Occupational therapy and other researchers in the United States (US), United Kingdom (UK), Europe and Australia have found that age-related changes and the greater likelihood of multiple chronic diseases put older drivers at an increased risk for unsafe driving behaviours or crashes (Mollenkopf et al 2002, Di Stefano and Macdonald 2005, Mitchell 2008, Classen et al 2009). Although people, 65 years and older, are living longer and driving longer (Foley et al 2002, Mollenkopf et al 2002, Mitchell 2008), the predictability of on-road driving performance is still under investigation, making the identification of at-risk drivers difficult (Insurance Institute for Highway Safety [IIHS] 2003). Apart from these person and measurement complexities, the safety of older drivers is also challenged by environmental features, such as roadway designs (Council and Zeeger 1992).
The US Federal Highway Administration (FHWA) proposed guidelines for highway design to increase the safe driving ability of older drivers (Staplin et al 2001). These guidelines include assessments and recommendations applicable to four categories of roadway design features: those appropriate for intersections, interchanges, roadway curvature /passing zones and construction work zones. In Britain, the comparable organisation to the FHWA is the Highways Agency. This is the public sector agency responsible for England's motorways and trunk (national) roads. Scotland and Wales have similar organisations, which fall under the devolved administrations for roads in their areas.
Although researchers examined, in a recent on-road study, the effect of intersection design (during the turn phase of an intersection) on the driving performance of older (65-85 years) and younger (25-45) drivers, little empirical evidence exists to support the effectiveness of the FHWA guidelines (Classen et al 2007). The present study deals exclusively with roadway intersection features in an urban area and presents the results from a follow-up study, examining the driving performance of young and older adults during the recovery phase of an intersection.
The researchers deemed this issue very important because (1) crashes and crash-related injuries and fatalities among older drivers are most prevalent in the northern and southern hemispheres at urban intersections; (2) despite the fact that road users in the UK drive on the left side of the road rather than the right side of the road as tested in this study, the outcome of environmental design may benefit drivers regardless of the side of the road they are driving on; and (3) the role of the environment on occupational performance is well articulated in models of the occupational therapy profession (Dunn et al 1994, Kielhofner 2002). As such, this study is in part conducted to inform occupational therapists about the potential benefits of environmental adaptation, specifically improved intersection design, as a facilitator for safer driving.
Characteristics and vehicle kinematics of driving manoeuvres at road intersections
Regardless of the side of the road that one is driving on, driving manoeuvres at intersections involve approach, turn and recovery phases. Using the FHWA guidelines (Staplin et al 2001) and an instrumented vehicle, an assessment was made (Classen et al 2007) of the turn phase at five pairs of improved and unimproved intersections. In general, this study showed that young and older adult participants alike benefited from roadways with safety features, suggesting that the FHWA guidelines are helpful for safer driving while making a turn for at least three of the five improved intersections studied.
After making a turn, a transitory action, the vehicle moves into the recovery phase. This phase is defined as starting at the end of the turn phase, where yaw (rotation about a vertical axis that passes through the car's centre of gravity and causing the car to swing off-course, also known as rate of turn) drops to a value of 0.05 radians /second and is measured for a distance of 500 feet (152.4 metres). Safe driving performance can be assessed by measuring stable vehicular states and the confidence of the driver. During a turn, yaw can reach a value as great as 0.75 radians/second, or even more during a very extreme turning manoeuvre.
The stable vehicular state is expressed as high lateral stability of the vehicle and can be determined by measuring lateral acceleration (side [g] forces) and magnitude of yaw (rate of turn measured in radians/second). Lateral acceleration and yaw are functionally related, with higher values indicating lesser lateral vehicular stability. Driver confidence is operationalised as a greater recovery speed (yet appropriate for the posted speed) and lesser forward acceleration. Lesser forward acceleration indicates more driver confidence because the driver begins the recovery phase at a higher speed and, therefore, does not need to accelerate as much. These measures can also be observed through measures of dispersion (maximum and average) and central tendency (variance and root mean square [RMS]).
Recovery responses may be influenced by lead vehicles, opportunities for passing other vehicles and other road activity (for example, a pedestrian crossing the street away from the intersection). The geometrics and traffic controls for each intersection type vary, leading to somewhat differing expectations of performance during the recovery phase. As a follow-up from the previous study, and using the FHWA guidelines for improved driving performance (that is, high vehicle stability and driver confidence), four intersection pairs (as opposed to five in the preceding study) are described. These are extended receiving lane, right turn (turning into approaching traffic) with channelisation, left-turn (turning across traffic) offset and no acute turn angle, as evaluated in the recovery phase. The separate lane signals with protected left turn phase intersection pair was not included, because it was confounded by geometrics (Classen et al 2007).
Geometrics of intersections and expectations in recovery performance
Detailed geometrics have been published (Classen et al 2007) but the main features are summarised below for the reader's benefit.
