Effect of Navigation Problems,Assessment Location,and a Practice Test on Driving Assessment Performance for People with Alzheimer’s Disease

Abstract

Background:

People with Alzheimer’s disease may be required to undertake clinical and on–road assessments to determine fitness to drive. The manner in which on–road assessments are conducted with drivers who do and do not have navigational problems may affect the outcome.

Objectives:

Investigate the effect of 1) navigational difficulties, 2) location of assessment (un/familiar area) and assessment order, and 3) undertaking a second assessment (practice), on passing an on–road driving assessment.

Methods:

Forty-three drivers undertook an Occupational Therapy-Driver Assessment Off Road Assessment (OT-DORA) Battery which included the Drive Home Maze Test (DHMT). Participants with/without a history of navigational problems were randomly allocated into three groups: 1) Unfamiliar/then familiar area assessment; 2) Unfamiliar/unfamiliar; 3) familiar/unfamiliar. An on–road assessment protocol was used including over 100 expected behaviors at nominated points along the directed route. For familiar area assessments, the driver self-navigated from their home to shops and services. A pass/fail decision was made for each assessment.

Results:

A generalized linear mixed effects model showed neither location, nor practice affected passing the on–road assessment. Participants with navigational problems were six times less likely to pass regardless of route familiarity and direction method, and the DHMT was a significant negative predictor of passing.

Conclusion:

Drivers with Alzheimer’s disease who have navigational problems and are slow to complete the DHMT are unlikely to pass an on–road assessment. However, navigation and maze completion skills may be a proxy for an underlying cognitive skill underpinning driving performance.

Keywords

Alzheimer’s disease automobile driver examination automobile driving occupational therapy

INTRODUCTION

For many individuals, successful aging-in-place and lifestyle maintenance relies on being able to drive a car. However, normal age-related changes in vision and reaction times combined with medical conditions associated with aging such as stroke and arthritis have the potential to impact on safe driving. One of the most widely studied conditions that impacts driving is dementia [1 –3]. There are currently over 5.5 million Americans living with dementia [4, 5] of which the most common types are Alzheimer’s disease (AD), vascular dementia, Lewy body disease, and frontotemporal dementia [6]. This figure is expected to reach over 80 million by 2050 [7]. This research focuses on people with AD, which accounts for approximately 70% of people with dementia.

Cognitive impairment, as experienced in people with AD, is an important factor leading to driving cessation [8], and while a portion of drivers will need to relinquish this activity, there is also a need to support those who can drive safely to continue to do so for as long as possible [2]. The prevalence of individuals with probable AD who drive has recently been estimated at around 23% [9]. Since driving is an overlearned skill, it may be possible for some individuals with AD to continue to drive safely for some time after a diagnosis and undergo periodic re-assessment to ensure the continued safety of the individual and the community [1, 10]. Early seminal research by Friedland and team [11] found that many drivers can continue to drive for as long as 4 years after a diagnosis of dementia. Drawing on Michon’s hierarchy of driving behaviors [12], this can be explained by the fact that operational (e.g., motor ability to drive a vehicle) and even many tactical behaviors and skills (e.g., stopping at red signals) are overlearned and can be performed automatically. However, as noted by Dickerson and colleagues [13], it is the strategic level skills, which include problem solving, planning, and navigation, that are more challenging for drivers. Therefore, people with dementia are able to ‘automatically’ drive the vehicle at operational and tactical levels and can devote their cognitive capacity to the ever-changing road environment and strategic driving. It is most likely that it is when a driver’s cognitive skills have declined to the point where they cannot operate at a strategic level, that safe driving is no longer possible. Making decisions regarding fitness to drive based on a diagnosis of AD and degree of cognitive impairment in the absence of an on–road assessment where a determination of how a client is performing at an operational, tactical, and strategic level is made, may lead to premature driving cessation or continuation of unsafe driving [1 , 14].

Many people who have a new or suspected diagnosis of AD are referred to an occupational therapy driver assessor (OTDA). Internationally, the guidelines regarding fitness to drive for people diagnosed with dementia vary, but generally acknowledge the need for occupational therapy assessment if a practical assessment is required. The US Driver Fitness Medical Guidelines National Highway Traffic Administration [15], page 13, state “A driver who is identified as having a cognitive problem may be required to undergo certain tests that will evaluate problems associated with driving. The primary test is the on–road evaluation but there are tests that are administered in an off-road setting that are also used to evaluate certain cognitive functions necessary for safe driving. These evaluations may be conducted by the Department of Motor Vehicles or by other professionals such occupational therapists or driving rehabilitation specialists.” A significant increase in the demand for occupational therapy driver assessment and rehabilitation services which cater for the needs of people with dementia is anticipated due to the projected incidence of dementia internationally.

