LITERATURE REVIEW 

INTRODUCTION

The purpose of the literature review was to update the Wilkinson and Moskowitz (2001) unpublished report Polypharmacy & Older Drivers‑ Literature Review. As such, it was limited to literature published since 2001 addressing polypharmacy, drugs, and older drivers.  Literature was also reviewed to identify techniques used to measure/test driving skills. 

Searches were conducted in Embase (which includes MedLine), PsycINFO, TRIS, and SafetyLit Weekly for recent literature (2001 to 2004) with any of the following keywords:

  • Polypharmacy
  • Multiple medication use
  • Prescription drugs
  • Over-the-counter drugs
  • Drug use
  • Older drivers
  • Older adults
  • Elderly persons
  • Community-dwelling elderly
  • Medication review
  • Brown bag method
  • Medication management
  • Medication compliance
  • Medication adherence
  • Medication persistence
  • Medication event monitoring
  • Self medication
  • Medication side effects
  • Driver impairing medications
  • Driver impairing medical conditions
  • Inappropriate prescribing
  • Contraindicated drugs/medications
  • Geriatrics
  • Gerontology
  • Medicaid/Medicare pharmacy claims
  • Veterans administration pharmacy benefits management database
  • Pharmaceutical claims
  • Prescription database
  • Medication usage/tracking
  • Medication dispensing technology
  • Motor vehicle crashes
  • Motor vehicle accidents
  • Dose administration aids
  • Driving skills
  • Driving ability
  • Driving performance
  • Falls or falling

 

Approximately 1,600 abstracts were identified in the original searches. A review of titles indicated that many reports focused on medications for resistant conditions. Eliminating the terms “resistance” and “resistant” and limiting the search to human subjects age 65 and older, and abstracts to those in English, reduced the search results to 365 potentially relevant articles.   In selecting articles for review, only those describing populations of likely drivers were included (i.e., “community-dwelling” populations, as opposed to residents of nursing homes or residents of group homes who suffer from developmental disabilities or conditions such as schizophrenia).  Systematic differences in prescribing and in monitoring compliance in controlled settings such as nursing homes and hospitals limit the generalizability of information from these populations to community-dwelling older persons.  In addition, research on multiple medication effects was primary over research on single medication effects; however, reports presenting the effects of single medications on driving performance were included to the degree that the research focused on the effects of the drug on driving performance or falling in older community-dwelling people. A set of 300 articles resulted from this first screen.

The selection of reports for review underwent further refinement with the assistance of two project consultants1. This screening process produced a final review set of 143 articles identifying medication use and measuring medication adherence; measuring driver performance; and polypharmacy and older persons.  

The literature review has been published as a stand-alone document entitled, Literature Review of Polypharmacy and Older Drivers: Identifying Strategies to Study Drug Usage and Driving Functioning Among Older Drivers (Lococo and Staplin, 2006).  It is divided into three major sections to address the information needs of the current project: 1) Medication/ Polypharmacy Effects on Older Persons; 2) Methods of Measuring/Monitoring Medication Usage; and 3) Methods to Measure Driving Performance.  Although prescription medications were the major focus of this review, over-the-counter medication use by older persons is included where it was reliably documented in the literature. Additional areas that received attention in the literature review on polypharmacy include: how polypharmacy among older adults impacts areas other than driving (such as falls), and older persons’ use of alcohol in combination with other medications—as alcohol use was not the focus of this project.  A summary of this literature follows.

MEDICATION/POLYPHARMACY EFFECTS ON OLDER PERSONS

The first section of the literature review describes medication use in the older community-dwelling population (not hospitals or nursing homes) in the United States and other countries.  Within the population of community-dwelling older persons in the U.S., more than 90 percent of persons age 65 or older use at least 1 medication per week; more than 40 percent use 5 or more different medications per week; and 12 percent use 10 or more different medications per week (Gurwitz, 2004).   Although the number of medications used to define “polypharmacy” ranges in research studies from 2 to 10 (Lee, 1998), the findings of Allard, Hébert, Rioux, Asselin, and Voyer (2001) that the consumption of 3 or more drugs per day increases the risk of functional decline in elderly people by 60 percent deserve attention. Decreases in functional ability brought on by polypharmacy have been associated with an increased risk of motor vehicle crashes (LeRoy, 2004), raising public health and safety concerns. 