Extended receiving lane (manoeuvre 1)
The extended receiving lane (a forgiving road shoulder) allows the driver to make a wider sweep (larger radius) turn. It is expected that the increased turn radius will result in stable lateral control, small deviations from the most desirable path of travel, higher yaw (because of the increased curvature) and higher speed when exiting the turn, but reaching a stable maximum speed sooner (Classen et al 2007). Recovery from the tighter turn (unimproved intersection) is expected to exhibit higher side forces, slower speed and increased muscular effort, resulting in greater positional errors.
Right turn with channelisation (manoeuvre 2)
Channelisation with an acceleration lane reduces differences in roadway speeds among vehicles and allows drivers to merge into the intersecting road safely. The presence of channelisation during the recovery phase may help drivers to match their speeds better with approaching vehicles, resulting in a lowered forward acceleration; maintain higher speeds to achieve maximum speed quicker; and obtain lateral stability within the mainstream quicker.
Left-turn offset (manoeuvre 3)
This intersection feature (when the green light phase is active or when the signal changes from green to amber for the turning driver) helps to increase the sight distance and gap acceptance of drivers. So, seeing smaller yaw and lateral acceleration and achieving appropriate speed quicker is expected.
No acute turn angle (manoeuvre 4)
The improved intersection is the 90[degrees] junction of two roadways, whereas the unimproved condition constitutes the acute angle junction (<75[degrees]), requiring a sharper turn. Cognitive and perceptual difficulties may be imposed for drivers at acute-angled intersections because of increased steering input requirements and more head and eye movements, a challenge for some older drivers (Yee 1985). Higher lateral stability, lesser acceleration and higher speed for drivers negotiating the improved condition during the recovery phase are expected.
Table 1 (adapted from the Classen et al 2007 publication to reflect recovery phase features) provides a summary of the four intersections, with a description of each intersection type, its characteristics, intended effects, and expected differences in driving performances during recovery.
Hypotheses and purpose
Following the FHWA guidelines (Staplin et al 2001) and the positive performance effect experienced by drivers negotiating the turn phase of the improved intersections (Classen et al 2007), whether these improved driving performances would be evident during the recovery phase of negotiating each of the improved intersections was examined. Specifically, it was hypothesised that the vehicle kinematics for the improved intersections during the recovery phase would show:
1. Improved vehicle stability expressed as lower yaw and lateral acceleration, except for manoeuvre 1 (extended receiving lane) where higher maximum yaw consistent with the increased curvature associated with the forgiving shoulder of this lane was expected
2. Higher driver confidence expressed as higher speed, lesser time to achieve maximum speed and lesser forward acceleration
3. For the age comparison of the recovery phase from both improved and unimproved intersections, older drivers were expected to: (a) exhibit greater variability of lateral forces indicating more deviation from a linear path; (b) drive at slower speeds indicating less driver confidence and more driver caution; and (c) exhibit lesser forward acceleration than younger drivers who generally drive at faster speeds.
Thus, using kinematics measures from an instrumented vehicle recorded during on-road evaluations, the effects of improved versus unimproved intersections for the recovery phase were investigated, and determined if negotiation of the improved intersections was safer for both older (65-85 years) and younger (25-45) drivers.
Participants from North Central Florida were recruited using paid advertisements in newspapers, flyers distributed to ageing service centres (for example, Area Agency on Ageing), health clubs, apartment complexes, community centres, open houses held at the University of Florida's Gator-Tech Smart House and word-of-mouth referrals. The University of Florida's Institutional Review Board approved the research plan. Before enrolling in the study, all participants who met the inclusion criteria completed a telephone survey and an informed consent form.
The inclusion criteria were having a valid US driver's licence, age (young = 25-45 years; older = 65-85 years), vision acuity (20 /70 both eyes and 20/40 in one eye in case of blindness in one eye) and mental status (a score of [less than or equal to] 24 on the Mini Mental State Exam [Folstein and Folstein 1975] and completing the Trailmaking Test, Part B [Reitan 1958] in less than 3 minutes). The exclusion criteria were having seizures within the past year and having major psychiatric or physical disorders influencing functional status.
A total of 71 healthy adults participated in the study, of whom 39 were young (mean = 33.54, SD [+ or -] 5.77 years of age) and 32 were older adults (mean = 74.19, SD [+ or -] 5.94 years of age). Of the 71 participants, the 32 older adults comprised 13 (40.6%) females and 19 (59.4%) males and the 39 younger adults comprised 24 (61.5%) females and 15 (38.5%) males.