Comprehensive occupational therapy driver assessment involves both a clinical driving evaluation and on–road assessment [16, 17]. The purpose of the clinical driving evaluation is to establish the client’s sensory, physical, and cognitive strengths and weaknesses that may impact driving [17] and to identify any clients who would be unsafe to take on the road. The on–road assessment is usually undertaken in a dual controlled vehicle with a trained driver instructor present. The role of the driver instructor is to give directions and take control of the vehicle if safety is compromised. Following completion of the assessment, and depending on regulatory requirements, a recommendation concerning fitness to drive is made. The recommendation may include suspension or cancellation of the license, issuing of a conditional license (for example to drive within a restricted local area if the assessment has been conducted in the driver’s familiar area), or to retain a full license. In the jurisdiction where this research was conducted, recommendations made by OTDAs are enacted in 90% of cases by the licensing authority.

Many OTDAs anecdotally believe that a driver’s performance may be significantly better in their familiar area than in an unfamiliar area. The key differences between the two types of assessment are outlined in Table 1. If the driver is successful in their familiar area only, then an area restriction is placed on their license. Lovell and Russell [16] examined the value of routine familiar area reassessment for drivers with dementia who had been referred for testing by their physician or family member to the licensing authority. Twenty participants with a diagnosis of dementia were recruited from a memory clinic and undertook a standard occupational therapy driver assessment. While 15 participants failed the initial open on–road assessment, 10 went on to pass assessments in their familiar area. However, there is little evidence to support whether an improved performance on subsequent testing is related to a practice effect rather than the location of the assessment. While most OTDAs offer a 5 to10 min familiarization drive at the start of the assessment prior to scoring client performance, many OTDAs question if clients do better if they have a practice session of 40–50 min. A Canadian study examined the impact of an education program on driver’s performance on an on–road assessment [18] and found that all of the older participants (although none had dementia) improved on the second assessment conducted two months after the initial assessment, regardless of involvement in the education program. They concluded that practice can improve test outcome. Similarly, Leung et al. [19] found that the opportunity to practice reduces drivers’ levels of anxiety, which is known to impact negatively on performance.

Table 1

Key differences between familiar area and unfamiliar area on–road assessments

Key difference	Unfamiliar area assessment	Familiar area assessment
Where the assessment commences from	Driver clinic	The participant’s home
Who gives directions during the assessment (navigation requirements)	Driver instructor. Participants with navigational difficulties may find this an advantage	Participant self directs. Participants with navigational difficulties may find this a disadvantage
Route familiarity	Unfamiliar	Familiar: Participant drives only to familiar and regularly driven local places
License type awarded for pass outcome	Unrestricted license	Restricted license (boundaries set to local familiar area)

In addition to the question of whether improved performance is due to location or practice effects, none of these studies have taken into account the impact of navigational problems on drivers’ performance. For example, if a driver has navigational problems, their performance may be significantly better in an unfamiliar area assessment where directions are given, in comparison with a familiar area assessment where the driver needs to self-navigate. Navigation involves the preparation and decision-making processes involved in planning to travel along a route from one point to another [20] and therefore occurs at the strategic level of Michon’s Model [12]. Navigational difficulties may be associated with poorer driving performance as a person cannot get to where they want to go [21]. In addition to the declines associated with increased age, dementia is likely to compound the diminished abilities experienced by elderly drivers due to deficits in spatial memory and cognitive mapping as well as failing to reference landmarks as navigational guides [22]. These navigational abilities are thought to be primarily subserved by hippocampal circuits, consistent with the components of neural circuitry compromised over the course of AD [21]. In their review paper, Lithfous and colleagues [22] concluded that navigational deficits are observed in individuals with AD over and above general cognitive decline attributable to the disease. Although research has been undertaken on the validity of directed versus self-navigated driving during on–road assessments [23], we could not locate any research specifically examining if people who have navigational problems are more likely to fail an on–road driving assessment.

