Visual Attention Cut Points Predicting Fitness to Drive in People With Parkinson’s Disease

BACKGROUND: Parkinson’s disease (PD), a progressive, neurodegenerative disorder, is characterized by motor and nonmotor symptoms of which visual attention most significantly impacts fitness to drive (Crizzle, Classen, & Uc, 2012). Visual attention is a predictor of on-road outcomes in PD drivers, but small sample sizes, variability in methods, and heterogeneity in on-road outcomes detract from scientific rigor and thus best clinical decision making (Crizzle et al., 2012; Uc et al., 2006). We also do not know if impaired visual attention occurs early in PD and if it is useful in screening for at-risk PD drivers (Classen et al., 2015; Devos, Ranchet, Akinwuntan, & Uc, 2015).

PURPOSE: This study determined if visual attention, measured by the Useful Field of View^® (UFOV), is an early and persistent impairment in PD drivers and a valid predictor of failing an on-road assessment. We hypothesized that the UFOV Risk Index (RI) is more consistently impaired in PD drivers and leads to greater predictive validity in failing an on-road assessment, regardless of disease severity (mild/moderate vs. severe) when compared to healthy control participants (HC).

METHOD: We prospectively assessed 101 PD and 138 HC drivers with a comprehensive driving evaluation. Using logistic regression, we (1) predicted pass–fail on-road outcomes as a function of gender and age in PD and HC. To each within-group analysis, we added the (2) UFOV RI as a covariate and evaluated its contribution to on-road outcomes, beyond gender and age, by comparing the area under the curve (AUC) of the receiver operating characteristic (ROC) curves for each analysis. We (3) constructed the ROC for UFOV RI alone to compute sensible cut points. Finally, we indicated the (4) AUC for UFOV RI for PD Severe vs. PD Mild/Moderate. We used SAS Version 9.2 (SAS Institute, Cary, NC), with p < .05 (two-tailed) for all analyses.

RESULTS: (1) Logistic regression results suggested the UFOV RI was a significant predictor of on-road outcomes in PD (odds ratio [OR] = 2.209; confidence interval [CI] [1.507, 3.238]; p < .0001). For HC the significant predictors were age (OR = 1.194; CI [1.046, 1.362]; p = .0085) and the UFOV RI (OR = 2.411; CI [1.338, 4.346]; p = .0034). For (2) HC the ROC curves including (AUC = .891; CI [.811, .972]; SE = .041) and not including (AUC =.828, CI [.701, .954]; SE =.064) the UFOV RI were similar, but for PD the ROC curve including the UFOV RI (AUC =.781; CI [.687, .875]; SE =.048) suggested better predictive accuracy than without UFOV RI (AUC = .615; CI [.502, .728]; SE =.058). For HC, the difference between the two ROCs was not significant (p = 26), whereas for PD, it was highly significant (p = .003). The difference between the PD and HC differences in AUCs was also not significant (p = .206). (3) The best cut point for UFOV RI was 3 or higher (sensitivity = .70, specificity = .77). (4) The AUC for UFOV RI for PD Severe was .875 (CI [.655, 1]; SE = .112) and for PD Mild/Moderate was .786 (CI [.676, .896]; SE = .056) with no significant difference (p = .92).

CONCLUSION: The predictive accuracy of the UFOV RI, combined with age and gender, increased significantly for PD (62%–78%) but not for HC (83%–89%). The UFOV RI, combined with age and gender, is a significant predictor of on-road outcomes in PD and HC, but the difference score is greater in PD. Thus, as hypothesized, the UFOV RI is more consistently impaired in PD than HC drivers. Although the UFOV RI emerged as a meaningful predictor of on-road outcomes for PD, with cut point 3 yielding the best combination of sensitivity and specificity, it did not predict a difference in on-road outcomes between PD Severe and PD Mild/Moderate.

IMPACT STATEMENT: These findings have immediate implications for practice as they enhance the knowledge of occupational therapists pertaining to PD and driving and provide them with cut points, derived from ROC curves, to improve clinical reasoning in identifying at-risk drivers with PD.

References

Classen, S., Holmes, J., Alvarez, L., Loew, K., Mulvagh, A., Rienas, . . . He, W. (2015). Clinical assessments as predictors of primary on-road outcomes in Parkinson’s disease. OTJR: Occupation, Participation and Health, 35, 213–220. https://doi.org/10.1177/1539449215601118

Crizzle, A. M., Classen, S., & Uc, E. Y. (2012). Parkinson disease and driving: An evidence-based review. Neurology, 79, 2067–2074. https://doi.org/10.1212/WNL.0b013e3182749e95

Devos, H., Ranchet, M., Akinwuntan, A. E., & Uc, E. Y. (2015). Establishing an evidence-base framework for driving rehabilitation in Parkinson’s disease: A systematic review of on-road driving studies. Neurorehabilitation, 37, 35–52. https://doi.org/10.3233/NRE-151239

Uc, E. Y., Rizzo, M., Anderson, S. W., Sparks, J. D., Rodnitzky, R. L., & Dawson, J. D. (2006). Impaired visual search in drivers with Parkinson’s disease. Annals of Neurology, 60, 407–413. https://doi.org/10.1002/ana.20958