. With this information, health care professionals will be better equipped to counsel their patients who have medical conditions about the effects of their conditions on driving, and will have the knowledge to support suggestions about potential restrictions on when and where they should drive to remain as safe as possible. Health care professionals should also emphasize to their patients the importance of informing the DMV of medical conditions that effect driving performance, for their own safety.

The following caveats regarding epidemiology data collection methods also apply. First, eligible driving days was the exposure measure; actual or estimated miles driven data were not obtained in this analysis. Next, the results are dependent upon the drivers who reported their medical conditions during the study period and not on all drivers who have medical conditions. The proportion of drivers with medical conditions who report them to the Utah Driver License Division is unknown. Also, the extent to which health care professionals assign functional ability levels according to the Medical Conditions program specifications is unknown. For example, a driver who has been assigned a functional ability rating that requires a restriction may shop around for a professional who will assign a more favorable rating, thus allowing him or her to drive unrestricted. Finally, the compliance rates for restricted drivers were not obtained at the time of crash or citation, nor did the analysis take into account the number of drivers who are repeat offenders.

Notwithstanding these methodological limitations, this pending publication represents the most comprehensive analysis to date of the relationship between type and severity of medical conditions and associated risks of adverse driving events.

Based on the Salzberg and Moffat (1998) findings, it appears that the process used by Washington State to identify older drivers who are at an increased crash risk (e.g., referral by physicians, law enforcement, family/friends, and licensing personnel) does in fact detect individuals who have significantly poorer driving records than their age-matched peers. Also, the requirement to undergo a re-exam in a familiar area and the consequent tailoring of restrictions serves to (generally) lower their crash risk to a level that does not pose any more of a safety hazard to the public than that of the general driving population. But, the program has differential effects for differing medical conditions. Positive outcomes are shown for drivers with diabetic retinopathy, cataracts, cardiovascular conditions and diabetes mellitus. On the other hand, licensing restrictions did not lower the crash risk of drivers with macular degeneration; and, for drivers with neurological, psychiatric, or stroke/cerebral vascular conditions, the obtained reductions in crash risk still left these drivers at a 3- to 4-fold greater risk level when compared to the control group drivers. Finally, drivers with psychiatric and stroke/cerebral vascular conditions continued to have a crash risk higher than that of the overall population of licensed drivers.

Two points should be considered in generalizing the results of the Washington State study to other populations. First, there are no actual measures of driving exposure. Second, comparisons are made between drivers with certain medical conditions (a subset of the special exam group) and the control group as a whole. Since the overall age distribution for all study subjects was 12.5 percent under age 60, 40.4 percent between ages 60 and 80, and 47.3 percent over age 80, and since the incidence of many medical conditions increases as age increases, it is possible that the control group of drivers could be younger than any given subset of drivers who were selected for the analysis because they presented a particular medical condition.

References:

Diller, Cook, Leonard, Reading, Dean, and Vernon (in press)

Foley, Wallace, and Eberhard (1995)

Hemmelgarn, Suissa, Huang, Boivin, and Pinard (1997)

Hu, Trumble, Foley, Eberhard, and Wallace (1998)

Koepsell, Wolf, McCloskey et al. (1994)

Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994)

Owsley, Allman, Ball, and Smoot (1998)

Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998)

Owsley, McGwin, and Ball (1998)

Owsley, Stalvey, Wells, and Sloane (1999)

Salzberg and Moffat (1998)

Sims, Owsley, Allman, Ball and Smoot (1998)

Stewart, Moore, Marks, May, and Hale (1993)

Preliminary Analysis of Drivers with Medical Conditions Compared to Control Group Drivers, Presented as Relative Risk (per 10,000 eligible licensed days) for Specified Driving Events (Citation, All Crashes, At-Fault Crashes).

Excerpted from: Diller, E., Cook, L., Leonard, D., Reading, J., Dean, J.M., and Vernon, D (in press). Evaluating Drivers with Medical Conditions in Utah, 1992-1996. NHTSA Tech. Report, Contract DTNH22-96-H-59017. Preliminary Report, Utah CODES Project.


Functional Ability Category License Restriction Status and Number of Drivers in Each Group: Not Restricted (FA Levels 3,4,5) vs Restricted (FA Levels 6-11) Citations All Crashes At-Fault Crashes
Diabetes & Other Metabolic Conditions Not Restricted (n=10,069) 1.04 1.41* 1.58*
Restricted (n=358) 1.40 1.43 1.79
Cardiovascular Not Restricted (n=18,990) 0.77** 1.14* 1.15*
Restricted (n=160) 1.57 1.61 1.72
Pulmonary Not Restricted (n=2,615) 0.88** 1.25* 1.35*
Restricted (n=244) 0.50 0.93 1.63
Neurologic Not Restricted (n=887) 0.93 1.67* 2.27*
Restricted (n=194) 0.77 1.40 1.51
Epilepsy & Other Episodic Conditions Not Restricted (n=2,620) 1.03 1.81* 2.11*
Restricted (n=775) 1.05 1.55* 2.47*
Learning, Memory, & Communication Not Restricted (n=107) 1.31 2.49* 3.57*
Restricted (n=6) 11.76* zero rate zero rate
Psychiatric or Emotional Conditions Not Restricted (n=6,763) 1.24* 1.65* 1.96*
Restricted (n=305) 0.83 1.97* 2.97*
Alcohol & Other Drugs Not Restricted (n=143) 2.37* 1.88* 2.33*
Restricted (n=24) 5.83* 4.21* 5.75*
Visual Acuity Not Restricted (n=10,363) 1.37* 1.49* 1.70*
Restricted (n=1,535) 1.37* 1.39* 1.72*
Musculoskeletal Abnormality or Chronic Medical Debility Not Restricted (n=370) 1.23 1.66* 1.92*
Restricted (n=32) zero rate 4.25 10.63*
Functional Motor Impairment Not Restricted (n=214) 1.42* 1.18 1.87*
Restricted (n=13) zero rate zero rate zero rate

* Differences in rates between medical conditions and control groups are statistically significant, with higher rates for medical conditions group.

* *Differences in rates between medical conditions and control groups are statistically significant, with higher rates for control group.

zero rate: there were no adverse driving events in one of the driver groups, so a rate could not be calculated.

Driving Records (number of incidents per 100 drivers per year) for Control Group Drivers and Drivers with Medical Conditions Who Were Required to Take a Special Driving Exam.

(Excerpted from: Salzberg and Moffat, 1998: Washington State Department of Licensing Special Exam Program - An Evaluation.)

Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control (n=449) 3.8180 1.1650 7.5087 2.2614
All Conditions: Failed Exam (n=69) 12.4224 .0000

(license canceled)

15.7350 .0000

(license canceled)

All Conditions: Passed Exam (n=380) 7.0677 3.2389 13.3835 5.2632
Cataracts (passed exam: n=45) 5.0794 2.0513 15.2381 2.0513
Diabetic Retinopathy (passed exam: n=14) 12.2449 .0000 8.1633 2.1978
Macular Degeneration (passed exam: n=71) 3.2193 3.4670 6.4386 5.2004
Diabetes Mellitus (passed exam: n=27) 6.3492 1.1396 8.4656 2.2792
Cardiovascular Conditions (passed exam: n=47) 7.2948 1.9640 20.6687 2.6187
Neurological Conditions (passed exam: 20) 8.5714 3.0769 17.1429 7.6923
Psychiatric Conditions (passed exam: n=46) 12.4224 4.6823 23.6025 8.0268
Stroke/Cerebral Vascular Conditions (passed exam: n=21) 5.4422 4.3956 8.1633 7.3260

Note: For comparison purposes, in Washington State during 1996, there were 140,215 total collisions and 4,037,543 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers during this one-year period.

Risk Ratios for Identified Medical Conditions [ref. Notebook sections IA1(a) through IA1(l)]


Condition Study Results
Cataracts Owsley, Stalvey, Wells, and Sloane (1999) Significant association between cataract and crash involvement:

Adjusted for driving exposure...RR=2.48, 95% CI=1.0-6.14*
Adjusted for impaired health...RR=2.49, 95% CI=1.0-6.27

Diabetic Retinopathy/ Diabetes Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998)

Koepsell, Wolf, McCloskey et al. (1994)

Crash risk 5 times greater with diagnosis of diabetic retinopathy.... 95% CI = 1.13-21.8

Injury-crash risk odds ratios (OR) for older drivers = 2.6 for diabetes mellitus (any); 5.8 for diabetics treated with insulin; 3.1 for diabetics treated with oral hypoglycemic agents; 8.0 for diabetes and coronary heart disease together.

Glaucoma Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998) Significant association between glaucoma and crash risk....RR=5.20, 95% CI = 1.19-22.72

Males- RR=9.81 Females-RR=5.14

Owsley, McGwin, and Ball (1998) Crash risk cases 3.6 times more likely to report glaucoma than controls
Hu, Trumble, Foley, Eberhard, and Wallace (1998) Association between highway crashes and glaucoma significant only for older males (OR=1.7)
Foot Abnormalities Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994) Association between 3+ foot abnormalities and adverse driving events RR=2.0, 95% CI = 1.0-3.8
Falls Sims, Owsley, Allman, Ball and Smoot (1998) Significant association between crash involvement and having fallen in the past two years: OR=2.6, 95% CI=1.1-6.1*
Persistent Back Pain

Hu, Trumble, Foley, Eberhard, and Wallace (1998)

Foley, Wallace, and Eberhard (1995).

Association between crash risk and persistent back pain significant for combined gender (6,553 female person-years and 5,414 male person-years); RR=1.25 for 3,000; 6,000, and 12,000 miles driven annually. RR = 1.54 for 9,000 and 18,000 annual miles.