This research aimed to investigate the impact of different testing procedures used by occupational therapists and other driver rehabilitation specialists when conducting on–road assessments with drivers who have AD. Specifically, the research aimed to determine the effect of: 1) navigational difficulties; 2) location of assessment (unfamiliar area or familiar area) and ordering of the assessments (unfamiliar area assessment first or second); and 3) opportunity to undertake a second assessment (practice) on passing or failing an on–road driver assessment, among people with AD.

METHODS

Research design

This study used a stratified cross-over design. Stratification of participants was initially based on self and significant other report of navigational difficulties. Although the Santa Barbara Sense of Direction (SBSD) scale [24] can reliably identify self-reported navigational problems in a normal population, this scale is quite lengthy and relies on memory and insight for accurate reporting and was therefore unsuitable for use in the current study. Participants underwent assessments on a rotating system to ensure occupational therapy driver assessors (OTDAs) were not biased from the results of the clinical driving evaluation as to the possible outcome of the on–road assessment (as detailed below). The overall outcome of each driver assessment was classified as pass or fail.

Participants

Consecutively admitted participants consented to this study from the Austin Health driving clinics. Participants were invited to the study if they met the following inclusion criteria: 1) diagnosis of AD (made by a physician specializing in geriatric medicine); 2) current holder of an unconditional driver license; and 3) currently driving, defined as reporting driving at least once a fortnight.

Participants were excluded from the study if they had previously undertaken a fitness to drive assessment, required an interpreter, or had a history of other medical conditions which may impact driving performance such as a stroke, Parkinson’s disease, or amputation.

Procedure

Ethics approval was obtained from the Institutional Review Boards from both Central Queensland University and Austin Health. Although all drivers in the study had been required to undergo a driver assessment by the licensing authority, they also volunteered to participate and were consented into the study. Data were collected by OTDAs with a qualified driver instructor in a vehicle with dual controls. Participants were given a choice of manual or automatic transmission, and underwent clinical driving evaluation (up to 90 min) and two on–road assessments (approximately 45 min). Participants with and without a history of navigational problems, respectively, were randomly allocated into one of the three groups, where each assessment was conducted on a separate day. Each participant underwent a clinical driver evaluation and then:

Group 1: unfamiliar area assessment, familiar area assessment.

Group 2: unfamiliar area assessment, another unfamiliar area assessment. If the participant did not pass the second unfamiliar area assessment, a familiar area license assessment was offered. This was not required for the research, but was clinically required to ensure participants had a fair and typical experience of being offered a familiar area assessment if they failed an unfamiliar area assessment.

Group 3: familiar area assessment, unfamiliar area assessment.

The final possible group of familiar area assessment, followed by another familiar area assessment was not used since the effect of practice was assessed in the other groups. A rotating system was used to ensure therapists were not biased from the results of the clinical driving evaluation as to the possible outcome of the on–road assessment, as follows: Therapist A- clinical driving evaluation, Therapist B- first on–road, and Therapist C- second on–road. Therapist A conducted a third on–road (if required) and collated reports, made the final recommendation, and provided feedback to the driver. The same driver instructor and car were used for all on–road assessments for each participant.

Measures

As no objective measure of navigational problems could be located that was suitable to use with clients with potential memory and insight deficits, the presence of navigational problems was determined by asking the driver, their family/significant others, and reviewing referral source information concerning three problem areas: 1) experience of getting lost when driving, 2) being gone longer than expected, and 3) not being able to identify where the car was parked. If three or more instances of way finding difficulties were identified across the three areas and three sources of information, the client was classified as having navigational problems. It was reasoned that one or two instances of a problem may indicate typical difficulties any older person might encounter. In this study, almost all participants identified as having navigational problems had three or more instances of problems. Two clients had one instance only of a problem and these clients were not deemed to have navigational problems. The clinical driving evaluation was conducted using the Occupational Therapy-Driver Assessment Off Road Assessment (OT-DORA) Battery [17,25, 17,25]. The OT-DORA Battery includes collection of information concerning car driving and medical history, and then comprehensively assesses sensory, physical, and cognitive skills over a 90-min period. These skills are assessed through a range of standardized tests included in the Battery such as the Confrontation Visual Field Tests [26], and Test of Proprioception Screen [27] for sensory skills, Motricity Index (MI) [28], Simulated Accelerator Brake Test (SABT) [25], and Right Heel Pivot Test (RHPT) [25] for motor skills, and the Road Law and Road Craft Test (RL&RCT) [29,30, 29,30], Mini-Mental Status Examination (MMSE) [31], and the Drive Home Maze Test (DHMT) [10,32,33 , 10,32,33] for cognitive skills. The DHMT requires drivers to complete a simple maze on a standard US letter page, and the completion time, including any planning time, is recorded. Completion time varies with age and disability; however, a sample of 34 healthy men with a mean age of 30.25 years (SD 13.13) completed the DHMT in an average of 20.81 s (SD 6.27) [29]. Completing the DHMT requires visuo-constructional, executive functioning, planning, decision making, and attention skills, which are also are also required for safe driving. Although the DHMT does not specifically identify if a client has navigational problems, the test does seem to provide an indication that a person is experiencing difficulty with skills such as visuo-constructional skills which are also required for navigation.

