Do Impulse Control Disorders Impair Car Driving Performance in Patients with Parkinson’s Disease?

Abstract

Background:

Based on data regarding the prevalence of Parkinson’s disease (PD), the prevalence of impulsive control disorders (ICD) in PD, and the percentage of PD patients driving a car, it has to be assumed that at least 50,000 PD patients with ICD in Germany actively drive a car. However, these patients might be at risk for unsafe driving due to ICD-related dysfunctions such as failure to resist an impulse or temptation, to control an act or other altered neurobehavioral processes.

Objective:

This study determines the influence of ICD on driving ability in PD.

Methods:

We prospectively compared driving simulator performance of 23 PD patients with and 23 matched patients without ICD. ICD had to be socially compensated and presence was defined clinically for primary and questionnaire-based (QUIP-RS) for post-hoc analyses. Furthermore, between-group comparisons of driving-relevant neuropsychological tests were executed.

Results:

Except from a lower blinking frequency when changing lanes, overall driving safety of patients with ICD did not differ significantly from those without—regardless of the clinical or QUIP-RS-based ICD definition. ICD severity did not correlate with driving performance, but the latter correlated significantly with mean reaction times and certain neuropsychiatric tests (MoCA, TMT-A, TAP-M “flexibility” and DBQ “error”).

Conclusion:

Clinically compensated ICD does not seem to impair driving safety in PD patients. Rather, cognitive and attentional deficits appear to be clinical markers for driving uncertainty.

Keywords

Parkinson’s disease impulse control disorders car driving driving safety

INTRODUCTION

For many patients with Parkinson’s disease (PD), driving a car is essential to maintain at least residual mobility and independency. Higher age, impaired cognition or vision, and, to a lesser extent, severe motor impairment and disease duration have been described as relevant subject-related risk factors for impaired driving in PD [1 –5]. Medication-related risk factors for driving in PD are fatigue, daytime sleepiness, and sudden onset of sleep [3 , 6–8].

Impulse control disorders (ICD) might also impair driving ability in PD patients [9]. A common key symptom of ICD is the failure to resist an impulse or temptation, to control an act or specific behavior, which is ultimately harmful to oneself or others and interferes in major areas of life functioning [10]. ICDs occur frequently as adverse effects of dopaminergic replacement therapy (DRT) and are more likely related to dopamine agonists (DA) than to levodopa (LD) therapy [11]. Especially tonically elevated dopamine levels and stimulation of D2 receptors have been found to prevent learning from negative decision outcomes, resulting in impulsive behavior and impairment to “hold the horses” [12]. In line, preclinical studies demonstrated that dopamine receptor D2/D3 agonists impair behavioral processes such as risk-taking behavior or preference for uncertainty which likely play a role in ICD. Additionally, altered neurobehavioral processes in PD patients with ICD, such as modified processing of reward or impaired decisional bias are associated with altered activity in cortical and subcortical areas [11]. These likely are reasons for described risky behavior in PD patients [13, 14] and might result in impairment not only “to hold the horses” but also “to hold the horsepower” when driving a car. A few clinical studies showed no or a rather weak negative association of levodopa equivalent dosage (LED) and driving ability [4 , 15]. Self-administered high dosage of LD was described anecdotally as reason for unsafe and compulsive, risk-seeking, “reckless” driving in two male PD patients [16].

ICD have been found with a point prevalence of 13.6 to 19.7% [17, 18] and with an even increased frequency up to 32.9% within a five year follow-up [18]. Given an estimated number of PD patients in Germany of up to 420,000 [19] and considering data showing that 60% of PD patients actively drive a car [20], at least 50,000 PD patients with ICD likely drive a car. This might implicate a relevant driving safety issue. However, a systematically approach assessing the influence of ICD on driving is lacking so far.

With this study we aimed to assess driving performance and safety in PD patients with compared to those without ICD and hypothesized that ICD might impair driving safety.

MATERIALS AND METHODS

Subjects

Two groups of PD patients were recruited from our movement disorders outpatient clinic from February to August 2019: one group with clinical ICD (“PD-ICD”) and one without a present or past history of ICD (“PD-control”). Patients were included into the PD-ICD group in case of self-reported or third-party medical history of ongoing individually abnormal impulse control.

Our clinical definition of ICD based on the American Psychiatric Association’s Diagnostic and Statistical Manual (DSM-5) that included ICDs in the chapter of “Disruptive, Impulse-Control, and Conduct Disorders” as a dysregulation of self-emotional and behavioral control [11]. Accordantly, in our study, patients had to present with at least one typical “behavioral addiction” [11] such as pathological gambling (PG), hypersexuality (HS), compulsive buying/shopping (CB), or binge eating (BE) [10, 18]. Patients might additionally suffer from other ICD-related impulsive–compulsive behaviors (ICB) as consequences of DRT such as dopamine dysregulation syndrome (DDS), punding, hobbyism, walk-abouts, or hoarding [11]. We did not include patients with isolated ICB because they may have different neural substrates and risk factors from typical ICDs [21]. We aimed to investigate subjects with an assumed generally altered underlying pathophysiologic process with higher vulnerability or inclination for impulsive behavior. Therefore, inclusion into the PD-ICD group was based on the presence and not on the severity of the ICD, that was additionally estimated by the screening physician according to a self-developed 3-items Clinical Rating Scale (CRS) as mild (=no or slight impact on social life and with complete compensation), moderate (=interfering impact on social life but with sufficient compensation), or severe (=marked impact on social life but with still acceptable compensation). Expression of ICD in the PD-ICD group had to be socially and relationally compensated, i.e., acceptable for the individual and the social environment. Subjects presenting with socially uncontrolled ICD and urgent need for substantial change of medication were excluded due to ethical reasons. Objective ICD according to the Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease-Rating Scale (QUIP-RS) was additionally assessed [22]. Both groups were matched for age, gender, and motor-symptom severity according to the MDS Unified Parkinson’s Disease Rating Scale, part III (MDS-UPDRS-III). Table 1 shows the detailed inclusion and exclusion criteria.

Table 1

Inclusion and exclusion criteria of the study

Inclusion criteria:
• PD diagnosis in accordance with the “UK Parkinson’s Disease Society Brain Bank Clinical Diagnostic Criteria”
• Valid driver’s license (German Class B)
• Active car driving with an average amount of at least 20 km per week and at least one drive per month during the last 6 months
• Age between 18 and 80 years
• Patients with and without ongoing socially controlled impulse control disorder (ICD) according to [10]
Exclusion criteria:
• Present history of socially uncontrolled ICD with need to acute therapeutical intervention
• Cognitive impairment (Montreal Cognitive Assessment [25] <21/30 points as recommended in [70])
• Severe motor impairment (defined according to the MDS Unified Parkinson’s Disease Rating Scale, part III (MDS-UPDRS-III) with a
• score ≥50)
• Significant visual field loss in the finger perimetry test
• Severe simulator sickness (e.g., vomiting) that rules out further study participation

The studies’ inclusion (upper part) and exclusion criteria (lower part) are presented here.