In a cohort study of nearly 28,000 Medicare+Choice enrollees cared for by a multispecialty practice (an ambulatory clinic setting) during a 12-month study period between 1999 and 2000, researchers found that 75 percent of the sample received prescriptions for 6 or more prescription drugs (Gurwitz, Field, Harrold, Rothchild, Debellis, Seger, Cadoret, Fish, Garber, Kelleher, and Bates, 2003).  Forty-nine percent of the sample was prescribed medications in four or more categories.  Combinations of medication use were not reported; however, the specific prescription medication categories and percentage of enrollees receiving prescriptions were as follows: 


  • Cardiovascular (53.2%)
  • Antibiotics/anti-infectives (44.5%)
  • Diuretics (29.5%)
  • Opioids (21.9%)
  • Antihyperlipidemic (21.7%)
  • Nonopioid analgesics (19.8%)
  • Gastrointestinal tract (19.0%)
  • Respiratory tract (15.6%)
  • Dermatologic (14.8%)
  • Antidepressants (13.2%)
  • Sedatives/hypnotics (12.9%)
  • Nutrients/supplements (12.3%)
  • Hypoglycemics (11.5%)
  • Steroids (9.7%)
  • Ophthalmics (9.6%)
  • Thyroid (9.4%)
  • Antihistamines (9.2%)
  • Hormones (9.1%)
  • Anticoagulants (7.0%)
  • Muscle relaxants (5.4%)
  • Osteoporosis (5.3%)
  • Antiseizure (3.4%)
  • Antigout (3.2%)
  • Antineiplastics (2.8%)
  • Antiplatelets (1.3%)
  • Antipsychotics (1.2%)
  • Antiparkinsonians (0.9%)
  • Alzheimer disease (0.9%)
  • Immunomodulators (0.04%)

The prevalence of potentially inappropriate medication use in this population has been found to range from 6 percent to 21 percent.  Potentially inappropriate medications are those that: (1) should generally be avoided in persons 65 years or older because they are either ineffective or they pose unnecessarily high risk for older persons and a safer alternative is available, and (2) medications that should not be used in older persons known to have specific medical conditions (Fick, Cooper, Wade, Waller, Maclean, and Beers, 2003).  Data examined by Aparasu and Mort (2004) indicate that the most prevalent of the inappropriately prescribed prescriptions to community-dwelling individuals in the U.S. include: amitriptyline and doxepin (antidepressants); chlordiazepoxide, diazepam, and flurazepam (benzodiazepines); chlorpropamide (an anti-diabetic/hypoglycemic); dipridamole (an anti-platelet agent); hydroxyzine (an antihistamine); meprobamate (an anti-anxiety drug); oxybutynin (an antispasmodic agent for the urinary tract); propoxyphene (a narcotic analgesic); non-steroidal anti-inflammatory drugs; and barbiturates.  Appendix A in this Final Report provides detail about specific potentially inappropriate medications found in the literature review to be commonly prescribed for older, community-dwelling individuals, their classification, and side effects relevant to a discussion of safe driving ability.

Studies on the effects of medications, and the physiological changes that occur with age and affect how older people metabolize their medications are examined next.  A range of physiological changes that may affect drug metabolism occurs with age.  The liver plays a central role in the termination of drug action and has, therefore, been well studied.  Liver size or volume and hepatic blood flow both decrease with age. Other physiological changes that occur with aging include reduced body mass and basal metabolic rate, reduced proportion of body water, increased proportion of body fat, decreased cardiac output, altered relative tissue perfusion, decreased plasma protein binding, reduced gastric acid production and gastric emptying time, and reduced gut motility and blood flow (Herrlinger and Klotz, 2001). Although the effect of aging on human drug metabolism has been much studied, few generalizations about how aging affects human drug metabolism have emerged.  Although some measures of drug metabolism are diminished in the elderly, significant interindividual variability in drug metabolism, drug action, and adverse reactions characterizes the elderly population (Herrlinger and Klotz, 2001; Schmucker, 2001; and Kinirons and O’Mahony, 2004).
 
