Update: Evaluation and Management of Driving Risk in Dementia

1. Report of the Quality Standards Subcommittee of the American Academy of Neurology Donald J. Iverson, MD, FAAN; Gary S. Gronseth, MD, FAAN; Mark A. Reger, PhD; Sherrilene Classen, PhD, MPH, OTR; Richard M. Dubinsky, MD, MPH, FAAN; Matt Rizzo, MD, FAAN Update: Evaluation and Management of Driving Riskin Dementia

2. If you have questions, comments, or feedback regarding this slide presentation, or would like to modify the contents for presentation in a lecture, please contact guidelines@aan.com

3. Presentation Objectives • To review the evidence regarding the usefulness of certain information sources in predicting driving capability among patients with dementia • Patient demographic characteristics • Driving history • Cognitive testing • To determine the efficacy of driving risk reduction strategies • To present evidence-based recommendations

4. Overview • Background • Gaps in care • AAN guideline process • Analysis of evidence, conclusions, recommendations • Recommendations for future research

5. Background • Driving skills deteriorate with increasing dementia severity,1and patients with mild dementia, as a group, are higher-risk drivers.1 • However, recent studies2-4report that as many as 76% of these patients are still able to pass an on-road driving test (ORDT) and can safely drive.3 • Faced with these facts, clinicians caring for patients with dementia seek to identify those patients with cognitive impairment who may be at higher risk for unsafe driving, without unnecessarily restricting those who are safe drivers.

6. Background, cont. • Clinicians’ predictions of driving performance, when based primarily on the Mini-Mental State Examination (MMSE), result in either no correlation5 or have a relatively low sensitivity for identifying an unsafe driver.3 • When elements of the driving history and additional cognitive testing are considered along with MMSE scores, the predictions are more accurate6 but still maintain only moderate degrees of both sensitivity and specificity. • This parameter, which updates a 2000 AAN parameter, seeks to identify historical features that are associated with increased driving risk.

7. Gaps in Care • The benefit of additional neuropsychological testing in predicting driving performance is uncertain. • Practitioners do not know what actions (if any) to take once a risky driver is identified. • Practitioners need to consider multiple cognitive tests—not just select ones—for evaluating driving in patients with dementia. • Driving is a multidimensional task—it’s hard to evaluate by isolating aspects of the task or using just one measure

8. AAN Guideline Process • Clinical Question • Evidence • Conclusions • Recommendations

9. Clinical Questions • The first step in developing guidelines is to clearly formulate questions to be answered. • Questions address areas of controversy, confusion, or variation in practice. • Questions must be answerable with data from the literature. • Answering the question must have the potential to improve care/patient outcomes.

10. Complete Search Review abstracts Review full text Select articles Relevant LiteratureSearch/Review Rigorous, Comprehensive, Transparent

11. AAN Classification of Evidence • All studies rated Class I, II, III, or IV • Five different classification systems: • Therapeutic • Randomization, control, blinding • Diagnostic • Comparison to gold standard • Prognostic • Screening • Causation

12. AAN Level of Recommendations • A = Established as effective, ineffective or harmful (or established as useful/predictive or not useful/predictive) for the given condition in the specified population. • B = Probably effective, ineffective or harmful (or probably useful/predictive or not useful/predictive) for the given condition in the specified population. • C = Possibly effective, ineffective or harmful (or possibly useful/predictive or not useful/predictive) for the given condition in the specified population. • U = Data inadequate or conflicting; given current knowledge, treatment (test, predictor) is unproven. • Note that recommendations can be positive or negative.

13. Translating Class to Recommendations • A = Requires at least two consistent Class I studies.* • B = Requires at least one Class I study or two consistent Class II studies. • C = Requires at least one Class II study or two consistent Class III studies. • U = Studies not meeting criteria for Class I through Class III.

14. Translating Class to Recommendations, cont. * In exceptional cases, one convincing Class I study may suffice for an “A” recommendation if 1) all criteria are met, 2) the magnitude of effect is large (relative rate improved outcome >5 and the lower limit of the confidence interval is >2).

15. Applying This Processto the Issue We will now turn our attention to the guidelines.

16. Clinical Questions 1. How strongly are global measures of dementia severity associated with decreased driving ability? • To what extent are patients and their caregivers able to assess driving ability and risk? • Which elements of the driving history are associated with decreased driving ability? • Which neuropsychological tests provide additional prognostic information? • Are there any interventions that reduce driving risk?

