| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Karlene Ball, PhD, Department of Psychology, Edward R. Roybal Center for Research on Applied Gerontology, University of Alabama at Birmingham, HM 100, 1530 3rd Avenue South, Birmingham, AL 35294-2100. E-mail: kball{at}uab.edu
 |
Abstract |
|---|
Key Words: Older drivers • Driver screening • Crash prediction
The relationships between basic research, applied research, and translational research are not always clear cut. Basic research refers to laboratory-based studies with precise control over experimental variables. Applied research can take many forms but usually involves the testing of phenomena or interventions in clinical or field site settings. Translational research involves the further step of implementing the research results in real-world settings, often under the control of an implementing agency. In these cases, the research findings may be used to make decisions about a treatment or may have some other behavioral consequence. Although basic research is traditionally thought to precede applied research, which typically precedes translational efforts, the order in which these forms of research occur is affected by multiple factors. Indeed, much basic research has arisen from the need to establish control over variables observed as a result of problem-driven or applied research.
The Roybal Centers, which are named as Centers for Research on Applied Gerontology, have focused on applied studies but have also included basic and, more recently, translational research. For example, the Roybal Center at the University of Michigan has conducted basic investigations of the relationship between age-related cognitive changes and patient behavior in medical settings. Results are being applied to the development of principles for designing medical instructions and devices. These principles will be disseminated to health care providers and medical technology developers in order to translate the findings into procedural changes. The Boston University Roybal Center has tested a fear-of-falling intervention among a patient population at risk for disablement (applied research) and has succeeded in implementing this program at other medical centers across the United States, thereby translating the research into practice.
This article reviews the history behind our own ongoing Roybal Center study of driving competence that is being conducted within the Maryland Motor Vehicle Administration, an example of translational research that stems from a decade of both basic and applied research. The article focuses on the challenges faced in moving this research from the laboratory to the real world.
|
Background |
|---|
|
Asking Applied Research Questions |
|---|
|
Identifying Early Field Collaborators |
|---|
|
Expanding to Field Research |
|---|
The importance of giving presentations and placing peer-reviewed publications in diverse outlets should also be mentioned as a key ingredient in helping to form collaborations and contacts with outside businesses and agencies involved in driving research. Presentations to mixed groups of researchers and service personnel, such as the audience in attendance at meetings of the Transportation Research Board and its Committee on Safety and Mobility of Older Drivers, were instrumental in establishing contacts that resulted in collaborations on the Maryland project. Publications from our Roybal Center, and the resulting national media coverage, also led to numerous contacts with individuals in organizations such as AARP, law enforcement, other state departments of motor vehicles, the American Automobile Association, hospitals, clinics, driving rehabilitation services, state highway administrators, senior centers, insurance companies, automobile manufacturers, and other companies interested for a variety of reasons in the assessment of older persons. These contacts led to invitations to make presentations at meetings that would not typically be attended by members of the scientific community, but which provide an opportunity to establish further contacts for applied research venues.
As an outgrowth of contacts such as these, field collaborators with common interests in the safety and mobility of older drivers were identified. Specifically, Dr. Robert Raleigh of the Maryland MVA and Dr. Loren Staplin of Transanalytics, Incorporated (funded through a contract from the National Highway Traffic Safety Administration, or NHTSA), joined forces with investigators from UAB and Western Kentucky University to conduct a largescale study of a brief screening battery designed to predict crash and other mobility outcomes among older drivers in the state of Maryland. Thus, this translational field study was jointly funded by NIA (through the UAB Roybal Center), NHTSA, and the Maryland MVA. The study has screened over 5,000 participants, enrolled 2,207, conducted first annual follow-up interviews with a subset of the sample, and is in the process of conducting second annual follow-up interviews. Preliminary study results with respect to crash risk and other mobility outcomes have been reported at national scientific and other professional meetings, and a paper reporting these outcomes will soon be submitted for scientific review (Ball, Roenker, et al., 2002).
|
Developing the Study Battery |
|---|
The first step for the study investigators was to become acquainted with existing science concerning performance-based predictors of crash risk among older adults. Our own research over the previous decade informed the battery selection process, along with a thorough search of pertinent scientific and medical literature. In essence, the team of study collaborators compiled and discussed potential battery components, weighing each with respect to its reliability and validity, crash prediction potential, availability of age-appropriate normative data, ease of administration, and, perhaps above all, the cost and benefit yielded for each minute of test time. Indeed, the need to keep the battery brief was the overriding principle dictating which tests were ultimately included in the battery. In some cases, brief subtests of more extensive instruments were selected for inclusion in the final screening battery.