During the on–road assessment, occupational therapist driver assessors used the standard on–road assessment protocol used in the licensing jurisdiction where the research was conducted to document driver behaviors and to describe any driver instructor interventions. This protocol includes over 100 expected behaviors at nominated points along the route, for example, ‘check mirrors and signal prior to turning’. The familiar area assessments commenced at participant’s homes and the assessment route was mapped out in advance (based on discussion with the client) and required the client to drive to their local shops, local doctor, and one or two other local services (depending on their distance) so that the whole route would require 40 min of driving. Client familiarity with the area was ascertained through having lived in the area for a minimum of 3 years (all clients met this criteria), and driving at least every 2 weeks to shops and services. The unfamiliar area assessments commenced at the driver assessment clinic and were conducted in an area unfamiliar (or of very low familiarity) to the driver as ascertained through discussion. In some instances, a driver needed to commence the test at the clinic which may have been familiar, and was then directed to an unfamiliar area. Unfamiliar area assessments also included a small self-navigational component where drivers were asked at a set point on the route to navigate and pull in at the next gas station, which was approximately a mile away on the same straight stretch of two-lane highway. This was used to confirm if clients had been placed correctly in the navigational or no navigational problems group. The unfamiliar area test also required the client be directed to shops and services and to park the car; however, the drive included 2 and 3 lane highways, complex intersections with controlled and uncontrolled left turns (driving on the right side of the road), and managing merging traffic. In contrast, a familiar route usually only included single lane traffic, and simple controlled and uncontrolled intersections. The driver assessors collaborated on the final outcome decision by drawing on all information from the clinical driving evaluation and on–road assessments and provided feedback on the overall result to participants within a day of their final assessment. The outcome for each participant was documented as pass or fail, and the licensing authority was notified of the final outcome, as per usual jurisdiction requirements.

Data analysis

Data were analyzed using the R statistical programming environment [34]. A generalized linear mixed effects (GLME) model was used to analyze the repeated measure outcome via a random intercept for participants with a logistic link, using the lme4 package [35]. Initially, the three independent variables of prime interest in the research were entered into the model to guide clinical practice on whether OTDAs should test clients for navigational problems, which area to test clients in (familiar or unfamiliar), and whether clients benefit from practice. McNemar’s tests on paired binomials (a repeated measures form of chi square) were then used to investigate if any of the scored motor or cognitive tests from the OT–DORA Battery (MI, SABT, RLRCT, MMSE, and the DHMT), could further explain on–road test outcomes. The outcomes for the three groups were as follows:

Group 1: pass unfamiliar area/ pass familiar area OR fail unfamiliar area/ pass familiar area;

Group 2: pass unfamiliar area/ pass unfamiliar area OR fail unfamiliar area/ fail unfamiliar area (then pass or fail familiar area); and

Group 3: fail familiar area/ pass unfamiliar area OR pass familiar area/ pass unfamiliar area.

RESULTS

Data were collected from 43 drivers with a diagnosis of AD over a three-year period. Participants’ ages ranged from 60 to 86 years (M = 77.00; SD = 6.69) with 31 (72.1%) males and 12 (27.9%) females. Eleven participants (25.6%) were identified as experiencing navigational problems and 32 participants reported no difficulties with navigation. For the 11 participants with navigational problems, five had been randomly allocated into Group 1, one client into Group 2, and five clients into Group 3.