General procedure

In all subjects, a comprehensive test battery was applied, including tests for cognition, frontal executive function, handedness, non-motor symptoms and quality of life. Furthermore, questionnaires assessing driving-specific aspects were applied (for details see below).

We compared driving ability in a driving simulator between the clinically defined groups PD-ICD and PD-control. Additionally, we performed a similar post-hoc analysis in which the presence of ICD was defined according to the QUIP-RS [22].

Detailed procedure

Oral and written study explanations were given to all participants and written informed consent were obtained. Subjects were told to wear their regular vision aids best for driving. Demographic data and information on medication of both patient groups, such as LED [23] and intake of DA were gathered. Examination was performed in the medication “on” state (1-2 hours after intake of the patients’ individual medication). First, clinical scores and questionnaires were obtained for patient characterization. After technical instruction and a 5-min training session in the driving simulator, all patients performed the default main driving route with certain predefined challenges. Depending on individual speed the driving assessment took approximately 10–15 min. Afterwards, comprehensive neuropsychological tests were applied. Overall, study participation lasted about 2.5 hours with voluntary breaks if necessary.

Clinical scores and questionnaires

The following validated German versions of recommended clinical scores and questionnaires [24] were applied for patients’ characterization before driving challenge:

Edinburgh Handedness inventory

MDS Unified Parkinson’s Disease Rating Scale, part III (MDS-UPDRS-III)

Hoehn & Yahr Score (H&Y)

Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease-Rating Scale (QUIP-RS [22])

Montreal Cognitive Assessment (MoCA [25])

Frontal Assessment Battery (FAB [26])

2nd Revision of the Beck Depression Inventory (BDI-II)

Non-Motor Symptoms Questionnaire and Scale for Parkinson’s disease (PDNMS)

Parkinson’s Disease Questionnaire (PDQ-39)

Driver Skill Inventory (DSI [27])

Driver Behavior Questionnaire (DBQ [28])

Driving Behavior Survey (DBS [29])

Questions #9 and #25 of the Driving Habits Questionnaire (DHQ [30]).

Driving simulator setting

We used the driving simulator Foerst Model F10-P (Dr.-Ing. Reiner Foerst GmbH, Wiehl; https://www.fahrsimulatoren.eu/de/) providing a realistic driver’s cabin of a recreated Ford Fiesta (see Fig. 1). For further information, please see the Supplementary Material.

Fig. 1

Driving simulator. The driving simulator Foerst Model F10-P that was used in our study is shown here during a real-life situation with one of our participants (subject consented to publication) performing the testing session.

We evaluated the following six parameters (per person and drive) provided by the driving simulation system to assess the participants’ individual driving performance:

Mean driving time (mDT) in seconds (s): real time recorded from the default starting to the finishing point in seconds

Mean driving speed (mDS) in km/h: recorded within and outside of town at real time every few milliseconds from the start to the finishing (including queuing or waiting times)

Mean reaction time (mRT) in seconds (s): Time from appearance of an obstacle on the road to applying the brake [31]. As challenges, seven hazardous situations were simulated (e.g., pedestrians or animals crossing). The sum of all RTs was divided by the number of registered RTs.

Driving errors: Twelve typical driving errors were defined (see Table 2). Mean error rate is the average value of counted errors per group and per error category. Each driving error was assessed regarding its safety-critical relevance and weighted with a self-developed “severity factor” from 1 to 8 as a measure for driving error seriousness (see Table 2). For example, safety-critical driving errors endangering other road users were weighted higher than errors with lower impact on safety such as missing shoulder check.

Table 2

Driving errors, definition, and severity factor

Driving error category	Definition^*	Severity Factor
No response to an event – accident	Accident without applying the brakes	8
Accident	Accident with delayed activation of the brakes	7
Endangering other road users	Situation in which a person, animal or object can be harmed	6
Ignoring red light	Ignoring red light on traffic lights	6
Wrong side of the road	All 4 wheels are driving on the other/opposing side of the road	4
Leaving roadway	At least one wheel is crossing the road marking	3
Knocking over marker posts	Collision with a marker post	3
Driving too fast	Exceeding the permitted speed limit with >10 km/h	2
Driving a little slowly	Duration of whole ride longer than one standard deviation (SD) compared to the groups’ means	1^*
Driving way too slowly	Duration of whole ride longer than two SD compared to the groups’ means	2^*
Missing indicator and shoulder check	Activating the indicator and shoulder check are not conducted correctly	2
Missing indicator when making a lane change	Indicator activation is not conducted correctly when making a lane change	1
Missing shoulder check	Shoulder check has to be conducted with a head rotation of at least 20 degrees to the left or right side	1

Table 2 shows different driving errors with their definition and severity factor according to safety-clinical relevance. These seriousness-weighted driving errors are used to calculate the Driving Safety Score (DSS). ^*These two driving error categories were later combined for statistical analysis.

These parameters were detected and recorded automatically (log file) and in the case of errors the driver was warned applying a certain sound.

Driving Error Sum Score (DESS): The DESS is the sum of all driving errors independently of error seriousness.

Driving Safety Score (DSS): The DSS is a composite score considering both, frequency and seriousness of driving errors and serves as main parameter for the overall driving safety. The DSS is calculated as sum of severity-weighted driving errors.

Neuropsychological tests

Next to the driving session, participants had to absolve a battery of neuropsychiatric examinations including the following tests:

The German “Test battery for attentiveness testing” (TAP; [32]) with its version TAP-M (Mobility, vers. 2.1./2007) to computationally assess neurocognitive and attentional capabilities. We evaluated critical parameters of six driving-relevant subdomains: 1) flexibility (number of false reaction and RT), 2) alertness (median of RTs), 3) visual scanning (number of missed critical stimuli and median of RTs to non-critical stimuli), 4) Go-NoGo (number of false reaction and median of RTs), and 5) divided attention (number of missed signals). In these 5 subdomains, T-values <40 were defined to indicate an impaired and T-values >40 an unimpaired TAP performance. When one of the parameters per subdomain was suspicious, subjects were designated as having an impaired performance in this subdomain. For 6) distractibility (number of omissions in the conditions with and without distractors), impairment was assumed when frequency of omissions was higher in distracted compared to non-distracted conditions.

The Symbol Digit Modalities Test (SDMT [33]) to detect cognitive impairment with focus on attentional deficits by mainly measuring individuals’ information processing speed. Normative data base on age, gender, and educational level.

The Trail Making Test (TMT [34]) to screen for cognitive functions. Part A evaluates visuo-motor function and visual processing speed whereas part B assesses working memory, cognitive flexibility, executive functioning, and visual-spatial capabilities. Normative data depend on age and education.

The Wechsler Block-Tapping test (“Blockspanne” = WMS-R block span [35]) forward (fw) and backward (bw) is based on the Corsi block-tapping test and was used to measure the spatial short-term (fw) and working memory (bw). Normative data base on age.

Statistical analysis

Driving simulator data were recorded in a simulator-provided log file and transformed into a SPSS file, into which the results of clinical data, questionnaires/scores and neuropsychiatric tests were included afterwards. IBM SPSS version 18 was used for statistical analysis.

Categorial outcomes were compared using Pearson chi-squared test or in case of ordinal outcomes using linear-by-linear chi-squared tests.

