«ASSESSING OLDER DRIVER’S FITNESS TO DRIVE ALLOWING FOR A LOW MILEAGE BIAS: USING THE GRIMPS SCREENING TEST Koppel, S., Langford, J., Charlton, J., ...»
ASSESSING OLDER DRIVER’S FITNESS TO DRIVE ALLOWING FOR
A LOW MILEAGE BIAS: USING THE GRIMPS SCREENING TEST
Koppel, S., Langford, J., Charlton, J., Fildes, B., Frith, W. & Newstead, S.
Data from 244 older drivers in New Zealand have been used to demonstrate that older drivers
who travel low mileages are liable to have more crashes per distance driven than older drivers who travel higher mileages. The results showed that drivers travelling 50 km or less per week had a considerably higher per-distance crash rate than drivers travelling 100 km or more per week. Low mileage drivers also performed significantly worse on both a screening test of fitness to drive (the GRIMPS screening test) and on the New Zealand Older Driver Re- licensing Test (NZODORT). With reduced driving performance likely to be a major factor in explaining the association between extent of driving and crash involvement, the findings presented in this paper are valuable in identifying a small, more precisely defined target group for road safety countermeasures, while excluding most older drivers from any special safety scrutiny.
KEY WORDS: older drivers, crash risk, low mileage bias, functional impairments
However, while distance or mileage driven is generally seen as the most robust measure for demonstrating older drivers’ crash risk, this measure is increasingly being called into question. It has been long known that the relationship between travel distances and crash rates is not linear (Janke, 1991). That is, all else being equal, crashes will tend to rise at a low and decreasing rate as mileage increases beyond a certain point (Janke, 1991). That is, independent of age, gender or other demographic factors, drivers travelling more kilometres will typically demonstrate reduced crash rates per kilometre, compared to those driving fewer kilometres. Older drivers typically drive less distance per trip and have lower accumulated mileages. Janke subsequently warned licensing administrators against becoming overly alarmed about older drivers’ apparent high crash risks when comparisons were based on per distance crash rates, without controlling for different annual kilometres driven.
Hakamies-Blomqvist and her colleagues (2002) compared older and young middle-aged drivers’ crash rates, controlling for annual distances driven. When older drivers were compared with younger drivers who had equivalent driving exposure, there was no agerelated increase in crashes (both casualty and non-casualty) per distance driven. Thus, the commonly perceived age-related risk was due not to age per se but to yearly driving distances, a phenomenon that the authors called low mileage bias. “These findings cast serious doubt on any previous reports of age differences in accident risk per distance driven” (HakamiesBlomqvist et al, 2002, p. 274). After the original study by Hakamies-Blomqvist et al., these findings were independently replicated using French (Fontaine, 2003) and Dutch (Langford et al., submitted) data sets, lending credibility to the robustness of the low mileage bias.
Janke (1991) attributed the association between distance driven and crash rates to the different levels of mileage driven on different types of roads by the different mileage groups. For example, high mileage drivers typically accumulate most of their mileage on freeways and multi-lane divided roadways with limited access. By implication, low mileage drivers undertake a higher proportion of their driving on local roads and streets, with greater number of potential conflict points and hence higher crash rates per unit road distance. Janke noted that there were 2.75 times more crashes per mile driven on non-freeways than freeways.
Hakamies-Blomqvist et al. (2002) also looked to the possibility of different amounts of freeway and non-freeway driving as explaining the mileage/crash association – but also held open the possibility of low mileage drivers as being otherwise more accident-prone than their higher mileage counterparts.
One explanation for low mileage older drivers being relatively accident-prone runs thus.
Some older drivers in response to a perceived decline in driving performance restrict their driving as a safety and/or comfort measure (Eberhard, 1996; Evans, 1988; McGwin & Brown, 1999; Preusser et al., 1998; Smiley, 1999). These drivers would be expected to have more medical conditions and greater functional difficulties leading to reduced driving skills, relative to drivers with higher mileages – and intuitively, a higher probability of crashing.
The current study investigates the hypothesised relationship between mileage driven, crash risk and the possibility of functional impairment amongst different mileage groups.
This paper seeks to re-examine data from 244 older drivers in New Zealand:
to determine whether the association between distances driven and crash involvement • (the low mileage bias) holds true for a sample of New Zealand older drivers;
to determine whether low mileage older drivers can be identified by a range of • functional performance measures;
to determine whether low mileage older drivers can be identified by external measures • of driving performance (the NZODORT).
METHOD Participants1: At the time of conducting the study in New Zealand, drivers aged 80 years and older were required to undertake a biennial medical check and also complete the New Zealand Older Driver Re-licensing Test (NZODORT) in order to maintain their driving licence.
Participants for the current study were recruited from a Land Transport Safety Authority (LTSA) database of drivers who met the following inclusion criteria:
1 Participants were recruited for a larger study that sought to assess the validity of three leading screening tests of fitness to drive (Austroads, 2004). Full details of the participants, recruitment procedures and assessment procedures can be found in this publication.
were aged 80 years or older (or had their 80th birthday during the study);
• had undertaken a medical examination (as required for their licence renewal); and • were either about to undertake the NZODORT in the 3-month period following the • commencement of the study, or had completed the NZODORT in the 9 months prior to commencement of the study.