The final section in this segment of the of the literature review provides a general overview of medication use and crash risk by examining recently conducted epidemiological and experimental studies to determine the consequences of a single class of medication on the ability to drive safely. Potentially driver impairing (PDI) medications are associated with known effects on the central nervous system, blood sugar levels, blood pressure, vision, or otherwise have the potential to interfere with driving skills. Possible PDI effects include sedation, hypoglycemia, blurred vision, hypotension, dizziness, fainting (syncope), and loss of coordination (ataxia).   Single classes of medications that have been associated with an increased crash risk include the benzodiazepines, opioids, sedating antidepressants, narcotic analgesics, anti-anxiety agents, antihypertensive agents, skeletal muscle relaxants, antimanic agents (lithium), and anti-diabetic agents (LeRoy, 2004; Walsh, de Gier, Christepherson, and Verstraete, 2004; Fishbain, Cutler, Rosomoff, and Rosomoff, 2003; Leveille, Buchner, Koepsell, McCloskey, Wolf, and Gagner, 1994; Ramaekers, 2003; Szlyk, Mahler, Seiple, Vajaranant, Blair, and Shahidi, 2004). 

A few studies are reported that examine multiple medication use. However, there is a dearth of research on the effects of combinations of specific medications or even combinations of drug classes on driving ability per se.  However, in one recent and comprehensive pharmacy database analysis of multiple medicine use (LeRoy, 2004), higher percentages of crash-involved drivers were prescribed two or more prescriptions than non crash-involved drivers.  The most-frequently appearing drug combinations (in descending order of frequency) in the group of crash-involved drivers age 50 and older were:

  • Narcotics + Non-Steroidal Anti-Inflammatory Drugs (NSAIDs).
  • Skeletal Muscle Relaxants + NSAIDs.
  • Narcotics + Skeletal Muscle Relaxants.
  • Narcotics + Skeletal Muscle Relaxants + NSAIDs.
  • Narcotics + Antibiotics.
  • Gastric Acid Secretion Reducers + Narcotics.
  • Anti-Anxiety Drugs + Narcotics.
  • Serotonin Reuptake Inhibitor (SSRI) Antidepressants + Narcotics.
  • Narcotics + NSAIDs + Antibiotics.

MEASURING/MONITORING MEDICATION USAGE

Researchers across several decades have described patient compliance as “the best documented, but least understood health behavior” (Coons, 2001; Becker and Maiman, 1975).  Although a variety of methods to measure compliance exist, problems with validity and reliability are inherent with every one of them (Marinker, Blenkinsopp, Bond, et al., 1997).  Vik, Maxwell, and Hogan (2004) state that presently there is no generally accepted gold standard for measuring adherence.  This section of the literature review begins with a discussion of methods used to measure compliance, including the pros and cons of each method. These methods include: clinical judgment, patient’s self report, clinical response, biochemical measures, pill counts, pharmacy records, and electronic medication monitoring devices. A brief summary of the pros and cons of each method follows.

Physician’s Clinical Judgment.  Studies have shown that physicians’ clinical judgments of compliance are no more accurate than predicting compliance by chance (American Pharmacists Association/APhA, 2003; Bikowski, Ripsin, and Lorraine, 2001).

Patient’s Clinical Response. The APhA (2003) reports that for many medications, a patient’s clinical response to a medication is only weakly related to compliance.  In addition, while some studies suggest that the absence of predictable adverse side effects is correlated to noncompliance, the link between compliance and adverse effects has not been consistent in the literature.

Patient Self Report.  In many studies, information obtained using the patient self-report method has correlated positively to information obtained from random pill counts, pharmacy records, biochemical measures, and electronic medication monitors. In several studies reviewed by APhA (2003), querying patients about their compliance resulted in the detection of over 50 percent of patients with poor compliance, with a specificity of 87 percent. Congruence of pharmacy records and self-reported medication use is of importance because self-reports of medications are often used as a surrogate for health status or the presence of chronic diseases, and are important when studying medication compliance, polypharmacy, and drug interactions.  Self reports of medication are frequently used in population-based studies where pharmacy records are lacking or are expensive to obtain.  A particular self-report method—the brown bag method of medication review—has been found to provide a reasonable substitute for pharmacy records as a measure of current medications (Caskie and Willis, 2004).  