Using a repeated measures experimental design, the driving performance of older and young adults was examined through four pairs of intersections (improved versus unimproved). The pairs of intersections (manoeuvres) included the presence (improved) and absence (unimproved) of: manoeuvre 1--extended receiving lane; manoeuvre 2--higher speed roads with right-turn channelisation at an intersection; manoeuvre 3--left-turn offsets; and manoeuvre 4--skewed angle intersecting roadways.
The vehicle used for on-road evaluations was a 2004 Buick Century, instrumented to provide kinematics data that reflected vehicle control. The instruments employed in the vehicle, the data acquisition system and the derived kinematics measures were developed by these researchers in collaboration with the Department of Engineering.
The road course consisted of an urbanised and residential street network in a mid-size town in Florida. Embedded in this road course were eight test signalised intersections; four of these had been improved, consistent with recommendations in the FHWA design guidelines for improving the performance and safety of older drivers. On average, participants required about an hour to complete the road course. Travel through the intersections of the road course was the same for all participants. The course was designed to balance the order of exposure of the improved and unimproved intersections. For two of the intersection comparisons, the improved intersection preceded the unimproved one. For the remaining two pairs, the unimproved intersection preceded the improved one. Between the intersection pairs were a total of 19 non-test intersections that participants were required to negotiate in order to complete the road course. None of the comparison sites were adjacent to one another.
The driving evaluator, sitting in the passenger seat of the test vehicle, provided verbal instructions to the participants, well in advance (about 500 feet or 152.4 metres) of the next intersection. These instructions were clear and concise; for example, a typical instruction was 'At the next traffic light, please turn left'. Additional conversation in the test vehicle was limited to a minimum. The testing occurred during the summer of 2005, under the most perfect conditions (for example, optimal weather conditions, non-peak traffic hours and daylight hours).
The road course and manoeuvre location protocol of these intersections, based on the FHWA guidelines for intersections, were developed in collaboration with the Traffic Engineering Department, geometrically matched by the city engineer and a driving consultant, and pilot tested by this study's driving evaluators. The diagrams and descriptions for these manoeuvres and locations were previously published (Classen et al 2007).
Kinematics, or vehicle control responses, included yaw (rate of turn) (radians /second), lateral (g's) and forward (longitudinal) (g's) acceleration, speed (mph or kph) and time (seconds) to complete the recovery phase. Stability measures were obtained from two accelerometers, the geographical positioning system and an angular rate sensor in the car during road tests, and were computed by an engineer with algorithms using Matlab (Version 7.0.4) (2005) software programme.
Data collection and management
Participants completed a telephone interview conducted by the trained staff at the University of Florida's National Older Driver Research and Training Centre. An occupational therapist with specialty training in conducting comprehensive driving evaluations completed a brief clinical assessment and the on-road assessment in the instrumented vehicle. The evaluator used a standardised road assessment performance (Justiss et al 2006) sheet to record driving errors as the participants drove the road course. These data were published elsewhere (Classen et al 2007). Video footage, obtained from the four cameras, was used for the training of the four driving evaluators. High interrater reliability among the evaluators (intraclass correlation coefficient = 0.80 -1.00, p < 0.05) was attained. The kinematics data were entered and managed with Matlab (Version 7.0.4). All data were next exported to MS-Excel and SPSS Version 13.0 (2005) for analyses.
An a priori power analysis (alpha = 0.05, beta = 0.80, moderate effect size, 20% attrition rate) required 109 participants. Of the 98 participants that performed all aspects of the evaluation, 21 had missing data. The data from 71 (39 young, 32 older) participants were analysed. To test for the effect of age and road condition, the kinematics data were analysed using 2 x 2 repeated measures ANOVA; the within-subject variable was intersection condition (improved versus unimproved) and the between-subject variable was age (young versus older). The kinematics data used were maximum, root mean square (RMS) and variance of yaw; lateral acceleration; maximum and average of forward acceleration and of speed; and time of recovery.
Results and interpretation
Figs 1 and 2 present an example of the kinematics data analysed for the improved and unimproved intersections of manoeuvre 1 for one driver. The areas between the start and end of recovery are the sections of interest. In these figures, speed, seen towards the top of each of the graphs, increases more rapidly during the recovery phase for the improved intersection, whereas forward acceleration of this non-stopped vehicle continues to increase to a maximum and then drop to near zero as the vehicle approaches a constant speed (more evident for the unimproved intersection). As the driver was instructed to make a left turn at this intersection, these graphs show negative values of lateral acceleration and yaw. These two measures follow similar paths, moving towards zero values for both graphs, but they continue to display oscillations during the recovery phases (with larger oscillations noted for the unimproved intersection). The magnitudes of these deviations from zero allowed stability measures expressed as RMS and variances for both lateral acceleration and yaw to be obtained.
The results and interpretations for the significant findings of the improved and unimproved intersections are discussed, for each of the four manoeuvres and age groups.
