A MnDOT flagger holds up a sign to warn drivers entering the work zone.
The night a minivan barreled through his crew’s work-zone closure is one Erik Shelstad will never forget.
That evening, the experienced Minnesota Department of Transportation (MnDOT) operations supervisor was overseeing a mill-and-fill project on a three-lane interstate. Shelstad’s crew had set up a full stationary double-lane closure with all the necessary cones, barricades, arrow boards and signage; from start to finish, the work zone was more than a mile long.
“Our milling machine was in the process of cutting out a failed area of asphalt when the minivan, traveling at a high speed, drifted into our closure,” Shelstad said. “With the noise of the equipment, nobody heard the van as it began running over more than 1,000 ft of cones, but when it struck the rear of the mill, everyone heard that.”
The van smashed into the mill and continued on, though no one was quite sure how. “The vast majority of the sheet metal from the passenger side of the van was lying in our closure, along with the van’s A-pillar and a huge amount of glass,” he said. “The inch-thick steel plate at the rear of our mill was folded over like a playing card.”
A deadly problem
The minivan that plowed through Shelstad’s work zone was never found, and miraculously no workers were hurt that night. Yet far too often, similar stories have tragic endings.
Work-zone intrusions—in which vehicles breach the boundaries of roadway construction or maintenance operations—are a serious safety concern. From 2005 to 2010, 733 road workers were killed in work zones in the U.S.; about half of those deaths were the result of workers being struck by motorists, according to the Federal Highway Administration. Motorists themselves also are injured or killed by intrusion crashes.
To address this safety risk, it is critical to understand what contributes to work-zone intrusions. Yet little is known because the methods and standards for capturing data around these events are not well established.
“Whether or not a worker is injured, when a motorist enters a work zone it is still an intrusion,” said Sheila Johnson, MnDOT Traffic Services Engineer for Metro District Maintenance Operations. “We currently record the incidents where our people or equipment are impacted, but we haven’t had a way to quantify the entire threat by formally recording all intrusions.”
Road construction crews in a work zone along I-35E in St. Paul, Minn. According to FHWA data, about half of road-worker deaths in work zones are the result of workers being struck by motorists.
Filling the knowledge gap
Transportation researchers with the University of Minnesota’s HumanFIRST Laboratory set out to fill this gap, creating a system for road crew workers to report work-zone intrusions. The data collected could then be used to examine risk factors, provide feedback to workers and the transportation agency, and provide an empirical basis for future policy recommendations to the state.
Research associate Curtis Craig said that in aiming to make the system comprehensive yet efficient and user-friendly, the researchers needed to first learn about the work-zone crews—what they knew, the context of their work and how they carried it out. “We wanted to make sure we were testing the system in ways that reflect how they would use it in the real world,” Craig said.
Researchers interviewed workers across Minnesota in both urban and rural settings; Shelstad was one of the workers selected to provide feedback. “No one wants to sit down in the middle of a job or at the end of their shift and work their way through a document that is complicated or hard to understand,” Shelstad said. “Hopefully our input helped to make it as simple as possible while capturing the important information.”
Through the interview process, researchers found that workers understood an intrusion as a vehicle entering the area cordoned off by cones, but they felt it was practical to report an intrusion only when there was an actual increased risk to the workers on-site. “Whenever there were high risks, they were more likely to want to report it,” Craig said. The interview process also allowed researchers to determine what workers considered to be the primary reportable elements of an intrusion; these included layout, environmental conditions (e.g., weather), location, time, road condition, maneuvers made by an intruding vehicle and work-zone operation type.
Next, the research team at HumanFIRST extracted the qualitative information from these interviews and codified it into a task analysis of a prototypical intrusion-reporting sequence, which informed the initial design of the prototype intrusion report. During testing of the initial design, researchers asked participants (MnDOT maintenance crew and supervisors) to input either a researcher-generated intrusion scenario or an actual one from their experience—“and they all had experiences that they were scared by or that were very memorable to them,” Craig said.
In later phases of testing, video recordings of real intrusions on MnDOT roads were used. Workers and supervisors were asked to “think aloud” as they interacted with the interface and were timed as they completed the reports. Following the reports, supervisors and crew answered questions and filled out questionnaires to clarify the difficulty and usability of the intrusion report interface. “We wanted to make sure it wasn’t taking too much time out of their day and get a feeling for how usable the interface was,” Craig said.
The second phase of testing showed that workers liked the drop-down menus, the comprehensiveness of the reporting system, and that it was relatively quick to use, but that they struggled with whether they would use the report to record minor intrusions that they personally did not feel at risk for, “like a car coming into and out of the work zone and knocking over a few cones,” Craig said. “[Workers] could just go put the cones back up and get on with their workday. So that was an ongoing tension between what we wanted, which was to get as much data as possible, and what they felt they needed to provide.”
As a result, researchers revised the earlier reporting logic by splitting it into an immediate “minor” report and a more comprehensive “major” report for higher-risk incidents. Supervisors and crew also tested different modes of the interface with a laptop, a tablet and a paper form. Based on the results of testing, the final version of the work-zone intrusion reporting interface should be user-friendly and effective. Work crew supervisors noted that the final version also should provide a clear explanation and rationale, which would help them motivate their crews to reliably report intrusions.
“Reporting work-zone intrusions, especially minor ones, is often seen as a low priority by many supervisors and workers because they believe it won’t change anything,” Craig said. “To obtain reliable and consistent intrusion reporting, an effort must be made to inform work-zone supervisors and crew members of the benefits and consequences of intrusion reporting, and engage them in the decision-making process. Communication should be made on a widespread and consistent basis about the current and recent data on intrusions, the data trends, and what changes are being made as a result of the intrusion reporting.”
Making reporting a reality
According to Craig, MnDOT staff are currently reviewing ways in which the intrusion reporting system could be integrated into the agency’s operations. Once the system is integrated into operations, researchers believe the data could be used to propose policy at the state level, further enhance current work-zone practices, and communicate to work-zone crews how their reporting is having a real impact. The success of the reporting system will depend not only on workers using it, Craig added, but on a sustained dialogue between the users and the administrators of the system—this engagement and dialogue between crew and management could improve the safety culture.
Johnson believes putting this data-collection system into place has the potential to provide important benefits for highway workers and transportation agencies: “The information could lead to changes in traffic-control layouts, the possibility of doing more detours rather than lane closures, consideration of additional safety devices and equipment, additional cameras or enforcement in work zones, data for possible work-zone safety legislation, and general consideration for the balance between work-zone safety and traffic delays.”
Likewise, Shelstad hopes this information will lead to further decreases in work-zone intrusions that endanger the lives of him and his crew. “As data is gathered and patterns begin to emerge, more changes can be made in terms of even better work-zone setups, legal consequences for drivers that do enter a work zone, and more public awareness,” he said. “Safety is getting better every year, but we still see injuries and fatalities, so continuing to improve every aspect of work-zone safety should be of utmost importance.”