Significant association between episodes of back pain and increased risk for crashes in a sample of 1,791 drivers age 68+ (RR=1.4, p<.05)

Cardiac Conditions (Irreg. Heartbeat) Stewart, Moore, Marks, May and Hale (1993) Significant correlation between irregular heartbeat and crashes: OR=1.83, 95% CI=1.25-2.68
Feet/Legs Cold on Exposure to Cold Stewart, Moore, Marks, May and Hale (1993) Significant correlation between feet or legs cold on exposure to cold and traffic crashes: OR = 1.82, 95% CI = 1.17-2.82
Bursitis Stewart, Moore, Marks, May and Hale (1993) Significant correlation between bursitis and traffic crashes: OR = 2.18, 95% CI = 1.41-3.38
Renal Disease (Protein in urine) Stewart, Moore, Marks, May and Hale (1993) Significant correlation between protein in urine and traffic crashes: OR = 1.84, 95% CI = 1.25-2.72
Use of Antidepressant/

Antianxiety drugs

Hu, Trumble, Foley, Eberhard, and Wallace (1998); Hemmelgarn, Suissa, Huang, Boivin, Pinard (1997)

Significant association between antidepressant use and crash risk (males only). RR= 1.98

Significant association between half-life benzodiazepine use (within 1st week of use) and crash risk (RR=1.45, CI=1.04-2.03) .

RR for continuous use up to 1 yr significant (RR=1.26, CI=1.09-1.45). In contrast, no increased risk within first week of short-half-life benzodiazepines (RR=1.04, CI = 0.81-1.34) or with continued use (RR=0.91, CI=0.82-1.01)

*RR=Relative Risk; OR= Odds Ratio; CI=Confidence Interval

I.A. IDENTIFY OLDER PEOPLE WHO ARE AT HIGH RISK OF CRASHES


I.A.2. Driving and/or Functional Assessment Outcomes

(a) Physical Performance Deficits

(b) Sensory (Vision) Deficits
(c) Deficits in Visual Attention/Speed of Processing
(d) Perceptual Skills
(e) Memory/Cognition Deficits
(f) Navigation Errors on Road Test
(g) Discriminating Maneuver Errors on Road Test
(h) Decision-Making and Response Selection in Driving Simulators


1A2(a). Physical Performance Deficits


Summary:

Lower Limb Mobility: In a sample of 283 community-dwelling individuals age 72 to 92 (mean age=77.8), Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994) found that the timed performance test most strongly associated with adverse events (traffic crash, violation, stopped by police) in the year following testing was the rapid-pace walk (> 7 seconds versus < 7 seconds [relative risk=2.0, 95% confidence interval=1.0-3.8]). Nine percent of the faster walkers had adverse driving events, compared to 17 percent of the slow walkers. This difference was significant at the p<.05 level. In the activity domain, walking less than 1 block per day was associated with adverse events (relative risk [RR]=1.9, 95% confidence interval [CI]=1.1-3.5). Twenty-one percent of the subjects who walked less than 1 block per day had adverse driving events, compared to 11 percent of the subjects who walked 1 block or more each day. This difference was significant at the p< .05 level. Foot tap time showed a trend toward association with adverse events in the study, and is face valid as a measure of ability to move leg/foot from gas to brake pedal.

In the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that subjects who took longer than 7 seconds to complete the rapid-pace walk (walk 10 ft, turn around, walk 10 ft back) were 1.25 times more likely to be involved in a crash compared to subjects who could complete the walk in 7 seconds or less. The mean walk time for the crash-free drivers was 6.78 seconds, and the mean walk time for the crash-involved drivers was 7.12 seconds. Also, subjects whose alternating foot-tap time was 10 seconds or more were 2.61 times more likely to be in a crash, compared to subjects whose foot tap times were less than 10 seconds. The mean foot-tap time for the crash-free subjects was 6.6 seconds, and the mean foot-tap time for the crash-involved subjects was 7.1 seconds. This difference was significant at the 0.04 level. Subjects ranged in age from 68 to 89 (mean age=75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997).

Upper Limb Mobility: In a panel data analysis of remaining eligible drivers in 1993 (507 female drivers and 375 male drivers) who participated in the Iowa 65+ Rural Health Study from 1981-1993, older females who had difficulty extending their arms above their shoulders had an increased probability of being involved in a crash (Hu, Trumble, Foley, Eberhard, and Wallace, 1998). In other words, an older female with difficulty extending her arms above shoulder height is more than twice as likely to be crash involved than another female with no difficulty, given that both drive 6,000 mi/yr.

Sims, Owsley, Allman, Ball, and Smoot (1998) conducted a study of 174 drivers ages 55-90 (mean age 71.1). Case drivers had at least 1 state-recorded at-fault crash in the 6 years preceding the assessment (n=99) and controls had no state-recorded at-fault crashes in the prior 6 years (n=75). Results at the univariate level indicated that crash-involvement was significantly associated with difficulty reaching out (p=.042).

In the pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that subjects who could not raise their arms above shoulder height were 1.91 times more likely to be involved in a crash, compared to subjects who could perform this action.

Head/Neck Range of Motion: The behavior of drivers at simulated T-intersections was investigated to determine the relationships between the range of movement of the head and neck, the visual field, and the decision time for a simulated traffic maneuver (Hunter-Zaworski, 1990). Impairment was defined by a combined static range of movement of the head/neck and visual field of less than 285 degrees. Younger (ages 30-50) impaired drivers were able to compensate for their impairment (their decision times were not affected by their reduced head/neck flexibility), but older impaired drivers (ages 60-80) were not.

In a study of 125 community-living cohort of older persons who were active drivers (ages 77+), limited neck range of motion (RR = 6.1, CI = 1.7-22.0) was one of the factors independently associated with (self-reported) adverse driving events (crash, moving violation, being stopped by police during previous 5.75 years) in multivariate analyses adjusting for driving frequency (Marottoli, Richardson, Stowe, Miller, Brass, Cooney, and Tinetti, 1998). Range of motion of the neck was measured by having the subject stand against a wall, and turn his or her head to identify a number placed behind either shoulder.

In the pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that subjects who could not turn their heads (including upper torso) to view the time on a clock placed directly behind them were 1.38 times more likely to be involved in a crash compared to subjects who could perform this action.

Conclusions/Preliminary Recommendations:

Older drivers with reductions in physical flexibility and range of motion of arms, legs, and neck are at increased crash risk. Physicians and other health care providers should include physical performance measures in their assessments of geriatric patients, ask questions to determine driving habits and problems, and counsel older drivers about the consequences of limited mobility/flexibility on driving performance. In addition, they should recommend exercises to help improve strength and flexibility, make suggestions about where and when patients should drive, and refer patients to occupational/physical therapists for remediation or fitting with adaptive equipment, when appropriate. In addition, increasing the public's awareness about the effects of diminishing physical capabilities on driving performance should enable drivers to make their own responsible decisions.

References:

Hu, Trumble, Foley, Eberhard, and Wallace (1998)
Hunter-Zaworski (1990)
Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994)
Marottoli, Richardson, Stowe, Miller, Brass, Cooney, and Tinetti (1998)
Sims, Owsley, Allman, Ball, and Smoot (1998)

IA2(b). Sensory (Vision) Deficits


Summary:

Static Acuity. With respect to driving, static visual acuity has consistently been found to have weak relationships to traffic crashes and convictions. For example, in a large sample study investigating the relationship between visual function and crash rate, Burg (1967) reported that the three static visual tests evaluated in their protocol had the second strongest relationship with crashes, with dynamic acuity having the strongest relationship. These three correlations, ranging from -0.053 to -0.129, were small but significant given the large sample size (n > 17,000). In Marottoli, Richardson, Stowe, Miller, Brass, Cooney, and Tinetti's (1998) study of 125 community-living older persons who were active drivers (ages 77+), corrected near visual acuity worse than 20/40 (Risk Ratio = 11.9; 95% Confidence Interval 1.3 - 109.1) was one of the factors independently associated with (self-reported) adverse driving events (crash, moving violation, being stopped by police during previous 5.75 years) in multivariate analyses adjusting for driving frequency.

Meta-analysis across studies investigating acuity and crash risk confirms that there is a weak, but consistent relationship between these variables (Staplin, Ball, Park, Decina, Lococo, Gish, and Kotwal, 1997). While the overall comparison of effect sizes is significant (2=19.79, p=0.00054) these differences are largely due to the level of significance that varies with sample size. There are several reasons why one might not expect to find a strong relationship between acuity and crash rate. Good acuity is probably beneficial to driving in instances where the vehicle is stopped or moving at a slow rate, such as at an intersection or in a parking lot. It is of less benefit while driving at normal speeds. Furthermore, unlike real visual scenes that vary in complexity, contrast, and illumination, the stimuli used to measure static visual acuity are small, of high contrast, and of low complexity. Therefore, many have argued that this type of measure bears little resemblance to the visual requirements of driving, and should not be expected to be strongly tied to crash involvement. Studies that have correlated the on-road driving performance of older subjects and static acuity are described below.

In a study of 82 drivers (age 60-91) referred to CA DMV, correlations between static acuity score (20/20, 20/80, and 20/200) measured with MultiCAD (square wave gratings with vertical bars were used), and weighted errors on driving test were not significant. However, correlations between static acuity response time at each level of acuity and weighted error scores on driving exam were as follows: 20/40 time: r=.3395 (p<.004); 20/80 time: r=.4230 (p<.000); 20/200 time: r=.1970 (p<.090) (Janke and Eberhard, 1998; Janke and Hersch, 1997; Staplin, Gish, Decina, Lococo, and McKnight, 1998).

In McKnight and McKnight's (1998) study of 360 drivers age 62+, correlations between static visual acuity (measured with Automated Psychophysical Test [APT]) and observed driving performance were relatively low but significant; correlations between on-the-road performance and time to respond to the acuity stimuli (r=.30) were higher than acuity errors (r=.18).

Salzberg and Moffat (1998) evaluated the driving records of 380 older drivers who were referred to the Washington State Special Examination Program (and passed), and 449 control group drivers. This program is described in more detail in Section IA1(m) of the Notebook. Static acuity readings were available for 357 of these drivers. A "special exam" includes an in-depth interview, and an extended or specialized on-road drive test, typically conducted near the driver's residence. The most common outcome of the "special exam" is to impose driving restrictions (time of day, area, equipment).