Biomechanical Measures. Biochemical measures of patient compliance (blood and urine sampling) to measure drug presence and concentrations are objective, but not practical, convenient, or appropriate for most circumstances, and not always reliable (APhA, 2003). 

Pill CountsPill counts measure compliance by comparing the number of doses remaining in a container with the number of doses that should remain, if the patient’s compliance were perfect.  This method can provide an overestimation of compliance if the patient is aware that a pill count is going to be conducted—patients may remove excess doses and discard them.  Another drawback to this method is that it cannot verify that a dose removed from a container was actually consumed, or whether it was consumed at the correct time.  Pill counts are not suitable for medications taken on an as-needed basis (APhA, 2003).  

Pharmacy Records/Administrative Databases.  The use of pharmacy records to estimate compliance based on pharmacy refills, correlates favorably with electronic measurement, but shares some of the same problems that are intrinsic to pill counts (APhA, 2003).  A refill record will provide information about how much medication was dispensed in a given interval, but it can not validate that the medication was actually consumed or consumed at the correct time.  In addition, refill records cannot guarantee that a patient is not obtaining medications from pharmacies other than the one being monitored, and cannot guarantee that the patient is not stockpiling medications or sharing them with others.  Also, samples given to patients by their physicians would not show up in the pharmacy database, resulting in an overestimate of noncompliance.  Pharmacy records were used in some of the studies that employed the brown-bag method, as a means of identifying members at risk for polypharmacy for inclusion in those medication-review studies. Other researchers have used pharmacy databases to determine the relative frequency of various combinations of medications, and to conduct case-control studies of the use of medications and adverse outcomes such as motor vehicle crashes (LeRoy, 2004).  As noted by LeRoy, an advantage to the use of claims data is that it is not dependent on patient recall of medication and disease information. However, in interpreting results derived from analyzing administrative claims data, the following sources of error or influence must be considered: reporting error (including under-reporting); ascertainment error (correctly billed but incorrectly diagnosed); and detection bias (frequent visits yield increased opportunity to detect). 

Electronic Medication Monitoring DevicesElectronic medication monitoring devices use microchip technology to record and download data to a computer for review and analysis describing the actual data and time that a dose is removed from a container.  Although they are recognized as the “gold standard” of compliance assessment, the method has drawbacks similar to the pill count method and the pharmacy refill method—electronic medication monitoring devices cannot assure that a dose that was removed was actually consumed or administered correctly. In addition to errors introduced when a bottle is opened and no medication is taken, errors may be introduced when a patient removes two doses during one opening (one for the morning and one for the afternoon) and only one event gets recorded for two doses taken (APhA, 2003)..  Despite these limitations, electronically measured adherence has been more highly associated with clinical outcomes than self report (Liu, Golin, Miller, Hays, Beck, Sanandaji et al., 2001) and pill counts (Namkoong, Farren, O’Connor, and O’Mally, 1999). 

This section of the literature review also examines the factors affecting compliance with medication regimes. Medication-related factors associated with low compliance include: increases in the complexity, cost, and duration of medication regime; increases in the number of prescribed medications; and increases in the severity of side effects. Patient-related factors that correlate with low compliance include: limited access to health care; financial problems; communication barriers; and lack of social support. The prescriber-related factors that correlate with low compliance include: a poor prescriber-patient relationship; poor prescriber communication skills; a mismatch between the prescriber and patient regarding health beliefs; and a lack of positive reinforcement from the health-care provider.

