Systemic Safety Project Selection Tool

Element 1: The Systemic Safety Planning Process

Overview of the Systemic Safety Planning Process

The systemic safety planning process (Element 1 of the Systemic Safety Project Selection Tool) consists of the four steps shown in Figure 3. Each step can be scaled based on the availability of technical resources and the quality or quantity of data available to support different analytical approaches. The systemic safety planning process is similar to most common safety management processes. The systemic safety planning process involves identifying the problem, screening and prioritizing candidate locations, selecting countermeasures, and prioritizing projects.

The premise that makes systemic safety planning different from traditional network screening techniques is that it looks for similar issues across the roadway system rather than focusing on select locations with high crash histories or densities. The systemic safety planning process begins by looking at systemwide data to analyze and identify focus crash types (those representing the greatest number of severe crashes) and potential risk factors. As the downward arrows indicate, the approach then moves to a micro-level risk assessment of locations across the network, which then leads to selecting relevant mitigating countermeasures most appropriate for broad implementation across those locations, and prioritizing projects for implementation. The upward arrows indicate that the results of one step might suggest the need to return to a previous step and make adjustments before continuing the process.

FIGURE 3. Systemic Safety Planning Process

The following chapters (Identify Focus Crash Types and Risk Factors, Screen and Prioritize Candidate Locations, Select Countermeasures, and Prioritize Projects) further describe the four steps of the systemic safety planning process. Each chapter includes a discussion of the objective, data needs, tasks, and outcomes, along with examples illustrating the fundamental process for each step.

Data Needs

The systemic safety planning process builds on the Federal Highway Administration (FHWA) guidance to develop SHSPs and structure an HSIP using data-driven processes. The four-step process uses basic types of crash, roadway, and traffic volume data that are recommended by the FHWA for use in safety analysis efforts. Several agencies actively applying the systemic approach to safety and conducting research relative to systemic safety improvement also use these types of data. The following chapters describe in detail the specific types of data associated with each step in the systemic safety planning process. These data are recommended for each step because they represent elements that can impact safety performance of a facility. Although agencies might not currently maintain all these data elements, more detailed data provide the opportunity to more specifically identify facility types and risk factors.

The basic objective of the systemic safety planning process is to identify candidates for safety investment. Candidate identification results from comparing the actual conditions of segments, curves, and intersections with a set of observed characteristics associated with the locations where the focus crash types actually occurred. The data-driven process identifies the observed characteristics (risk factors) associated with the focus crash type. The systemic safety planning process uses selected risk factors to differentiate one segment from another, one curve from another, and one intersection from another in order to prioritize these facilities and give higher priority to locations where there is greater potential for future severe focus crashes. The data suggested to support the systemic approach are not intended to be prescriptive; some agencies currently use these types of data to successfully advance a systemic approach to safety. The systemic safety planning process can be adapted to meet agency-specific needs or crash reduction goals, consistent with the data available. The systemic planning process produces results with minimal levels of data; greater levels of data, however, support a more refined prioritization.

Identify Focus Crash Types and Risk Factors

Objective

The objective of the first step in the systemic safety planning process is to identify risk factors commonly associated with each focus crash type experienced across a system. This step involves reviewing systemwide (i.e., macro-level) crash data and identifying potential risk factors for further analysis. Identification of potential risk factors does not necessarily imply the observed characteristics "caused" the crashes; rather, the characteristics are useful to identify the situations in which the crashes occurred, so that similar "looking" locations can be identified and prioritized.

Data Needs

The identification of focus crash types and risk factors results from using a range of data elements. Following are the minimum recommended crash data elements:

• System type
• Crash type
• Facility type
• Crash location type
• Location characteristics

The FHWA's Model Inventory of Roadway Elements Fundamental Data Elements (MIRE FDE) identified for the HSIP is a good starting point for determining which roadway data to collect for a systemic analysis because it includes many of the roadway and traffic volume data elements recommended for use in this first step (FHWA, 2012). However, other road characteristics, such as shoulder width and type or roadway curvature, should be investigated when possible. The road and intersection data inventory is typically available in a geographic information system (GIS) or database, which should be linked with the crash records system. Ideally, the various data sets would incorporate the same time period. When volume data are from an earlier time period than crash data, first consider if changes have occurred that make the volumes no longer relevant (i.e., economic growth). If little change has occurred, agencies can consider using older traffic volumes. To compensate for greater change, agencies can estimate current traffic volumes by applying a growth factor appropriate for the region or jurisdiction.