For continuous outcomes, group comparisons were performed using parametric tests (t-test, univariate ANOVA). Error frequencies per error category were analyzed using a Poisson generalized linear mixed model with individuals as random intercepts corrected for driving experience (in km/week). In cases with unequal variances, the non-parametric Welch’s t-Test was used.

Correlations were performed using Pearson’s (linear variables) or Spearman’s (ordered variables) correlation coefficient to investigate associations between test results and driving performance. We abstained from correction for multiple testing as we did an explorative analysis with hypothesis-generating p-values.

Ethics

The study was approved by the local ethics committee of the medical council of Hamburg (reference number: PV3557) and is in accordance with the Declaration of Helsinki.

RESULTS

A total of 50 patients with PD were recruited and screened for eligibility. Four participants had to be excluded from the study: two because of a severe motor impairment with an MDS-UPDRS score >50 and two because of severe simulation sickness (nausea close to vomiting). Finally, 46 PD patients (23 with and 23 without clinically defined ICD) were included in the study. Patients were pair-matched by age and group-matched by gender and MDS-UPDRS III scores. All besides two patients (one left-handed and one ambidextrous) were right-handed. No study participant reported severe visual disorders.

Group differences in clinical scores and questionnaires

Clinical and demographic data as well as questionnaire results are given in Table 3. Both groups differed significantly regarding their driving experience (with more km per week in the PD-control compared to the PD-ICD group, T(44) = 2.459, p = 0.018) and their QUIP-RS ratings with—expectedly—higher scores in the PD-ICD group (T(44) = –4.047, p < 0.001). Concerning the QUIP-RS subscales, all subscales except “hobbyism” and “DDS” differed significantly with higher scores in the PD-ICD group (see Table 3 for statistics). All the other parameters did not show any significant between-group differences.

Table 3

Clinical scores and questionnaire results

Clinical parameters / scores	PD-ICD (n = 23)	PD-control (n = 23)	p
	(Mean±SD [Min- Max])	(Mean±SD [Min- Max])
Age (y)	58.17±8.49 [44–76]	58.19±8.78 [40–77]	0.993
Gender (male (%))	n = 19 (83%)	n = 21 (91%)	0.381^c
Clinical ICD diagnosis (n, %)	Single ICD: n = 17(74%)	–	–
	- PG: n = 8 (35%)
	- HS: n = 3 (13%)
	- CB: n = 3 (13%)
	- BE: n = 3 (13%)
	Multiple ICD: n = 6 (26%)
	- PG/BE: n = 1 (4%)
	- CB/BE: n = 1 (4%)
	- HS/CB: n = 1 (4%)
	- HS/BE: n = 1 (4%)
	- HS/PG: n = 2 (9%)
Physician’s Clinical Rating Scale (CRS) of ICD severity	mild: n = 11 (48%)	–	–
	moderate: n = 6 (26%)
	severe: n = 6 (26%)
Driving experience (km per week)	140.0±100.4 [20–450]	233.0±151.1 [25–700]	0.018 ^*
Edinburgh Handedness Inventory	92.7±11.1 [66–100]	88.0±39.6 [80–100]	0.586
Levodopa Equivalent Dose (LED, in mg)	881.8±512.6	813.7±323.3	0.592
	[100–2351]	[300–1582]
Intake of dopamine agonists (DA)	204.4±154.4 [0–580]	254.5±142.3 [0–600]	0.259
Mean LED of DA only (in mg)
Pramipexole	n = 4 (17%)	n = 12 (52%)	–
Ropinirole	n = 7 (30%)	n = 6 (26%)
Piribedil	n = 2 (9%)	n = 3 (13%)
Rotigotine	n = 10 (43%)	n = 3 (13%)
Apomorphine pen	n = 1 (4%)	n = 1 (4%)
More than one DA	n = 4 (17%)	n = 3 (13%)
No DA intake at study inclusion	n = 4 (17%)	n = 1 (4%)
MDS-UPDRS III	26.2±9.0 [10–44]	24.9±9.4 [7–43]	0.645
Hoehn & Yahr			0.079^b
stage 2, n (%)	9 (39%)	16 (70%)
stage 3, n (%)	13 (57%)	6 (26%)
stage 4, n (%)	1 (4%)	1 (4%)
QUIP-RS_ICD (mean)	17.0±9.4 [3–42]	7.4±6.6 [0–20]	<0.001 ^***
Pathologic Gambling (PG)	2.86±3.87 [0–12]	0.22±0.51 [0–2]	0.004^a ^**
Hypersexuality (HS)	5.39±3.10 [0–12]	3.04±3.26 [0–11]	0.016 ^*
Compulsive Buying (CB)	3.95±3.21 [0–13]	1.26±1.62 [0–5]	0.001^a ^**
Binge Eating (BE)	4.74±3.46 [0–12]	2.73±2.89 [0–8]	0.039^*
Hobbyism	4.6±3.53 [0–11]	3.04±2.65 [0–8]	0.097
Punding	3.86±2.97 [0–11]	1.82±1.99 [0–7]	0.009^**
Dopamine Dysregulation Syndrome (DDS)	4.30±3.87 [0–12]	2.35±2.62 [0–8]	0.052
QUIP-RS (beyond cut-off)	–
QUIP-RS_ICD	n = 18 (78%)	n = 10 (43%)
PG	n = 6 (26%)	n = 0 (0%)
HS	n = 5 (22%)	n = 2 (9%)
CB	n = 2 (9%)	n = 0 (0%)
BE	n = 8 (35%)	n = 2 (9%)
Hobbyism/Punding	n = 13 (57%)	n = 6 (26%)
DDS	n = n.a.	n = n.a.
QUIP-RS_IPQ	29.8±14.6 [3–56]	14.7±11.6 [1–39]	<0.001^***
Montreal Cognitive Assessment (MoCA)	27.9±2.3 [21–30]	28.5±1.5 [25–30]	0.296
Frontal Assessment Battery (FAB)	17.1±1.2 [14–18]	17.2±0.9 [15–18]	0.673
Beyond cut-off	n = 1	n = 0
Beck Depression Inventory, 2nd rev. (BDI-II)	10.3±6.1 [0–22]	9.9±7.6 [0–30]	0.831
No depression	n = 14 (61%)	n = 15 (65%)	–
Mild depression	n = 7 (30%)	n = 6 (26%)
Moderate depression	n = 2 (9%)	n = 1 (4%)
Severe depression	n = 0	n = 1 (4%)
Parkinson Disease Non Motor Symptom Scale (PDNMS)	10.0±4.7 [0–19]	7.9±6.1 [0–22]	0.209
Quality of live (PDQ-39)	23.7±12.8 [5.7–54.6]	17.45±13.15	0.111
Driver Skill Inventory (DSI)	85.65±19.33	87.17±18.252	0.785
Driver Behavior Questionnaire (DBQ)
Lapses	6.78±5.41	7.65±4.811	0.568
Errors	2.91±3.05	1.91±1.83	0.212
Violations	4.87±3.252	4.22±2.67	0.462
Driving Behavior Survey (DBS)
Anxiety-based performance deficits	12.52±4.68	11.48±3.87	0.415
Exaggerated safety and/or heightened caution	22.43±9.01	22.00±7.94	0.863
Hostile/aggressive behavior	10.13±4.01	11.26±4.04	0.347
Driving Habits Questionnaire (DHQ) - Subjective driving quality (question #9)
Poor	n = 0 (0%)	n = 0 (0%)	0.627^b
Fair	n = 1 (4%)	n = 1 (4%)
Average	n = 4 (17%)	n = 3 (13%)
Good	n = 15 (65%)	n = 19 (83%)
Excellent	n = 3 (13%)	n = 0 (0%)
Reported accident rate (question #25)	0–2 accidents / person	0 accident / person	0.061^c
	1 accident: n = 4 (17%)
	2 accidents: n = 1 (4%)