This section of the literature review also examines several factors that affect older persons’ willingness to participate in research and offers some lessons learned to help in recruitment of elderly patients into studies. Cooperation of the candidate subjects’ pharmacy and physicians increases subjects’ willingness to participate in studies (Ory, Lipman, Karlen, Gerety, Stevens, Singh, Buchner, Schechtman, and the FICIT Group, 2002).  Patients are more likely to undergo medication review if it is recommended by their physicians (Nathan, Goodyer, Lovejoy, and Rashid, 1999; Ory et al., 2002).  In addition, explaining the benefits of participating a medication review in recruitment letters to subjects has been shown to increase participation (e.g., “a medication review will ensure that you are taking only the necessary medications to control your medical conditions, the doses are correct, and the medications are safe to be taken together”) (MacRae, Lowrie, MacLaren, Kinn, and Fish, 2003; Farris, Ganther-Urmie, Fang, Doucette, Brooks, Klepser, and Kuhle, 2004).

METHODS TO MEASURE DRIVING PERFORMANCE

The literature review concludes with a section that synthesizes methods used in studies to measure driving performance, highlighting methods that appear to hold the greatest promise for evaluating the effects of drugs on driving performance while also acknowledging shortcomings and limitations that have been reported in the literature.

Road tests have long been considered the gold standard for measuring driving ability.  They have widely-recognized limitations, however.  In addition to  inconsistencies in the administration and scoring of the results, which presents a major challenge to standardized and objective assessments, as a rule examinees are not exposed to the most risky and demanding situations where driving errors that lead to crashes are most likely.  Nevertheless, Ramaekers (2003) asserts that actual driving tests are essential to conclusively define the potential impact of drugs on driving.  The literature review presents the advantages and disadvantages of naturalistic studies (driving in traffic) and controlled driving studies (driving on a closed course). 

The literature review also includes a discussion of the pros and cons of driving simulation to measure driving performance. Driving simulators have been touted as offering experimental control for driving performance evaluation; they have also been criticized for a lack of fidelity for many aspects of actual driving (and therefore poor generalizability to conditions outside the laboratory), as well as simulator sickness (particularly for older adults).  Another difficulty in evaluating the utility of this method is the ambiguity attached to the term “driving simulation.”  Testing systems bearing this label range from actual vehicles on motion platforms with fully-interactive control over high-resolution virtual environments, to “driving bucks”—a single seat with basic wheel and pedal controls for user inputs—that offer a more limited display of the driving environment and little or no motion, to desktop computer graphic/video presentations of selected driving scenarios for ‘part-task’ measurement (e.g., visual search, hazard detection).  This review considers three categories, or levels, of simulation that range from non-interactive, computer graphic and/or digital video visuals with no motion, to interactive, computer graphic visuals with full motion.    

As one additional perspective, a combination of approaches to investigate the effects of medications on the ability to drive has been advocated by Álvarez and del Río (2002).  These authors note that relevant psychological and functional capabilities can be analyzed using vigilance and performance tests, psychomotor test batteries, reaction tests, etc., but should be complemented by simulator and on-road studies.  Similarly, Keller, Kesselring, and Hiltbrunner (2003) offered the opinion that psychological tests in combination with an on-road test allow for a balanced judgment of a patient’s fitness to drive.  Their psychological assessment (including cognitive tests, a tracking task with divided attention, psychomotor tests, and assessments of impulse control and spatial orientation), when compared to the results of an on-road test, yielded consistent results in 88 percent of the cases studied (38 of 43 patients with neurological disabilities). Keller et al. (2003) also cite other researchers (Sundet, Goffeng, and Holt, 1995) in concluding that, while a psychological assessment sheds light on a subject’s perceptual efficiency, goal-oriented behavior, distractibility, psychomotor efficiency, and impulse control, a road test is key to demonstrate the subject’s capacity to retrieve from procedural memory the technical skills of handling a car, the capacity to allocate and shift attention, and to keep a general overview of concurrent events, all of which are required for safe and successful driving.


1 Dr. Robert Raleigh and Dr. Marion Anders served as consultants in this project.  Dr. Robert Raleigh recently retired as the Chief of the Maryland Medical Advisory Board.  Dr. Anders is a world renowned expert in interactions of single and multiple drugs.  He is recently retired as Chair of the Departments of Pharmacology and Physiology at the University of Rochester Medical School.