Kinematics data for manoeuvre 1 (Table 2) showed, as expected, significantly higher maximum yaw (F = 3.81; p = 0.05), maximum speed (F = 7.77; p = 0.01) and average speed (F = 30.41; p = 0.01) for the improved intersection, and significantly lower variance of yaw (F = 5.05; p = 0.03), maximum lateral acceleration (F = 6.59; p = 0.01) and average forward acceleration (F = 25.62; p = 0.01) for the improved intersection. Although no interaction effects were observed, age differences existed in maximum forward acceleration (F = 7.59; p = 0.01) and average speed (F = 5.00; p = 0.03), with the averages, as expected, higher for the older drivers. The RMS of lateral acceleration (F = 59.51; p = 0.01) and variance of lateral acceleration (F = 45.72; p = 0.01) showed statistical significance in the opposite directions to those initially expected. Compared with the unimproved intersection, the time for the improved intersection was unexpectedly significantly higher (F = 49.64; p = 0.01).
As expected, greater lateral stability was seen indicated by a lesser maximum acceleration, but greater variability of lateral acceleration immediately after the turn because of the wider sweeping turn. With the exception of time, the drivers performed as expected for most of the kinematics measures. For example, the higher maximum yaw value was a result of making a wider turn at the intersection with the wider receiving lane, while the lower yaw variance and lower maximum lateral acceleration are attributed to greater control during recovery. The maximum and average speeds were higher for the improved intersection because the extended receiving lane facilitated improved performance.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
The average forward acceleration was significantly less for the improved intersection, but comparisons of the average speeds through the intersection showed that the turn-phase speeds were about 25% higher (yet still appropriate for the posted speed) than those produced at the comparison site (Classen et al 2007). These sustained higher speeds required drivers at this location to accelerate less to achieve an appropriate speed. This is also the most likely reason that forward acceleration was less for recovery at this improved intersection.
The age differences, observed for higher maximum forward acceleration, are accounted for as older drivers produced lower speeds in the turn phase, necessitating an increased acceleration in the recovery phase. The lower average speed may be related to older drivers requiring additional processing time. The greater time for the improved intersection, not expected, may have resulted from a higher traffic flow pattern (Traffic Engineering Department, personal communications, 23 January 2006).
Kinematics data for the improved intersection of manoeuvre 2 (Table 2) demonstrated, as expected, a lower maximum (F = 57.70; p = 0.01), RMS (F = 344.36; p = 0.01) and variance of lateral acceleration (F = 8.45; p = 0.01), and decreased time (F = 12.46; p = 0.01) for recovery; a higher average forward acceleration (F = 163.80; p = 0.01) and higher maximum (F = 317.97; p = 0.01), and average of speed (F = 111.09; p = 0.01) for recovery. An age effect existed, the maximum of yaw being significantly higher for older drivers (F = 4.72; p<0.03). An interaction effect (age x intersection) was apparent for maximum forward acceleration (F = 3.99; p = 0.05), with older drivers having a higher mean (mean = 0.19, SD = 0.06) for the improved intersection and a lower mean (mean = 0.18, SD = 0.05) for the unimproved intersection, whereas younger drivers had a lower mean (mean = 0.17, SD = 0.03) for the improved intersection and a higher mean (mean = 0.18, SD = 0.04) for the unimproved intersection.
Except for the maximum and RMS of yaw which showed no significance, all measures in this manoeuvre were consistent with the expectations. The lower maximum lateral acceleration signified a smoother movement from the acceleration lane into the intersecting mainstream roadway, while the lower RMS and variance of lateral acceleration showed improved lateral control. The higher average forward acceleration, higher maximum and average speed and lesser time for recovery fulfilled the purpose of the acceleration lane: to facilitate higher speed more quickly.
The age effect for the improved intersection (higher maximum yaw for older drivers) suggests that older drivers have greater difficulty in making lane changes following the juncture. The interaction effect for the maximum forward acceleration showed that older drivers accelerated more during the recovery of the improved intersection as expected, while younger drivers accelerated more during recovery of the unimproved intersection. This finding is consistent with the literature showing caution among older drivers as a group and increased risk-taking behaviours among the younger drivers as a group (McGwin and Brown 1999, Patel et al 2000, Clarke et al 2005).
The data of manoeuvre 3 (Table 2) demonstrated for the improved intersection a significantly lower RMS (F = 17.05; p = 0.01) and variance of yaw (F = 30.10; p = 0.01), and a lower maximum (F = 37.51; p = 0.01), RMS (F = 31.86; p = 0.01) and variance of lateral acceleration (F = 38.42; p = 0.01). A significantly higher average forward acceleration (F = 13.92; p = 0.01) and higher maximum (F = 60.39; p = 0.01), and average speed (F = 32.98; p = 0.01), and higher time (F = 6.19; p = 0.01), for recovery was observed for the improved intersection, but maximum yaw was not significant. A significant age effect existed for the improved intersection for maximum forward acceleration (F = 13.39; p = 0.01), with the averages higher for the older drivers (mean = 0.14, SD = 0.02). No interaction effects were evident.
