Summary

The National Highway Traffic Safety Administration (NHTSA) reports that the number of bicyclists killed in traffic crashes has been steadily increasing since 2010 with an average of 779 killed between 2012 to 2016, 890 from 2017 to 2021, peaking at 966 killed in 2021. In response to these trends (and similar trends for pedestrians), communities throughout the United States are adopting policies and practices to improve safety for bicyclists. However, planners, engineers, and designers who are implementing bikeways need additional information about the safety performance of intersection treatments when assessing tradeoffs and making design decisions at intersections. The objective of this research, NCHRP 15-73 Design Options to Reduce Turning Motor Vehicle-Bicycle Conflicts, was to develop guidelines and tools for transportation practitioners to reduce severity of conflicts between bicyclists and turning motorists at controlled intersections and improve the safety and comfort of intersections for bicyclists. The foundation of the research is a state of the practice review that included a literature review, a summary of current design guidelines, and practitioner interviews that identified research needs and focus areas. This research focused on five different intersection treatments identified in the state of the practice review: conventional bike lane at intersection, separated bike lane at intersection, pocket bike lane, mixing zone, and protected corner. These treatments are focused on the design at the intersection, and they can be combined with various bikeway types on the segment (e.g., a conventional bike lane along a segment could transition to a separated bike lane at intersection treatment or it could transition to a Mixing Zone at the intersection). There are additional intersection treatments that are available for practitioners to implement at intersections that were not studied as part of this project, including bikeways with two-way bicycle traffic (e.g., two-way separated bike lanes and shared use paths), raised crossings, left-turn phasing, bicycle boxes, two-stage bicycle turn boxes, and roundabouts.

The first completed analysis was the macro-level crash analysis that examined bicycle crashes and injuries at a broad scale using state-level databases in California, Minnesota, and Texas. It provided a thorough understanding of bicycle crashes at intersections in varying land use contexts, including crash type (both left-hook and right-hook crashes), frequency, and severity. The findings from the macro-level crash analysis were used to help identify potential candidate sites for the subsequent micro-level crash analysis and video-based conflict analysis. The cities selected for these analyses (Austin, TX, Minneapolis, MN, New York City, NY, and Seattle, WA) were chosen primarily for the presence of treatment types and availability of high-quality safety and contextual data.

Following the macro-level crash analysis, the team implemented a tiered approach using three research methods at the intersection and approach level that clarified the relationship between key risk factors and intersection treatment designs in varying intersection contexts, and ultimately disentangled these relationships to provide substantive guidelines for practitioners.

• Micro-level crash analysis examined total bicycle crashes on the intersection approach, injuries, roadway design, and operational variables to clarify how exposure (i.e., volume) and various roadway design factors influence risk across a variety of facility types and site-specific contexts. Safety performance was evaluated at 573 intersection approaches; 233 crashes were examined across the study cities. Network-wide exposure models were available in Seattle, Minneapolis, and Austin; however, only a subset of New York City sites had available bicycle count data.
• Video-based conflict analysis examined both left- and right-turn conflicts (though most observed conflicts were right-turn conflicts) for individual intersections approaches to provide additional insight into the dynamics between bicycle safety and known risk factors such as turning volumes, speed, and design features. A total of 28 intersection approaches were studied using 2,167 hours of video data.

• Human factors study used a driving simulator to explore and validate design assumptions about interactions between right-turning vehicles and through bicycles, as well as cognitive behaviors of drivers. Forty participants completed 640 turns generating over eight hours of detailed driver performance data.

The research revealed insights into design-level questions uncovered in the state of the practice review. Importantly, it provided information on relative safety performance and informed design-related thresholds and guidelines. Each study in this research is robust and produced valuable information; however, the differing scale, focus, and limitations of each approach requires interpretation to synthesize the results. Taking these studies together, the synthesized results for each of the studied intersection treatment types are as follows:

• Conventional Bicycle Lane at Intersection – conventional bicycle lane at intersection had the second highest crash rate in the micro-level crash analysis and second highest predicted conflict frequency in the conflict analysis. The conflict analysis showed that severe conflicts at conventional bike lanes was more sensitive to increases in bike volumes compared to other treatments (i.e., conventional bike lanes are more likely to result in severe conflicts as bicycle volumes increase). Other studies clearly show that bicycle facilities are preferable to no facility for bicyclist comfort, so, while providing a Conventional Bike Lane at Intersection is preferable to providing no facility at the intersection, enhanced intersection treatments with better safety performance should be used where possible. In the Decision Tool and Design Guidelines, a conventional bicycle lane at intersection is only recommended once practitioners have made every effort to reallocate space to provide a Protected Corner or Separated Bike Lane at Intersection.
• Separated Bicycle Lane at Intersection – The design of this intersection treatment is similar to conventional bicycle lane at intersection, except it includes vertical separation elements in advance of the intersection and a horizontal offset (i.e., buffer) from the adjacent travel lane. By comparison, Separated Bike Lanes had almost half the number of severe conflicts as the Conventional Bike Lane at Intersections. Because of the varying safety results (i.e., high crash rates but low predicted number of conflicts) and because a common application of this treatment is at locations with lower turn volumes or bicycle volumes, it is not expected to have effective safety performance at higher turning volumes. At higher volume locations, full or partial phase separation may be needed to manage the conflicts and improve safety performance. In the Decision Tool and Design Guidelines, a Separated Bicycle Lane at Intersection, is one of two preferred treatments. In some conditions (depending upon the right-turning motor vehicle volume and anticipated bicycle volume), a leading interval or full phase separation is recommended with this treatment.
• Pocket Bike Lane – (Note: The Decision Tool and Design Guidelines uses the term “pocket bike lane.” Within the body of this report, the term “pocket bike lane” and “keyhole bike lane” are used interchangeably and refer to the same treatment.) The micro-level crash analysis suggested that pocket bike lanes were relatively safer than conventional bicycle lanes, showing they were associated with fewer crashes. However, the conflict analysis and human factors study provide additional insights. Specifically, the conflict analysis showed higher-severity conflicts and higher motorist speeds at these treatments, and the human factors study showed high-speed movements that degrade the benefits of the treatment type. Taken together, these findings indicate that this treatment type may not be appropriate for accommodating Interested but Concerned bicyclists, due to the need to consistently watch for and negotiate higher-speed conflicts with motorists for safe passage. Finally, the pocket/keyhole bicycle lane may be an acceptable design in limited situations to accommodate a right-turn lane as long as the risk factors identified (e.g., vehicle speed) are addressed, but it is otherwise not recommended. In the Decision Tool and Design Guidelines, a Pocket Bicycle Lane is only recommended in limited situations.