Crash and violation records of special exam group drivers were compared with that of the control group, for a period of 1.75 years before the exam, and 3.25 years after the exam (a 5-year period). Crash and violation rates were calculated to describe the number of incidents per 100 subjects per year, since the pre- and post-observation periods differed in length. The crash and violation rates for the exam group drivers by acuity score (20/20+, 20/40+, 20/100+, and 20/200+) who passed the "special exam" and the (entire) control group are presented below, for the pre-exam and post-exam period. For comparison purposes, in Washington State during 1996 there were 140,215 total collisions and 4,037,534 licensed drivers, yielding a rate of 3.47 collisions per 100 licensed drivers in a one-year period. It is important to note that approximately 60 percent of exam group drivers had other medical conditions; only 123 of the 380 drivers (32%) were referred to the program because a vision certificate was filed with the Department of Licensing. Other reasons for referral included: law enforcement noting signs of unsafe driving (3%); Licensing Service Representative noticing diminished capabilities (6%); medical certificate filed with Department of Licensing (15%); physician referral (5%); and because a driver failed the initial re-exam test (35%). Therefore visual acuity is confounded with other medical conditions, and no direct relationship with crashes or violations can be drawn.

Group Pre-Exam Collision Rate Post-Exam Collision Rate Pre-Exam Violation Rate Post-Exam Violation Rate
Control
(n=449)
3.8180 1.1650 7.5087 2.2614
Special Exam
No vision info. available (n=23)
4.9689 6.6890 22.3602 4.0134

Special Exam
20/20+ (n=44)

10.3896 1.3986 24.6753 3.4965
Special Exam
20/40+ (n=219)
7.8278 2.5290 12.0026 4.3555
Special Exam
20/100+ (n=45)
7.6190 2.0513 16.5079 6.1538
Special Exam
20/200+ (n=49)
1.1662 7.5353 2.3324 10.6750

What is interesting to note about the pre-exam crash rates is that the drivers with the best acuity (20/20) had the highest rates, and that drivers with the poorest vision (20/200) had the lowest crash rates. Obviously, drivers with 20/20 vision were not part of the special program because of poor vision. It is instructive to look just at drivers with 20/40-20/100 acuity, whose crash rates are about double that of the control group during the pre-exam period. The requirement to undergo a special exam and the consequent licensing restrictions had the effect (at least on the surface) of lowering their crash rates to a level that does not pose any more risk than the population of licensed drivers in the State of Washington. But the reduction still puts these drivers at twice the risk of control group (older) drivers. The authors explain the decline in crash risk for the control group (who did not have any intervention) as decreased driving exposure through increased self-restriction over the 5-year study period. The significant increase in the 20/200 group from the pre-exam to the post-exam period could be the result of increased exposure by these drivers who possibly misinterpreted the decision to allow them to retain driving privileges as positive feedback about their ability to drive safely.

Dynamic Acuity. Dynamic visual acuity (DVA), like static acuity, also declines with age (Burg, 1967; 1968; 1971), with some suggestion that the age-related declines in DVA are larger than for static visual acuity (Burg, 1966). Dynamic acuity reflects the ability to resolve the details of a moving target, and therefore it has been proposed that this measure of acuity should be more relevant to driving. Some activities that appear to rely on dynamic acuity are reading street signs while in motion, locating road boundaries when negotiating a turn, and making lateral lane changes. In these situations, greater speeds are associated with poorer DVA. The earlier studies on driving and the elderly that have assessed both static and dynamic acuity have indeed found that DVA is more strongly associated with crash risk than static acuity. However, the statistically significant correlations between dynamic visual acuity and crash rate have also been consistently weak (Staplin et al, 1998). For example, the correlation between DVA and crash rate for the older drivers, as reported by Hills and Burg (1977), was too low (r=0.054) to be of any practical significance for identifying at-risk drivers. As stated earlier, dynamic visual acuity has been found to be weakly associated with crash involvement in several correlational studies (Burg, 1968; Shinar, McDowell, and Rockwell, 1977; Laux and Brelsford, 1990). In a study of professional drivers over age 50, the top 10 percent with respect to dynamic acuity were found to have lower than average crash rates. The bottom 10 percent with respect to dynamic acuity were found to have higher than average crash rates (Henderson and Burg, 1974). In other studies, Shinar, Mayer and Treat (1975) noted that drivers found recently to be at fault in a crash had poorer dynamic visual acuity than a group of persons who had not been in a crash for 2 years. As with static acuity, however, the strength of the relationships is generally weak, and meta-analysis confirms the consistency of these findings that differ primarily due to sample size discrepancies (Staplin et al., 1998). Studies that have correlated the on-road driving performance of older subjects and dynamic acuity are described below.

In a study of 82 drivers (age 60-91) referred to CA DMV, correlations between dynamic acuity score (20/20, 20/80, and 20/200) measured with MultiCAD (square wave gratings with vertical bars were used, with a rate of movement across the screen of 12 degrees per second) and weighted errors on driving test were not significant. However, correlations between dynamic acuity response time at each level of acuity and weighted error scores on driving exam were as follows: 20/40 time: r=.3092 (p<.010); 20/80 time: r=.3256 (p<.005); 20/200 time: r=.3297 (p<.004). (See Janke and Eberhard; Staplin, Gish, Decina, Lococo, and McKnight, 1998).

In a study of 360 drivers age 62 and older, correlations between dynamic visual acuity (measured with APT) and observed driving performance were relatively low but significant; correlation between on-the-road performance and time to respond to the acuity stimuli was r=.24; correlation between on-road performance and acuity errors was r=.21 (McKnight and McKnight, 1998).

Static Contrast Sensitivity. Contrast sensitivity tests measure both the response to sharply-defined, black-on-white targets and those with grayer, less-distinct edges. Recent studies that have included contrast sensitivity as a predictor of driving crashes have shown that, while it is a slightly better predictor than acuity, the strength of the relationship is still relatively weak (r<0.25) (Ball and Owsley, 1991; Owsley et al., 1991; Ball et al., 1993). More recently, Hennessy (1995) studied 3,669 randomly-selected Class C license renewal applicants who were licensed in California for at least 12 years. Four driver age groups were studied: 26-39, 40-51, 52-69, and 70+. The 48-letter test designed by Pelli, Robson, and Wilkins, 1988, of contrast sensitivity at one spatial frequency was one of the independent measures examined. In this test, the contrast between letters and background decreases as one moves down and toward the right of wall-mounted chart, viewed at distance of 2 meters under normal room illumination. The letters from left to right and from top to bottom progressively fade out as if they must be read in thicker and thicker fog. Letters (in groups of 3) range from 90 percent contrast (upper left) to 0.5 percent contrast (lower right). Testing requires no more than 3 minutes. The dependent measure was the crash frequency during the previous 3-year period, extracted from the DMV database. Results showed that for all age groups combined, the contrast sensitivity test score was not significantly associated with total prior 3-year crash involvement when considered in isolation. There was a very small percentage of drivers age 70+ with good low-contrast acuity. Using a pass-fail criterion of 36 or more correctly identified letters as pass and less than 36 letters fail, Pelli-Robson specificity was 53 percent and sensitivity was 29 percent in predicting citations for age 70+ drivers; accuracy of predicting citation occurrence was 6.5 percent. For subjects ages 52-69, specificity was 65 percent, sensitivity was 19 percent, and positive prediction was 7 percent. Studies that have correlated older drivers on-road performance with static contrast sensitivity are described below.

In a study of 82 drivers (ages 60-91) referred to CA DMV, static contrast sensitivity response time for the high contrast 20/80 target measured with MultiCAD was significantly correlated with weighted error score on the driving test (r = .3884, p< .001). (See Janke and Eberhard, 1998; Janke and Hersch, 1997; Staplin, Gish, Decina, Lococo, and McKnight, 1998).

In Janke and Eberhard's (1998) study of 102 "referred" subjects aged 60-91 (34 of which were identified as probably being cognitively impaired to some degree) and 33 paid "volunteers" ages 56-85, the correlation between Pelli-Robson errors and weighted error score on a road test was significant (r=.4009, p<.0001) for combined referrals and volunteers (n=135). For the referral group only (n=102), the correlation between Pelli-Robson errors and weighted error score on the road test was also significant (r= .2069, p<.044).

In Brown, Greaney, Mitchel, and Lee's (1993) study of 1,475 ITT Hartford Insurance Co. policyholders (age 50-80+) divided into two groups based on the presence or absence of recent at-fault crashes, the Pelli-Robson Letter Sensitivity Chart consistently yielded the highest correlation to crashes in the sample during 1989-1991 (r=-0.11, p < 0.05).

In a study of 12,400 drivers ages 16 to 75+ in Pennsylvania, who came to Photo ID centers for license renewal, failure on the combined criteria that incorporates the current PennDOT standard (binocular acuity of 20/40 and horizontal visual field of 140 degrees) and a broadly defined contrast sensitivity criterion (scores below normal for 1 or more of the 3 spatial frequencies tested using Vistech contrast sensitivity gratings via an Optec 1000 vision tester) produced the strongest relationship linking poor vision and high crash involvement, especially for 66-75 and 76+ driver age groups (Decina and Staplin, 1993). Neither visual acuity nor horizontal field measures in isolation were significantly related to crash involvement. The study authors recommended periodic screening using the combined criterion, for drivers over age 55.

In McKnight and McKnight's (1998) study of 360 drivers age 62 and older, correlations between low contrast acuity (measured with APT) and observed driving performance were low but significant; correlations between on-the-road performance and time to respond to the acuity stimuli (r=.23) were higher than contrast sensitivity errors (r=.18).