17. Methods • MEDLINE, EBSCOhost, Ovid, Web of Science, Ageline, Dissertation Abstracts, WorldCat, AOA, TRIS, NTL, TRB, SafetyLit, ASA, NHTSA/USDOT • 1970 through December 2006 • A secondary bibliography search of all full-text articles • Relevant, fully published, peer-reviewed articles (see article for search terms)

18. Methods, cont. • At least two authors reviewed each article for inclusion. • Risk of bias was determined using the classification of evidence for each study (Classes I–IV). • Strength of practice recommendations were linked directly to levels of evidence (Levels A, B, C, and U). • Conflicts of interest were disclosed.

19. 6,000abstracts 502 articles Literature Review • Inclusion criteria: • Pertained to automobile driving • Had driving as an outcome measure • Involve persons with some type of neurological condition, which includes aging • Exclusion criteria: • Did not present primary research findings • Included drivers who had confounding medical conditions • Pertained only to driving with visual impairment that is not caused by some neurological disorder

20. AAN Classification of Evidencefor Diagnostic Accuracy • Class I: A cohort study with prospective data collection of a broad spectrum of persons with the suspected condition, using an acceptable reference standard for case definition. The diagnostic test is objective or performed and interpreted without knowledge of the patient’s clinical status. Study results allow calculation of measures of diagnostic accuracy. • Class II: A case control study of a broad spectrum of persons with the condition established by an acceptable reference standard compared to a broad spectrum of controls or a cohort study where a broad spectrum of persons with the suspected condition where the data was collected retrospectively. The diagnostic test is objective or performed and interpreted without knowledge of disease status. Study results allow calculation of measures of diagnostic accuracy.

21. AAN Classification of Evidencefor Diagnostic Accuracy, cont. • Class III:A case control study or cohort study where either persons with the condition or controls are of a narrow spectrum. The condition is established by an acceptable reference standard. The reference standard and diagnostic test are objective or performed and interpreted by different observers. Study results allow calculation of measures of diagnostic accuracy. • Class IV:Studies not meeting Class I, II or III criteria including consensus, expert opinion or a case report.

22. Analysis of Evidence Question 1: How strongly are global measures of dementia severity associated with decreased driving ability?

23. Conclusion/Recommendation Conclusion: • The Clinical Dementia Rating (CDR) is established as useful for identifying patients at increased risk for unsafe driving (two Class I and two Class II studies); however, a substantial number of patients with a CDR of 0.5-1 (41%-85%) will be found to be safe drivers by an ORDT. Recommendation: • For patients with dementia, the CDR scale is established as useful for identifying patients at increased risk for unsafe driving (Level A).

24. Conclusion/Recommendation Conclusion: • An MMSE score of ≤24is possibly useful in identifying patients at increased risk for unsafe driving (l Class II study). Otherwise, the correlation between MMSE scores and driving performance is unclear, as data are conflicting. Recommendation: • For patients with dementia, MMSE scores of ≤24 may be considered useful for identifying patients at increased risk for unsafe driving (Level C).

25. Analysis of Evidence Question 2: To what extent are patients and their caregivers able to assess driving ability and risk?

26. Conclusion/Recommendation Conclusion: • A caregiver’s rating of marginal or unsafe is probably useful in identifying unsafe drivers (one Class I study). Recommendation: • For patients with dementia, a caregiver’s rating of a patient’s driving ability as marginal or unsafe should be considered useful for identifying patients at increased risk for unsafe driving (Level B).

27. Conclusion/Recommendation Conclusion: • A patient’s self-rating of safe is established as not useful for determining that the patient is safe to drive (three Class I studies). Recommendation: • For patients with dementia, a patient’s self-rating of safe driving ability is established as not useful for identifying patients at increased risk for unsafe driving (Level A).

28. Analysis of Evidence Question 3: Which elements of the driving history are associated with decreased driving ability?

29. Conclusions Conclusions: • A history of a crash in the previous 1 to 5 years or a traffic citation the previous 2 to 3 years is possibly useful in identifying patients with decreased driving ability (1 Class II and 5 Class III studies). • A history of a crash is possibly more useful in identifying patients at risk for subsequent crashes than the presence of mild dementia alone (3 Class III studies).