From the OT-DORA Battery, participants’ scores on the MMSE [31], ranged from 14 to 28 (M = 22, SD 4.06), and scores on the Drive Home Maze Test (DHMT), ranged from 11.03 to 206.00 s (M = 67.02, SD 57.48). A total of 9 drivers failed overall (21%) and 34 drivers passed (79%), but some of these drivers failed an open area assessment and then went on to pass a familiar area assessment. Therefore, it was deemed more accurate to review the total number of on–road assessments taken, and pass and fail results for each. The 43 drivers undertook a total of 93 drives as follows: 36 drivers undertook two on–road assessments, and 7 drivers undertook three on–road assessments as they were randomized to undertake two unfamiliar area assessments and failed both and then needed to complete a familiar area assessment (thus requiring a total of three assessments). The comprehensive occupational therapy driver assessments were undertaken over as short a period of time as possible, allowing for client schedules such as holidays. From the first clinical evaluation through to the final on–road test (either two or three tests) took a minimum of 15 days and a maximum of 99 days with a mean of 42 days (SD 21.765). There were no differences between drivers who had navigational difficulties in terms of age (t(41)=1.661, p = 0.948), and the only clinical evaluation test variables that were just significantly different for drivers with and without navigational problems were MMSE (t(28) =−3.641, p = 0.056), and DHMT (t(41) = 1.968, p = 0.057).

Applying the GLME model, Table 2 demonstrates that neither location familiarity (including the ordering of the assessments) nor undertaking a second assessments (practice) had an effect on the probability of passing the on–road assessment. There was no interaction effect between navigation group and practice X²(1) = 0.08, p = 0.778. However, there was a significant effect for navigational problems. Participants with navigational problems were 1/ exp (-1.793) = 6.01 times less likely to pass than those without, after controlling for order and location familiarity. This is further supported by a cursory viewing of Table 3, where 29% (n = 20/70) of clients with no navigational problems failed, as opposed to 57% (n = 13/23) of clients with navigational problems. However, a more detailed review of Table 3 reveals that clients with navigational problems failed almost equally if they were tested in an unfamiliar area (directed) or familiar area (self-navigating). This suggests that drivers with navigational problems do not do better when directed compared to when directed in an unfamiliar area, and therefore appear to have another underlying skill deficit that is affecting performance. It must also be noted that the variable of navigational problems is confounded by area familiarity as clients can self-direct in familiar areas but must be directed in unfamiliar areas.

Table 2

Generalized Linear Mixed Effects Model showing the effect of variables on passing the on–road driving test

	Dependent variable
	Pass
Practice	−0.349
	(−1.363, 0.665)
Location–Unfamiliar Area	−0.506
	(−1.638, 0.626)
Navigational Problems	−1.793**
	(−3.265, −0.321)
Constant	1.960*
	(−0.004, 3.924)
Observations	86
Log Likelihood	−52.082p<0.1; *p<0.05

Table 3

Summary of outcomes for 43 drivers with and without navigational problems (this included a total of 93 drives as follows: 36 drivers undertook 2 tests, and 7 drivers who undertook 3 tests as they were randomized to undertake 2 unfamiliar area assessments and failed both and then needed to complete a familiar area assessment for practical purposes)

	No Navigational Problems		Navigational Problems
	Familiar area Self-directed	Unfamiliar area Directed	Familiar area Self-directed	Unfamiliar area Directed
Fail	4/26	16/44	7/12	6/11
on–road Assessment	15%	36%	58%	55%
Pass	22/26	28/44	5/12	5/11
on–road Assessment	85%	64%	42%	45%

As there were too many missing data points to enter all variables into a regression analysis (for example, a missing MMSE score would be missing three times if the driver underwent three on–road tests), each of the scored clinical driving tests from the OT–DORA (e.g., MI, SABT, RLRCT, MMSE, and the DHMT), were then examined individually to determine if any variable in isolation or in interaction with navigational problems, was predictive of the on–road assessment outcome. Only the DHMT time was a significant negative predictor of the probability of someone passing the on–road assessment (standardized B = −1.14, z = −2.219, p = 0.012), with slower (higher) time scores predictive of fail. For each standard-deviation increase in DHMT time taken, the participant was 1/exp (−1.1454) = 3.14 times less likely to pass the on–road assessment. A model including an interaction term between DHMT and navigational problems did not improve the model fit X²(1)=0.02, p = 0.880, indicating that the predictive utility of the DHMT did not differ between drivers with and without navigational problems.