We relied on existing measures with known psychometric properties in the Maryland study because this strategy allowed us to compare results obtained in this setting with results obtained under controlled testing conditions. The process of developing, piloting, and validating new or modified instruments was beyond the scope and resources of the study.
General considerations in battery selection or development include the incorporation of multimethod, multimeasure procedures to the extent possible within a given field setting, in order to avoid test-specific effects that are difficult to interpret. An additional consideration in developing the protocol is its vulnerability to order effects such as practice and fatigue that can affect performance on individual components of the battery and obscure results with respect to the battery as a whole.
In the Maryland study, the final screening battery that met the criteria discussed above was a 15-min assessment composed of performance-based and self-report measures of medical conditions, physical function, visual function, and cognitive function. Specifically, the battery was composed of the Useful Field of View (UFOV) Subtest 2 (Ball & Owsley, 1993); the Gross Impairments Screening Battery (Staplin, Lococo, Gish, & Decina, 2002), incorporating existing measures of cognitive function and physical function; and a questionnaire assessing health, medication usage, driving habits, and general mobility.
|
Training Field Personnel |
|---|
Our experience with automated testing led to a broader discussion of the influence of test administrators on successful data collection efforts. Must a protocol be completely computerized, or can tests be used that require some degree of judgment of the administrator? In general, the more automated that test instructions, stimulus materials, test protocol, test scoring, and data storage can become, the better. Nevertheless, the conclusion reached by our research group was that nonautomated tests requiring facilitation and explanation by the administrator can be feasible in an applied setting given careful test selection and tester training. In summary, substantial time and energy were devoted to the choice and refinement of each measure before any attempt was made to move into the field setting and train the field administrators.
Once testing technique has been refined, there are a number of obstacles to moving the polished test from the laboratory to the field. Probably the single largest problem is achieving consistency of test administration. During earlier stages of test development, we gave little thought to the participant–test administrator interaction because our test administrators were either ourselves or highly trained students and staff. The rigor of the scientific method and the need for consistency in the assessment process are part of our daily experience. These are skills that lie at the core of a research staff. It is unlikely that novice administrators identified in a field setting will have the same respect for scientific rigor that we take for granted in laboratory staff. Although potential test administrators can be informed of the need for such rigor, it is unlikely that such information will become an ingrained part of the test administration. It has been our experience that novice testers will modify the testing procedure over time, usually in ways that they believe make test administration easier for them. Changes in protocol most likely will be subtle and cumulative, thereby representing a slow evolution of the test procedure over time. Such changes may be important, such as changes in the content of instructions, or trivial. In either case, the testing procedures must be adapted in advance with the goal of minimizing the impact of the test administrators' potential to drift from protocol. This can be achieved by giving some thought to how staff will be trained. For example, we found videotape to be a useful tool for modeling proper test administration techniques. This tool shows correct performance of the test and can also serve as a reference source against which to point out procedural "evolutions." Second, we used unannounced quality control visits as an effective method for discovering changes in test administration. Not only were we able to use these visits to restore proper administration practices, but they also served to reemphasize the importance of the data being collected. Third, we used computer automation to minimize the role of the test administrator to the extent possible.
Finally, some thought was given to the personal characteristics of test administrators. Because access to personnel in a field setting is generally limited, it is likely that these same individuals will serve in a recruitment role in addition to a test administrator role. Therefore, their style of interaction with the general public is an important consideration.
|
Approaching and Consenting Participants in the Field |
|---|
Our attempts to minimize other obstacles to participation in the Maryland study included keeping the testing session as brief as possible and considering—and then trying to counter—other costs and factors that might prevent an individual from participating (e.g., perception of the study as a hassle). At the same time, we attempted to maximize the benefits of participating in the study. Individuals who agreed to participate were allowed to move to the front of the waiting line for procurement of their license, for example, as a personally relevant incentive. They were also presented with a rationale for the study, highlighting its importance from a perspective of public safety and its potential to enhance safe mobility if successfully implemented on a larger scale in the future. In this way, an appeal to altruistic motives was made.