30. Recommendations Recommendations: • For patients with dementia, a history of traffic citations may be considered useful for identifying patients at increased risk for unsafe driving (Level C). • For patients with dementia, a history of crashes may be considered useful for identifying patients at increased risk for unsafe driving (Level C).

31. Conclusions Conclusions: • In mixed-population studies of aged drivers and drivers with mild dementia, reduced driving mileage is possibly associated with an increased risk of poor driving performance (1 Class II, 1 Class III study). • In aged drivers, self-reported avoidance is possibly useful to identify drivers at increased risk (1 Class II study). • The absence of self-reported avoidance is possibly not useful for identifying safe drivers (1 Class II and 1 Class III study).

32. Recommendations Recommendations: • For patients with dementia, reduced driving mileage may be considered useful for identifying patients at increased risk for unsafe driving (Level C). • For patients with dementia, self-reported situational avoidance may be considered useful for identifying patients at increased risk for unsafe driving (Level C). • For patients with dementia, lack of situational avoidance may be considered as not useful for identifying patients at increased risk for unsafe driving (Level C).

33. Conclusion/Recommendation Conclusion: • Aggressive or impulsive personality characteristics are possibly useful to identify patients with increased driving risk (l Class II and 1 Class III study). Recommendation: • For patients with dementia, aggressive or impulsive personality characteristics may be considered useful for identifying patients at increased risk for unsafe driving (Level C).

34. Analysis of Evidence Question 4: Which neuropsychological tests provide additional prognostic information?

35. Conclusion Conclusion: • Comprehensive neuropsychological assessment is another means of assessing global cognitive impairment that may be complementary to that of a bedside examination and an informant interview. While neuropsychological testing itself may better define dementia severity, there is insufficient evidence to support or refute the benefit of neuropsychological testing in evaluating driving risk in patients with dementia.

36. Recommendation Recommendation: • There is insufficient evidence to support or refute the benefit, after controlling for the presence and severity of dementia, of neuropsychological testing for drivers with dementia (Level U).

37. Analysis of Evidence Question 5: Are there any interventions that reduce driving risk?

38. Conclusion/Recommendation Conclusion: • There is insufficient evidence to support or refute a benefit of interventional strategies for drivers with dementia. Recommendation: • There is insufficient evidence to support or refute the benefit of interventional strategies for drivers with dementia (Level U).

39. Clinical Context • Clinicians are obligated to identify conditions that may risk their patients’ or the public’s health. • Because there is no test result or historical feature that accurately quantifies driving risk, clinicians can make only qualitative estimates of driving risk. • Clinicians may present data showing that patients with mild dementia (CDR of 1) are at a substantially higher risk for unsafe driving and thus should strongly consider discontinuing driving. Note: Clinical context has been abridged. See the published guideline for the complete text.

40. Clinical Context, cont. • However, advocates for maintaining driving privileges may cite the wide CIs for relative risk and ORDT pass rates of 41%7to 76%3as evidence against a categorical recommendation for these patients to cease driving. • Such advocates do not want truly capable drivers to prematurely cease driving.8,9In that case, one may look for evidence of increased risk in an individual patient.10,11Consideration of these additional issues can result in a more accurate prediction of driving performance.

41. Clinical Context, cont. • A clinician may wish to integrate this information into an algorithm to obtain a qualitative estimate of driving risk. This algorithm should only be considered supplementary to the clinician’s judgment. Patients at higher risk may agree to surrender privileges. • For those who wish to continue driving, clinicians may consider referral for a professional or governmental driving evaluation, depending on state reporting laws. Patients who continue to drive should be reassessed at 6-month intervals.12

42. Algorithm: Evaluating Driving Risk

43. Clinical Context, cont. • Neuropsychological testing offers a means of assessing memory, spatial cognition, and executive functioning that is more sensitive than the MMSE or CDR. • While it seems intuitive that a more accurate determination of impairment in specific cognitive domains would result in a more accurate estimate of driving risk, there are no data at this time to support or refute this approach. • Additional medical conditions may also be relevant, but those issues are beyond the scope of this review.