Dynamic Contrast Sensitivity. In a study of 82 drivers (ages 60-91) who were referred to CA DMV, the correlation between dynamic contrast sensitivity response time for the high contrast 20/80 target (using MultiCAD) and weighted errors on the road test was significant (r=.2466, p<.049). The stimuli in this study consisted of square wave gratings with vertical bars, with a rate of movement across the screen of 12 degrees per second (see Janke and Eberhard, 1998; Janke and Hersch, 1997; Staplin, Gish, Decina, Lococo, and McKnight, 1998).

Conclusions/Preliminary Recommendations:

Contrast sensitivity is a visual capability, where deficits have been shown to be related to traffic crashes and poor driving performance in older drivers. Correlations between driving performance and contrast sensitivity response time (for correct responses) as well as percent correct responses have been found to be significant. Contrast sensitivity test slides are available for DMV-model vision screener devices, and would be easily implementable in a DMV vision screening protocol. Other presentation methods include wall charts and computer displays of test stimuli.

For acuity, it appears that the time to respond is more strongly related to driving performance than other dependent measures, such as percent correct. This result should be interpreted with caution, however, because as usual, the range of responses on this dependent measure (i.e., correctness) was restricted, and there was potential for substantial "noise" in the data from other sources of variance. Time to respond would be a more difficult measure to implement in a DMV setting unless stimuli were presented by computer.

References:

Ball and Owsley (1991)
Ball, Owsley, Sloane, Roenker, and Bruni (1993)
Brown, Greaney, Mitchel, and Lee (1993)
Burg (1966, 1967, 1968, 1971)
Decina and Staplin (1993)
Henderson and Burg (1974)
Hennessy (1995)
Hills and Burg (1977)
Janke and Eberhard (1998)
Janke and Hersch (1997)
Laux and Brelsford (1990)
Marottoli, Richardson, Stowe, Miller, Brass, Cooney, and Tinetti (1998)
McKnight and McKnight (1998)
Owsley, Ball, Sloane, Roenker and Bruni (1991)
Shinar, Mayer and Treat (1975)
Shinar, McDowell, and Rockwell (1977)
Staplin, Ball, Park, Decina, Lococo, Gish, and Kotwal (1997)
Staplin, Gish, Decina, Lococo, and McKnight (1998)

IA2(c). Deficits in Visual Attention/Speed of Processing


Summary:

A current and potentially most-promising area of inquiry relating crash risk to functional impairment is the study of visual attention deficits and underlying divided attention and speed-of-processing functions. Prominent among studies in this area are those addressing measures of information processing efficiency such as "useful field of view" and "channel capacity;" research summaries are presented below.

A prospective study was conducted with 294 older drivers (ages 56-90) to identify measures of visual processing associated with crash involvement by older drivers (Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley, 1998). This sample had been previously drawn for a case-control study by Ball et al. (1993) from the population of all licensed drivers in Jefferson County, Alabama age 55 and older. The subjects represent 3 crash categories (0 crashes, 1-3 crashes, and 4+ crashes during the previous 5-year period) and 7 age categories (55-59, 60-64, 65-69, 70-74, 75-79, 80-84, and 85+ years of age). Of the 302 subjects drawn for sample, 6 were excluded because they had ceased driving, and 2 did not complete the protocol. The study focused on the prospective 3-year follow-up of the 294 drivers who were assessed in 1990 to determine what visual characteristics were associated with future crash involvement. Subjects received the following sensory tests in 1990: Letter Acuity - ETDRS chart; Contrast Sensitivity - Pelli-Robson chart; Stereoacuity - TNO Test; Disability Glare - MCT-8000 (VisTech); Visual Field Sensitivity -Humphrey Field Analyzer 120-point program for central 60 degree radius field; and visual attention and visual processing speed - Useful Field of View. Impaired useful field of view (UFOV) was the only visual processing variable associated with increased crash risk. A significant, independent association with crash risk in 3-year follow-up was found for UFOV reduction of > 40 percent: RR=2.3; 95% CI=1.27-4.29. Of UFOV component scores, speed of processing (subtest 1) and selective attention (subtest 3) were NOT associated with crash occurrence. Impairment in divided attention (subtest 2) was significantly associated with a 2.3 fold increased risk of crashing (95% CI=1.24-4.38, p=0.01). For every 10 points of UFOV reduction, subjects had a 16 percent increase in crash risk. Estimates are that 24 percent of older driver crashes are due to UFOV reduction >40 percent.

Owsley, McGwin, and Ball (1998) studied 193 older drivers between ages 55-87 (mean=71 years) to identify visual risk factors for vehicle crashes by older drivers that result in injury. Univariate analyses showed that older drivers involved in injurious crashes were more likely to have UFOV reductions (Odds Ratio [OR]=5.3 for reductions of 23 to 40 percent; OR=16.3 for reductions of 41 to 60 percent; and OR=22.0 for reductions greater than 60 percent). Only two variables were independently associated with crash risk in the multivariate analyses: UFOV and glaucoma. UFOV reductions of 22.5-40 percent, 41-60 percent, and >60 percent were associated with 5.2, 16.5, and 21.1-fold increased risk of an injurious crash, respectively compared to those with reductions of less than 22.5 percent. This sample was a subset of the sample described above, consisting of 78 drivers (cases) who had at least 1 crash in the prior 5-year period that resulted in an injury to anyone in the involved vehicles, and 115 drivers (controls) who had no crashes in the same 5-year period. Excluded were 101 subjects who were involved in crashes where no injury was reported.

Goode, Ball, Sloane, Roenker, Roth, Myers, and Owsley (1998) studied 239 older drivers (ages 56-90) to examine the utility of a set of commonly used neuropsychological tests in comparison to the UFOV in predicting state-recorded, at-fault crashes in the prior 5 year period in a group of older drivers, a model using UFOV alone was significant (p<.001). This model was as predictive as a model using UFOV and traditional tests (MOMSSE, Trails, Wechsler Memory Scale subtest, and Rey-Osterreith Complex Figures). The classification success was 85.4 percent, with sensitivity of 86.3 percent and specificity of 84.3 percent. The estimated probability of crashing with a UFOV score of 20 was 22 percent; for a UFOV score of 60, the probability of crashing increased to 81 percent. The subjects in this study were recruited from the larger sample of drivers participating in the larger study (Ball et al., 1993). Of the original sample of 294 subjects, 251 received all of the cognitive tests. Those with poor visual acuity (n=12) were excluded (since those with acuity worse than 20/50 uniformly fail the first subtest of the UFOV).

Owsley, Ball, Sloane, Roenker, and Bruni (1991) studied 53 drivers ages 57-83 (mean age = 70), to determine whether incorporating eye health, visual function, UFOV (Visual Attention Analyzer), and mental status could predict the number of crashes in the sample. Only the mental status total score and UFOV were significantly related to state-reported crashes. The subjects were recruited from the Primary Care Clinic of the School of Optometry at the University of Alabama at Birmingham, had valid AL licenses, and drove at least 1,000 miles/year. Results indicated that only the UFOV was related to traffic citations. Subjects who failed the UFOV had 4.2 times more crashes than those who passed. For intersection crashes, subjects who failed the UFOV had 15.6 times more intersection crashes than subjects who passed. Subjects with high MOMSSE scores had 6.3 times more intersection crashes. Together, these variables predicted 29 percent of the variance in intersection crashes, R=.54, F(2,49) =9.8, p<0.001. For intersection crashes, UFOV had 26 correct rejections, 14 false alarms, 1 miss, 11 hits.

In Ball, Owsley, Sloane, Roenker, and Bruni's (1993) retrospective study of 294 drivers ages 55-90, UFOV and mental status were the only variables that had a direct effect on crash frequency, accounting for 28 percent of the variance in crash frequency. The test battery included tests described for Owsley, Ball, Sloane, Roenker, and Bruni (1991), plus the following cognitive tests assessing visuospatial abilities: Rey-Osterreith test; Trail-Making test; and the WAIS block design test. As a predictor, UFOV resulted in 142 hits, 18 misses, 25 false-positives, and 109 correct rejections. Of the 25 false-positives, 19 were subjects who reported avoiding driving in general, avoided driving alone, and/or avoided left turns, thus minimizing their driving exposure. Removing these people from the data set increases the correlation between UFOV and crash frequency from r=0.52 to r=0.62. UFOV had high sensitivity (89%) and high specificity (81%); mental status had sensitivity and specificity values of 61 percent and 62 percent, respectively.

Another, ongoing research study, has yielded results showing UFOV's relationship to performance during an on-road driving evaluation. Of the clients who passed the UFOV test (less than 40 percent reduction in UFOV), the majority pass the on-road evaluation, and of the clients who failed the UFOV test (have more than a 40 percent reduction in UFOV), the majority fail the on-road evaluation. Of the 23 drivers who passed the UFOV, 18 passed the on-road, 4 failed on-road, and 1 is pending. Of the 25 drivers who failed UFOV screening, 6 passed the on-road evaluation, 16 failed the on-road test, and 3 are pending (pers. comm., Tom Kalina, Bryn Mawr Rehab, 10/97).

Hennessy (1995) conducted a study using 3,669 randomly-selected Class C license renewal applicants, licensed in California for at least 12 years, and unable to renew by mail. Four driver age groups were studied: 26-39, 40-51, 52-69, and 70+. Subjects age 70+ showed high variability in visual divided attention ability (subtest 2) and perceptual reaction time (subtest 1 PRT). There was a very small percentage of drivers age 70+ with very good total UFOV. Test scores had small but statistically significant predictive value (2.9%) for subjects age 70+. After adjusting for gender, age, and exposure, total UFOV scores explained 0.9 percent of the variance in crash involvement, PRT explained 0.9 percent and divided attention explained 0.9 percent. The association with crashes for subjects in the 70+ age group was stronger, with total UFOV accounting for 4.1 percent of the variation in crashes, PRT accounting for 4.1 percent of the crashes, and divided attention accounting for 4.3 percent of the crashes in the oldest age group. UFOV was not predictive of crashes in the 3 younger age groups. Of 285 subjects age 70+, 84 (29%) scored poorly. Thirty-six of the 285 subjects had a crash, and of the 36, 13 (36%) scored poorly on the UFOV. Thus UFOV sensitivity was 36 percent, specificity was 71 percent, and positive predictive accuracy was 15.5 percent. For citation occurrence, sensitivity was 28 percent, specificity was 70 percent, and positive predictive accuracy was 12 percent.