DISCUSSION

Forty–three drivers undertook a total of 93 on–road assessments in this study. It was found that neither location (including the ordering of the assessments) nor undertaking a second assessment (practice) had an effect on the probability of passing the on–road assessment. However, participants with navigational problems were six times less likely to pass the on–road assessment than those without, regardless of their familiarity with the area and whether they were self–directed or directed. It was also found that one of the clinical assessments, the Drive Home Maze Test (DHMT) was a significant negative predictor of someone passing the on–road assessment for all drivers both with and without navigational problems. For each standard-deviation increase in DHMT time taken, the participant was three times less likely to pass the on–road assessment. Given that drivers with navigational problems were almost equally likely to fail the on–road test regardless of whether they were directed or self–directed, and the lack of an interaction effect between navigation problems and DHMT, it seems that identifying if the client has navigation problems and the DHMT are both detecting an underlying cognitive/ visuospatial skill that is impacting on safe driving that we are not yet directly assessing.

In this study, 58% of drivers with navigational problems failed the familiar area assessment, compared to 15%, who had no known navigational problems. A familiar area (self-directed) assessment route requires the driver to be actively engaged in multiple tasks such as visual scanning, visuospatial awareness, divided attention, planning, and judgment which is thought to require considerably more cognitive skill than following directions from a driver instructor, as occurs during an unfamiliar area assessment [13, 23]. In the research undertaken by Lovell and Russell [16], the highest number of driver instructor interventions was also reported when drivers were undertaking the self-navigated component of the on–road assessment.

However, 55% of drivers with navigational difficulties also failed when they were given directions (a similar percentage to drivers with navigational problems who self-directed) and 36% of drivers without navigational problems who were directed similarly failed the unfamiliar area assessment. Although navigational skill assessed in this manner is always confounded by familiarity with the area, this result seems to suggest that the fail rate is not related to navigation skills nor familiarity but that problems with some other cognitive skills/s are limiting safe driving, such as visuospatial awareness, divided attention, and/or planning [22].

It is therefore possible that current comprehensive driver assessments are not directly assessing the skill/s that are preventing these participants from safe driving and further research is required to identify and assess these underlying constructs. However, in the short term, having information about a driver’s navigational skills and capacity to complete the DHMT is providing us with ‘proxy’ information that can be used clinically. A reliable and valid assessment of navigational problems that can be used with people who have AD requires development, but until this occurs drivers and significant others can be questioned about navigation problems as was done in this research. Clients can choose to undertake an unfamiliar area on–road assessment (which needs to include a section that is self-navigated) or a familiar area on–road assessment (which is entirely self-navigated) and the client can be licensed accordingly. The inclusion of a self-navigated component during an on–road assessment is routinely used in OTDA on–road assessments and challenges the client at the strategic and tactical levels in Michon’s model and highlights deficits which may contribute to unsafe driving [13]. The current practice of offering clients an unfamiliar area assessment followed by a familiar area assessment can also continue. OTDAs can have increased confidence that when these two assessments are undertaken, there is very little practice effect on the driver’s performance.

The DHMT was found to be a significant predictor of on–road assessment outcome, as has been previously shown in other samples [10, 32]. This test has high compliance among drivers given the face validity of showing the route for a car to drive home. Scores of over 100 s have been reported to predict a fail outcome in a general sample of clients with mixed diagnosis [17], and scores over 63 s have previously predicted a fail outcome among drivers with AD [10]. In this research, only three drivers failed who had scores faster (lower) than 90 s.

One of the limitations of this study was that the factors of familiarity with route and whether the route was directed by the driver instructor or self-directed, were confounded. These factors are difficult to tease apart on the road, as drivers must be directed in unfamiliar areas, although a small self-navigated component (drive to the nearest gas station) can be included. Therefore, we cannot be certain if a driver’s performance was influenced by how familiar they are with the environment or whether they have navigational difficulties. As noted above, development of a clinical evaluation of navigational skills in the future may assist tease out these factors. Other limitations of this study include the relatively small number of participants identified as having navigational problems. In addition, we had to ask participants and their family if they had navigational problems as we could not locate a standardized, objective test of navigational problems. While we were able to test for the impact of navigational problems, location and order of tests as well as practice, we were not able to add the test results from the OT-DORA battery into an overall regression model due to several missing data points. With a more complete data set, it is also possible that some of the other clinical evaluation tests from OT-DORA such as the RL&RCT or the MMSE may be predictive of on–road test outcome. Furthermore, this research included only drivers formally diagnosed with AD. Although this is the largest group of people with dementia, the research reported in this paper requires replication with drivers who have other forms of dementia and also with drivers who have been identified as having mild cognitive decline, but do not yet have any diagnoses.

