Chapter 3:
Functional Limitations Review

The project activities covered in this chapter established the limits of scientific knowledge relating to age-related functional limitations (and medical conditions associated with diminished functional abilities) which are likely to result in driving impairments, and the influence of such changes on the likelihood of crash involvement. A review of recent and in-progress studies was conducted to identify and contrast alternative tests and procedures for assessing drivers' functional ability, and their relationship to crash risk or to performance measures of driving competency. To provide a concise overview while maximizing the accessibility of the findings drawn from this large body of literature, a tabular format was selected to present the initial product of this research--the Annotated Research Compendium of Age-Related Functional Impairments and Driving Safety--which was later incorporated into the Safe Mobility for Older People Notebook.

Using the Compendium as a resource, a consensus among prominent researchers and experts in specific domains of functional assessment was sought regarding impacts of measured deficits on driving performance, taking into account the varying demands of the driving task in different settings and circumstances. This activity was undertaken using an iterative, structured survey approach known as a "Delphi" exercise, with separate groups addressing each of three domains: (1) sensory processes/vision; (2) attentional and perceptual processes; and (3) physical and medical factors and dementia. The output of this effort was a set of tables prioritizing the types of functional tests to perform in a screening program.

Safe Mobility Notebook and Compendium of Research Findings

The review of the literature describing older driver diminished capabilities as they relate to unsafe driving performance leading to increased crash risk began with a review of three comprehensive reports on the topic by Janke (1994), a draft manuscript of Hu, Trumble, Foley, Eberhard, and Wallace (1998), and Staplin, Ball, Park, Decina, Lococo, Gish, and Kotwal (1997). In addition, e-mail and telephone contacts were made with the researchers of the projects identified below requesting updates on research in-progress, throughout this task.

• "Develop Performance Assessment Techniques" (NHTSA).
• "Identifying At-Risk Older Drivers" (Andrus Foundation).
• "Physician Assessment Tools" (Andrus Foundation).
• "Role of Cognitive Style in Driving Skills" (Andrus Foundation).
• "Validation of the Senior Driver Research Inventory" (Ontario Ministry of Transport).
• "Analysis of the Useful Field of View" (Ontario Ministry of Transport).
• "Evaluation of a Behavioral Intervention to Reduce Crash Involvement and Injuries in Older Drivers" (National Institute on Aging).
• "Dementia and Driving Performance" (National Institute on Aging).
• "Effects of a Cognitive Training Intervention on Crash Involvement" (National Institute on Aging).
• "Effects of Cognitive Training or a Simulator-Training Intervention on Actual On-the-Road Driving" (National Institute on Aging).
• "Elderly Driver Referral" (Centers for Disease Control).
• "Driving Ability and Car Crashes in Old Age and Dementia" (Centers for Disease Control).
• "The Safe Older Driver: Sensory and Medical Characteristics" (Andrus Foundation).
• "Predictors of Safe and Unsafe Driving in the Elderly" (Andrus Foundation).
• "Longitudinal Study of Health Status and Driving Risk of Older Drivers" (Centers for Disease Control).

The research studies cited in the Staplin et al. (1997) report were performed between 1960 and 1992, and where possible, the technique of meta-analysis was used to facilitate the integration of the large body of findings relating functional capability to crash risk. This document also includes text discussing the results and implications of the cited research findings and was useful for the determination of what sensory, perceptual, and physical capabilities are important for safe driving, but not necessarily how these capabilities should be measured. The literature review performed by Janke (1994) is a comprehensive report that describes normal impairments associated with aging, medical conditions associated with aging, instruments used for assessing functional abilities necessary for driving, and licensing programs for older drivers. Survey data and crash data were the basis of the Hu et al. (1998) document, used to generate a model to identify factors placing older drivers at risk for crashes, and factors that relate to driving cessation.