We also attempted to capitalize on staff strengths in the area of recruitment by selecting, to the extent possible, individuals who could rapidly establish rapport with older adults and competently guide them through the informed consent process. A combination of strategies therefore provided the framework for introducing the field study and carrying it out in a unique setting.
|
Characterizing Consenters Versus Decliners |
|---|
|
Collection of Follow-Up Telephone Data |
|---|
|
Managing the Data |
|---|
Furthermore, the collaborators on this project represent a variety of scientific disciplines as well as the fields of medicine, industry, and government. Consequently, the amount and nature of prior training in the areas of data management and statistics among this group is diverse. More to the point, each stakeholder in the project has had a different investment in the study outcomes. Not surprisingly, then, a great deal of discussion has ensued as to the appropriate site for maintenance of the data, the appropriate analytic approaches to the data, the appropriate cut points to use for each test in the screening battery in order to maximize sensitivity and specificity, and the appropriate interpretation of the study results. Real-world challenges have included lively discussions relative to sharing data, questioning of biases associated with each stakeholder's perspective, and questioning of statistical results and conclusions based on them. Nevertheless, these differences and challenges have made for a fruitful project in which no one investigator is able to dictate a "right" way to handle the data. In some respects, the compromises reached resemble the peer review process that can only serve to enhance the enterprise of research.
|
Changing Medical Practice, Public Perception, and Public Policy |
|---|
There is also growing evidence that speed of processing training can result in both improved UFOV and improved driving skills. Ongoing studies are examining the effects of speed of processing training, relative to various types of sham training, on driving simulator performance, driving habits, on-road performance in an instrumented car, and other mobility outcomes. Preliminary findings suggest that speed of processing training not only improves UFOVperformance but may transfer to other cognitive functions as well as to certain aspects of driving (Edwards et al., 2002; Roenker, Cissell, Ball, Wadley, & Edwards, in press; Rumble, Edwards, Clay, Woodley, & Ball, 2001).
There are, however, numerous obstacles encountered in attempting to take research results such as these and infuse them into public policy or medical practice. In the case of implementing research findings at the level of medical practice, it can be difficult to quantify the associated costs and benefits. For example, what are the health care costs associated with mobility loss and to whom do those costs accrue? How much money does continued mobility save the health care industry, and what cost is incurred for training programs that result in this savings benefit?
Translating research outcomes into changes at the level of legislation is a slow process. Once research outcomes indicate that a change in policy is needed, an institution such as the MVA may have procedures in place to prepare legislative proposals for consideration by the state assembly. First, the nature of these proposals must be determined. For example, the MVA might consider the benefits of a mandatory testing program for older drivers versus a recommendation for further use of voluntary testing programs. Once the content is decided and the legislative proposal is drafted, adoption of the legislation tends to occur only after years of well-publicized research findings and promotion of the legislation by interested parties—parties with funds. Thus, the process of instituting various licensing procedures, including adoption of more sensitive screening tests, is difficult and time consuming and is made even more difficult due to the fact that in many cases changes must occur state by state rather than at a national level. Each state prefers to do its own driving evaluation, and the most popular way of conducting that evaluation is of course "the way it's always been done here." Most states resist being the first to adopt something new. Thus, there appears to be a natural resistance to change that works against implementation of new programs, despite their scientific basis.
Another obstacle to implementation of study findings is that the implementing agency may be afraid of the public's perception of the change. Thus, it may be necessary to work on changing public perceptions before instituting policy change. In some cases, collaborations with various media outlets to establish a public information campaign may be helpful.
In the special case of driver evaluation, acceptability of driver testing for the driver him- or herself is a related issue. Again, the acceptability of an evaluation will be contingent on what is at stake given the outcome of that evaluation. There are a number of potential strategies available to minimize the impact of visual, cognitive, and physical impairments on driver safety among older adults (e.g., treatment of eye disease, cognitive training programs, physical therapy). If the outcome of evaluation and remediation is that the length of time that an older driver may continue to drive safely is maximized, then the initial evaluation may be viewed in a more positive light.
To address this very complex public health issue, a comprehensive approach is needed. Currently, mandatory driver retesting is conducted in many states for those drivers who are referred to a medical advisory board due to concern on the part of family and friends, poor driving records, or a particular medical diagnosis. This approach should not be abandoned, but it could be improved. One concern with respect to any mandatory retesting program is that it may lead some potentially fit older drivers to stop driving rather than submit to retesting. Therefore, as discussed above, any such program needs to be accompanied by a strategic plan for informing the public of the testing procedures to be used and of the remediation options that will be made available to those with impairments that are amenable to rehabilitation.
