When it comes to improving the safety of its buses on the roads, King County Metro takes a reactive approach, when it should be proactive, a new audit of the transit agency has found.

While Metro conducts investigations of all collisions and injuries, it reserves its closest examinations for incidents deemed preventable, while focusing less on near collisions and those found to be out of the hands of a bus driver.

By filtering out incidents based on outcomes rather than the circumstances that led to them, Metro misses the opportunity to take a more holistic view of how serious injuries and deaths occur, the team from the King County Auditor’s Office found.

The office examined 10,000 incidents from 2018 through 2021.

“The difference between a fatality, a serious injury and a near-miss are tenths of seconds in some situations,” said deputy county auditor Ben Thompson. The goal should be “trying to focus less on the outcome, because there’s so much randomness in that, and more on the circumstances that led to that encounter.”

Under the current system, only incidents deemed preventable undergo a “root cause analysis.” The events are assigned a minor, major or serious classification based on the outcome of the collision. Staff then follow up with operators to provide training or discipline. The categorization also simplifies analysis of incident data.

Just 20% of collisions or injuries were deemed preventable. Roughly 40% of incidents did not result in either injury or property damage, and therefore did not receive the same level of analysis. Another 40% were found to be “non-preventable” by the operator.


The auditor’s team singled out occasions when passengers fall under a bus, which is often serious or deadly, as an example of incidents deemed “non-preventable.” This can occur when someone grabs on to the back door as a bus is leaving. By not probing these incidents more deeply, Metro could be missing something that could be changed, the auditors wrote. It also means that Metro is less able to analyze large swaths of data about incidents not determined to be preventable.

“The exclusion of these types of incidents from further analysis makes it more difficult for Metro Transit to find patterns and design measures that could help prevent similar incidents in the future,” the audit reads.

Rebecca Frankhouser, Metro’s director of safety and security, emphasized that Metro does follow-up investigations on all incidents. Nevertheless, she agrees with the audit’s conclusions that the agency could learn more about what leads to collisions or other incidents by broadening its analyses of root causes. The issue is staffing, she said.

“We would love to do root cause analysis on all of them, but it’s about using our resources appropriately,” she said.

The auditing team also focused on training, which is currently only prescribed after a safety incident. As a result, some drivers may go years without any follow-up or refresher training, which could create “operator complacency.” When a driver is assigned training, its connection to a recent incident can make the course feel punitive, the auditors found.

Thompson said the auditing team is “not trying to sensationalize” the problem at hand. Metro is following best practices identified by the Federal Transit Administration and the system remains relatively safe when compared with other modes of transportation, Thompson said.


“Metro is largely doing a good job with this and it’s really the case of being able to make improvements, expand what they’re doing to deepen and really understand what works,” he said.

The auditor’s office is making several recommendations, including analyzing all collisions and near-collisions, and scheduling regular trainings not tied to specific incidents. The auditors also recommended more details of every safety incident be included in the county’s databases for analysis. For example, while Metro staff identified operator fatigue as a potential cause of collisions, fatigue is not a field in the agency’s data sets and therefore can’t be screened en masse by the safety division.

“If the division had more efficient access to data like safety complaints and traffic violations, it would be better able to recognize potential hazards and recommend appropriate response measures to prevent future incidents,” the auditors wrote.

The auditor’s office also noted one change King County Metro is considering that could reduce drivers’ incentive to speed. Metro is considering shifting from a fixed-time schedule to intervals, so riders could expect a bus every, say, 10 minutes, but not the exact time of its arrival. Drivers would then feel less compelled to speed up to meet a specific time.

Responding to the audit, the county executive’s office agreed with all of the recommendations. King County’s chief operating officer, Dwight Dively, wrote that the “implementation of your recommendations will help us continue our work to improve our operations.”

Improving data collection could take some time as the county may need to overhaul its current systems and create new data sources, he said. He also said making trainings routine for all drivers presents a challenge.

“Metro has almost 2,600 transit operators and the logistics of getting everyone through annual refresher training has had mixed success in the past,” he wrote.