The significantly lower RMS and variance of yaw, and maximum, RMS and variance of lateral acceleration, are consistent with smoother control movements during the recovery period for the improved intersection. The data for maximum yaw behaved as expected, resulting in no significant changes. The significantly higher average forward acceleration, and maximum and average speed, conform to the expectations of the improved intersection. The increased time for recovery in the improved intersection was not expected but, upon further investigation, it was found that, compared with the unimproved intersection, denser traffic flow was evident at the improved intersection (Traffic Engineering Department, personal communications, 23 January 2006). The significant age effect (higher means for older driver) for maximum forward acceleration (improved intersection) may indicate that the older drivers attempted to minimise risk by getting up to speed quicker, but the magnitude of these differences is very small.
The data of manoeuvre 4 (summarised in Table 2) demonstrated for the improved intersection a significantly lower RMS (F = 4.31; p = 0.04) and variance of yaw (F = 16.19; p = 0.01), maximum (F = 120.74; p = 0.01) and average forward acceleration (F = 683.09; p = 0.01), and maximum speed (F = 51.65; p = 0.01). The higher maximum (F = 71.95; p = 0.01) and RMS (F = 126.36; p = 0.01) of lateral acceleration, and greater time (F = 365.85; p = 0.01) for the improved intersection, were not expected. Maximum yaw showed no significance for intersection type and, although not significant, average speed approached significance (F = 3.76; p = 0.06), but in the favour of the unimproved intersection. A significant age effect was apparent for maximum forward acceleration (F = 6.21; p = 0.01), with the averages higher for the older drivers.
The significantly lower RMS and variance of yaw, for the improved intersection, indicated smoother control movement during the recovery period. The significantly lower maximum and average forward acceleration, and maximum speed, for the improved intersection is an interesting finding, indicating that the drivers sped up on the intersection with the acute angle after making the turn. The higher maximum and RMS of lateral acceleration, and greater time for the improved intersection, indicated decreased control and confidence during the recovery phase. The greater time for recovery on the improved intersection is partly explained by the speed limit being 30 miles per hour (48.28 kph) in the improved intersection, versus 45 miles per hour (72.42 kph) for the unimproved intersection, accounting for the doubling of time to drive through the recovery zone as well as increased traffic density at the improved intersection (Traffic Engineering Department, personal communications, 23 January 2006). The significant age effect for maximum forward acceleration may be a function of older drivers attempting to minimise risk by getting up to speed quicker.
Table 2 provides a summary of the kinematics data for the four manoeuvres (n = 71; 39 young, 32 older). This table also indicates the percentage of measures with superior performance for improved intersection, as well as the percentage of measures with superior performance for the older group. The most sensitive measures to detect differences between improved and unimproved intersections were the maximum and RMS of lateral acceleration; average forward acceleration; maximum speed; and time. The most sensitive measures to detect differences between older and younger adults were maximum forward acceleration, with maximum yaw and average speed detecting some age-related differences.
Implications of the study for occupational therapy
The effects of four improved versus unimproved intersections were quantified, showing that the improved intersections facilitated safer (stable vehicular states and confidence) driving for the recovery phase, and established that negotiating the improved intersections was safer for both older (65-85 years) and younger (25-45) drivers. As such, empirical information on the benefits of environmental design to the occupational performance (safe driving) of older drivers at urban intersections was provided. Second, environmental enhancements (improved intersections) benefited not just the older drivers, but the safe driving of younger drivers as well. As such, occupational therapists may assume a role of advocacy with city planners and engineers as they decide to implement design guidelines.
How the kinematics measures supported (or not) the FHWA guidelines for each intersection, and for each age group, are summarised.
The FHWA guidelines for the extended receiving lane elicit the desired effect, as reflected by the kinematics measures. However, from the 11 measures, an age effect in only two of the measures was observed, indicating that young and older adults alike may benefit from these design features.
With the exception of maximum of yaw, the FHWA guidelines for the right-turn channelisation with acceleration lane prove to elicit the desired effect as it pertains to the kinematics measures. Although an interaction effect was detected, the differences were small and only one measure showed statistically significant age differences, suggesting that this guideline may benefit both older and younger drivers.
With the exception of maximum forward acceleration, the FHWA guidelines for the left-turn offset are supported by this study's findings and showed the desired effect as it pertains to the kinematics measures. Older drivers in general are at a disadvantage in that their gap acceptance abilities tend to be poorer than those of younger drivers (Staplin et al 2001). Although an age effect in maximum forward acceleration was detected, the practical significance is small, suggesting that this intersection design guideline benefits older and younger drivers alike.