• Mixing Zone – The micro-level and conflict analysis indicated effective safety performance (i.e., lowest crash rates and lowest predicted number of conflicts). The human factors study also indicates high levels of motorist attention, which can contribute to safe operation. However, it is likely that more risk-tolerant bicyclists use bikeways with mixing zones because they require more interaction with motorists. In design and operation of mixing zones, care should be taken to maximize bicyclist comfort and comprehension of expected behavior for both bicyclists and motorists by implementing treatments to reduce speeds, such as shortening the right-turn lane and adding flex posts to shorten the entry point for motorists. In the Decision Tool and Design Guidelines, a Mixing Zone is only recommended where right-turning motor vehicle volumes are high and practitioners have made every effort to reallocate space to provide a right-turn lane and a Protected Corner or Separated Bike Lane at Intersection.
• Protected Corner – (Note: The Decision Tool and Design Guidelines uses the term “Protected Corner” which refers to the treatment of one intersection approach with elements of a protected intersection. In the body of this report, the terms “Protected Corner,” “Protected Intersection,” and “Offset Intersection” are used to refer to the same treatment) This treatment type was associated with the largest bicyclist volumes, underscoring preferences for greater separation between bicycle and motorist traffic. Despite these higher bicycle and motor vehicle turning volumes, the micro-level and conflict analysis results indicate that offset or protected intersections perform better (in terms of crash rates and high-severity conflicts) than conventional bicycle lanes in all cities and better than all other treatment types in New York. Based on this research, this treatment appears preferable to conventional bicycle lanes and pocket/keyhole bicycle lanes and comparable to mixing zones from a safety performance perspective. However, the comparable result with mixing zones is likely a product of more risk-tolerant bicyclists (i.e., those comfortable enough to interact with motor vehicle traffic) being the majority of users at mixing zones. In the Decision Tool and Design Guidelines, a Separated Bicycle Lane at Intersection is one of two preferred treatments. In some conditions (depending upon the right-turning motor vehicle volume and anticipated bicycle volume), a leading interval or full phase separation is recommended with this treatment.

The research culminated with the development of the Decision Tool and Design Guidelines, a stand-alone document that provides a framework for decision making and the assessment of tradeoffs for intersection treatments. Conflicts between bicyclists and right-turning motorists was the focus of the original research, so right-turn conflicts is also the focus of this decision tool. The document acknowledges the significance of left-hook crashes and includes references to the existing body of knowledge and state of the practice for managing left-turn conflicts. It combines the findings of the research on the relative safety performance of intersection treatments completed as a part of NCHRP 15-73 and previous research related to bicyclist comfort, specifically the preference for separation – either physically or in time. Taking this research and previous research on comfort together, separated bike lanes (with physical separation extending to the intersection), protected corners, and phase separation (at higher vehicle and bicycle volumes) were prioritized in the decision tool. The guide includes a set of decision-making principles that emphasize: 1) safe system approach, 2) considering bicycle design users in design decisions,, and 3) using design flexibility and engineering judgement. The tool is a flow chart that relies primarily on total bicyclist volume (either existing or anticipated) and motorists turning volumes; both were the risk factors prominent in this research. The supporting materials include additional discussion on strategies for reallocating space and additional considerations for phase separation (i.e., heavy vehicles, intersection skew, and presence of transit). Finally, supplemental design guidelines provide recommendations for mitigating known safety concerns for each intersection treatment, including a shared lane, conventional bike lane at intersection, separated bike lane at intersection, pocket bike lane, mixing zone, and protected corner.

The report is organized into ten chapters. Chapter 1 provides an overview and lays out the objectives of the research. Chapters 2, 3, and 4 cover the state of the practice (i.e., Literature Review, Summary of Design Guidelines, Practitioner Interviews). Chapters 5 through 9 include the research methods and findings beginning with Site Selection, followed by one chapter on each of the research methods (i.e., Crash Analysis, Video-Based Conflict Analysis, Human Factors Study), and a chapter on Synthesis and Summary of Results that combines the key findings from each of the analysis approaches and recommends future research. Finally, the report concludes with a brief Chapter 10 that explains the process for developing the decision tool, design guidelines, and training materials.