Clinical scores, demographic data and questionnaire results including the driving questionnaires of PD patients with (PD-ICD) and without (PD-control) impulse control disorder (ICD) are given as mean±standard deviation and minimum and maximum, number of patients (n) and percentage (%). Statistical significance is marked with ^*: ^*p < 0.005, ^**p < 0.01, ^***p < 0.001. Most comparisons were performed using t-Tests. ^aResult of Welch-Test. ^bResult of chi-squared Test (linear by linear association). ^cResult of Pearson Chi-squared Test.

Also regarding results of the driving questionnaires (DSI total scores, DBQ subscales, DBS subscales and DHQ questions #9 and #25) no significant group differences could be found between the PD-ICD group and the PD-control group.

Group differences in driving simulator performance

Results of the main driving parameters are given in Table 4 and did not differ significantly between the PD-ICD and the PD-control group (all p > 0.05).

Table 4

Driving performance results

Driving performance parameters	PD-ICD (n = 23)	PD-control (n = 23)	Statistics (F-value, degree of freedom, p)
	(Mean±SD [Min- Max])	(Mean±SD [Min- Max])
Mean driving time (mDT) in ms	575.6±90.3 [423.7–868.1]	597.54±118.6 [443.1–848.5]	F(1,43) = 0.504, p = 0.482
Mean reaction time (mRT) in ms	1.61±0.2 [1.37–1.97]	1.58±0.2 [1.26–2.01]	F(1,43) = 0.159, p = 0.692
Mean driving speed (mDS) in km/h	40.99±5.9 [27.01–54.55]	40.60±7.7 [27.63–54.31]	F(1,43) = 0.051, p = 0.823
Driving errors	See Fig. 2 for details
Driving Error Sum Score (DESS)	3.17±2.04 [0–8]	4.08±3.15 [0–13]	F(1,43) = 1.076, p = 0.305
Driving Safety Score (DSS)	15.52±10.47 [0–38]	20.08±13.35 [0–55]	F(1,43) = 0.782, p = 0.381

Driving simulator results (univariate ANOVA, corrected for driving experience) of PD patients with (PD-ICD) and without (PD-control) impulse control disorders (ICD) are given as mean±standard deviation and minimum and maximum. There were no statistically significant group differences.

Fig. 2

Estimated marginal means of error rates per driving error category per group. Estimated marginal means of error rates per driving error category are given for both groups without (PD-controls, left) and with (PD-ICD, right) impulse control disorders (ICD). *The only significant difference was found for the category “Missing indicator when lane change” (p = 0.029) with no such error type in PD-controls, but in the PD-ICD group (estimated mean = 0.179, CI 0.073–0.437).

General linear mixed model analysis revealed that in both patient groups, most errors belonged to the driving error categories “Accident” followed by “Endangering other road users” (see Fig. 2). There was an interaction group*error category (p < 0.001), but no mean effect of group per se (p = 0.746). The only group difference was found for the error category “Missing indicator when lane change” (p = 0.029) with no such error type in PD-controls, but in the PD-ICD group (estimated mean = 0.179, CI 0.073–0.437).

Regarding the number of patients with or without errors per category, both groups only differed in the “Missing indicator when lane change” condition (chi-squared p = 0.036; Fig. 3).

Fig. 3

Number of patients without or with one or more errors per category. Number of patients with (left, total n = 23) and without (right, total n = 23) impulse control disorders (ICD) that showed no error (light gray) or one/more errors (dark grey). Significant results are marked with * (p < 0.05).

Group differences in neuropsychological tests

We did not find a significant between-group difference in any of the performed neuropsychological tests (all p > 0.05; see Supplementary Table 1).

Correlations between clinical and neuropsychological measures and Driving Safety Score

Table 5 gives an overview of all correlations between the DSS and clinical and neuropsychological data of all participants (n = 46). Significant positive correlations with DSS were found for the driving related tests DBQ subtest “Errors” (p = 0.042), the TMT part A (p = 0.045) and for the mean reaction time in the driving simulator (p = 0.004) meaning that participants with higher DSS composite scores also showed higher errors in the DBQ, needed more time to perform the TMT-A (visual processing) and had longer reaction times in the driving simulator. Negative correlations were found for the MoCA (p = 0.033) and the TAP-M subtest “flexibility” (p < 0.001 1 ), indicating that lower MoCA scores (poorer cognitive function) and more impairment in the flexibility subtest were associated with a higher DSS meaning poorer driving performance.

Table 5

Correlations with Driving Safety Score (DSS)

Parameters	r-value with confidence intervals	p
Age	0.041 (–0.180–0.266)	0.787
Driving experience (km per week)	0.196 (–0.109–0.458)	0.192
Levodopa Equivalent Dose (LED in mg)	0.174 (–0.075–0.495)	0.246
LED of Dopamine agonists in mg	–0.003 (–0.268–0.261)	0.984
MDS-UPDRS III	0.014 (–0.219–0.260)	0.926
QUIP-RS_ICD	–0.118 (–0.353–0.153)	0.433
Montreal Cognitive Assessment (MoCA)	–0.314 (–0.542–(–0.021))	0.033
Frontal Assessment Battery (FAB)	–0.240 (–0.466–0.016)	0.108
2nd Revision of Beck Depression Inventory (BDI-II)	0.025 (–0.275–0.334)	0.870
Parkinson’s disease Non-Motor Symptoms Questionnaire and Scale (PDNMS)	0.145 (–0.290–0.480)	0.338
Driver Skill Inventory (DSI)	0.134 (–0.196–0.446)	0.373
Driver Behavior Questionnaire (DBQ) Lapses	0.140 (–0.205–0.499)	0.353
Driver Behavior Questionnaire (DBQ) Errors	0.302 (–0.104–0.566)	0.042
Driver Behavior Questionnaire (DBQ) Violations	0.259 (0.049–0.499)	0.082
Driving Behavior Survey (DBS) Hostile/aggressive behavior	–0.093 (–0.327–0.166)	0.537
Test battery for attentiveness testing, version mobility (TAP-M)
Flexibility	–0.535 (–0.734––0.287)	<0.001^a
Alertness	0.096 (–0.210–0.390)	0.529^a
Distractibility	–0.033(–0.318–0.232)	0.831^a
Visual Scanning	–0.039 (–0.240–0.312))	0.800^a
GoNogo	–0.249 (–0.536–0.059)	0.100^a
Divided attention	0.136 (–0.176–0.441)	0.372^a
Symbol Digit Modalities Test (SDMT)	–0.139 (–0.389–0.180)	0.362
Trail Making Test (TMT) part A	0.300 (0.002–0.555)	0.045
Trail Making Test (TMT) part B	0.048 (–0.226–0.356)	0.754
Wechsler Block-Tapping test (WMS-R) block span forwards	–0.010 (–0.258–0.222)	0.947
WMS-R block span backwards	–0.250 (–0.497–0.051)	0.097
Mean driving time (ms)	0.263 (–0.131–0.529)	0.078
Mean reaction time (ms)	0.421 (0.209–0.646)	0.004
Mean driving speed in km/h	–0.087 (–0.384–0.271)	0.568