Various measures, maximum of yaw, variance of lateral acceleration and average speed, did not perform as expected on the intersection with roadways intersecting at 90[degrees]. However, in the analysis of the turn phase at these intersections (Classen at al 2007), vehicle stability was shown to be greater at the improved intersection. As had been expected, both maximum lateral acceleration and maximum yaw were lower during turning at the right angle intersection (the improved intersection). Other measures used in the turn phase (maximum combined acceleration, and longitudinal and lateral acceleration) also significantly favoured the improved intersection (Classen et al 2007). Although maximum speed and maximum forward acceleration showed no differences between the improved and unimproved intersections and considering the stability results from both the turn and recovery phases, it was concluded that superior lateral control was maintained during the turn phase of the 90[degrees] intersection. At the unimproved intersection, the expectations were for drivers to slow down when negotiating this more severe turn, but they did not do so. However, they maintained their speed and then sped up after making the turn and exhibited good lateral stability.
Accordingly, there is a mixed set of results indicating that for turn phase, probably the more critical phase from a safety viewpoint, better control took place at the improved intersection. It is therefore premature to conclude that this intersection guideline was not effective. The improved intersection was somewhat confounded by differing speed zones and traffic flow. An age effect for only one (maximum forward acceleration) of the 11 measures was detected, suggesting that this intersection design guideline benefits both older and younger drivers.
Limitations of the study
Certain limitations must be considered. Although an order effect may exist, it is unlikely that the arrangement of the intersections influenced the performance of drivers in this study. Compared with manoeuvres 1, 2 and 3, manoeuvre 4 demonstrated less effectiveness for the FHWA intersection guideline application. Matching the two road conditions for all variable factors is difficult in a real life situation. Although all of these kinematics measures are considered indicators of safety, it is not known what the relationships are between these indicators and real life crash situations. As the apparently healthy older adult has been selected, this study's findings cannot be generalised to the general population of older adults.
Although the results of kinematics measures for manoeuvre 4 did not support the FHWA guideline in the recovery phase, manoeuvres 1, 2, and 3 showed effectiveness of the impact of the FHWA intersection guidelines (Staplin et al 2001) for the recovery phases. From the kinematics measures, different from the turn-phase results (Classen et al 2007), an age effect was observed in each improved intersection for at least one of the measures in the recovery phase favouring older drivers, but practical differences were small and it was concluded that both older and younger drivers benefited from these intersection design guidelines. The study was powered for 109 participants, but completed data for 71 participants were available. Thus, type 2 error could be evident in the study; that is, not finding a statistically significant effect if one really existed.
Implications for practice and research
It is suggested that there are three direct implications for practice. Occupational therapists may educate their clients on the importance of environmental design. That is, they may discuss beneficial roadway design elements so that clients may choose safer roads to travel on, and avoid intersections that may pose a risk for unsafe driving. Occupational therapists may also advocate for updating driver education programmes to elucidate the benefits of roadway design elements. Finally, occupational therapists working in driver safety may advocate for policies to regulate institutionalising of improved intersection design guidelines across jurisdictions because these intersection design guidelines benefit most adult drivers.
Three ideas for occupational therapy research are also suggested. A replication of the concept of this study, that is, testing intersection design guidelines, but applied to the UK driving environment, may generate valuable information on the plausibility of environmental design on the driving performance of older and younger drivers. Such studies, when shown to be effective, will position occupational therapy researchers in the UK to inform city planners and engineers to consider intersection design guidelines as they develop roadways. Moreover, the conceptual and analytical framework used in this study may be useful for researchers in the UK to test intersection design guidelines from Britain's Highways Agency, or to test environmentally appropriate design guidelines specific to the British context (for example, roundabouts).
* The kinematics data showed that implementing the suggested FHWA guidelines in road geometry and traffic controls was helpful for safer (stable vehicular states and confidence) driving in three of the four improved intersections.
* This benefited older and younger adults alike.
What the study has added
This study brings empirical intersection design and safety information to occupational therapists, suggesting that enhancements in the environment may have an impact on safer driving, particularly at urban intersections.
Clarke DD, Ward P, Truman W (2005) Voluntary risk taking and skill deficits in young driver accidents in the UK. Accident Analysis and Prevention, 37(3), 523-29.
Classen S, Shechtman O, Stephens B, Davis E, Bendixen R, Belchior P, et al (2007) The impact of roadway intersection design on driving performance of young and senior adults. Traffic Injury Prevention, 8(1), 69-77.
Classen S, Winter S, Lopez EDS (2009) Meta-synthesis of qualitative studies on older driver safety and mobility. Occupational Therapy Journal of Research: Occupation, Participation and Health, 29(1), 24-31.
Council FM, Zeeger CV (1992) Accident analysis of older drivers and pedestrians at intersection--Task B working paper. Publication no. DTFH61-91-C-00033. Washington, DC: Federal Highway Administration.