The systemic safety planning process focuses on severe crashes (defined as fatal plus serious injury crashes) to be consistent with national safety programs. Furthermore, focusing on severe crashes reduces the scale of the effort and resources required to gather supplemental data if needed.

The data-gathering effort is an iterative process; as such, all data do not need to be collected at the beginning of the process. For example, the focus facility types selected will determine the necessary roadway and volume data to be collected to identify potential risk factors.

Process

The first step of the systemic safety planning process—identifying and understanding the risk factors commonly associated with the focus crash types—includes a series of analyses involving three tasks (Figure 4). The first task refines the problem identification from all crashes to a few focus crash types for an entire system or subregion. Further investigation through additional analyses attains a more comprehensive understanding of the facility type and risk factors (i.e., roadway characteristics) commonly linked with the severe focus crash types.

FIGURE 4. Systemic Safety Planning Process: Tasks to Identify Focus Crash Types and Risk Factors

Task 1: Select Focus Crash Types

Task 1 in the step to identify focus crash types and risk factors involves conducting a systemwide analysis of crash types to select those representing the greatest potential to reduce fatalities and severe injuries. This effort typically results in identifying the crash types that represent the greatest number of severe crashes across the system being analyzed.

A good starting point for identifying focus crash types is a state or regional SHSP, which documents emphasis areas (i.e., focus crash types) for the state or region's safety program. These emphasis areas were likely identified through a data-driven process, which is a primary principle of the systemic safety planning process. State agencies with completed Roadway Departure and Intersection Safety Implementation Plans can identify these as focus crash types for their systemic safety programs. As such, the analysis summaries contained in the plans are useful for the subsequent tasks that identify focus facilities and risk factors. Agencies can also refer to safety plans in which emphasis areas specific to their jurisdiction have been identified through a similar data-driven approach.

If the state or regional SHSP does not cover the roadway type or region of interest, then referencing the emphasis areas identified in the American Association of State Highway and Transportation Officials (AASHTO) SHSP is useful to determine what types of crashes require further investigation. Depending upon their needs, an agency can add additional emphasis areas to the list of 22 emphasis areas in the AASHTO SHSP (AASHTO, 2005). Some common additions include winter weather crashes, animal crashes, and driveway/access crashes. If a state SHSP added emphasis areas to AASHTO's list, then, at a minimum, those same emphasis areas should be included in this task. In addition to a systemwide assignment of crashes into the emphasis areas, subregional and jurisdictional analyses are an important consideration. The systemwide analysis can mask important localized issues—for example, differences between rural and urban locations or between state and local highway systems. Following are some typical sublevel summaries for severe crashes by emphasis area:

• State-maintained highways compared with local agency roads
• Jurisdictional subdivisions (e.g., county, city)
• Rural areas or districts compared to urban/suburban areas or districts
• Areas with unique geography, such as mountainous areas
• Areas with strong seasonal travel patterns, such as tourist destinations

Since the systemwide analysis identifies crash types for an aggregate crash data set, the results are not as sensitive to specific locations as are the results of a subregion analysis in which the crash data set is more focused on a specific geographical area. Comparing the results of the systemwide analysis to subregion analyses highlights those crash types and contributing factors that may be substantial only in a localized area. For example, intersection crashes could be the top crash type in all districts within the state, while pedestrian and bicycle crashes represent a sizable portion of severe crashes only in the urbanized district. In this case, the safety program manager will not identify pedestrian and bicycle crashes as a top emphasis for the state; however, the crash data justifies encouraging the urban district manager to develop a pedestrian and bicycle safety program.

Example 1 illustrates the data-driven selection of focus crash types by the New York State Department of Transportation (NYSDOT). Using crash data from the state's Safety Information Management System (SIMS) database, NYSDOT identified focus crash types by disaggregating the crash data by emphasis area and facility type. This example highlights how roadway systems in different jurisdictions or subregions require different priorities.

Depending on the focus crash type, further data disaggregation may be necessary to more specifically define a systemic safety problem. For example, Thurston County began their systemic planning effort with a focus on lane departure crashes that had been formulated from previous safety-related studies. Example 2 illustrates the disaggregation of the County's lane departure crash data into various data elements available from the crash reports.

Return to top

Task 2: Select Focus Facilities

After selecting the focus crash types, Task 2 answers the question Where are the crashes occurring? A "crash tree" diagram is the recommended approach to answer this question. The crash tree can have a number of different formats, depending on agency capabilities and data availability. One such example is to begin the crash tree with the total number of severe crashes at the highest level. Each subsequent level separates the severe crashes by facility type.

