A large-scale double-blind verification of the system showed an excellent agreement of manual and automated analysis and proved the system to work dependably.
Although the AUC of 0.9 in our original report for both training and blinded verification was promising, we had to eliminate some markers, such as HSP90, due to preanalytic bias leading to site-to-site differences.
While drivers in the Control group showed no change in blind spot verification (12.5% vs. 13.8% of verification for the Pre- and Post-training sessions, respectively), drivers in the Feedback group increased the verification of their blind spot (32.3% vs. 64.9%, for the Pre- and Post-training sessions, respectively).
Moreover, by participating in active practice sessions in the simulator, they were able to engage the necessary processes leading to appropriate driving behaviors when tested on-road (i.e., increased frequency of blind spot verifications and earlier verifications).
The course included specific sections with graphical support to inform participants about traffic regulations, mirror and blind spot verification when changing lanes and vehicle control.
It is worth noting that blind spot verification was specifically reviewed in the refresher program and the importance of this visual search action prior to changing lanes was emphasized with specific graphical support.
In addition, there were potential sources of bias, such as the lack of a description of blind test verification and clarity as to whether the same clinical data were available when test results were interpreted as would be available when the test is used in practice.
Study quality was defined as a clear description of sampling, data collection, study design, blinding, (partial) verification and missing data.
At the Pre-training evaluation, although we randomly assigned subjects to each group, there was a significant group difference for the frequency of blind spot verifications during lane changes (32.3% vs. 12.5% for Feedback and Control, respectively).
The analysis of visual search strategies during lane changes revealed that the driving-specific Feedback group increased drastically the frequency of blind spot verifications (from 32.3% to 64.9% of the lane changes) whereas the Control group did not.
