Rasch Regression Predicts Driving Capability

Rasch Regression for the Older Driver Data

|-----+-----+-----+-----+-----+-----+-----+-----+-----|
      0  :    1  :   2  :    3   :    4                   CLOCK
    0  :   3:  4:  8 :  9:10:11 12 13:  14:   15  :  16   VISUAL
|-----+-----+-----+-----+-----+-----+-----+-----+-----|
incapable           0     :      1            capable  DRIVING DIAGNOSIS
|-----+-----+-----+-----+-----+-----+-----+-----+-----|
-4   -3    -2    -1     0     1     2     3     4     5
  X                   X  X X X X X  X   X     X      X CAPABLE DRIVERS
                      X  X X     X  X   X     X      X
                      X  X          X                X
                         X          X                X
                                    X
                                    X
                     S            M            S
           X       X  X  X   X X X                     INCAPABLE DRIVERS
           X       X  X  X   X X X
           X       X  X      X   X
                      X      X
                 S       M       S

The conventional way to assess the driving ability of older drivers is to administer a standard driving test. Less expensive diagnostic alternatives are being sought as replacements. A representative sample of 49 elderly drivers, referred for a driving diagnosis and possible license renewal, was given two paper-and-pencil tests of cognitive functioning: the Motor-Free Visual Perceptual Test (VISUAL) and the Clock Drawing Task (CLOCK). Both instruments have been used as "screening tool[s] in identifying those who are unfit to undergo an on-road evaluation" (Korner-Bitensky et al., p. 253). The data were analyzed with Wright's Rasch regression model. The VISUAL and CLOCK variables were co-calibrated and anchored to define a two-test unidimensional, reproducible, and stable yardstick. The Driving Diagnosis was then added as a dichotomous item.

The KEYMAP illustrates that a person with the score of 11 and higher on the VISUAL test or the score of 3 and 4 on the CLOCK test would have a high probability of being diagnosed as capable of driving. Here we find 18 patients diagnosed as capable drivers and 9 patients diagnosed as incapable drivers. Similarly, people with score of 10 and below on the VISUAL test or the score of 1 and 2 on the CLOCK test would have a high probability of being diagnosed as incapable of driving. Here we find 10 patients diagnosed as capable drivers and 12 patients diagnosed as incapable drivers.

The persons, however, are mainly clustered in the area between 0 and 1 on the Driving Diagnosis, that is, within approximately 1 logit of the half-score point. This lends itself to the conclusion that more than two thirds of the participants could not be diagnosed with greater than 75% accuracy.

The map also shows that the predictive potential of the two tests is different. For the CLOCK test, even the highest score of 4 does not predict a person's driving capability with 90% accuracy. Therefore, this test may not be a reliable screening tool for identifying capable drivers. For the VISUAL test, the ruler is longer, but almost all the drivers cluster in the middle. A person needs to score between 0 and 4 to be accurately diagnosed as cognitively incapable, or to score between 14 and 16 to be accurately diagnosed as capable.

Svetlana Beltyukova, Dan Cipriani, Song Yan, Tina Ughrin & Christine Fox, The University of Toledo

Korner-Bitensky, N. A., et al. (2000). Visual testing for readiness to drive after stroke: A multicenter study. [VISUAL]. Medical Journal of Physical Medicine and Rehabilitation, 79, 253-259.

Shulman, K. I. (2000). Clock-drawing: Is it the ideal cognitive screening test? [CLOCK]. International Journal of Geriatric Psychiatry