At a minimum, the crash tree analysis should include separation by urban and rural, ownership (state and local), segment and intersection, segment type, and intersection control type. This minimum level of detail allows for the refinement of facility types, which is useful to focus the identification of risk factors (i.e., characteristics associated with the locations where the focus crash types are occurring) and select relevant countermeasures.

Depending on the data set and the scope of the analysis, examples of potential combinations include the following:

• Segment versus intersection
• Segment type, including freeway, multilane, two-lane, and one-way (typical segment divisions)
• Intersection control type, including signalized, unsignalized, and uncontrolled (typical intersection divisions)
• On tangent versus on curve
• High-speed versus low-speed
• Presence of street lighting
• District or region

Examining the crash tree leads to the identification and selection of the facility types where the focus crash types most frequently occur. States with a Roadway Departure or Intersection Safety Implementation Plan may already have crash trees and could use these plans to identify focus facility types for either their roadway departure or intersection crashes. Example 3 shows a crash tree diagram for the New York state roadway system.

If data or resources are not available to complete the remaining tasks and steps of the four-step systemic safety planning process, then an agency can still move forward with selecting and implementing countermeasures to address the focus crash types on the focus facility type. After identifying the focus crash type and facility type, a systematic deployment (deploy everywhere), instead of systemic (deploy at locations with greatest risk), of cost-effective countermeasures is a possibility. While a systematic deployment requires no further analysis to prioritize locations, the approach requires more funding since all locations are improved.

Return to top

Task 3: Identify and Evaluate Risk Factors

Task 3 further defines the facility types selected in the previous task by documenting the most common characteristics of the locations where crashes occurred. For example, if the previous tasks suggested a focus on road departure crashes on rural two-lane segments, then this task might indicate that these crashes were over-represented on roads with poor edges, curvilinear alignment, and volume within a specific Average Annual Daily Traffic (AADT) range. Tasks 1 and 2 relied on data typically available in the crash records system. Task 3 is the first point where road and intersection inventories are likely needed to provide additional levels of detail to support the data analysis.

The first decision to be made for this task is determining which characteristics will be evaluated—in other words, the potential risk factors for the systemic network screening. The initial characteristics are further evaluated before selecting the final characteristics that represent the risk factors. Selecting characteristics for evaluation might be based on several considerations, but it is important that they can indicate greater potential for severe focus crashes to occur. For example, the shoulder surface type might be a risk factor for roadway departure crashes in a curve but not for right angle crashes at intersections. If not certain whether a characteristic might be a suitable risk factor, then first reference safety research before evaluating that risk factor for a specific jurisdiction. Example 4 summarizes the research process the Kentucky Transportation Cabinet (KYTC) conducted to identify potential risk factors.

Return to top

Additionally, the availability of the data element in existing databases, the ability to quickly gather data if not already available, and the applicability to the focus crash type and facility type (e.g., intersection characteristics do not need to be investigated for a systemic lane departure program) influence the selection of characteristics to review as potential risk factors. For data elements in an existing database, concerns about quality or the amount of time since the last update also may influence the decision of whether to evaluate that data element as a potential risk factor. Data that do not reflect current conditions could lead to implementing a countermeasure that is no longer appropriate or the best choice for addressing the focus crash type.

The list of potential risk factors suggests characteristics to be examined for curves, segments, and intersections. These characteristics are based on a review of published research and professional experience, many of which are inputs used by the predictive methods in the AASHTO Highway Safety Manual (HSM), and the degree to which many of these attributes might affect crash frequency is available within the HSM (AASHTO, 2010). The FHWA's Crash Modification Factor (CMF) Clearinghouse is another source for information on the relationship between location characteristics and their potential to contribute to a severe focus crash type (refer to the Appendix for more information about the CMF Clearinghouse). Table 1 shows examples of potential risk factors based on focus crash types.

After potential risk factors have been identified, they are evaluated to determine whether the characteristics exhibit a relationship to future crash potential. Only those that positively demonstrate a relationship should be selected as risk factors. Two approaches to evaluate these relationships are using descriptive statistics and reviewing characteristics using CMFs from published research.