Brown, Greaney, Mitchel, and Lee (1993) studied 1,475 ITT Hartford Insurance Co. policyholders for whom past driving histories were available through insurance records. They were divided into two groups based on the presence or absence of recent at-fault crashes. Driver age ranged between 50 and 80+. The Visual Attention Analyzer was employed; the overall score from the three subtests--speed of information processing, divided attention, and a measure of distractibility--was used to describe useful field of view loss. Results showed that 42 percent of the sample had an at-fault crash between 1989-1991. The correlation between performance on the UFOV test and at-fault crashes (r=0.05) was significant (p<0.05). The low correlation was explained by the possibility that because participants were recruited through their insurance company (as opposed to being recruited through an eye clinic and offered a detailed eye exam, as were the subjects in the Ball et al. [1991] study), drivers who were less confident in their driving skills may have elected not to participate for fear that their insurance rates could be affected. Also, a noisy, crowded test environment was described which may have yielded unrepresentative visual attention measures.

Another study using WayPoint (a proprietary test measuring channel capacity or information processing rate) and Subtest 1 of UFOV (measuring speed of processing) examined 101 licensed drivers (39 females and 62 males) ages 72-90, with a mean age of 78.3 (Janke and Hersch, 1997). In this study, WayPoint was administered twice (in succession) to see if drivers with presumed cognitive impairment either failed to improve from the first administration to the second, or did not improve as much as subjects without presumed cognitive impairment. The scoring system determines: (1) channel capacity or information-processing rate, defined as the average speed per exercise on the first administration over two of the exercises; and (2) high vs low risk of preventable and non preventable collisions, reflecting the driver's situational awareness. An on-road driving exam was given by the project driving instructor (owner/operator of a driving school in San Francisco) based on the California Driving Performance Evaluation (DPE). Average time per exercise on the first administration of WayPoint was significantly related to road test weighted errors (r=.37) as was channel capacity (r=.35). Using only WayPoint 1 average time and UFOV subtest 1 as predictors of weighted error score on the road test yielded multiple R = .428; adjusted R2=0.166.

As reported by the test's developer, in six studies with 102 drivers age 20-60, WayPoint correctly classified 72 percent as high or low crash-risk drivers, missed 18 percent of the high crash-risk drivers, and falsely labeled 9.2 percent of the drivers as high-risk when they were actually low risk. Also, results of a study with emergency response (ER) trainees showed that errors on WayPoint were (1) directly related to technical errors on the ER course (a high speed drive circuit), (2) directly related to line-of-travel errors, and (3) positively correlated with lap speed. On the non-emergency test, WayPoint errors were positively correlated with driving errors and with the number of traffic cones contacted on the obstacle course (Cantor, 1995).

In Marottoli, Richardson, Stowe, Miller, Brass, Cooney, and Tinetti's (1998) study of 125 community-living older persons who were active drivers (ages 77+), poor performance on a visual attention task (< 48 correct on a number cancellation task, RR=3.0, CI=1.2-7.8) was one of the factors independently associated with (self-reported) adverse driving events (crash, moving violation, being stopped by police during previous 5.75 years) in multivariate analyses adjusting for driving frequency. The number cancellation task involved marking out all of the numbers in a row that matched a circled number at the far left-hand side of the row, within a given amount of time.

Conclusions/Preliminary Recommendations:

Older drivers with 40 percent or greater impairment in their useful field of view--which stems from declines in visual sensory function, visual processing speed, and/or visual attentional skills--appear to be at an increased crash risk. Broadly speaking, there is a strong case that age-related visual processing impairments, particularly in the ability to divide attention, are directly related to future crash risk. Based on the success to date of predicting crashes, it is recommended that the UFOV protocol (or a related procedure validated on the same measurement construct) be incorporated as a diagnostic test of cognitive deficits which predict driving impairments for license renewal applicants; in particular, the evaluation of divided attention (one of the UFOV subtests) is recommended. Quick and inexpensive assessments of gross deficits in attentional and information processing abilities also appear quite valuable; the traditional Trails protocol (see also discussion in next section of Notebook) and derivative techniques using paper-and-pencil or computer-based methods are most promising.

References:

Ball, Owsley, Sloane, Roenker, and Bruni (1993)
Brown, Greaney, Mitchel, and Lee (1993)
Cantor (1995)
Goode, Ball, Sloane, Roenker, Roth, Myers, and Owsley (1998)
Hennessy (1995)
Janke and Hersch (1997)
Marottoli, Richardson, Stowe, Miller, Brass, Cooney, and Tinetti (1998)
Owsley, Ball, McGwin, Sloane, Roenker, White, and Overley (1998)
Owsley, Ball, Sloane, et al. (1991)
Owsley, McGwin, and Ball (1998)

IA2(d). Perceptual Skills
(Visual Search, Spatial Integration, Gap/Headway Judgment)


Summary:

Visual Search. In a study of 3,238 drivers age 65 and older, who applied for renewal of North Carolina driver's license, performance on a paper-and-pencil test of general cognitive function (Trails A and B), measuring speed of visual search, attention, mental flexibility, and motor function was correlated with crash involvement in the preceding 3-year period (Stutts, Stewart and Martell, 1996, 1997). Trails A Results: Correlational coefficient with number of crashes = 0.065 (p<0.001). Subjects who scored in best quartile had 47 percent fewer crashes (.037 crash involvements per year) than drivers who scored in the worst quartile (.054 crash involvements per year). Trails B Results: Drivers in the poorest decile of performance had a predicted average annual crash rate of 1.5 times that of drivers in the highest decile of cognitive performance. Correlational coefficient with number of crashes = 0.072 (p<0.001). Annual crash involvements increased with increasing (poorer) cognitive scores.

In a study of 105 drivers ages 65-88 (Tarawneh, McCoy, Bishu, and Ballard, 1993), only the Trail-Making Part B test showed a significant correlation to performance on an on-road driving task, with a correlation coefficient of -0.42 (p<.0001). The correlation between Trails A and driving performance was -0.03 (p<.7329). Trails B showed the highest correlation of all factors (visual, visual perception, cognitive, range of motion) included in the analysis.

In a study of 39 drivers (21 with Alzheimer's disease) to determine fitness to drive for neurological patients, performance on Trails B was a significant predictor of simulator crashes, with an odds ratio of 30.19 (Rizzo, Reinach, McGehee, and Dawson, 1997).

In a study of 20 drivers age 55 and older, who were administered 11 assessment tests and an on-road driving test, 6 subjects were classified as below minimum standards in driving performance (a total of 19 or more errors on the NY State Driving Exam) (Cushman, 1988, 1992). These six subjects scored more poorly on Trails B (mean Trails B total time =130.5 s) than the subjects whose on-road driving performance was at least adequate (mean total Trails B time = 93.07 s). The Trail-Making Test (Part B) was the only test that was significantly correlated with driving performance for all subjects (r=0.61, p<0.01).

In the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that subjects who took 5 minutes or longer to complete the Trails B protocol were 1.41 times more likely to be crash involved, compared to subjects who completed this test in less than 5 minutes. The mean time to complete the Trails B protocol was 161.14 seconds for the crash-free drivers, and 180.57 seconds for the crash-involved drivers. Subjects ranged in age from 68 to 89 (mean age = 75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997).

A modified and automated version of Reitan's (1958) Trail-Making Test (Part A) has been developed. In this test, 14 numbers are presented on a computer monitor arranged randomly against the background of a traffic scene, as observed by the driver through the windshield of a car. The subject must touch the numbers (touch screen display) in numerical order as rapidly and accurately as possible. Timing is done by the computer. This test was used in a study of 69 subjects ages 60-91 who were referred to the California DMV for reexamination, and 31 paid "volunteers" ages 56-85, recruited through signs posted at study site or by word of mouth (Janke and Eberhard, 1998; Janke and Hersch, 1997). An on-road driving exam was given based on the California Driving Performance Evaluation (DPE). The referral group performed significantly worse than the volunteer group (correlation between Auto-Trails time and Group = .405, p<.05). Auto-Trails mean time for referrals was 24.26 seconds; for volunteers, mean time was 16.91 seconds. Auto-Trails time correlated significantly with weighted error score on the road test, for combined referrals and volunteers (r=.4523, p<.000) and for referrals only ( r=.3748, p<.002). Auto-Trails time did not discriminate the cognitively impaired referral subjects from the cognitively unimpaired referral subjects.

Spatial Integration. In Tarawneh, McCoy, Bishu, and Ballard's (1993) study of 105 drivers ages 65-88, among the visual perception factors, Visual Closure response-time score (from the MVPT) correlated significantly with an on-road driving performance measure (correlation coefficient =-0.38). As percent of correct responses increased on the visual perception tests, performance on the driving test increased; as the reaction time scores increased, performance on the driving test decreased.

In a study of 42 patients with Alzheimer's Disease (mean age = 72.2 years) and 81 normal elderly controls (mean age = 69.1 years), driver simulator performance measures correlated strongly with Visual Memory immediate scores, and Visual Closure subscore of the Motor-Free Visual Perception Test for both AD and control subjects (Keyl, Rebok, Bylsma, Tune, Brandt, Teret, Chase, and Sterns (manuscript under review).

In the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that subjects who made 3 or more errors on the MVPT Visual Closure subtest were 1.7 times more likely to be crash involved, compared to subjects who made 2 errors or less. The mean number of incorrect items was 1.91 for the crash-free drivers and 2.62 for the crash-involved drivers. This difference was significant at the 0.002 level. Subjects ranged in age from 68 to 89 (mean age = 75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997).