Correlations (Pearson or Spearman^(a)) of the Driving Safety Score (DSS) with clinical and neuropsychiatric data of all participants (n = 46) are shown with r-values, 95% confidence intervals and p-values. Significant correlations are marked bold. LED, Levodopa equivalent dose; MDS-UPDRS III, MDS Unified Parkinson’s Disease Rating Scale, part III; QUIP-RS_ICD, Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease-Rating Scale, ICD subscore.

Furthermore, no significant correlations were found for the LED of Dopamine agonists and the QUIP-RS ICD or IPQ scores or the Clinical Rating Scale (CRS; all p > 0.20).

Correlations between neuropsychological measures and mean driving speed (mDS)

As—especially an increased—mean driving speed (mDS) might be an indirect indicator of an impulse control disorder and because this variable is not included in the DSS, also correlations for mDS were calculated (both groups combined, n = 46). Significant positive correlations were found for the mDS with the MOCA (r = 0.329, CI: 0.047–0.545, p = 0.025) and the SDMT (r = 0.391, CI: 0.097–0.611, p = 0.008) indicating that patients with better cognition and faster information processing speed drive faster.

Significant negative correlations with the mDS were found with age (r = –0.437, CI: –0.635–0.187, p = 0.002), and with the DBQ subtests Error (r = –0.392, CI: –0.618–(–0.045), p = 0.007) and, by trend, with the TMT parts A (r = –0.252, p = 0.095) and B (r = –0.290, p = 0.053) indicating that older PD patients and those who needed more time to perform the TMT A and B, drive slower.

Post-hoc evaluation of driving performance when defining ICD groups with QUIP-RS

For the explorative post-hoc analysis of driving performance, the presence of ICD was defined according to results of the QUIP-RS and grouped participants as follows: 1) Patients with ICD: QUIP-RS scores ≥10 (PD-ICD^score, n = 27) and 2) Patients without ICD: QUIP-RS scores <10 points (PD-control^score, n = 19; see Supplementary Table 3). Comparable to the clinical ICD categorization, this approach did not show significant group differences for any of the evaluated driving parameters (mean driving time with p = 0.376; mean reaction time with p = 0.460; DESS with p = 0.804; DSS with p = 0.933). Even when only comparing participants with very high QUIP-RS scores ≥20 (n = 11) and subjects with values below the cut-off score <10 (n = 19), we did not find any significant group differences regarding the driving parameters (all p > 0.05).

Comparison of clinical and QUIP-RS-based categorization of ICD

In the clinically defined PD-ICD group (n = 23), 18 patients had a QUIP-RS score >10 points, indicating ICD. On the other hand, in the clinically defined PD-control group (n = 23), 8 patients scored >10 points and one patient without clinical signs or history of ICD even scored >20 points in the QUIP-RS score. As we took the comprehensive clinical DSM-5 based diagnosis (clinical examination and history) as gold standard to detect ICD, we found the QUIP-RS to have a sensitivity of 78.3% (18/23) and specificity of 60.9% (14/23) for indicating ICD. When relying on the screening tool QUIP-RS as diagnostic instrument, our clinical diagnosis would have a sensitivity of 85.2% (n = 23/27) and a specificity of 73.7% (n = 14/19) to detect ICD. Thus, the QUIP-RS might be rather over-sensitive with more false-positive results (27 compared to 23 according to clinical assessment).

DISCUSSION

Our study indicates that PD patients with clinically and socially compensated ICD drive comparably safe compared to matched PD patients without ICD, at least within a simulator setting. This result applies to both, defining ICD clinically or questionnaire-based using the QUIP-RS score. Patients with ICD missed more often to blink when changing lanes. This might indicate a more careless driving. However, performance of combined execution of blinking and shoulder check as well as overall driving safety was not touched so that the relevance of this result seems to be not crucial.

To our knowledge, this is the first time that a controlled study is available examining the driving behavior of PD patients with ICD. Results suggest that PD patients with socially controlled ICD should not be judged differently from PD patients without ICD in daily life regarding driving ability—an aspect with a high impact on daily routine care considering the high prevalence of both, ICDs [11] and active drivers in PD patients [20]. Furthermore, a test battery applying MoCA, TAP-M “flexibility” subscore, TMT A and the DBQ subscore “errors” might be a suitable screening tool to detect PD patients at risk for driving.

We initially hypothesized a higher risk for unsafe driving in PD patients with ICD as shown anecdotally [16], because ICD might lead to impaired learning from negative decision outcomes, resulting in impulsive as well as risk-taking behavior [12 –14]. For the treating physician it is important to know if these ICD patients need special advice with respect to driving ability. To assess the latter, we used a certified driving simulator [31]. Simulator performance has been proven to reflect real life driving ability [1] and is suitable to monitor on-road driving impairments in PD patients [2]. We assessed speed, driving duration and driving errors in key safety traffic situations which have been approved as to reflect safety-related parameters in traffic simulation models [3]. The application of a safety-score weights for the severity of driving errors and accounts for high frequency/low severity and low frequency/high-severity errors. This is meaningful for determining crash risk in PD, because especially low frequency/high-severity errors lead to car crashes [20]. Furthermore, applying rural and urban driving scenarios reflect everyday driving.

Group comparisons (PD-ICD vs. PD-control)

Both PD patient groups, with and without clinically defined ICD did not differ regarding clinical parameters or comprehensive neuropsychiatric testing applied. Only the mileage per week differed with less driven kilometers in the ICD compared to the control group. However, as we controlled for it, we found that it does not bias study results. Additionally, patients with ICD still drove regularly with a mean of 140 km per week. Second, assuming that lower mileage might be associated with lower driving experience, patients with ICD in our study would even be at higher risk for accidents compared to controls. In fact, accident rate was not higher in the ICD group. Third, mileage was previously described to not reflect driving experience properly. While people driving low mileages tend to accumulate much of their mileage on congested city streets with more complex traffic situations, high-mileage drivers typically spend most of those miles on freeways or other divided multilane highways with less demands on driving ability [36]. Noteworthy, both groups in our study caused a relative high number of accidents in comparison to the other possible error categories that may also be a result of the “artificial compilation” of menacing situations in the driving simulator which has to be further examined (and compared with healthy subjects) in the future. To date, it is reported controversial as to whether PD patients have a higher accident rate than healthy controls [3 , 38].

