| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||
Departments of 1 Internal Medicine and 5 Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut.
2 Geriatrics and Extended Care Section, 3 Cooperative Studies Program Clinical Epidemiology Research Center, and 4 Cooperative Studies Program Coordinating Center, VA Connecticut Healthcare System, West Haven, Connecticut.
Address correspondence to Richard A. Marottoli, MD, MPH, Yale University School of Medicine, Section of Geriatrics, 20 York Street, TMP 15, New Haven, CT 06504. E-mail: Richard.Marottoli{at}ynhh.org
 |
A |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. This randomized controlled trial with blinded endpoint assessment enrolled 126 community-living drivers 70 years old or older who were recruited from clinic and community sources. Treatment assignment was concealed until eligibility was established. Participants randomized to intervention underwent two 4-hour classroom and two 1-hour on-road sessions focused on common problem areas of older drivers. Controls received modules directed at vehicle, home, and environmental safety. A knowledge test and driving performance were assessed at baseline and 8 weeks. On-road driving performance was assessed by an experienced evaluator in a dual-brake–equipped vehicle in urban, residential, and highway traffic. Driving performance was rated on a 36-item scale with potential scores from 0 to 72 (higher score better). The knowledge test included 20 road knowledge and eight road sign questions, scored from 0 to 28 correct.
Results. The least squares mean change in road test score relative to baseline was 2.87 points higher in the intervention than in the control group (p =.001). The least squares mean change in knowledge test scores relative to baseline was 3.45 points higher in the intervention than in the control group (p <.001).
Conclusions. An education program consisting of classroom and on-road training targeted to common errors of older drivers enhanced performance on knowledge and on-road tests. Such interventions offer older drivers the potential to continue driving safely longer and to maintain their out-of-home mobility.
Thus, identifying ways to enhance driving capability would be a valuable tool to maintain safety and out-of-home mobility. Earlier work in this area has focused on identifying individuals at risk for driving difficulties (12–16), although few interventions have been developed targeting specific risk factors (17,18). However, the benefit of these interventions is limited to individuals who have the risk factor in question. Focusing instead on directly improving driving capabilities may benefit a broader range of individuals.
Few studies have evaluated the effectiveness of classroom-based or on-road driving training for older individuals. McKnight and colleagues (19) found no statistically significant differences in the number of crashes between groups of older drivers who had or had not received classroom training, although there was a low response rate to crash questions. Janke (20) detected a slight increase in state-recorded crashes, but a decrease in violations, among individuals receiving classroom training. Bedard and colleagues (21) found no difference in on-road driving performance following a classroom program, but did find improvement in driving performance in a follow-up study using both classroom and on-road instruction (22).
The current study used a randomized controlled design to determine whether an intervention including classroom and on-road instruction could enhance the driving performance of individuals with driving difficulties at baseline.
|
METHODS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Participants
Participants recruited from the general medical clinics of the VA Connecticut Healthcare System (VACHS; 84% of participants) site had their baseline assessment and started their on-road assessment at VACHS. Participants recruited from a variety of community sources (prior studies, senior housing complexes, senior centers, health fairs, flu clinics, exercise groups, advertisements; 16%) had their baseline assessment and started their on-road assessment at the university site. A single road route through New Haven and West Haven, Connecticut (with entry points at VACHS and the university site), was used.
Inclusion criteria were age 70 years or older, driving at least once a week, valid Connecticut driver's license, English speaking, having a phone, absence of medical conditions that might deteriorate during the course of the study (e.g., dementia and other neurodegenerative disorders, metastatic cancer) or acute medical illness at the time of screening, binocular distance visual acuity of 20/70 or better, Mini-Mental State Examination score of 24 or better (23), no driver education course in the past year, and an on-road assessment score between 40 and 65 (on a 72-point scale).
Baseline Assessment
Participants underwent a baseline assessment of health, functional status, sociodemographic factors, and driving practices developed in earlier studies (12,13,18). Initial inter-rater reliability among research associates was assessed along with periodic quality control checks during the study. Inter-rater reliability on a sample of 15 volunteers was excellent with 
values for categorical variables from 0.76 to 1.0 and intraclass correlation coefficients for continuous variables from 0.89 to 1.0.
Health factors ascertained included chronic conditions, medication use, and alcohol use. Elements of function assessed included vision, cognition, and physical ability. Sociodemographic factors included age, gender, and education. Driving practices included driving frequency, mileage, driving circumstances, and self-reported adverse driving events (12). A knowledge test was administered including 20 road knowledge questions from the AAA Driver Improvement Program (24) and eight road sign questions used in our earlier studies, summed for a possible total score of 28 correct.