CONCLUSION

Ensuring the independence and safety of individual road users with deteriorating health, together with the road safety of all road users is an important goal for our community. It is vital that drivers in the early stages of AD who are able to continue to drive safely are supported to maintain this independence [1, 2]. The aims of this research were to determine the effect of order of test and test location on pass or fail on–road assessment outcome for drivers both with, and without, navigational problems. The results suggest that opportunity to have a second assessment (practice), and the ordering of the tests had no effect on the probability of passing. Clinically, these are valuable findings as OTDAs can continue to offer a short familiarization drive knowing that neither further practice nor test order will affect the outcome. It was also found that drivers who had navigational problems and slower DHMT were less likely to pass the on–road driver assessment. However, the navigational results were confounded by driver familiarity with the test location. Drivers with navigational problems who self-navigated in a familiar location were equality likely to fail as when they were directed in an unfamiliar location. Drivers who had no navigational problems were more likely to fail in an unfamiliar route than when they were directed, suggesting that its only helpful to be in a familiar area for drivers without navigational problems. Since this research was not able to fully tease apart navigation skills from driver familiarity with an area, further research is required to identify a way to separate and measure these constructs. Further research to develop a standardized objective assessment of client navigational skills in a clinical environment may assist in identification of clients who will fail an on–road assessment. However, there may be another underlying cognitive or visuospatial skill that is being indirectly assessed when identifying a client’s navigation skills and completing a maze test such as the DHMT and further research to identify and directly assess this factor (or complex interaction of factors) is required.

Footnotes

ACKNOWLEDGMENTS

This work was supported by a grant from the Alzheimer’s Australia -Victoria Dementia Research Foundation. Thanks are extended to the occupational therapy driver assessors and clients who volunteered to participate in the research.

Authors’ disclosures available online ().

References

Dobbs

, Carr

, Morris

(2002) Evaluation and management of the driver with dementia. Neurologist 8, 61–70.

Barco

, Baum

, Ott

, Ice

, Johnson

, Wallendorf

, Carr

(2015) Driving errors in persons with dementia. J Am Geriatr Soc 63, 1373–1380.

Austroads (2016) Assessing fitness to drive. Sydney.

Alzheimer’s Association (2017) Alzheimer’s Association Annual Report, Alzheimer’s Association, Chicago, IL.

Institute for Dementia Research and Prevention (2017) Frequently Asked Questions. http://www.idrp.pbrc.edu/faq.htm. Accessed on August 20, 2018.

Brodaty

, Woodward

, Boundy

, Ames

, Balshaw

(2011) Patients in Australian Memory Clinics: baseline characteristics and predictors of decline at six months. Int Psychogeriatr 23, 1086–1096.

, Goodkind

, Kowal

(2016) U.S. Census Bureau, International Population Reports, P95/16-1, An Aging World: 2015, U.S. Government Publishing Office, Washington, DC, 2016. https://www.census.gov/content/dam/Census/library/publications/2016/demo/p95-16-1.pdf. Accessed on November 24, 2016.

Anstey

, Windsor

, Luszcz

, Andrews

(2006) Predicting driving performance over 5 years in older adults. Psychological well-being and cognitive competence are stronger predictors than physical health. J Am Geriatr Soc 54, 121–126.

Amjad

, Roth

, Samus

, Yasar

, Wolff

(2016) Potentially unsafe activities and living conditions of older adults with dementia. J Am Geriatr Soc 64, 1223–1232.

10.

Unsworth

, Chan

S-P

(2016) Determining fitness to drive among drivers with Alzheimer’s disease or cognitive decline. Br J Occup Ther 79, 102–110.

11.