Gap Judgment/Headway. In a study of 82 "referred" subjects ages 60-91 (26 of whom were identified as likely having cognitive impairment) where the ability of subjects to rapidly detect changes in the relative motion of their own versus other vehicles was measured, cognitively impaired referrals had a significantly higher error proportion (they did not brake in 47.3% of the trials where the lead vehicle braked ahead and the brake lights were visible) compared to cognitively unimpaired referral subjects (who did not brake in 21% of the trials). Also, the correlation between proportion of errors on trials where brake lights were visible, and weighted error score on an on-road drive test, was significant (r=.2801, p<.013). (See Staplin, Gish, Decina, Lococo, and McKnight, 1998; Janke and Eberhard, 1998).

McKnight and McKnight (1998) evaluated the on-road driving performance of 402 drivers age 62 and older. Approximately two-thirds of the subjects were referred to the licensing agency for reexamination based upon reports of deficient driving incidents, and the balance were incident-free volunteers. The road test was based on the Driver Performance Evaluation (DPE) developed by the California Department of Motor Vehicles (see Hagge, 1994). The incident-involved drivers tended to underestimate gap size (stating that they could safely enter gaps of less than 6 seconds) even though they erred on the safe side in the gaps that they actually entered.

Conclusions/Preliminary Recommendations:

Tests measuring visual perception, speed of visual search, and ability to sense changes in angular motion (i.e., cues to the speed and distance of other vehicles) have been shown to predict driving performance in simulators, on the road, and prior crash rate, and also have the ability to distinguish cognitively impaired individuals from unimpaired individuals. It is recommended that a Trail-Making protocol be implemented in driver screening for relicensing.

References:

Cushman (1988, 1992)
Engum, Lambert, Womac, and Pendergrass (1988)
Goode, Ball, Sloane, Roenker, Roth, Myers, and Owsley (1998)
Janke and Hersch (1997)
Janke and Eberhard (1998)
Keyl, Rebok, Bylsma, et al. (submitted)
McKnight and McKnight (1998)
Rizzo, Reinach, McGehee, and Dawson (1997)
Staplin, Gish, Decina, Lococo, and McKnight (1998)
Stutts, Stewart and Martell (1996, 1997)
Tallman, Tuokko, and Beattie (1993)
Tarawneh, McCoy, Bishu, and Ballard (1993)


IA2(e). Memory/Cognition Deficits


Summary:

In a panel data analysis of 507 female drivers and 375 male drivers who participated in the Iowa 65+ Rural Health Study from 1981-1993, having impaired cognitive ability (low score on word recall test) was a risk factor that determined the probability of an older male being involved in a crash (Hu, Trumble, Foley, Eberhard, and Wallace, 1998). Foley, Wallace, and Eberhard (1995) interviewed 1,791 drivers in this cohort, and found that drivers who could remember fewer than 3 of the 20 words given in a free-recall memory test had an increased crash risk (Relative Risk = 1.4, Confidence Interval: 1.1 to 1.9, p< 0.05).

In a study of 37 drivers age 65 and older in a case group (suspensions + crashes) and 37 matched controls (no suspensions or crashes), cases had significantly lower immediate memory task performance (p=.010) compared to matched controls (Johansson, Bronge, Lundberg, Persson, Seideman, and Viitanen, 1996; Johansson, 1997). Immediate memory was tested by a 5-item recall test, where the subject was required to name and recall 5 objects viewed on a desk after a 10-minute period (the items were not listed in this review). The delayed recall score was 1 point per correct item. Comparison of the 23 case subjects with crashes and the 29 control subjects with no crashes in the past 5 years showed that the crashed drivers had poorer 5-item recall (p<.003).

In a study of 360 drivers age 62 and older, measures of short-term and delayed short-term memory (measured with the Automated Psychophysical Test [APT]) showed fairly strong correlations between accuracy and safe driving and response time and safe driving. The correlations were significant, and ranged from 0.22 to 0.34 (McKnight and McKnight, 1998).

Hunt, Morris, Edwards, and Wilson (1993) administered the Logical Memory subscale of the Wechsler Memory Scale, which assesses immediate or delayed recall of verbal ideas presented in two paragraphs, read aloud by the experimenter. Each subject then drove for 1 hour on a pre-designed route using urban streets and highways, that included common driving situations (stop signs, traffic signals, left turns at intersections, entering and exiting an interstate highway, changing lanes, merging, diagonal and parallel parking). Subjects drove in low volume conditions. A gestalt "pass/fail" rating was given by each observer in the vehicle. In a sample of 13 healthy elderly controls (mean age = 73.5) 12 subjects with very mild dementia (mean age = 72.5) and 13 subjects with mild dementia (mean age = 73.4), the correlation between the pass/fail outcome on the road test and performance on the Logical Memory test was significant at the p<.0009 level.

In a study of 146 drivers age 65 and older (mean age = 72.0), three tests: the Brief Test of Attention (numbers), Trails A, and the Serial Sevens item in the Mini Mental State Examination (MMSE, see section IC2b(i) of the Notebook) were most strongly associated with crashes (Keyl, Rebok, and Gallo, in press). Patients who had poor performance on more than one of these tests had a 6.2-fold increase in crash occurrence in the previous two years. [A precaution by Lindal and Stefansson, 1993 regarding gender differences: when women use serial 7's they obtain much lower scores on the MMSE than if they use backward spelling, and conversely, men receive a lower score if they use backward spelling as opposed to serial 7's]

Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994) found that persons with borderline cognitive impairment (MMSE score of 23-25) were more likely to have adverse events (traffic crash, violation, or stopped by police) in the year following examination than those with higher or lower scores (relative risk 2.0, 95% CI, 1.1-3.7). The authors examined the components of the MMSE individually and by cognitive domain (orientation, memory, attention, language, and visuospatial ability), and found that the item most closely associated with adverse events was impaired design copying (24% of persons who could not correctly copy the intersecting pentagons had events compared with 8% of those who could [relative risk 3.0, CI, 1.6-5.6]).

In the recently completed pre-pilot study conducted in Salisbury, Maryland for the NHTSA "Model Driver Screening and Evaluation Program" project, the present Notebook authors found that inability to recall three short words was related to crashing (Odds Ratio = 1.52). Subjects ranged in age from 68 to 89 (mean age = 75.7); 131 of the 363 subjects were involved in at least 1 crash in the previous 6-year period (1991-1997).

Conclusions/Preliminary Recommendations:

Impaired cognitive ability, measured using immediate and delayed recall, is associated with increased crash risk and poorer on-road driving performance in older people. The inability to count backwards by 7's (ability to perform a mental function) is also related to increased crash risk in older drivers, but may have a gender bias.

References:

Hu, Trumble, Foley, Eberhard, and Wallace (1998)
Hunt, Morris, Edwards, and Wilson (1993)
Foley, Wallace, and Eberhard (1995)
Johansson (1997)
Johansson, Bronge, Lundberg, Persson, Seideman, and Viitanen (1996)
Keyl, Rebok, and Gallo (in press)
Lindal and Stefansson (1993)
Marottoli, Cooney, Wagner, Doucette, and Tinetti (1994)
McKnight and McKnight (1998)


IA2(f). Navigation Errors on Road Test


Summary:

In a study of 75 subjects ages 60-91 who were referred to CA DMV for reexamination (26 of whom were identified as probably being cognitively impaired to some degree), and 31 volunteers ages 56-85, cognitively impaired referrals had significantly more "confusion errors" than cognitively nonimpaired referrals. Confusion (concentration) errors occurred when subjects were unable to proceed to the field office at the end of the drive test, or drove past the street on which the field office was located and did not recognize their error. This particular measure was the only on-road driving performance measure where there was a difference between the performance of cognitively impaired and cognitively nonimpaired drivers. (See Janke and Eberhard, 1998; Janke and Hersch, 1997).

McKnight and McKnight (1998) evaluated the on-road driving performance of 402 drivers age 62 and older. Approximately two-thirds of the subjects were referred to the licensing agency for reexamination based upon reports of deficient driving incidents, and the balance were incident-free volunteers. The road test was based on the Driver Performance Evaluation (DPE) developed by the California Department of Motor Vehicles (see Hagge, 1994). Navigation tasks included remembering a series of directions (turning at named streets and following a sequence of turns) and maintaining spatial orientation in order to drive around a block. The instructions given to subjects for the location-finding task were, "Please proceed until (name street) and turn left/right onto (name street again)." The directions given to travel around the block in order to end up at a specified location and traveling direction were, "In a moment, I'll ask you to make a right turn. When I do, please turn right and then make a series of right turns around the block, ending up on this same street, going in the same direction." The correlation between navigation errors and unsafe driving incidents was significant (r=0.41). The incident-involved drivers performed more poorly than the incident-free drivers on the on-road navigation tasks.

Conclusions/Preliminary Recommendations:

A destination-finding task should be included in on-road driving tests tailored to detect possible cognitive impairment among older drivers who are referred for reexamination, or to determine the extent to which cognitive impairment has progressed to the point where driving is not recommended (as in the third-year post Alzheimer's disease onset).

References:

Janke and Eberhard (1998)
Janke and Hersch (1997)
McKnight and McKnight (1998)

 

IA2(g). Discriminating Maneuver Errors on Road Test


Summary:

Older and cognitively impaired drivers, like all drivers, commit many common errors both during the stage of information acquisition and in the execution of vehicle control movements that appear to have little bearing on the likelihood of crash involvement--or rather, that the variance that can be accounted for by differences in these behaviors will always be lower than that accounted for by situational factors (Staplin, Gish, Decina, Lococo, and McKnight, 1998). In the study by Staplin et al., almost all of the older drivers (n=62) failed to look both ways before entering intersections to execute a through maneuver during the green (permissive) phase, and instead, treated their movement as one that was protected. Such "common," or nondiscriminating errors are therefore poor candidates for the validation of screening indices, or for identifying individuals deserving one sort of intervention or licensing action from another. Dobbs (1997) similarly has advocated the segregation of nondiscriminating from discriminating (or hazardous) errors in the development and application of screening instruments for driving competency.