Assessment of Driving Performance
The road test was based on the Connecticut Department of Motor Vehicles test and assessed a wide range of driving abilities in different settings including: an off-road portion (parking lot maneuvers); low, medium, and high traffic density areas; and highway segments. There were two entry points depending on recruitment site. Parameters assessed included speed, lane changes, merging, observance of signs and signals, interaction with traffic, and operation of vehicle controls. A nonhighway alternative was used for individuals who did not drive on highways. Evaluations were performed by one of two experienced driving assessors in a dual-brake vehicle over the same route, at the same time of day, under reasonable weather conditions. Participants were graded on each of 36 components on a 0- to 2-point scale ranging from poor to intermediate to safe performance with standardized criteria for rating each item (25). Individual item scores were summed to create a possible total score ranging from 0 (worst) to 72 (best). Inter-rater reliability of the on-road assessment was established in a sample of 10 older volunteer drivers. Weighted values for the 36 elements ranged from 0.34 to 1.0 (all but two were > 0.58), and the overall intra-class correlation coefficient for the total driving score was 0.96.
Based on our earlier study (18), a road test score between 40 and 62 was required for study eligibility in order to exclude potentially unsafe drivers or better drivers unlikely to benefit from the intervention. The scores of initial participants were monitored. As the population and distribution of scores differed from the earlier study, the upper limit was raised to 65 early in the study, still allowing for detection of the target change of four points with intervention.
Intervention
Participants randomized to the intervention group received a combination of 8 hours of classroom and 2 hours of on-road instruction.
Classroom instruction.-- The content of the classroom instruction sessions was based on the AAA Driver Improvement Program (24), the literature, and common on-road errors encountered in our earlier study (18). In that study population, the most common driver errors included: inadequate scanning to sides/rear; not using seat belts, mirrors, or directional signals; not maintaining safe following distances; problems backing up; and poor left turns, lane changes, and speed regulation. Both classroom and on-road training sessions provided strategies to counteract or compensate for these problems. Classroom sessions focused on chapters of the AAA Driver Improvement Program (24). Classroom topics included: driving risk, developing good visual habits, communication, adjusting speed, safety margins, driving emergencies, vehicle features, alcohol, medications, and aggressive drivers.
On-road instruction.-- On-road instruction was based on the literature and common errors encountered in our earlier study, supplemented with specific techniques for enhancing safety. The first session dealt with errors made by the driver in the baseline assessment. The second session reviewed these points and addressed common errors of older drivers in general.
Areas addressed in on-road training included: seat position, mirror adjustment, visibility, safety belts, viewing strategy, speed control, safe following distance, stopping distance, vehicle position, lane changes, turning, signaling, intersections, right of way, K-turns, backing up, highway driving, and communication. Assessment and on-road training routes were distinct, but had similar features and traffic patterns. On-road training was performed by one individual experienced in on-road instruction and distinct from the on-road driving assessor and the AAA-certified classroom instructor.
Control
Two modules were presented by a trained research assistant to control participants. The modules were accompanied by materials written at an eighth grade level and illustrated with simple line drawings. The home/environment safety module was conducted in participants' homes and included formal home safety evaluations covering topics like fall and trip hazards, lighting, handrails, throw rugs, chair design, and different locations like bathrooms, kitchens, bedrooms, and stairs. The vehicle safety module covered topics like tire pressure, lights, mirrors, vehicle maintenance, emergency equipment, crime prevention, and pedestrian issues.
Outcomes
Both intervention and control groups repeated the on-road assessments and knowledge tests at 8 weeks. The assessors were blinded to treatment group. After the last intervention or control session and again just before the final assessment, participants were reminded to not mention anything about their group assignment or experiences in the intervening 8 weeks. The assessors were instructed to not ask about group assignment and to quickly redirect participants if any discussion suggested group identification. The primary outcome was change in driving performance at 8 weeks compared to baseline. Change in knowledge test scores at 8 weeks compared to baseline was the secondary outcome.
Informed Consent
The study was approved by the Yale School of Medicine Human Investigation Committee and the VACHS Human Subjects Subcommittee. Written informed consent was obtained from all participants.
Statistical Methods
Randomization.--
Randomization was stratified by recruitment site using a permutated block scheme and equal allocation to treatment groups. Because treatment was given in groups, participants were randomized into groups of up to seven to the two treatments arms. Treatment allocation was assigned only after eligibility was established.