The information contained in these documents, while thorough, was unwieldy for the purposes of the current project. To try to simplify the task of prioritizing measurement needs and potentially useful tools and procedures to meet them, research-in-progress and research completed within the past 10 years bearing on the effectiveness of tests vis-à-vis the driver performance-versus-safety relationship was synthesized and reformatted into tabular format. This table became the Annotated Research Compendium of Age-Related Functional Impairments and Driving Safety.

The Annotated Research Compendium contains six columns with the headings: Test; Subjects; Procedure/Test Description; Where Applied; Findings; and Researchers. Each row in the table is dedicated to a description of one test that has been applied in a single research study to evaluate how predictive it is of the research subjects' driving performance. There are multiple rows for tests that have been employed by multiple researchers. It was frequently the case that a researcher employed a battery of tests in one study; but wherever possible, each test was broken out and is presented separately in its own row in the Compendium.

The information contained in these reviews was used as the starting point for a "Delphi" exercise, described below, that sampled the opinions of experts in each of three "functional area working groups" regarding the relationships between diminished capability and driving safety.

It must be noted that the identification of fair, accurate, and administratively-feasible screening procedures to detect functionally-impaired drivers was the first--but not the only--information need addressed in this project task. Loss of mobility is a serious health and quality of life issue for older people. Accordingly, the Model Program recommendations emerging from this research must also address a host of related issues, most notably the potential for remediation of impairing conditions; the need for education and counseling for older drivers faced with restriction or cessation of driving; and the availability of safe and accessible alternative transportation options. The literature review and synthesis culminating in the Compendium also revealed a great deal of pertinent information about these related topics. This information was assembled, together with the Annotated Research Compendium, into one comprehensive volume titled the Safe Mobility for Older People Notebook. This document was intended to serve as a resource, providing a snapshot of current knowledge and practices circa 1999, which could support a broad array of program initiatives.

Notebook topic areas include:

• Identification of high-risk older drivers.
• Counseling and remediation of at-risk drivers.
• Public information and educational tools to support program implementation.
• Mobility options and alternatives to transportation for seniors.

Coverage of topics in the Notebook within each of these areas is provided in 70 subtopic discussions, containing three sections each: (1) a summary of outcomes in relevant research studies and implementation efforts; (2) conclusions/preliminary recommendations pertinent to the development of a national Model Program; and (3) references identifying data sources. The Safe Mobility for Older People Notebook may be accessed using a link on NHTSA's website at www.nhtsa.dot.gov/people/injury/olddrive/safe.

Expert Panel Recommendations

A key activity in this task was to pinpoint functional deficits which can be most confidently linked to a "significant increase in motor vehicle crash risk." Three Functional Area Working Groups were established, and then half a dozen nationally renowned experts were sought as members of each group. The three Working Groups focused respectively on: (1) sensory (visual) processes; (2) attentional and perceptual (cognitive) processes; and (3) medical factors/dementia. The membership of each group is presented in table 1.

The collective judgments of experts in each Working Group were obtained in a Delphi exercise, based on an iterative, confidential ranking and sorting of responses. This procedure was chosen because of its three main characteristics: anonymity of groups; interaction with controlled feedback; and statistical summaries of group response (Gustafson et al., 1975).

The Delphi exercise was conducted via fax and e-mail. Prior to initiating the exercise, a request was made of each member to provide a candidate list of critical dimensions of functional capability for which they believed it was most likely that deficits would lead to increased crash risk. This input was used to create a matrix of functional capability-by-driving task demand for the Delphi exercise. The functional abilities emerging from this inquiry are listed in table 2.

Next, four driving situations were defined, in recognition that varying task demands will result in varying probabilities of driving error--and, presumably, risk of a crash--for a given level of functional deficit. Two key variables influence task demand: (1) traffic speed, and (2) the overall complexity of the traffic operations in the setting. The first variable is straightforward; it determines the time constraints on a driver for all required information processing operations needed to result in safe and effective vehicle control. The second variable is less well-defined, but is assumed to include the perceptual "texture" of the environment (e.g., visual clutter along the roadway); the number and spatial distribution of other road users (potential conflicts); the level of adjacent land development, with the associated problems of uncontrolled access into the traffic stream; and, the complexity of the roadway geometry and operational rules. Fundament-ally, this scheme was designed to account for gross differences in the nature and amount of information that must be processed while driving.