The LED of dopamine agonists was comparable between both groups with only numerically more PD patients applying rotigotine in the PD-ICD group. This is remarkable as previous studies found a higher risk for ICD in PD patients with compared to those without DA [39] and higher daily doses of DA in PD patients with ICD compared to those without [40]. However, it is imaginable that patients at risk for ICD already got lower doses of DA than those who are not.

Interestingly, neither driving performance nor neuropsychiatric tests did show any significant group differences between PD-ICD and PD-control subjects, although it has previously been found, that ICD may interfere with various neuropsychiatric functions (that are also relevant for the complex activity of car driving), such as derangements of visuo-spatial long-term memory and frontal lobe functions (visuo-spatial reasoning, set-shifting, cognitive flexibility, spatial working memory and spatial planning [41]). Intriguingly, the severity of cognitive impairment seems to decrease in the order multiple ICDs and hypersexuality > compulsive eating > pathological gambling [42]. Furthermore, hypersexuality might be associated with prefrontal and memory dysfunctions, whereas pathological gambling rather seems to be related to frontal dysfunction [42]. We excluded patients with manifest dementia (MoCA <21), which might explain the overall comparably small number (n = 6) of patients with multiple ICD and the higher rate of pathologic gamblers (n = 8) compared to subjects with hypersexuality (n = 3). This could be a reason for lack of significant neuropsychiatric group differences in our study. Nevertheless, literature regarding neuropsychiatric function impairments in ICD are controversial ranging from worse [43] to comparable [44, 45] to even better functions [46] compared to non-ICD PD patients. This has to be further examined in future studies.

Correlations

Slower reaction times are associated with reduced driving performance in the normal population and evident especially in elderly compared to younger drivers [47]. In PD patients, reaction times (RTs) are usually slower compared to healthy subjects due to impaired motor (timing of movement) as well as premotor processes (onset of movement) [48, 49]. In line, across both PD groups, slower RTs were correlated with higher DSS—indicating slower RTs being a risk for weaker driving performance. Driving speed did not differ between both patient groups. Noteworthy, PD patients with ICD did not drive faster than patients without ICD as might be expected. Overall mean driving speed was slightly above 40 km/h which seems to be adequate as a relevant part of driving was within town with stops at crossings or traffic lights and with challenges demanding to slow down or stop while driving in rural areas. Earlier studies are inconsistent regarding speed in PD patients compared to controls and suggested either overspeeding (=driving faster than posted speed limit, [50]), underspeeding (=driving slower than posted speed limit, [51]) or no differences [52].

To date, there is no defined standardized test battery for PD patients sufficiently predicting driving safety [6]. Therefore, we performed a comprehensive neuropsychological testing to identify potential predictive markers to predict driving ability. Independently of the presence of ICD, we found driving performance being worse in PD patients with lower cognition. This was expected, as cognitive impairment is a main negative predictor for driving safety not only in PD [1] but also for the normal population [53]. While in our study, the total MoCA score was associated with driving outcomes, others found only the MoCA subdomains “attention” and “visuospatial/executive” to be best predictive [54].

The TAP-M “flexibility” subscore might be another useful indicator for driving safety in daily routine as it is a fast and simple set-shifting task in which the participant has to press a button to react to certain target stimuli with two different levels of difficulty. Results were rated either as normal (1, T > 40) or impaired (0, T < 40). The test demands attentional and executive functions that are often impaired in PD patients [55 –58]. We found a low performance being correlated with a high DSS, indicating impaired overall driving safety. We did not find other TAP parameters such as the subscores “alertness” [59] or “divided attention” [60, 61] being correlated with driving safety. This is noteworthy, because these subscores have been described in the TAP-M manuscript to be helpful parameters for driving performance in healthy subjects vs. subjects with neurologic diseases or in different age groups.

Additionally, the TMT part A, a short and easy test evaluating visuo-motor function and visual processing speed, could be indicative for poorer driving performance and should also be considered as a screening tool to predict unsafe driving [62, 63], but may have a lack of precision distinguishing competent from incompetent drivers as has been shown before [64] and might rather be sensitive for age [65].

Of our driving specific questionnaires, only the DBQ subscore “errors” showed a positive correlation with the DSS, indicating that this item might be the most sensitive screening test to detect PD patients at risk for driving. However, earlier data in non-PD subjects regarding the predictive value of the DBQ for driving safety are inconsistent [66 –69].

Limitations

Number of participants in this study was relatively low (n = 46) and patient samples were rather young (mean age of about 58 years). However, clinically (especially neuropsychiatrically) well-characterized PD patients were included and basic characteristics (age, gender, motor symptom severity, and cognition) were comparable between both groups. As higher age is a major risk factor for reduced driving ability, mean lower age of our participants might reduce age-based risk of biasing driving results. A clinical ICD definition was used primarily to reflect daily routine, where questionnaire or score-based testing of potential ICD is not common. To assess the risk for possible “false group classification”, we performed a post-hoc analysis with assessing ICD applying the QUIP-RS score. Both approaches showed high sensitivity and specificity in detecting ICD and identical results regarding the influence of ICD on driving. However, mainly due to ethical reasons (e.g., urgent need for medication adaptation), we excluded instable patients with clinically or socially decompensated ICD and focused on PD patients with a socially controlled ICD. Thus, results have to be interpreted carefully—especially on an individual level as the severity of a patients’ ICD can change at any time—and further evaluation of driving safety in socially decompensated ICD is necessary in the future.

We did not explicitly examine visual functioning but asked for it and (besides problems with short or distance vision corrected by vision aids) no patient reported severe vision problems.

Finally, the test battery load was high which might have promoted tiredness. However, we prioritized the sequence of our tests and ensured prompt performance of driving session in the medication “on” state to ascertain best possible driving conditions.

Conclusion

PD patients with socially compensated ICD show similar driving simulator performance compared to those without ICD and should therefore not be advised differently regarding car driving safety issues. In daily routine, a test battery applying MoCA, TAP-M “flexibility” subscore, TMT part A and the DBQ subscore “errors” might be a suitable screening to detect PD patients at risk for driving. Abnormalities in these scores should persuade the doctor to suggest further driving ability assessment, including a standardized simulator or on-road driving test.

Footnotes

ACKNOWLEDGMENTS

We thank Gerhard Schön, Center for Experimental Medicine & Institute of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, for additional statistical advice and Beate Schönwald, Department of Neurology, University Medical Center Hamburg-Eppendorf, for her support with patient recruitment.

CONFLICT OF INTEREST

The authors have no conflict of interest to report regarding this manuscript.

The supplementary material is available in the electronic version of this article: .

A post-hoc analysis dividing all participants into those with unimpaired (T > 40) and impaired (T < 40) TAP-M subscores revealed a significant group difference only for the flexibility subscore (see ).

References

, Rizzo

(2008) Driving and neurodegenerative diseases. Curr Neurol Neurosci Rep8, 377–383.

, Rizzo

, Anderson

, Dastrup

, Sparks

, Dawson

(2009) Driving under low-contrast visibility conditions in Parkinson disease. Neurology73, 1103–1110.