Di Stefano M, Macdonald W (2005) On-the-road evaluation of driving performance. In: JM Pellerito, ed. Driver rehabilitation and community principles and practice. St Louis, MO: Elsevier Mosby, 255-74.
Dunn W, Brown C, McGuigan A (1994) The ecology of human performance: a framework for considering the effect of context. American Journal of Occupational Therapy, 48(7), 595-607.
Foley DJ, Heimovitz HK, Guralnik JM, Brock DB (2002) Driving life expectancy of persons aged 70 years and older in the United States. American Journal of Public Health, 92(8), 1284-89.
Folstein M, Folstein S (1975) 'Mini-mental state': a practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189-98.
Insurance Institute for Highway Safety (2003) Fatality facts, older people. Arlington, VA: IIHS. Available at: http://www.highwaysafety.org/safety_ facts/fatality_facts/older_people.htm Accessed on 28.01.04.
Justiss MD, Mann WC, Stav W, Velozo C (2006) Development of a Behind-the-Wheel Driving Performance Assessment for older adults. Topics in Geriatric Rehabilitation, 22, 121-28.
Kielhofner G (2002) A model of human occupation: theory and application. 3rd ed. Baltimore, MD: Lippincott Williams and Wilkins.
Matlab (Version 7.0.4) (2005) (R14) Copyright 1984-2005 by MathWorks, Inc. USA: MathWorks, Inc.
McGwin G Jr, Brown DB (1999) Characteristics of traffic crashes among young, middle-aged, and older drivers. Accident Analysis and Prevention, 31(3), 181-98.
Mitchell CGB (2008) The licensing of older drivers in Europe: a case study. Traffic Injury Prevention, 9(4), 360-66.
Mollenkopf H, Marcellini F, Ruoppila I, Szeman Z, Tacken M, Kaspar R, et al (2002) The role of driving in maintaining mobility in later life: a European view. Gerontechnology, 1(4), 231-50.
Patel DR, Greydanus DE, Rowlett JD (2000) Romance with the automobile in the 20th century: implications for adolescents in a new millennium. Adolescent Medicine, 11(1), 127-39.
Reitan RM (1958) Validity of the Trail Making test as an indicator of organic brain damage. Perceptual and Motor Skills, 8, 271-76.
SPSS for Windows (2005) Release Version 13.0.1. Chicago: SPSS Inc.
Staplin L, Lococo K, Byington S, Harkey D (2001) Guidelines and recommendations to accommodate older drivers and pedestrians. Publication no. FHWA-RD-01-051. Washington, DC: US Department of Transportation, Federal Highway Administration.
Yee D (1985) A survey of the traffic safety needs and problems of drivers age 55 and over. Needs and problems of older drivers: survey results and recommendations. Falls Church, VA: American Automobile Association, Foundation for Traffic Safety.
Sherrilene Classen, (1) Orit Shechtman, (1) Burton Stephens, (1) Ethan Davis, (1) Desiree Lanford (1) and William Mann (1)
(1) University of Florida.
Corresponding author: Dr Sherrilene Classen, Assistant Professor, Occupational Therapy, and Director, Institute for Mobility, Activity, Participation, University of Florida, PO Box 100164, Gainesville, Florida, United States 32610-0164.
Reference: Classen S, Shechtman O, Stephens B, Davis E, Lanford D, Mann W (2009) The impact of intersection design on the driving performance of adults in the recovery phase of a turn. British Journal of Occupational Therapy, 72(11), 472-481.
[c] The College of Occupational Therapists Ltd.
Submitted: 5 November 2008.
Accepted: 29 June 2009.