44. Clinical Context, cont. • Qualitative risk estimates are a familiar part of clinical practice. However, clinicians may be less comfortable making such judgments in a legal context.13When the threshold for “likely” impairment is low14or unclear, some clinicians may choose to report borderline cases. • In some states, doing so may leave them open to civil litigation.13This practice parameter cannot operationalize these types of subjective statutory requirements; it is intended for use in a clinical setting to assist in an evidence-based estimate of driving risk.

45. Future Research • Future studies may wish to evaluate the appropriate weighting of risk factors using, for example, discriminant function analysis12,15to develop a composite method of rating risk in patients with mild dementia. • It is recommended that future studies of individual cognitive domains (e.g., attention, executive function): a) emphasize simple bedside tests (e.g., Trail Making Test), b) control for both the presence and severity of dementia to identify independent predictors of unsafe driving, c) report results in terms of relative risk based on the presence of a risk factor or cutoff score, rather than in terms of correlation coefficients,16and d) report using STARD-D criteria.

46. Future Research, cont. • Rules for determining when to revoke driving privileges are ultimately made by society through the legislative process. By accurately quantifying risk with well-designed prospective studies using validated outcome measures, researchers can ensure that legislators are provided with the best science with which to inform their decisions.

47. References • Dubinsky RM, Stein AC, Lyons K. Practice parameter: risk of driving and alzheimer’s disease (an evidence-based review). Neurology 2000;54:2205-2211. • Grace J, Amick MM, D’Abreu A, Festa EK, Heindel WC, Ott BR. Neuropsychological deficits associated with driving performance in parkinson’s and alzheimer’s disease. Journal of the International Neuropsychological Society 2005;11:766-775. • Brown LB, Ott BR, Papandonatos GD, Sui Y, Ready RE, Morris JC. Prediction of on-road driving performance in patients with early alzheimer’s disease. J Am Geriatr Soc 2005;53:94–98. • Ott BR, Heindel WC, Papandonatos GD, et al. A longitudinal study of drivers with Alzheimer’s disease. Neurology 2008;70:1171-1178. • Fox GK, Bowden SC, Bashford GM, Smith DS. Alzheimer’s disease and driving: prediction and assessment of driving performance. JAGS 1997;45:949-953. • Ott BR, Anthony D, Papandonatos GD, et al. Clincal assessment of the driving competence of patients with dementia. JAGS 2005;53(5):829-833. • Hunt LA, Murphy CF, Carr D, et al. Environmental cueing may affect performance on a road test for drivers with dementia of the alzheimer type. Alzheimer Disease and Associated Disorders 1997;11(Suppl 1):13-16.

48. References, cont. 8. Fonda SJ, Wallace RB, Herzog AR. Changes in driving patterns and worsening depressive symptoms among older adults. J Gerontol B Psychol Sci Soc Sci 2001;56(6):S343-351. • Edwards JD, Perkins M, Ross LA, Reynolds SL. Driving status and three-year mortality among community-dwelling older adults. J Gerontol A Biol Sci Med Sci 2009;64(2):300-305. • Parker D, McDonald L, Rabbitt P, Sutcliffe P. Elderly drivers and their accidents: the aging driver questionnaire. Accid Anal Prev 2000;32:751-759. • Rimmö P-A. Aberrant driving behaviour: homogeneity of a four-factor structure in samples differing in age and gender. Ergonomics 2002;45(8):569-582. • Fitten LJ, Perryman KM, Wilkinson CJ, et al. Alzheimer and vascular dementias and driving: a prospective road and laboratory study. JAMA 1995;273(17):1360–1365. • American Association of Motor Vehicle Administrators. Jurisdictional Procedures, Practices & Programs. http://www.aamva.org/knoweldgecenter/Driver/At-Risk/JurisProceduresPractAndProgs.htm. Accessed November 30, 2007.

49. References, cont. • California Code Regulations, Title 17, Division 1, Chapter 4, Subchapter 2.5, Section 2802. http://ccr.oal.ca.gov • De Raedt R, Ponjaert-Kristoffersen I. Predicting at-fault car accidents of older drivers. Accid Anal Prev 2001;33:809-819. • Molnar FJ, Patel A, Marshall SC, et al. Clinical utility of office-based cognitive predictors of fitness to drive in persons with dementia: a systematic review. J Am Geriatr Soc 2006;54:1809-1824. For a complete list of references, please access the full guideline at www.aan.com/guidelines

50. Questions/Comments