The study was conducted in Wellington, New Zealand from February 2001 to May 2002.
Data sources: Driver Survey: Participants completed a survey which included items covering demographic measures, ratings of self-reported driving performance and information on travel patterns, health status, medical conditions and functional performance and selfreported crash history as a car driver. Crash history was defined as involvement in a crash in the last 2 years where (i) the car was moving (ii) caused occupant injury or vehicle damage (ii) not on private property (iv) either ‘at fault’ or not. The survey was administered in an interview style and took approximately 10-15 minutes to complete.
Driving performance: Participants in the present study underwent a driving assessment as part of the standard re-licensing procedures in New Zealand. Results of the driving assessment were made available by the licensing authority with participants’ approval. The NZODORT was developed by LTSA and has been used in New Zealand since May 1999. There are three parts to the test, each of which has a set of performance criteria. Part 1 is a sequence of four driving manoeuvres including leaving the kerb, driving straight, turning left and returning to the kerb, to ensure a ‘warm-up’ period for the driver. Parts 2 and 3 are more detailed, with the difference between the two parts being the level of ‘traffic demand’, i.e., speed, traffic, density and type of manoeuvre. Driving manoeuvres assessed in Part 2 include turning left and right (unburdened and at stop and give-way signs), straight through roundabout, and hazard detection. For the hazard detection task, drivers are required to pull to the side of the road at certain points and are asked to identify potential hazards on the road ahead. Part 3 assesses the following driving manoeuvres: driving through a shopping/commercial precinct;
turning right across traffic and at stop and give-way signs; driving straight. Drivers are permitted multiple attempts on the NZODORT. However, for the purpose of this study, the on-road results were scored as “pass” or “fail” on the first attempt. That is, those who had two or more attempts before passing were scored as a “fail”.
Screening Test: Participants in the present study also completed the Gross Impairments Screening Battery of General Physical and Mental Abilities (GRIMPS) (Scientex, Washington). This paper and pencil test measures a number of skills and abilities that are believed to be important for the driving task including tests of cognitive and gross motor functioning. These skills are believed to be at risk of decline in older age and it is believed that such a decline may place a driver at an increased risk of crashing. The GRIMPS test is comprised of 11 sub-tasks. Performances for each task are scored (timed, number of errors etc,) as either “average or above” or ‘below average’, using criteria established by the test developers. A description of each sub-task and the criteria required to score “average or above” is listed below.
1. Rapid-Pace Walk – Participants were required to walk along a line (approximately 10 feet long) on the ground as quickly as they could. This task was timed. Participants who completed the task in 7 seconds or less scored an “average or above” rating.
2. Foot-Tap Test – Participants were required to sit in a chair and tap their foot from side to side (five times on each side of a folder). This task was timed. Participants who completed the task in 8 seconds or less scored an “average or above” rating.
3. Cued Recall – The assessor read three short words (e.g. bed, apple, shoe) to the participant and asked the participants to repeat the three words. Following this, the assessor told the participant that they would be asked to remember the same three words later in the assessment. Participants scored an “average or above” rating for this task if they could recall all 3 items after the first presentation.
4. Arm Reach – Participants were asked to raise both arms (one at a time) as high over their head as they could. Participants who could lift each arm so that their elbow was shoulder height or above scored an “average or above” rating for this task.
5. Head/Neck & Upper Body Rotation – Participants were required to sit in a chair with a lap sash seat belt on and were then asked to look over each shoulder and read the time on a clock face held up behind them. Participants who could turn and correctly read the presented time scored an “average or above” rating for this task.
6. Motor-Free Visual Perception – Participants were shown 11 incomplete objects and were then required to identify what each object would look like if the object was completed from a selection of choices. Participants who made 2 or less errors on this task scored an “average or above” rating.
7. Delayed Recall – Participants were asked to remember the three words presented to them earlier in the “cued recall” task. Participants scored an “average or above” rating for this task if they could correctly recall 2 or 3 of the items that were presented earlier.
8. Scan Test – Participants were required to stand at arms length in front of the assessor and, without moving their head, scan a chart held in front of them and name all of the shapes that were presented. Scan patterns were scored in the following categories: normal (clockwise, by rows etc); erratic (all shapes were identified but in a haphazard way) or neglect (two or more shapes weren’t identified at all). Participants who demonstrated a “normal” scanning method scored an “average or above” rating for this task.
9. Trails A - Participants were required to connect a series of numbers in ascending order from 1 through to 8 on a sheet of paper as quickly as they could. This task was timed.
Participants who completed the task in 30 seconds or less scored an “average or above” rating.
10. Trails B - Participants were required to connect a series of numbers and letters in ascending order (e.g. 1-A-2-B-3-C…) on a sheet of paper. This task was timed. Participants who completed the task in 210 seconds or less scored an “average or above” rating for this task.
11. Visual Acuity – In the last task, participants were required to read a series of letters (high and low contrast) on a chart. Participants with an acuity score of 20/40 (i.e., able to read line 5 without errors) and who had a high/low contrast difference score of 0, 1 or 2 scored an “average or above” rating for this task.
The following scores were derived (i) individual sub-test scores (i.e., “average or above” or “below average”; (ii) individual raw scores (where appropriate); and (iii) a total score (maximum score = 11).