Dobbs (1997) studied 279 drivers in three groups:

176 patients referred to a clinic with suspected decline in mental abilities (majority were diagnosed with Alzheimer's) with mean age of 72 years;
70 mature healthy drivers volunteered for the research (mean age = 69 years);
33 young healthy controls also volunteered (age range 30-40; mean age = 36 years).

A two-part road test was administered by 2 experienced driving instructors from the Canadian Automobile Association. Testing was conducted in a mid-sized American car equipped with dual brakes. The first part was a closed course on paved streets with curbs, but was undeveloped allowing traffic to be restricted and signs to be placed as desired. The open road test consisted of 37 maneuvers, required 40 minutes to administer, and was conducted on commercial and residential streets, and an urban freeway. Maneuvers were selected to maximize those implicated in older-driver crashes. Some instructions for downstream maneuvers were given; other maneuvers required planning (e.g., a lane change prior to a turn); and some maneuvers required working memory skills (e.g., turn left after two blocks). There was also a "take me to" instruction.

Definition and scoring of errors was as follows.

Hazardous or potentially catastrophic driving errors: errors committed by drivers who are no longer competent to drive (e.g., wrong-way on a freeway, stop at green light), and would result in a crash if examiner did not intervene or traffic did not adjust.

Discriminating driving errors: potentially dangerous errors that signal declining driving skill (e.g., poor positioning on turns and straightaways, observational and scanning errors, and overcautiousness).

Non-Discriminating driving errors: errors made equally often by good and bad drivers, reflecting bad habits as opposed to declining ability (e.g., "rolling" stops and speed errors). Drivers are not penalized for non-discriminating errors. Discriminating errors are documented and scored in terms of their severity (5, 10, or 51 points).

Hazardous errors were renamed as Criterion errors and their commission results in an automatic fail. A combined criterion of one or more criterion errors and/or discriminating point total exceeding criterion, results in a failure on the road test.

Using the joint criterion, all of the young normal drivers passed the road test, approximately 95 percent of the mature control group drivers passed the road test, and only 25 percent of the cognitively impaired (patient) group passed the road test.

In McKnight and McKnight's (1998) study that compared the on-road driving performance of incident-involved and incident-free older drivers [see Notebook section IA2(f)], the incident-involved drivers did more poorly on the following measures: intersection visual search (sharing attention); path maintenance through turns; maintaining a constant speed; positioning the car at intersections and merges; and navigating correctly. They also tended to err on the side of over-caution by driving slowly through turns, on straight stretches, and when changing lanes, as well as rejecting safe gaps at intersections.

Researchers who have compared the driving performance of cognitively-impaired (mild dementia) older drivers and healthy older controls have found that older cognitively impaired drivers make the following errors (Hunt, 1991; Hunt, Morris, Edwards, and Wilson, 1993; Hunt, Murphy, Carr, Duchek, Buckles, and Morris, 1997a, 1997b; Cooper, Tallman, Tuokko and Beattie, 1993; Dobbs, 1997; Janke and Hersch, 1997):

Stopping at green lights
Making sudden stops for no apparent reason
Coasting to near stop in moving traffic
Failure to check blind spot
Delay in changing lanes when an obstacle appeared
Drifting into other lanes
Wrong lane prior to left or right turn
Wrong lane after left or right turn
Impulsive and unsafe left turn
Attempted left turn when not allowed
Attempted left turn on red
Inappropriate decision-making ('judgment') in traffic
Failed to yield right-of-way
Misinterpretation of traffic signs
Failure to move over or stop for ambulance
Collisions or near collisions on hazard avoidance tasks
Collisions/near collisions with median
Wrong-way maneuvers
Getting lost in familiar areas
Require repeated step-by-step directions
Require verbal cues to signal when changing lanes throughout the driving task
Signaling late (when they did signal)
Driving while pressing the brake and accelerator simultaneously
Failing to realize why other drivers honked at them

Conclusions/Preliminary Recommendations:

Driving errors demonstrated by cognitively impaired older drivers differ from the types of errors that many drivers, both good and poor, commit (bad habits as opposed to cognitive decline). Therefore, road tests developed to determine driving competency (older driver re-exams) should include the conditions and maneuvers shown to be problematic to drivers with cognitive decline, and scoring of errors (number and severity) should be such that drivers are not penalized for making errors that do not discriminate impaired from unimpaired drivers. The test must be traffic interactive, performance based, and examine cognitive behaviors.

References:

Cooper, Tallman, Tuokko, and Beattie (1993)
Dobbs (1997)
DrivAble Testing, Ltd. (1997)
Hunt (1991)
Hunt, Morris, Edwards, and Wilson (1993)
Hunt, Murphy, Carr, Duchek, Buckles, and Morris (1997a, 1997b)
Janke (1994)
Janke and Hersch (1997)
McKnight and McKnight (1998)
Staplin, Gish, Decina, Lococo, and McKnight (1998)


IA2(h). Decision-Making and Response Selection in Driving Simulators


Summary:

(Note: See Notebook section IC2(b)iv for a description of driving simulators)

Schiff and Oldak (1993) found performance differences (EasyDriver) between 109 older subjects (ages 55-95) and 61 younger subjects (ages 15-54) that included:

Slower driving speeds by older subjects, particularly in the poor visibility conditions and under headlight glare conditions;
Longer (but not significant) simple reaction time (RT);
Longer RT's to traffic events such as braking in response to lead vehicle brake lights, a pedestrian, and the basketball (dusk) scenarios;
Late braking by 40-90 year olds in response to a school bus pulling into their lane; and
Lack of response by a substantial number of older subjects to the tennis ball and basket ball (dusk) scenarios.

Using GAR score as a criterion, multiple regression analyses were performed to determine which scenarios would best predict driving performance. A Global Accident Risk (GAR) score was the dependent measure, which consisted of the total number of reported at-fault crashes for each driver, with the addition of up to 3 more points for self-reported medical or driving problems (dizziness, attentional lapses, severe arthritis, poor vision, and poor vehicle control). The resulting range of scores was 0-13. Regression analysis were performed separately for older and younger subjects using 65 years as the criterion age split. For the older subjects, RTs from hit pedestrian, tennis ball, basketball (dusk) and city brakes yielded an R=.47, accounting for 22 percent of the variance in GAR scores. For young subjects, schoolbus, hit pedestrian, and tennis ball yielded an R=.41, accounting for 16 percent of the variance.

In Szlyk, Brigell, and Seiple's (1993) study of 6 subjects with hemianopic visual field deficits (ages 53-80, mean 71 years) and 7 older controls (ages 62-83, mean 70), simulator performance measures of effectiveness (MOEs) included: mean speed (in mi/h); average slowing and stopping to traffic signals; number of lane boundary crossings; mean break pedal pressure; mean gas pedal pressure; number of simulator crashes; lane position; steering angle and vehicle angle to the road. Six staged driving simulator challenges required visuocognitive/motor skills to avoid a crash; three of these were intersections with cross traffic. Two of the four older subjects who had real-world crashes also had the longest slowing times, the longest stopping times, and the most crashes in the driving simulator.

In a study 82 older subjects ages 60-91 (26 of whom were identified as probably being cognitively impaired to some degree) who were referred to the CA DMV for reexamination, the proportion of errors on simulator trials where the driving video (MultiCAD) showed a threat vehicle entering the driver's path from the periphery at 15 degrees (divided attention trials) was significantly correlated with weighted error score on an on-road drive test (r=.2430, p<.043). A gross measure of the number of errors made in the driving video (angular motion sensitivity trials) significantly correlated with weighted error score on the road test (r=.3462, p< .002). In addition, the correlation between proportion of errors on trials where brake lights were visible and weighted error score on the drive test was significant (r=.2801, p<.013). (See Staplin, Gish, Decina, Lococo, and McKnight, 1998; Janke and Hersch, 1997)

Conclusions/Preliminary Recommendations:

Ecologically valid stimuli (realistic views of the driving environment) yield predictive assessments of the cognitive and visual motor components required in driving. A simulation of apparent motion of self through a three-dimensional environment (even if simulated on a two-dimensional screen) which contains the visual scene complexity associated with the actual driving environment is important for simulator measures for predicting actual driving performance. Simulators are recommended for pre-testing drivers recovering from strokes, cerebral vascular accidents, and those with progressive cognitive disorders, to determine their progress and whether it is safe to assess them on the road. They may also be beneficial in highlighting risks for drivers, who may not acknowledge diminished capabilities, and as an educational tool in a rehabilitation environment.

References:

Janke and Hersch (1997)
Schiff and Oldak (1993)
Staplin, Gish, Decina, Lococo, and McKnight (1998)
Szlyk, Brigell, and Seiple (1993)


I.A. IDENTIFY OLDER PEOPLE WHO ARE AT HIGH RISK OF CRASHES


I.A.3. Avoidance of High Risk Situations and Other Compensatory Behaviors


Summary:

Data from the 1990 Fatal Analysis Reporting System (FARS) and the 1990 Nationwide Personal Transportation Survey (NPTS) show that people age 75 and older are involved in more fatal crashes per mile driven than people of any other age group (Massie and Campbell, 1993; Massie, Campbell, and Williams, 1995). But, because they drive relatively few miles each year, their fatal involvement rate per licensed driver is only slightly above the overall rate. While the per capita fatal involvement rate for people age 75 and older is lower than for people of all ages combined, this may be explained in part by the fact that relatively lower percentages of people in older age cohorts hold valid licenses (approximately 50 percent of women age 75+ and 80 percent of men age 75+, compared to 84.5 percent of all women and 92 percent of all men in the population of driving age).