Meindorfner

, Korner

, Moller

, Stiasny-Kolster

, Oertel

, Kruger

(2005) Driving in Parkinson’s disease: Mobility, accidents, and sudden onset of sleep at the wheel. Mov Disord20, 832–842.

Devos

, Vandenberghe

, Nieuwboer

, Tant

, Baten

, De Weerdt

(2007) Predictors of fitness to drive in people with Parkinson disease. Neurology69, 1434–1441.

Worringham

, Wood

, Kerr

, Silburn

(2006) Predictors of driving assessment outcome in Parkinson’s disease. Mov Disord21, 230–235.

Buhmann

, Gerloff

(2013) Autofahren bei morbus Parkinson [Driving with Parkinson’s disease]. Akt Neurol40, 315–320.

Klimkeit

, Bradshaw

, Charlton

, Stolwyk

, Georgiou-Karistianis

(2009) Driving ability in Parkinson’s disease: Current status of research. Neurosci Biobehav Rev33, 223–231.

Ghorayeb

, Loundou

, Auquier

, Dauvilliers

, Bioulac

, Tison

(2007) A nationwide survey of excessive daytime sleepiness in Parkinson’s disease in France. Mov Disord22, 1567–1572.

Ranchet

, Devos

, Uc

(2020) Driving in Parkinson disease. Clin Geriatr Med36, 141–148.

10.

Gatto

, Aldinio

(2019) Impulse control disorders in Parkinson’s disease. A brief and comprehensive review. Front Neurol10, 351.

11.

Marques

, Durif

, Fernagut

(2018) Impulse control disorders in Parkinson’s disease. J Neural Transm (Vienna)125, 1299–1312.

12.

Frank

, Samanta

, Moustafa

, Sherman

(2007) Hold your horses: Impulsivity, deep brain stimulation, and medication in parkinsonism. Science318, 1309–1312.

13.

Grover

, Lill

, Kasten

, Klein

, Del Greco

, Konig

(2019) Risky behaviors and Parkinson disease: A mendelian randomization study. Neurology93, e1412–e1424.

14.

Savica

(2019) Livin’ on the edge: Risky behaviors and Parkinson disease: Genes, environment, or both?Neurology93, 653–654.

15.

Wood

, Worringham

, Kerr

, Mallon

, Silburn

(2005) Quantitative assessment of driving performance in Parkinson’s disease. J Neurol Neurosurg Psychiatry76, 176–180.

16.

Avanzi

, Baratti

, Cabrini

, Uber

, Brighetti

, Bonfa

(2008) The thrill of reckless driving in patients with Parkinson’s disease: An additional behavioural phenomenon in dopamine dysregulation syndrome?Parkinsonism Relat Disord14, 257–258.

17.

Weintraub

, Koester

, Potenza

, Siderowf

, Stacy

, Voon

, Whetteckey

, Wunderlich

, Lang

(2010) Impulse control disorders in Parkinson disease: A cross-sectional study of 3090 patients. Arch Neurol67, 589–595.

18.

Corvol

, Artaud

, Cormier-Dequaire

, Rascol

, Durif

, Derkinderen

, Marques

, Bourdain

, Brandel

, Pico

, Lacomblez

, Bonnet

, Brefel-Courbon

, Ory-Magne

, Grabli

, Klebe

, Mangone

, You

, Mesnage

, Lee

, Brice

, Vidailhet

, Elbaz

, Group

(2018) Longitudinal analysis of impulse control disorders in Parkinson disease. Neurology91, e189–e201.

19.

Heinzel

, Berg

, Binder

, Ebersbach

, Hickstein

, Herbst

, Lorrain

, Wellach

, Maetzler

, Petersen

, Schmedt

, Volkmann

, Woitalla

, Amelung

(2018) Do we need to rethink the epidemiology and healthcare utilization of Parkinson’s disease in Germany?Front Neurol9, 500.

20.

Buhmann

, Vesper

, Oelsner

(2018) Driving ability in Parkinson’s disease. Fortschr Neurol Psychiatr86, 43–48.

21.

Weintraub

, David

, Evans

, Grant

, Stacy

(2015) Clinical spectrum of impulse control disorders in Parkinson’s disease. Mov Disord30, 121–127.

22.

Weintraub

, Mamikonyan

, Papay

, Shea

, Xie

, Siderowf

(2012) Questionnaire for impulsive-compulsive disorders in Parkinson’s disease-rating scale. Mov Disord27, 242–247.

23.

Tomlinson

, Stowe

, Patel

, Rick

, Gray

, Clarke

(2010) Systematic review of levodopa dose equivalency reporting in Parkinson’s disease. Mov Disord25, 2649–2653.

24.

International Parkinson and Movement Disorder Society (MDS), MDS-Recommended Rating Scales, https://www.movementdisorders.org/MDS/Education/Rating-Scales/MDS-Recommended-Rating-Scales.htm.

25.

Nasreddine

, Phillips

, Bedirian

, Charbonneau

, Whitehead

, Collin

, Cummings

, Chertkow

(2005) The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc53, 695–699.

26.

Dubois

, Slachevsky

, Litvan

, Pillon

(2000) The FAB: A Frontal Assessment Battery at bedside. Neurology55, 1621–1626.

27.

Ostapczuk

, Joseph

, Musch

, Dickers

(2019) DSI. Driver Skill Inventory - deutsche Version [Verfahrensdokumentation aus PSYNDEX Tests-Nr. 9007727 und Fragebogen]. Leibniz-Zentrum für Psychologische Information und Dokumentation (ZPID) (Hrsg.) Elektronisches Testarchiv.

28.

Reason

, Manstead

, Stradling

, Baxter

, Campbell

(1990) Errors and violations on the roads: A real distinction?Ergonomics33, 1315–1332.

29.

Clapp

, Olsen

, Beck

, Palyo

, Grant

, Gudmundsdottir

, Marques

(2011) The Driving Behavior Survey: Scale construction and validation. J Anxiety Disord25, 96–105.

30.

Owsley

, Stalvey

, Wells

, Sloane

(1999) Older drivers and cataract: Driving habits and crash risk. J Gerontol A Biol Sci Med Sci54, M203–211.

31.

Muttray

, Breitinger

, Goetze

, Schnupp

, Geissler

, Kaufmann

, Golz

, Letzel

(2013) Further development of a commercial driving simulation for research in occupational medicine. Int J Occup Med Environ Health26, 949–965.

32.

Zimmermann

, Fimm

(2007) Testbatterie zur Aufmerksamkeitstestung (TAP 2.1) Fa. Psytest, Herzogenrath.

33.

Smith

(1973) Symbol Digit Modalities Test (Manual). Western Psychological Services.

34.

Reitan

(1992) Trail Making Test, Reitan Neuropsychology Laboratory, Tucson, AZ.

35.

Wechsler

(1997) WMS-III: Wechsler Memory Scale Administration and Scoring Manual, Psychological Corporation.

36.

Janke

(1991) Accidents, mileage, and the exaggeration of risk. Accid Anal Prev23, 183–188.

37.

Dubinsky

, Gray

, Husted

, Busenbark

, Vetere-Overfield

, Wiltfong

, Parrish

, Koller

(1991) Driving in Parkinson’s disease. Neurology41, 517–520.