Thus, the classification scheme for driving situation was as follows:

1. High speed, high complexity (e.g., high-volume urban/suburban arterials, expressways, and freeways).
2. High speed, low complexity (e.g., low-volume suburban and rural freeways).
3. Low speed, high complexity (e.g., downtown streets and high-volume suburban arterials with heavy commercial development).
4. Low speed, low complexity (e.g., low-volume streets in less developed suburban and residential environments).

A "functional criterion matrix" was then created for each Working Group, described by columns labeled according to the functional abilities within each domain, from table 2, and rows labeled by the driving situations listed above, which were the same for each Group. Each member of each Working Group was provided with the appropriate functional criterion matrix, the Annotated Research Compendium, sections from the Staplin et al. (1997) and Janke (1994) documents, and a set of instructions including descriptions of the four roadway types and their associated driving task demands (e.g., vehicle control requirements, roadway information sources, and potential conflicts).

By design, the Delphi was to proceed in a series of step, as follows:

• Step 1. Instructions were provided to each Group member: Please make an entry in every cell in your functional criterion matrix to indicate a level of decline for the indicated functional ability that 'more often than not' will lead to a significant increase in crash risk, for the identified driving situations. It was hoped that Group members would take into account differences in task demands from one driving situation to another, and tailor their responses accordingly. Group members were also encouraged to provide a rationale for their responses.
• Step 2. Group members' responses within each cell of the matrix were ordered from maximum to minimum values. Measures of central tendency and dispersion were identified to summarize the range of responses received from each Group.
• Step 3. Group members were provided with the measure of central tendency from Step 2, as a benchmark, plus the dispersion in the Step 2 responses. They were then asked to submit a new response, with an explanation if it deviated from the benchmark. In this manner, the range of responses as narrowed, with the goal of a single, consensus value.

At Step 2, a criterion was adopted whereby over 50 percent of respondents must enter a value in at least one cell under a given column heading, for that functional ability to be retained in the Delphi exercise. Therefore, if more than half of the respondents left the column blank, citing either insufficient knowledge to provide a defensible input, or stating that in their opinion the capability in question a) was not relevant, b) could not be measured in a valid or reliable manner, or c) was redundant with another measure included in the table, that column was rejected for further consideration. Using this criterion, dynamic acuity and dynamic contrast sensitivity were dropped from the Sensory (Visual) Processes Group matrix; and proprioception and somatosensory processes, and strength and range of motion were dropped from the Medical Factors/Dementia Group matrix. The Delphi responses derived from Step 3 for each of these respective Working Groups, for the functional abilities remaining under consideration at this stage of the exercise, are presented in appendix A.

Unfortunately, the responses generated within the domain of Attentional/Perceptual Processes did not permit the Delphi to advance beyond Step 1. More than half of the members of this Group provided responses that may be paraphrased as "insufficient data to pick a 'best measure,' let alone a specific score" for the functional abilities included under this heading.

Even so, the Delphi exercise provided valuable insights that assisted with planning of the Maryland Pilot Study. Specifically, the research team, in conjunction with NHTSA, was influenced by these findings to concentrate the more ambitious data collection activities to follow in this project on the areas where experts' opinions indicated the greatest deficit of knowledge. A goal for continuing research efforts in these areas was thus established: to build a database describing population norms (i.e., for older drivers) and to enable analyses measuring the strength of relationship with traffic safety for, first, attentional/perceptual processes and second, physical strength and range of motion, as highlighted in this work.

In addition, candidate measures for assessing functional status were suggested by the experts participating in the Delphi exercise, for each of the abilities emerging as a column heading in the functional capabilities matrix. These suggestions, together with the exhaustive survey of related work comprising the Annotated Research Compendium, provided the basis for a preliminary selection of measurement (screening) procedures for use in the Maryland Pilot Study.

BACK  |  CONTENTS  |  NEXT