Visual‐cognitive tools used to determine fitness‐to‐drive may reflect normal aging

Abstract

Background

Scores on many visual‐cognitive tools are proposed as indicators of fitness‐to‐drive. A purported feature of some tools and one believed to be important is that they are ‘age‐independent’. Specifically, scores are not correlated with age and poor scores represent a pathological process rather than normal aging. Yet, we know that several cognitive abilities are associated with age. One potential reason for the apparent age‐independence of some tools is that focusing on older drivers leads to ‘range restriction’, a statistical issue that reduces the magnitude of correlations when values for one variable are restricted to a smaller range than naturally occurs. Hence, the purpose of this study was to investigate whether age is correlated with scores on visual‐cognitive tests when we examine the full age range.

Methods

We recruited 114 drivers aged 18 to 89-years (mean: 42.30 ± 26.50-years). Participants completed several visual‐cognitive tools often used to examine fitness‐to‐drive (Trail Making Tests A and B, Attention Network Test and ‘useful field of view’).

Results

Correlations between age and test scores for drivers 65-years and older only ranged from 0.03 to 0.48. With the whole age range, correlations ranged from 0.56 to 0.84. We also compared ordinary Pearson correlations among visual‐cognitive tests scores to the corresponding partial correlations after removing the effect of age. Whereas ordinary Pearson correlations ranged from 0.40 to 0.69, partial correlations ranged from 0.01 to 0.30.

Conclusion

Test scores may reflect age‐associated normal biological changes. These results have implications for predicting fitness‐to‐drive among older drivers and suggest caution in using these scores.

Key words:

• age
• cognition
• driver's vision
• tests

Acknowledgements

The authors thank Terry Willie and the staff at Admiral Training for their contribution and support and all participants who volunteered their time to make this study possible. This work was supported by a grant from AUTO21 – Networks of Centres of Excellence (A501‐AST) and the Canadian Institutes of Health Research.