Descriptive statistics. Descriptive statistics are used to compare the proportion of locations where the location characteristics exist with the percentage of severe crashes. For example, consider a scenario where an analysis of an intersection database identifies only 10 percent of rural unsignalized intersections are skewed, yet more than 40 percent of the severe rural unsignalized crashes occurred at skewed intersections. This indicates skew is overrepresented at rural unsignalized locations experiencing a severe crash. By extension, skew would be a good choice for one risk factor when prioritizing locations. Another option is to compare the crash density of locations exhibiting the characteristic with locations without the characteristic. Where the crash density differs between locations with and without the characteristic, then that characteristic could be used to differentiate between locations for prioritization.

Reviewing characteristics using CMFs from published research. The safety effectiveness of design features and traffic characteristics can be used instead of descriptive statistics to select risk factors. For example, intersection skew has been shown to increase crash frequency (CMF greater than 1.0) while intersection lighting has been proven to reduce crashes (CMF less than 1.0). By applying CMFs documented in safety research results, an agency could select intersection skew and the absence of intersection lighting as risk factors in lieu of local crash analysis. Furthermore, combining the findings and CMFs from research with the results of the local descriptive statistics might increase confidence in the final selection of risk factors. For example, if a county agency has a relatively small data set to evaluate risk factors for severe intersection crashes, then the descriptive statistics might not be conclusive for skew and lighting. However, CMFs for skew and lighting might reinforce or confirm the patterns seen in the local descriptive statistics, providing greater confidence that these characteristics are risk factors for a future severe crash.

Example 5 depicts the descriptive statistics application used by NYSDOT to identify curve radii as a risk factor.

As illustrated in Example 6, Thurston County also used descriptive statistics to compare the proportion of severe curve-related roadway departure crashes on various functional classifications of roadways to the proportion of those functional classifications represented on the entire county roadway system.

Return to top

Outcome

Completing the first step of the systemic safety planning process provides an understanding of the focus crash types, focus facilities, and risk factors that represent the core of the developing systemic safety program. This information informs the risk assessment process in the next step of the systemic safety planning process.

Screen and Prioritize Candidate Locations

Objective

The objective of the second step in the systemic safety planning process is to develop a prioritized list of potential locations on the roadway system (e.g., segments, curves, intersections) that could benefit from systemic safety improvement projects. The results of this step answers the question Are all elements of my system equally at risk for the focus crash types or are some more at risk than others?

The process to screen and prioritize candidate locations evaluates specific sites (i.e., individual segments, curves, and intersections) at the micro-level. The tasks in this step involve reviewing all elements from the focus facility types and assigning a level of risk to each element. Risk is inferred from the presence of crashes and risk factors, which are roadway and traffic features commonly associated with the locations where the focus crash types occur.

Data Needs

The screening and prioritization process requires two primary types of data. The first type is site-specific crash information, including severity, type, and any contributing factors relevant to the focus crash types and facility types. The second data type is the basic features of the road system, especially geometric or traffic elements selected as risk factors in the previous step of the systemic safety planning process. This information is available in from variety of sources, including electronic databases of roadway features and traffic characteristics, archives of as-built plans, aerial photography and street views, video logs, traffic control device inventories, and traffic flow maps.

Process

As Figure 5 illustrates, screening and prioritizing candidate locations involves three tasks: selecting the locations or elements of the roadway system to review, verifying selected crash risk factors, and prioritizing these elements based on the presence of the risk factors at each location. In this process, the more risk factors present infers a higher priority as a candidate for safety investment.

FIGURE 5. Systemic Safety Planning Process: Tasks to Screen and Prioritize Candidate Locations

Task 1: Identify Network Elements to Analyze

Task 1 identifies the elements from the focus facility types, which represent the locations where the focus crash types tend to occur. For spot-based applications such as curves and intersections, all relevant locations are identified first. For segment applications, corridors are split into elements with consistent design (i.e., cross-section) to establish homogeneous segments. Additionally, the segment elements should have uniform traffic and design characteristics whenever possible, especially with respect to the selected risk factors. Homogeneous segments allow for consistent application of risk assessment and the same countermeasure to be selected for the entire element. For example, a rural two-lane corridor might be divided into segments extending from the limits of one small rural community to the limits of the next community. The Highway Safety Improvement Program Manual provides additional guidance for identifying network elements (FHWA, 2010).

After identifying the network elements to analyze, assemble the information in a format useful for conducting the risk assessment and documenting the results. Refer to the Minnesota Department of Transportation Case Study for an illustration of how to document identified network elements.

