Virtual driving assessments could forecast crash risk in young drivers

Young drivers who have recently been licensed have a higher-than-average risk of crashes, especially in the first several years from licensure. A major contributor to this increased risk is the lack of adequate competence in driving skills.

A new paper recently published in Pediatrics explores the possibility of predicting post-licensure crash risk by a virtual driving assessment (VDA) performed at the time of licensure.

Study: Driving skills at licensure and time to first crash. Image Credit: Ground Picture / Shutterstock.com

High risk of young drivers

Drivers between the ages of 15 and 20 years comprise only about 5% of all licensed drivers, according to 2020 statistics. However, 12% of all crash drivers are between 15-20 years of age, with 9% of drivers within this age range dying in crashes. The crash risk is highest in the immediate post-licensure period, slowly reducing within a few years to that of experienced adult drivers.

Some potentially useful interventions to reduce this risk include mandatory on-road training for young drivers or graduated driver licensing. The cost, time, and non-targeted nature of these programs has led to their poor implementation in most United States regions.

In contrast, considerable research indicates that poor driving skills are crucial reasons for crashes among inexperienced drivers. An earlier study showed that VDA could predict on-road testing failures due to both vehicle control skills and driving skills; however, this study did not discuss which skills were at fault. This was then used to identify clusters of driving skills, like speeding, tailgating, and control, into four major classes, including no issues, minor issues, major issues, and significant issues with dangerous behavior.

What did the study show?

The current study included about 17,000 young drivers below the age of 25 years in Ohio. All study participants performed a virtual driving assessment (VDA) immediately before their licensing examination, which they passed to obtain their license.

The Ready-Assess VDA is a ∼15-minute-long self-directed workflow including a driving route that incorporates common serious crash risk scenarios and varied settings (urban and suburban), physical road features, and other road potential hazards.”

The time from the VDA was recorded within the nearest 15 days from the first crash. Drivers were categorized based on their measured skills and classified as no issues, minor issues, major challenges, and significant issues with dangerous behavior class had a higher crash risk. The classification depends on the degree of control and risky driving behavior.

The results were analyzed after compensating for test center location, licensing age, sex, and socioeconomic status. In the fully adjusted analysis, the best driving was observed among the no issues drivers, whose crash risk was an average of 10% lower than the average risk. This was so with all associated skill clusters; however, only cautious driving was statistically significant at a 17% reduced crash risk.

A significantly increased risk of 7% was associated with the speeder, tailgater, and rule breaker cluster in the minor issues class. Thus, this cluster, which accounts for less than 1% of drivers, belongs to one of the major classes.

Drivers who were licensed at 18 years old were associated with a higher crash risk of 18% above average, perhaps because of the lack of mandatory pre-licensure driver education and training at this age. This risk increased by almost 40% with the skill cluster skilled average.

The greatest risk was among those exhibiting major issues with dangerous behavior, who had an 11% higher crash risk than the average. All skill clusters were associated with an increased crash risk.

Significantly increased risk was observed among newly licensed drivers with risky, no-control, and jackrabbit behavior, with a 24% increase in crash risk. Jackrabbit behavior refers to sudden and jerky movements of the vehicle.

Another high-risk cluster was drivers who were extremely aggressive or reckless, with a 7% increase in crash risk.

The difference in relative risk between these categories increased over time rather than remaining constant. This was true with three skill clusters, comprising below-average control with a nearly 50% increase in risk at 1-1.5 years post-licensure, the quick with controlled braking cluster where the 18-month risk rose by almost 75% from almost-average levels at baseline, as well as the risky, poor control, jackrabbit cluster, in which the risk at 12-18 months more than doubled from the already raised risk at licensure, declining to 53% higher than the baseline risk after 18 months.

What are the implications?

The study confirmed the presence of a significant association between the driving skills of a newly licensed driver and the time to the first crash. The no issues new drivers were the safest, while among those with minor issues.

Conversely, some skill clusters in the major issues clusters were not associated with significantly raised crash risks. The age of licensing was related to crash risk in the 18-year age group only, perhaps due to the absence of pre-licensure mandated driver education and training with an initial six-month learner’s permit.

With this information, programs can be developed that provide personalized and targeted interventions.”

This data could help provide feedback at the time of licensure and develop skills that could avoid crashes in the early driving period for novice drivers by helping them with greater situational awareness and hazard identification, with appropriate preventive action.