Driving and Parkinson's Disease

Abstract & Commentary

By Melissa J. Nirenberg, MD, PhD, Assistant Professor, Neurology and Neuroscience, Weill Cornell Medical College. Dr. Nirenberg participates in clinical research trials for Boehringer-Ingelheim.

Synopsis: A simple screening test can predict which patients with Parkinson's disease (PD) may have trouble driving a motor vehicle.

Source: Devos H, et al. Predictors of fitness to drive in people with Parkinson disease. Neurology 2007; 69:1434-1441.

Physicians frequently are asked whether it is safe for a patient with Parkinson's disease (PD) to drive a motor vehicle, but the physician may have limited information with which to make such an assessment. Many factors have the potential to impact driver safety, including motor, cognitive, visuospatial, and psychiatric symptoms, as well as sleep disorders and medication usage. The primary goal of this study was to create a short clinical screening battery to predict fitness to drive in patients with PD.

The authors compared 40 PD patients with an equal number of age- and gender-matched control subjects, all of whom were regular drivers and possessed valid driver's licenses. Subjects with a clinical dementia rating scale of > 1 or impaired binocular visual acuity were excluded from the study, as were PD patients with unpredictable motor fluctuations or more advanced disease (Hoehn and Yahr stage > 3).

Patients completed self-reported driving and accident histories, self-appraisals of their fitness to drive, Epworth Sleepiness Scale scores, contrast sensitivity testing, complex figure drawing (Rey diagram), and the Clinical Dementia Rating (CDR) scale. They also were tested with a formal driving simulator; PD patients were evaluated exclusively when they were in the "on" state. Activities of daily living and motor symptoms were evaluated with the Unified Parkinson's Disease Rating Scale (UPDRS) parts II and III, respectively. PD patients underwent a standardized 3-4 hour evaluation by Belgian Center for Fitness to Drive Evaluation and Car Adaptations (CARA), including visual and cognitive evaluations and an on-road driving test. CARA scores (dichotomized as pass/fail) were used as the "gold standard" against which the predictive value of other assessments was compared.

The authors used logistic regression analysis to create a model that would predict fitness to drive outcome based on the CARA. A combination of disease duration, contrast specificity, CDR, and UPDRS motor score provided the best model (R2 = 0.52), correctly classifying 90% of patients (n = 36), with a sensitivity of 91% and specificity of 90%. When the driving simulator score was added to the clinical model, the sensitivity remained the same, but the specificity increased to 100% (R2 = 0.60). Patient self-perception of fitness to drive correlated poorly with CARA scores, with 20% of PD patients (n = 8) misjudging their abilities and most (n = 5) overestimating their fitness to drive.


Determination of fitness to drive is a critical issue in the care of patients with PD and other neurological disorders. In this study, a combination of four easily measured, objective clinical assessments — disease duration, contrast sensitivity, CDR, and UPDRS motor score — were most predictive of driving fitness in PD (as defined by CARA score). In contrast, patient self-perception was a poor predictor of fitness to drive. Study limitations include the small sample size and exclusion of patients with advanced disease or dementia, which may limit the practical value of the results. The findings provide useful information for physicians who care for PD patients, and underscore the need for larger, prospective studies to test this model as a predictor of real world driving performance.