Through performance of three, complementary efforts in this project—a literature review, brainstorming session with subject matter experts, and focus groups with older drivers—a base of information and opinion has emerged that will guide future research in the area of medication use and driving functioning.  The conclusions and recommendations that follow are keyed to specific topics and research design issues highlighted in the objectives and statement of work for this project.  Specifically,

  • What is our current understanding of the effects of drugs/medications, and combinations of medications, on crash risk/crash involvement, and which should be the focus of continuing research on polypharmacy and driving functioning?

  • What are the most feasible and reliable means of measuring/monitoring drug/medication usage—including over-the-counter drugs—by older drivers?

  • What are the most practical and valid ways of assessing the impact of drug/medication use on actual driving performance?

  • What are the potential barriers to the participation of older persons in studies of drug/medication use and driving functioning, and how might they be overcome?


Our conclusions in this area rest upon the preceding review by Wilkinson and Moskowitz (2001) and the update for recent (2001 through 2004) reports conducted in the present literature review.  An exploratory analysis of an administrative claims database by LeRoy (2004), which linked patient-level pharmacy information with codes denoting injuries resulting from motor vehicle crashes, also was instrumental in shaping our current understanding of potentially driver impairing (PDI) medications.  For crash-involved drivers age 50 and older, the combinations that now appear to be of greatest concern are:

  • Narcotics/narcotic analgesics plus
    • Non-steroidal anti-inflammatory drugs (NSAIDs)
    • Skeletal muscle relaxants
    • Anti-anxiety drugs
    • Selective serotonin reuptake inhibitor (SSRI) antidepressants
    • Antibiotics
    • Gastric acid secretion reducers

  • Narcotics/narcotic analgesics plus NSAIDs plus
    • Skeletal muscle relaxants
    • Antibiotics

A number of single (classes of) medications that have been linked to increased crash risk, including antihypertensive agents and anti-diabetic agents, have not shown significant associations with motor vehicle crashes/injuries in combinations with other (prescription) drugs in the limited cases that have been examined to date.  From our review of the literature, we would notconclude that sufficient evidence exists to exclude these classes from future polypharmacy and driving studies, however.

This perspective was ratified by the comments of experts participating in the brainstorming session.  They were also in agreement that the proper focus for continuing work in this area should be on classes of medications—rather than the total number of  drugs—that a person is taking. 

The brainstorming session participants, having been provided in advance with the literature review findings, further prioritized drugs/medications for future research based on their exposure in the (community dwelling) population of older persons.  Our recommendations in this area accordingly are to concentrate on combinations including:

  • Alpha blockers and other drugs that affect blood pressure (anti-hypertensive medications). 
  • Sedating drugs such as the benzodiazepines, tricyclic antidepressants, and opioids.
  • Drugs that affect blood sugar levels (anti-diabetic agents and drugs that could potentiate hypoglycemic effects).

A research strategy of utilizing epidemiological studies and/or analyses of large (e.g., administrative claims) databases to select and prioritize drugs and combinations of drugs for future empirical studies was broadly recommended by the brainstorming session participants; we endorse this approach, too, in principle, although its feasibility remains to be demonstrated.  We also recommend that over-the-counter (OTC) drugs be included in future study designs to the extent that an OTC drug has been shown to impair functional ability(ies) needed for safe driving, or has been associated with significantly elevated crash involvement through database analyses or epidemiological research.

Finally, there was support in the literature and among the experts in the brain-storming session that the interaction of medications and alcohol canbe properly excluded as a primary focus of continuing polypharmacy research.  It is recommended that potential subjects in such studies be screened to identify and restrict the participation of those who drink alcohol in addition to their medication regime. 


The literature review indicated a variety of ways in which researchers might learn which medications an (older) person is using, each with its own strengths and limitations.  Taking practicality and feasibility into account, in addition to the accuracy of results likely to be obtained using competing methods considerably narrows the list of approaches that can be recommended to measure/monitor drug use by research study participants.

The overriding conclusion in this area is that, to obtain complete and reliable information from older persons about their medication regimes, the researcher must simultaneously address a number of specific and essential concerns:

  • Privacy – Older persons will divulge information about the drugs they take only to someone who they believe will keep it in confidence, who will not communicate about it to anyone (including a family member, or their own physician) without their consent.  It is therefore important to rely on face-to-face interactions, which can provide such assurances, whereas mailed surveys or telephone contacts cannot. 

  • Security – Older persons do not want a stranger coming to their residence if it can be avoided. 

  • Credibility – Older persons prefer to discuss the medications they are taking with a recognized authority in this area, one who can provide advice and answer questions as well as receive information.