38.

Ritter

, Steinberg

(1979) Parkinsonism and fitness to drive. MMW Munch Med Wochenschr121, 1329–1330.

39.

Grall-Bronnec

, Victorri-Vigneau

, Donnio

, Leboucher

, Rousselet

, Thiabaud

, Zreika

, Derkinderen

, Challet-Bouju

(2018) Dopamine agonists and impulse control disorders: A complex association. Drug Saf41, 19–75.

40.

Weintraub

, Siderowf

, Potenza

, Goveas

, Morales

, Duda

, Moberg

, Stern

(2006) Association of dopamine agonist use with impulse control disorders in Parkinson disease. Arch Neurol63, 969–973.

41.

Santangelo

, Vitale

, Trojano

, Verde

, Grossi

, Barone

(2009) Cognitive dysfunctions and pathological gambling in patients with Parkinson’s disease. Mov Disord24, 899–905.

42.

Vitale

, Santangelo

, Trojano

, Verde

, Rocco

, Grossi

, Barone

(2011) Comparative neuropsychological profile of pathological gambling, hypersexuality, and compulsive eating in Parkinson’s disease. Mov Disord26, 830–836.

43.

Voon

, Thomsen

, Miyasaki

, de Souza

, Shafro

, Fox

, Duff-Canning

, Lang

, Zurowski

(2007) Factors associated with dopaminergic drug-related pathological gambling in Parkinson disease. Arch Neurol64, 212–216.

44.

Djamshidian

, O’Sullivan

, Lees

, Averbeck

(2011) Stroop test performance in impulsive and non impulsive patients with Parkinson’s disease. Parkinsonism Relat Disord17, 212–214.

45.

Bentivoglio

, Baldonero

, Ricciardi

, De Nigris

, Daniele

(2013) Neuropsychological features of patients with Parkinson’s disease and impulse control disorders. Neurol Sci34, 1207–1213.

46.

Siri

, Cilia

, De Gaspari

, Canesi

, Meucci

, Zecchinelli

, Pezzoli

, Antonini

(2010) Cognitive status of patients with Parkinson’s disease and pathological gambling. J Neurol257, 247–252.

47.

Doroudgar

, Chuang

, Perry

, Thomas

, Bohnert

, Canedo

(2017) Driving performance comparing older versus younger drivers. Traffic Inj Prev18, 41–46.

48.

Low

, Miller

, Vierck

(2002) Response slowing in Parkinson’s disease: A psychophysiological analysis of premotor and motor processes. Brain125, 1980–1994.

49.

Berry

, Nicolson

, Foster

, Behrmann

, Sagar

(1999) Slowing of reaction time in Parkinson’s disease: The involvement of the frontal lobes. Neuropsychologia37, 787–795.

50.

Crizzle

, Myers

(2013) Examination of naturalistic driving practices in drivers with Parkinson’s disease compared to age and gender-matched controls. Accid Anal Prev50, 724–731.

51.

Classen

, Brumback

, Monahan

, Malaty

, Rodriguez

, Okun

, McFarland

(2014) Driving errors in Parkinson’s disease: Moving closer to predicting on-road outcomes. Am J Occup Ther68, 77–85.

52.

, Rizzo

, Johnson

, Dastrup

, Anderson

, Dawson

(2009) Road safety in drivers with Parkinson disease. Neurology73, 2112–2119.

53.

Apolinario

, Magaldi

, Busse

, Lopes

LDC

, Kasai

JYT

, Satomi

(2009) Cognitive impairment and driving: A review of the literature. Dement Neuropsychol3, 283–290.

54.

Ma’u

, Cheung

(2020) Ability of the Maze Navigation Test, Montreal Cognitive Assessment, and Trail Making Tests A & B to predict on-road driving performance in current drivers diagnosed with dementia. N Z Med J133, 23–32.

55.

Allcock

, Rowan

, Steen

, Wesnes

, Kenny

, Burn

(2009) Impaired attention predicts falling in Parkinson’s disease. Parkinsonism Relat Disord15, 110–115.

56.

Dujardin

, Tard

, Duhamel

, Delval

, Moreau

, Devos

, Defebvre

(2013) The pattern of attentional deficits in Parkinson’s disease. Parkinsonism Relat Disord19, 300–305.

57.

Dirnberger

, Jahanshahi

(2013) Executive dysfunction in Parkinson’s disease: A review. J Neuropsychol7, 193–224.

58.

Lange

, Bruckner

, Knebel

, Seer

, Kopp

(2018) Executive dysfunction in Parkinson’s disease: A meta-analysis on the Wisconsin Card Sorting Test literature. Neurosci Biobehav Rev93, 38–56.

59.

Niemann

, Hartje

(2007) Neurokognitive funktionen und fahreignung. Z Epileptol20, 184–196.

60.

Burgard

(2005) ahrkompetenz im Alter - Die Aussagekraft diagnostischer Instrumente bei Senioren und neurologischen Patienten. Medizinische Fakultät der Ludwig-Maximilians-Universität zu München.

61.

Akinwuntan

, Feys

, De Weerdt

, Baten

, Arno

, Kiekens

(2006) Prediction of driving after stroke: A prospective study. Neurorehabil Neural Repair20, 417–423.

62.

Choi

, Lee

, Oh

(2016) Cut-off point for the trail making test to predict unsafe driving after stroke. J Phys Ther Sci28, 2110–2113.

63.

Seong-Youl

, Jae-Shin

, S

(2014) Cognitive test to forecast unsafe driving in older drivers: Meta-analysis. NeuroRehabilitation35, 771–778.

64.

Vaucher

, Herzig

, Cardoso

, Herzog

, Mangin

, Favrat

(2014) The trail making test as a screening instrument for driving performance in older drivers; a translational research. BMC Geriatr14, 123.

65.

Selander

, Wressle

, Samuelsson

(2020) Cognitive prerequisites for fitness to drive: Norm values for the TMT, UFOV and NorSDSA tests. Scand J Occup Ther27, 231–239.

66.

Davey

, Wishart

, Freeman

, Watson

(2007) An application of the Driver Behaviour Questionnaire in an Australian organisational fleet setting. Transp Res Part F Traffic Psychol Behav10, 11–21.

67.

Cordazzo

, Scialfa

, Bubric

, Ross

(2014) The driver behaviour questionnaire: A north American analysis. J Safety Res50, 99–107.

68.

DeLucia

, Bleckley

, Meyer

, Bush

(2003) Judgments about collision in younger and older drivers. Transp Res Part F Traffic Psychol Behav6, 63–80.

69.

Blockey

, Hartley

(1995) Aberrant driving behaviour: Errors and violations. Ergonomics38, 1759–1771.

70.

Pinto

TCC

, Santos

MSP

, Machado

, Bulgacov

, Rodrigues-Junior

, Silva

, Costa

MLG

, Ximenes

RCC

, Sougey

(2019) Optimal cutoff scores for dementia and mild cognitive impairment in the Brazilian version of the Montreal Cognitive Assessment among the elderly. Dement Geriatr Cogn Dis Extra9, 44–52.