Sample size.-- Based on our earlier study, a clinically meaningful difference in the primary outcome was defined as a change of four points (18). The sample size to detect this difference between intervention and control groups was 126 participants for a type I error of 5% (two-sided), 80% power, road test score standard deviation of 7.5 points (prior study data), and 10% adjustment for losses.
Statistical Analysis
All analyses were conducted according to original treatment assignment. The primary outcome was the change in road test score from baseline to the 8-week follow-up measurement. Eight participants did not complete follow-up driving evaluations, mainly due to driving cessation. Therefore, two analyses were conducted: one excluding participants with missing data (complete case analysis) and one using multiple imputation methods to replace missing road test scores (intention-to-treat analysis). For imputation analyses, 20 complete data sets were generated and combined using standard methods (26,27).
A linear mixed regression model was used to analyze the effect of treatment (intervention relative to control) on the primary outcome for both the complete case and intention-to-treat analyses. Treatment comparisons were adjusted for study design—recruitment site (VACHS vs university) and road test examiners at baseline and follow-up—and baseline road test score (28). Because enrollment occurred over several years, to control for possible temporal effects on driving frequency and mileage, enrollment date was included in the multivariable model. Because treatment sessions were given in groups of participants, the model included a random intercept and the nesting of individuals within instructional groups. Treatment effects were summarized as least squares means (i.e., means that have been adjusted for covariates), and statistical significance was assessed by the t statistic. Model goodness-of-fit was assessed using information criteria, residual analysis, and regression diagnostic tools.
The effect of treatment on the secondary outcome, the knowledge test, was analyzed similarly. Because 12 participants assigned to intervention and 22 assigned to control had missing knowledge tests, multiple imputation was used to replace missing values. Most participants had missing knowledge tests early in the study because of an administrative oversight that was later corrected. Similar to the analysis of the primary outcome, complete case and intention-to-treat analyses (with imputed values) were conducted. SAS (version 9.1.3; SAS Institute, Cary, NC) was used for all analyses, and a p value of.05 (two-sided) was used for all tests of significance (29).
|
RESULTS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Adherence
All 118 participants who completed the follow-up assessment completed all 8 hours of classroom training and the two 1-hour driving lessons.
|
DISCUSSION |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
The strength of this study is its design, a randomized controlled trial of classroom and on-road instruction with blinded endpoint assessment. In addition, all participants were volunteers from clinic sites and the community at large rather than being referred because of concern for driving difficulties. The intervention was designed to address common errors of older drivers, and a wide range of drivers could potentially benefit from it. Thus, the study findings have broad potential applicability.
There are several limitations of the study. Although the difference of 2.87 points between intervention and control groups achieved statistical significance, the clinical implications of this difference are unclear. In our previous study, a one-point increase in the driving score was equal to a 3.3% decrease in crash occurrence over 2 years of follow-up (our unpublished data). Thus, an improvement of 2.87 points would equal a 9.5% decrease in crash risk. The clinical significance of the 3.45-point, between-group difference in knowledge tests scores will be determined when prospective crash and moving violation data collection on the cohort is completed. As indirect evidence, the correlation between change in knowledge test scores and change in road test scores was 0.34 (p <.001), suggesting that improvement in knowledge test scores was associated with improvement in road test scores. In addition, because of the modest sample size, it is difficult to distinguish individual driver factors that may have contributed to the between-group difference. Another limitation was the considerable overlap (by design) between the content of the classroom and on-road sessions, thereby making it difficult to determine the extent to which each contributed to the improvement detected.
Despite these limitations, the findings are encouraging. Although much has been made of the potential safety risk of older drivers and the factors that contribute to this risk, relatively little attention has been paid to strategies to enhance driving performance. Using an easy to implement and broadly applicable intervention, we demonstrated a statistically significant improvement in driving performance among intervention participants relative to controls. Future studies need to determine those elements of the intervention that produced the greatest effect and the participants who received the most benefit in order to optimize the efficiency of the intervention. In addition, potential synergistic effects with other interventions, such as those directed at drivers with limitations in physical ability that have been found to enhance driving performance (18), need to be explored.
Summary
The current findings offer encouragement to a broad spectrum of drivers that a relatively easy to implement intervention can enhance their driving performance and potentially prolong their safe driving years, thereby maintaining their activity level and mobility. The availability of effective interventions will help to encourage drivers and clinicians to engage in discussions about this important safety issue.
|
A |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
We thank Sharon Jennings for her assistance in preparing the manuscript.
|
F |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Received February 14, 2007
Accepted July 6, 2007
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