Task 2: Conduct Risk Assessment

In Task 2, agencies assess the risk of the systems and specific locations selected for analysis. The purpose of this assessment is to identify and document roadway and traffic characteristics pertaining to the selected risk factors, as well as total crash history, if applicable, for each location. For purposes of the Systemic Tool, the term "risk assessment" means the effort to collect the roadway/traffic data and to characterize the potential for a severe focus crash to occur at a given location or along a given segment of roadway based on certain characteristics present at these network elements.

The roadway and traffic characteristics could be available in a database format or can be collected as part of a field review of the road system. Roadway and traffic characteristics might also be collected in a virtual environment with field confirmation, as appropriate. Agencies with experience developing systemic approaches have found that roadway features can be documented in the office using video logs, aerial photography, and traffic volume data bases. Also, local agencies can provide some basic information about design features and traffic control devices.

The process to document the presence of a risk factor at a given network element involves recording if the risk factor is present or not. Presence can be indicated with a "1" or an asterisk or some other identifier that can be tabulated. Summing the number of risk factors present provides the ability to compare and prioritize network elements. The MnDOT Case Study at the end of this section provides an example of the documentation produced for a risk assessment.

After identifying and documenting the network elements for the prioritization process, consider if the selected risk factors are feasible for a systemwide analysis. A risk factor is not useful to prioritize locations if it is present at every location on the focus facility. For example, speed limit may need to be removed as a risk factor if it is the same for all roadway segments identified for the screening process. Also, curve radii may not be an appropriate risk factor if every curve on the focus facility type has the same radius. After considering, remove or add risk factors if necessary. This revision may be necessary if risk factors were selected using CMFs or a sample of locations on the focus facility. The MnDOT Case Study contains an example of the determination of risk factor feasibility.

Task 3: Prioritize Focus Facility Elements

Task 3 generates the prioritized lists of segments, horizontal curves, and intersections based on the presence of the selected risk factors—the more risk factors present, the greater the potential for the focus crash type and the higher the priority as a candidate for safety investment. Systemic analysis approaches might weigh risk factors equally, which simply means the more risk factors present, the higher the location's priority. However, risk factors also can be given relative weights. In this case, particular risk factors—if found to have a stronger association with locations where severe focus crashes occurred—can be given larger weights. The values for relative weights may be high/medium/low or based on integers that infer a higher level of confidence in the weights. Example 7 illustrates how Thurston County weighted their selected risk factors based on their level of confidence in the data representing that risk factor.

Agencies can adjust the thresholds (the number of risk factors present or a weighted value) used to identify high-priority candidates based on the level of funding anticipated for implementation. To set the threshold, agencies should identify more candidate locations than can be implemented in a given year. There are two reasons for this. First, the effort creates a backlog or shelf life for the analysis so that it does not have to be completed annually. Second, locations might drop off the list as a result of field conditions or where additional time or funding is needed to implement the selected countermeasure. Having identified more candidates than can be implemented in one year provides flexibility to ensure all available funding can be allocated to priority projects. The MnDOT Case Study contains an example illustration of how risk assessment results are useful to establish a threshold for candidate locations.

The final result of Task 3 is a prioritized list of focus facility elements that represents the locations with the greatest potential for the focus crash type to occur. At a minimum, the list includes information about the element location and the risk assessment score. Example 8 shows the eight horizontal curves on the Thurston County Rural Major Collector system that have the highest risk scores and, therefore, represent the highest priority locations for systemic safety improvements. This list shows the value of systemic safety planning—more than half of these locations with risk factors present do not have a documented severe crash. However, the risk assessment indicates these locations have common characteristics with locations where severe crashes have occurred and, therefore, have the potential for a future severe crash.

Outcome

The outcome of this step is a risk assessment and ranking of the focus facility elements. The elements of the focus facilities carried forward to the next phase represent the locations with greatest potential for safety improvement.

Return to top

Select Countermeasures

Objective

The objective of the third step in the systemic safety planning process is to assemble a small number of low-cost, highly effective countermeasures to be considered for project development at candidate locations. The countermeasures are selected based on research related to the systemwide data analysis results from the first step (Identify Focus Crash Types and Risk Factors), and specific jurisdictional experience, policies, and practices regarding potential countermeasures.

