The Performance Analysis of Driving Ability (P-Drive) by Patomella (2004), is a comprehensive standardized driving assessment designed to determine an individual’s fitness to drive and is based on three theoretical assumptions: The Model of Human Occupation activity analysis as well as psychological theory about a person’s decision making while on road. Originally developed for use with a driving simulator, the P-Drive, has been adapted to real life driving situations and is composed of 27 items that form 4 subgroups: (1) maneuvering, (2) orientation, (3) ability to follow regulations, and (4) attending and acting on different stimuli in the traffic context (Patomella & Bundy, 2015). During assessment the subject is graded on his or her ability to safely and effectively engage in task items that, as a whole, lead to safe driving and include such items as steering, using turn signals, wayfinding, trip planning, obeying stop and speed regulations, and problem solving behind the wheel (Patomella & Bundy, 2015). Scoring requires the clinician to grade the safety and quality of observed performance as the subject drives along a predetermined route or simulated program if using a driving simulator. Each item is scored along a 4-point criterion-referenced scale (1 to 4) where a score of (4) suggests that the person is competent and safe, (3) they are questionable, (2) problematic, and a score of (1) denotes an incompetent and unsafe performance. A total score of 108 is possible with higher scores indicting better driving ability. The P-Drive can be completed in 30 to 60 minutes.
A study of 85 volunteer drivers (mean age: 76 years) determined that the median fail score for 18 subjects was 78 (range: 74 to 81), while for those that passed (n = 67) it was 94 (range: 90 to 96) (Selander, 2012). The most challenging items were observing signs, attending to the left and right, following speed regulations, and giving the right of way, while the most frequent errors were made in obeying the speed limit, changing gears, use of indicators, and attention to the left and right (Selander, 2012). A subsequent study of 99 subjects with neurological impairment referred for driving evaluation determined that the P-Drive was able to separate drivers into 4 different strata with a reliability of r = 0.92, while at a cutoff score of 81 the sensitivity was 93% and specificity was 92% (Patomella & Bundy, 2015). Using the same cutoff score, the positive predictive value was 0.95 and the negative predictive value was 0.90 (Patomella & Bundy, 2015). A similar study of 101 subjects post-cerebrovascular accident (CVA) determined that the P-Drive was also able separate subject driving ability into different strata at a reliability of only r = 0.84 (Patomella, Tham, & Kottorp, 2006). The same study’s investigation of criterion dimensionality found that the first component accounted for 64% of the variance indicating good uni-dimensionality, whereas the second component explained 6% of the variance (Patomella et al., 2006). When researchers analyze components of a scale the first or main one should explain at least 20% of the variance and the second one should not explain more than 5% of the variance.
Findings suggest that, overall, the P-Drive is a valid and stable assessment tool with sound psychometric properties that can differentiate people with diverse driving abilities across a number of patient populations including those related to age. Since it was developed by an occupational therapist, it could also be considered specific to that profession.
Researchers have noted inherent problems with driving tests in general that can effect standardization, and those that affect the P-Drive in particular are variations in the choice of driving routes, as each clinician develops his or her own, the choice of vehicle, as well as individual differences in how testing is conducted. Also, the study by Patomella et al. (2006) of persons post-CVA, discussed earlier, determined that although most items displayed goodness of fit at the 95% confidence interval, controlled speeding did not fit model expectations at only 0.54, suggesting that item may need to be reworked. The same study found that a score of 4 was obtained 56% of the time and a score of 1 only 7% denoting a possible ceiling effect. Finally, P-Drive is still considered a work in progress and although its administration is discussed at length in several articles an examiner’s manual is not yet available.
If using a simulator, the subject is allowed to practice for up to 15 minutes to get used to the controls of the car, such as the pedals, gear, and steering. Similarly, familiarity is allowed in the vehicle prior to actual driving. Route creation for both simulated and on-road assessment requires the clinician to design one that includes specific challenging traffic situations based on a list from the manual to include such items as yielding and stop regulations, problem solving tasks using unstructured instructions (i.e., find your way back to the gas station where we recently parked), as well as driving in different situations, at different speeds, in rural areas, on highways, and in city traffic (Patomella & Bundy, 2015). Scoring requires the clinician to grade the subject’s ability as they relate to the 27 task items of the assessment which may include such observations as the participant collided with a car coming from the right as they had no right-of-way (simulator); the subject was able to pay attention to the right and was able to successfully give the right-of-way when necessary; or the subject was able to control speed by adapting it appropriately up to and through a crossing (Patomella et al., 2006).
• Using pedals
• Controlling speed
• Keeping distance
• Yielding right-of-way
• Attending and acting straight ahead
• Attending and acting to left
• Attending and acting to right
• Attending and acting to mirrors
• Attending and acting to fellow road users
Adapted from Patomella, A. & Bundy, A. (2015). P-Drive: Implementing an assessment of on-road driving in clinical settings and investigating its internal and predictive validity. American Journal of Occupational Therapy, 69(4), 4.