Friedland

, Koss

, Kumar

, Gaine

, Metzler

, Haxby

, Moore

(1998) Motor vehicle crashes in dementia of the Alzheimer’s type. Ann Neurol 24, 782–786.

12.

Michon

(1985) A critical view of driver behavior models: what do we know, what should we do? In Human behavior and traffic safety, Evan

, Schwing

, eds, Plenum Press, New York.

13.

Dickerson

, Davis

, Staplin

(2014) Can clinicians, researchers, and driver licensing officials build a shared vocabulary? Occup Ther Health Care 28, 188–193.

14.

Berndt

, Clark

, May

(2008) Dementia severity and on road assessment; Briefly revisited. Australas J Ageing 27, 157–160.

15.

National Highway Traffic Safety Administration. (NHTSA) & American Association of Motor Vehicle Administrators (AAMVA) (2009) Driver Fitness Medical Guidelines https://www.ems.gov/pdf/811210.pdf. Accessed on August 20, 2018.

16.

Lovell

, Russell

(2005) Developing referral and reassessment criteria for drivers with dementia. Aust Occup Ther J 52, 26–33.

17.

Unsworth

, Baker

, Taitz

, Chan

, Pallant

, Russell

, Odell

, Coulson

(2012) Development of a standardised Occupational Therapy Driver Off Road Assessment (OT-DORA) battery to assess older and/ or functionally impaired drivers. Aust Occup Ther J 59, 23–36.

18.

Bedard

, Isherwood

, Moore

, Gibbons

, Lindstrom

(2004) Evaluation of a re-training program for older drivers. Can J Public Health 95, 295–298.

19.

Leung

, Deanne

, Taylor

, Bilokas

(2009) Anxiety in driving assessment of individuals with cognitive impairment. Disabil Rehabil 31, 1700–1708.

20.

Golledge

(1992) Place recognition and wayfinding: making sense of space. Geoforum 23, 199–214.

21.

Head

, Isom

(2010) Age effects on wayfinding and route learning skills. Behav Brain Res 209, 49–59.

22.

Lithfous

, Dufour

, Despres

(2013) Spatial navigation in normal aging and the prodromal stage of Alzheimer’s disease: Insights from imaging and behoral studies. Ageing Res Rev 12, 201–213.

23.

Mallon

, Wood

(2004) Occupational therapy assessment of open-road driving performance: validity of directed and self-directed navigational instructional components. Am J Occup Ther 58, 279–286.

24.

Hegarty

, Richardson

, Montello

, Lovelace

, Subbiah

(2002) Development of a self-report measure of environmental spatial ability. Intelligence 30, 425–447.

25.

Unsworth

, Pallant

, Russell

, Odell

(2011) OT–DORA Battery: Occupational Therapy Driver Off-Road Assessment Battery, AOTA Press, Bethesda, MD.

26.

Elliot

, North

, Flanagan

(1997) Confrontation visual field tests. Ophthalmic Physiol Opt 17, S17–S24.

27.

Cox

(1989) Elderly drivers’ perceptions of their driving abilities compared to their functional motor skills and their actual driving performance. Phys Occup Ther Ger 7, 51–82.

28.

Collin

, Wade

(1990) Assessing motor impairment after stroke: A pilot reliability study. J Neurol Neurosurg Psychiatry 53, 576–579.

29.

Unsworth

, Pallant

, Russell

, Germano

, Odell

(2010) Validation of a test of road law and road craft knowledge with older or functionally impaired drivers. Am J Occup Ther 64, 306–315.

30.

Unsworth

, Pallant

, Russell

, Odell

, Coulson

(2011) Inter-rater reliability of the Road Law and Road Craft test as part of the OT-DORA Batter for off-road Driver assessment. Br J Occup Ther 74, 394–398.

31.

Folstein

, Folstein

, McHugh

(1975) Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

32.

Krishnasamy

, Unsworth

(2011) Normative data, preliminary inter rater reliability and predictive validity of the Drive Home Maze Test. Clin Rehabil 25, 88–95.

33.

Krishnasamy

, Unsworth

(2012) Validity of the Occupational Therapy Drive Home Maze Test for right and left handed test takers. Accid Anal Prev 49, 274–277.

34.

Team RC (2000) R language definition. R foundation for statistical computing, Vienna, Austria.

35.

Bates

, Sarkar

, Bates

, Matrix

(2007) The lme4 package. R package version 2, 74.