A controversial question that has been widely debated is whether age alone should trigger a mandatory driving evaluation. Mandatory retesting for functional impairment at licensing agencies would obviously be quite costly if all licensed drivers were retested. Thus, discussions have turned to whether retesting could or should be limited to only those drivers over a certain age. For example, legislation was introduced in June 1999 in the California State Assembly that would have required all drivers aged 75 and above to demonstrate driving proficiency through both written and road tests in order to renew their driver's licenses. This proposed legislation did not pass, and the proposed trigger for reexamination was changed from an age criterion to a criterion based on prior crashes, violations, or referral for evaluation.
One perspective on this question is that the reexamination of functional abilities needed to drive safely should be viewed as a form of preventive medicine. According to this view, the occurrence of functional decline occurs with a much higher prevalence in older age groups, so screening for such decline should not begin until older age, just as screening for certain medical conditions in the absence of any symptoms is routinely based on attaining the age at which risk increases. This approach would allow for baseline measurement of individual performance against which any indication of decline could be noted at a future date.
A specific example of an age-related functional decline that is predictive of driving outcomes is decline in the useful field of view. Both UFOV reduction and crashes are more prevalent with increasing age (Ball, Roenker, & Bruni, 1990; Transportation Research Board, 1988); however, UFOV reduction is substantially better than chronological age at distinguishing between drivers who are at risk for crash and those who are not (Ball et al., 1993). The probability of a hit (accurate prediction of crash) versus a false alarm (inaccurate prediction of crash) for independent predictor variables was calculated by Ball and colleagues (1993). Results demonstrated that the UFOV test (d' = 2.27) was superior not only to age (d' = 0.58) but to vision parameters (d's ranging from 0.24 for acuity to 0.67 for contrast sensitivity) and mental status tests (d' = 0.59) in accurate prediction of crash, with sensitivity of 89% and specificity of 81%. Thus, if an individual was crash-involved, the likelihood of his or her having a UFOV reduction greater than 40% was 0.89, whereas the likelihood of an individual who was not crash-involved having intact UFOV (
40% reduction) was 0.81.
This illustration of accounting for age-related functional declines versus age alone takes on additional importance when one considers the implications of misclassifying drivers. What happens to drivers wrongly classified as impaired or high risk for crash when they are not? Individuals' autonomy and quality of life could be jeopardized by such misclassifications if they led to automatic revocation of the driver's license. Therefore, a classification of a driver as high risk should lead to further assessment, with driving outcomes contingent on the results of multiple performance indices. Conversely, what are the ramifications of wrongly classifying drivers as safe when they are impaired? In these cases, the individuals' safety, as well as public safety, are at stake. Thus, errors made in either direction could prove costly both for the individual and for society.
The challenges of implementing research outcomes such as those described here into community practice, state law, or public policy are formidable. Much groundwork has been laid—a process that has taken time, effort, and considerable resources. In our quest to disseminate research on older drivers and translate the results into policy-level changes with the potential to improve quality of life through sustained mobility, we have tried to establish as many collaborations as possible with industry, government, and older adult advocacy groups. We have recently sought funding in partnership with small business to develop educational videotapes for physicians and other health care providers in order to more directly inform this group of the older driver research and engage them in the practice of evaluating and intervening on behalf of high-risk patients. The Roybal Center program provided the impetus and the dollars for our evolution from basic to applied and translational research. It is our hope that this investment will ultimately reap benefits for older adults by enhancing their safe mobility and extending their independence for as long as possible.
|
Footnotes |
|---|
1 Department of Psychology and Edward R. Roybal Center for Research on Applied Gerontology, University of Alabama at Birmingham, Birmingham, AL.
2 Department of Psychology, Western Kentucky University, Bowling Green, KY.
Received for publication July 15, 2002. Accepted for publication September 16, 2002.
|
References |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. J. Czaja and J. Sharit Practically Relevant Research: Capturing Real World Tasks, Environments, and Outcomes Gerontologist, March 1, 2003; 43(90001): 9 - 18. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
TOP
Abstract
Background