Table 1. Major characteristics of four intersection types Characteristics of Improved Guideline in FHWA improved intersection report intersections Manoeuvre 1: A minimum receiving Provides more space Extended receiving lane width of 3.6 m for vehicles lane (12 ft) accompanied entering by a shoulder of intersection to make 1.2 m (4 ft) minimum wide turn without width. running off road. Manoeuvre 2: Right-turn Because no stopping Right turn with channelisation with is required at the channelisation and an acceleration lane intersection, allows acceleration lane to provide for the protection from acceleration approaching characteristics of vehicles. passenger cars. Manoeuvre 3: Unrestricted sight When green light Left-turn offset distances and signal phase is corresponding present, improved left-turn lane sight distance offsets are afforded by this recommended whenever improvement allows possible in the more rapid design of opposite acceleration and left-turn lanes at higher speed during intersections. turn. Manoeuvre 4: For new facilities Compared with acute Intersection with where right-of-way angle intersection, roadways intersecting is not restricted, stability and speed at 90[degrees] all intersecting control are roadways should meet adequate. at a 90[degrees] angle. For new facilities or redesign of existing facilities where right-of-way is restricted, intersecting roadways should meet at an angle of no less than 75[degrees]. Improved Intended effects intersection Manoeuvre 1: * Reduces sharp turns into intersecting Extended receiving lanes with approaching traffic lane * Reduces run-off-road incidents * Helps maintain speed, thus reducing rear-end collisions. Manoeuvre 2: * Speed reduction during turns is relatively Right turn with small channelisation and * Turning vehicle is able to gain acceleration lane appropriate speed quicker * Reduces speed differences between in-stream traffic and entering vehicles. Manoeuvre 3: * Better vision of opposing traffic is Left-turn offset intended to improve gap acceptance decisions thus * Reduced acceptance of dangerous short gaps * Reduced delays to queue traffic in turn lane. Manoeuvre 4: * Driver performances are expected to be Intersection with acceptable roadways intersecting * When compared with the acute angle at 90[degrees] intersection, crashes should be less (particularly rear-end collisions). Improved Expected differences in intersection performances during recovery Manoeuvre 1: * Greater lateral stability (indicated by a Extended receiving lesser maximum acceleration), with greater lane variability of lateral acceleration because of the wider sweeping turn * Yaw may be greater with less variability * Higher speeds can be achieved more quickly, due in part to greater speed achieved during the turn phase * Forward acceleration may be less in obtaining these higher speeds. Manoeuvre 2: * Greater lateral stability (indicated by Right turn with maximum lateral acceleration) channelisation and * Less variability (indicated by root mean acceleration lane square [RMS] and variance of lateral acceleration) * Smaller differences in yaw and yaw variability are possible with the acceleration lane (but if speed is higher for the improved intersection differences would probably cancel out) * Greater driver confidence reflected by higher speed and forward acceleration. Manoeuvre 3: * Greater lateral stability achieved quicker Left-turn offset with less variability * Because the angles of the roadways at these intersections are the same, and assuming same speeds and radius of the road, no difference is expected between the yaw measures * Greater driver confidence is expected by higher speed and greater forward acceleration. Manoeuvre 4: * Lateral stability and speed control should Intersection with be average for this typical right-angle roadways intersecting intersection at 90[degrees] * When compared with a degraded intersection design, stability and speed control are expected to be better * However, the unusual acute angle intersection may increase alertness of drivers and thus affect both driver control and confidence after the turn phase. Table 2. Summary table for kinematics data for four manoeuvres, n = 71 ([n.sub.young] = 39, [n.sub.old] = 32) Max. RMS Variance yaw yaw yaw (radians/ (radians/ (radians/ sec) sec) sec) Manoeuvre 1 I > U I = U I < U Y = O Y = O Y = O Manoeuvre 2 I = U I = U I = U Y < O Y = O Y = O Manoeuvre 3 I = U I < U I < U Y = O Y = O Y = O Manoeuvre 4 I = U I < U I < U Y = O Y = O Y = O % of measures with superior 25 50 75 performance for improved (75) (50) (25) intersection (No difference) % of measures with superior 25 0 0 performance for older group (75) (100) (100) (No difference) Max. RMS Variance Max. lat. lat. lat. forward acc. acc. acc. acc. (g) (g) (g) (g) Manoeuvre 1 I > U I > U I > U I = U Y = O Y = O Y = O Y < O Manoeuvre 2 I < U I < U I < U I = U * Y = O Y = O Y = O Y = O* Manoeuvre 3 I < U I < U I < U I = U Y = O Y = O Y = O Y < O Manoeuvre 4 I > U I > U I = U I < U Y = O Y = O Y = O Y < O % of measures with superior 100 100 75 25 performance for improved (0) (0) (25) (75) intersection (No difference) % of measures with superior 0 0 0 75 performance for older group (100) (100) (100) (25) (No difference) Ave. Max. Ave. Time forward speed speed (sec) acc. (mph) (mph) (g) Manoeuvre 1 I < U I > U I > U I > U Y = O Y = O Y < O Y = O Manoeuvre 2 I < U I > U I > U I < U Y = O Y = O Y = O Y = O Manoeuvre 3 I > U I > U I > U I > U Y = O Y = O Y = O Y = O Manoeuvre 4 I < U I < U I = U I > U Y = O Y = O Y = O Y = O % of measures with superior 100 100 75 100 performance for improved (0) (0) (25) (0) intersection (No difference) % of measures with superior 0 0 25 0 performance for older group (100) (100) (75) (100) (No difference) RMS = root mean square; Max. = maximum; Lat. = lateral; Acc. = acceleration; Ave. = average; I = improved; U = unimproved; Y = young; O = old; = is equal; > is significantly greater; < is significantly smaller; * = interaction effect. Note: Manoeuvre 2, max forward acceleration: although the main effects have no statistical differences, an interaction effect is evident for age x intersection, as indicated by *.…