At the same time, an analysis of driving and travel patterns between 1983 and 1990 showed that drivers in age categories 65 and older drove at least 30 percent more in 1990 than in 1983, at an annual increase of 4 percent. Older drivers continued to concentrate their driving between 9:00 a.m and 4:00 p.m. Analysis of the fatality rates by day and night showed that the highest daytime rates were for drivers age 75 and older, while the highest nighttime rates were for drivers age 16-19. For older drivers, the nighttime rate is 1.1 times the daytime rate, while for the youngest drivers, it is 6.1 times the daytime rate (Massie and Campbell, 1993; Massie, Campbell, and Williams, 1995).

A panel data analysis found that although annual miles driven is the single most influential risk factor in crash involvement for older male and female drivers, the influence of mileage on the likelihood of being involved in vehicle crashes is significantly smaller in men than in women (Hu, Trumble, Foley, Eberhard, and Wallace, 1998). For female drivers, the amount of annual driving and limitation in gross mobility (inability to raise arms above shoulder height) were the two significant risks in older women being involved in crashes. For males, being employed and cognitively disabled, having a history of glaucoma, and using anti-depression drugs amplify the likelihood of being involved in vehicle crashes. Use of antidepressants by male drivers is the second most important risk next to the amount of annual driving, doubling the risk compared to drivers who do not use antidepressant drugs. After controlling for the amount of annual driving, men who are cognitively impaired (low score on word recall test), are 40 percent more likely to be involved in a crash than men who are not; cognitive ability is irrelevant in older females being involved in crashes.

In driving habits surveys, older drivers report driving fewer miles and avoiding demanding driving situations compared to younger drivers (Tallman, Tuokko, and Beattie, 1993; Janke and Eberhard, 1998; Gutman and Milstein, 1998). In one study, drivers with the highest avoidance scores were those who performed most poorly on an on-road exam, but avoidance score did not discriminate between cognitively impaired and unimpaired drivers (Janke and Hersch, 1997). In another study, drivers who were more visually and/or cognitively impaired tended to report more avoidance and less exposure (e.g., avoid night driving, high-traffic roads, rush-hour traffic, high-speed interstates, driving alone, making left-hand turns across traffic, driving in the rain; and reported driving fewer days per week); however, relationships between mental status and the avoidance items were weaker than those between visual function and avoidance (Ball, Owsley, Stalvey, Roenker, Sloane, and Graves, 1998). In this study, older drivers with cataracts (n=83) reported more avoidance of driving on high-traffic roads, in rush-hour traffic, on high-speed roadways, alone, and in the rain than drivers with no eye health problems (n=126); however, drivers with cataracts did not report higher levels of avoidance of driving at night and making left turns. Older drivers with age-related macular degeneration (n=19) reported higher overall avoidance than the no-eye-disease group for all avoidance categories. Older drivers with multiple impairments (visual and cognitive) restricted their driving to a larger extent and in more situations than those with visual impairments alone, or those who were functionally normal. Drivers with higher numbers of crashes in the prior 5-year period reported more avoidance of driving in the rain, making left turns, and driving during rush hour.

In a sample of 3,238 drivers age 65 and older who were administered a battery of visual and cognitive assessment tests, the prevalence odds of reduced driving exposure were higher for the cognitive function variables than for the visual function variables, and higher for males than for females (Stutts, 1998). Men who scored in the lowest quartile on the Trail-Making A and Short Blessed tests (cognitive measures) were 6 to 9 times more likely to report driving less than 3,000 miles per year than men scoring in the highest quartiles, and women with low scores were three times more likely to report driving less than 3,000 miles than women with higher scores. The effect of reduced high-contrast visual acuity was greater at higher age levels than at lower age levels. A model developed to predict high risk avoidance (not driving after dark, during rush hour or in heavy traffic, on expressways or interstate highways, on busy multi-lane roads, in rain, or other bad weather conditions) found that the cognitive and visual function measures were associated less strongly with avoidance of particular driving situations than with an overall reduction in mileage. Also, in this model, the odds ratio for the cognitive function measures were only slightly higher than those for the visual function measures.

Although Marottoli and Richardson (1998) found that individuals who drove more miles were more likely to rate themselves as being better drivers than their peers, results of their study showed that on-road driving performance and history of adverse driving events were not related to drivers' ratings of self confidence in their driving ability. The subjects in their study were 125 active older drivers age 77 and older, 40 percent of whom reported a history of adverse driving events. In terms of self-ratings of driving ability, none of the 125 participants rated themselves as being worse than other drivers. Of the 50 participants with a history of an adverse driving event, 34 (68%) rated themselves as being better or much better than other drivers their age; this is identical to the proportion of individuals who rated themselves as better or much better and had no history of adverse driving events. In addition, all nine individuals who were rated by a driving therapist as having moderate or major difficulties on a road test, rated their driving ability at least as good as their peers, and 3 of the 9 rated their ability as better or much better than their same-age peers. Of the 125 drivers, 34 (27%) had a discrepancy in their self-rating of ability (i.e., they had adverse driving events or were rated by a driving evaluator as being poor drivers, but they rated their driving ability as better than that of their peers). The authors state that this indicates a lack of awareness, as these drivers may exceed their limitations and place themselves and others at risk.

Turning to a consideration of whether older drivers know when to stop driving, Stutts, Wilkins, and Schatz (submitted) found that older drivers think they will be the first to know when they should stop driving, and most seniors have not considered the possibility that they may not realize when it is time for them to stop driving. The majority of the focus group participants indicated that seniors do not plan for the possibility that they could outlive their driving ability. This information was obtained through focus group discussions with 44 older drivers who had recently stopped driving (half of the group) or believed that they may stop driving within two years.

Stutts et al. also reported that men are particularly reluctant to stop driving, and often deny any deterioration in their driving skills. Some seniors continue to drive "in spite of everything," regardless of physician recommendations against driving and injury-producing, at-fault crashes. On the other hand, there is a subset of older drivers, typically women, who give up driving prematurely. Generally, these drivers never really enjoyed driving, are uncomfortable in today's driving environment, and have a spouse who drives. Although an event like a hospitalization may trigger their decision to stop driving, often they just drive less and less until they no longer feel comfortable behind the wheel.

Wilkins, Stutts, and Schatz (submitted) conducted one-hour, on-road evaluations of eight senior women who wanted to drive more, but had either voluntarily stopped driving or voluntarily drove infrequently (once per week or less). Subjects were screened by telephone to eliminate those who had a vision or other health problem that prevented them from driving more. The evaluations were provided at no cost to the participants, and were conducted by a certified driving instructor under the auspices of a local driving school. Driving evaluations began at each woman's home. The instructor completed a standard evaluation form (Miller Road Test), and provided each woman verbal feedback describing her driving performance, and whether additional practice and/or driving lessons were recommended. When contacted for a telephone follow-up interview, all of the women described the evaluation as a useful experience, and several indicated that it had given them confidence in their driving ability; these women indicated that they were driving more as a result of the evaluations. All three of the women who had previously ceased driving indicated that they planned to resume driving, at least enough to maintain their skills. The authors state that although it is unknown how much the women would be willing to pay for an evaluation and lessons, such a countermeasure would help keep older women on the road safely, longer than they would without such intervention. Education for older drivers as a rehabilitation procedure is described in Notebook section IC3a(i).

Conclusions/Preliminary Recommendations:

Mileage-based crash risk increases with age, and this risk can be offset to some degree by self-regulation (driving less frequently and fewer miles, and under less demanding conditions). Many older drivers (who are aware of diminished abilities) compensate for age-related functional declines at the strategic level by planning to avoid rush hour or nighttime driving, and at the tactical level by adjusting speed (driving slower) and accepting larger gaps at intersections. Sensory and physical declines are easier to identify and compensate for (and potentially correct) than are cognitive declines. But, recent research has shown that the prevalence of undetected eye disease increases with age (Decina and Staplin, 1993; Shipp, 1998). Possibly more serious is the driver with diminished cognitive decline. Drivers with dementia overestimate their capabilities (Cushman, 1992) and may not restrict their driving to times and situations that reduce risk (they don't compensate because they are not aware of their decline). Janke and Eberhard (1998) found that the amount of avoidance reported in a driving habits questionnaire did not discriminate between cognitively impaired referral subjects and cognitively unimpaired referral subjects. Also, drivers who have no access to alternative transportation and who live alone may be more likely to drive in situations, even when they realize they are at higher risk; reports from older driver focus groups consistently indicate that when there is no choice but to drive to get to doctor appointments, grocery shopping, etc, they will do so.

As reported by Stutts (1998), while approximately half of the drivers in the lowest quartiles of performance on a cognitive function test (Trail-Making A and B), reported driving less than 3,000 miles per year, the other half of this population is driving over 3,000 miles per year, and 20 percent of the entire sample reported driving more than 10,000 miles per year. While many older drivers with cognitive and visual impairments limit their driving exposure, self-regulation alone does not adequately protect the public's health.

A program that provides materials to help older drivers assess their own capabilities and provides tips for reducing driving risk must be accompanied by a coordinated effort that includes health-care professionals, individuals in the community who come in contact with older persons, DMV counter personnel, and law enforcement officers to ensure that older drivers remain safely mobile. In addition, driving evaluations and on-road lessons may help provide confidence in driving ability for older drivers who are fit to continue to drive but cease or restrict driving prematurely.

References:

Ball, Owsley, Stalvey, Roenker, Sloane, and Graves (1998)
Cushman (1992)
Decina and Staplin (1993)
Gutman and Milstein (1988)
Hu, Trumble, Foley, Eberhard, and Wallace (1998)
Hu, Young, and Lu (1993)
Janke and Eberhard (1998)
Janke and Hersch (1997)
Marottoli and Richardson (1998)
Massie and Campbell (1993)
Massie, Campbell, and Williams (1995)
Ranney (in press)
Shipp (1998)
Stutts, Wilkins, and Schatz (submitted)
Stutts (1998)
Stutts, Stewart, and Martell (1996)
Tallman, Tuokko, and Beattie (1993)
Wilkins, Stutts, and Schatz (submitted)


Continue to next section | Return to Table of Contents