With this in mind, it is recommended that researchers do not attempt to learn about an older person’s medications by initiating direct contacts with a son/daughter or spouse, or with the individual’s doctor.  The best approach emerging from the focus group discussions is a form of the “brown bag” method—which was also rated highly by the brainstorming session experts—where an older research participant is asked to bring all of his/her medications (including over-the-counter drugs) to an office or other neutral, non-threatening location to be inventoried by an appropriate professional.  Pharmacists (active or retired) are the best choice in this regard, although nurses also may have sufficient perceived authority in this area, and offer a strong “comfort level.”  A member of the research team who can provide the required assurances of privacy and confidentiality should also be present during the brown-bag interview.

If possible, multiple methods should be employed to identify what medications an older person is taking.  Specifically, the combination of accessing an administrative claims (pharmacy) database and using a brown-bag interview may be recommended.  The information obtained from the pharmacy database will provide a historical record of the individual’s (prescription) drug regime.  It can serve as a screen in subject recruitment, and provides a valuable point of reference for confirming a subject’s expected drugs and dosages during a specific study period.

A self-report of all drugs taken within twelve hours preceding a driving evaluation by a consenting older research participant should be obtained.  Corroboration of his/her drug usage preceding the driving evaluation by a spouse or family member is also desirable—which, it may be noted, is markedly different than beginning the inquiry into an older person’s drug regime through such contacts.


The literature review generated descriptions of various measurement approaches to investigate the relationship between (multiple) drug usage and driving functioning, which were discussed and rated by experts in the brainstorming session.  The top candidates emerging from these project activities included in-vehicle evaluations, on both a closed course and (with certain safeguards) in actual traffic; instrumented vehicle studies to observe driver behavior under more “naturalistic” conditions; and clinical measures of the effect of medications on specific functional abilities needed for safe driving.

The focus group discussions with older drivers refined our understanding of what are likely to be acceptable methodologies with this cohort.  One conclusion from these discussions is that older persons have a strong preference for using their own vehicles, and especially if an evaluation is to be conducted in actual traffic.  A closed driving course was also clearly preferable to an on-road evaluation among the focus group participants, because of safety concerns; and this concern was magnified by the prospect of using an unfamiliar vehicle (e.g., a dual-control car) and/or having a stranger (an occupational therapist or other driving evaluation specialist) riding with them.  The responses of the older discussants to an instrumented vehicle methodology were split, with some objecting to the “big brother” aspect of unobtrusive monitoring, while others very much appreciated the prospect of driving their own cars without an evaluator physically present.

Our resulting recommendations for measuring driving performance with older persons taking (multiple) medications include the use of:

  • Off-road (clinical) assessments of functional abilities – These may be performed using paper-and-pencil methods but computer-based methods are recommended if feasible, to promote standardization in test administration and scoring and to permit the use of specialized cognitive tests (e.g., processing speed) that require precise timing.  The functional measures derived from prior NHTSA research in this area4 are recommended.

  • In-vehicle assessments on a closed course with a driving evaluator – These behind-the-wheel measures should be performed in a dual-control vehicle, as a safety precaution; because they are not exposed to actual traffic conditions it is likely that this will be acceptable to a majority of older research participants.  The driving evaluator should be an OT or other driving evaluation specialist, using an evaluation protocol that is accepted by an applicable certifying body, institution, or organization.

  • Assessments under actual traffic conditions in an instrumented vehicle – The older person’s own car should be instrumented for this purpose.  To heighten acceptance of this method, it should be emphasized that individuals need only drive as they would “normally” (i.e., apart from any involvement with the study) and that it is not necessary for another person to accompany them.

An across-the-board recommendation, regardless of the methodology employed, is that researchers evaluate the effects of medications on driving performance among older persons only for those drugs that individuals are already taking prior to their enrollment as test subjects.


The conclusions to be drawn in this area rest primarily on the comments of older persons participating in the focus group discussions held in Pennsylvania and Florida.  On the positive side, there was a general willingness on the part of discussants to become involved in research that can improve safety for older drivers, including studies that require them to divulge information about the drugs they are taking and/or to have their driving performance evaluated.  Their participation remains contingent upon a number of explicit requirements, however, as referenced in the following recommendations.

To encourage the participation of community-dwelling older persons in future investigations of the effect of drugs on driving functioning, researchers should:

  • Provide absolute assurances that information and data obtained in the study will remain confidential, with all individual participants remaining anonymous in reports of the study’s results.

  • Provide absolute assurances that participation in the study will have no impact on the license status or driving privilege of any individual, subject to the laws of the State in which the study is conducted.

  • Strictly observe informed consent procedures that specify to study participants all of the parties to whom the results of their driving evaluations may be made available.

  • Provide incentives, preferably cash but possibly coupons for services at local stores or restaurants, that represent fair compensation for the amount of time a participant must commit to the study.

  • Clearly explain the benefits, to society and to themselves, that are expected to result from the older driver’s participation in the planned study.

4cf. Model Driver Screening and Evaluation Program (vol. 3), DOT HS 809 581.  May be accessed at