Data Needs

The National Cooperative Highway Research Program (NCHRP) Report 500 series is a suggested source of information for assembling the initial list of safety countermeasures (TRB, Various). These 24 volumes provide an extensive list of countermeasures for each of AASHTO's safety emphasis areas plus they provide insight about implementation costs. Recent countermeasure effectiveness information is available in the HSM and the CMF Clearinghouse. The National Highway Safety Traffic Administration's (NHTSA's) Countermeasures That Work report provides effectiveness information about behavioral countermeasures (NHTSA, 2013). These resources provide information about the expected effectiveness of the countermeasures, and documentation of implementation and maintenance costs. Other information needed to support countermeasure selection is gathered from the crash, roadway, and traffic data assembled in the previous two steps.

The information documented for the countermeasures might indicate whether the countermeasure is proven or tried, the expected safety effectiveness of the countermeasure, estimated implementation and maintenance costs, and consistency with agency policies and practices. In addition to assisting with countermeasure selection, this information is useful to the safety planning process because it provides:

• Documentation of proven, effective, or tried countermeasures which support a level of confidence about forecasted crash reductions.
• Effectiveness measures which assist in computing an expected crash reduction associated with deployment across a system.
• Implementation costs which are critical in the safety investment decision-making process.
• A final check of deployment feasibility in relation to agency policies and practices.

FIGURE 6. Systemic Safety Planning Process: Tasks to Select Countermeasures

Process

Countermeasure selection involves assembling a comprehensive list of potential countermeasures and evaluating each countermeasure to determine a select few most appropriate for systemic deployment. Figure 6 shows the three primary tasks conducted to define the safety countermeasures for implementation at the candidate locations.

Task 1: Assemble Comprehensive List of Countermeasures

Task 1 involves assembling a comprehensive list of the safety countermeasures associated with each focus crash type. The list is assembled after reviewing the latest research and other available information (e.g., agency experience or engineering judgment) to identify those countermeasures with the greatest potential to address the focus crash types. As previously mentioned, the NCHRP Report 500 series, HSM, and CMF Clearinghouse are suggested starting points for the research effort. A state SHSP or other local safety plans also might include countermeasures already defined for particular focus crash types. Other resources, such as FHWA's illustrated guide sheets for 77 intersection countermeasures, may be used for developing the initial list of countermeasures (FHWA, Office of Safety). An agency's existing maintenance program could also be a source for identifying potential countermeasures. For example, a countermeasure could be to replace missing or damaged traffic signs with signs that meet current retroreflectivity standards. The MnDOT Case Study provides an illustration of the process to assemble a comprehensive list of countermeasures.

Task 2: Evaluate and Screen Countermeasures

The second task in the countermeasure selection process is to evaluate and screen the initial list of countermeasures based on documented safety effectiveness at reducing the focus crashes, implementation and maintenance costs, and consistency with the agency's policies, practices, and experiences. Implementing countermeasures that are proven to be effective at reducing crashes provides agencies the highest possible level of confidence that their investment will have a positive outcome—a reduction in the focus crashes. The HSM and the CMF Clearinghouse provide the most current information about countermeasure effectiveness. Evaluation results from previous countermeasure implementation efforts (discussed in Element 3 of the Systemic Safety Project Selection Tool) offer additional critical information for evaluating and selecting future countermeasures. The key is to follow the evidence and select the countermeasures with the greatest potential to address the focus crash type not just at a single location, but across the prioritized focus facility elements.

The screening process should consider maintenance requirements for the countermeasures because a countermeasure cannot perpetually provide the potential for crash reduction if it is not maintained or replaced when necessary. Thus, agencies should consider their ability to perform and fund routine maintenance of countermeasures implemented through a systemic deployment program. A countermeasure is likely not going to be a viable improvement and should be removed during the screening process if it will not be maintained. The MnDOT Case Study provides an illustration of how to evaluate and screen countermeasures.

Agencies can also screen potential countermeasures by performing a benefit-cost analysis of various countermeasures applicable for a typical section that represents the focus facility type. A spreadsheet tool is useful to automate the benefit-cost ratio calculations. Example 9 illustrates the benefit-cost analysis used to identify countermeasures for roadway departure crashes on bridges in Salem County, New Jersey. When performed in this manner, a benefit-cost analysis can help to determine the most cost-effective way to spend the funding allocated to systemic improvements.

Task 3: Select Countermeasures for Deployment

The result of the countermeasure screening process is the selection of one or more countermeasures for each focus crash type. This short list would include primarily low-cost countermeasures, along with a few higher-cost countermeasures. Agencies use this list of countermeasures to develop safety projects at specific locations across their roadway system. Although an agency might choose to deploy one particular countermeasure, there might also be locations where that countermeasure is not an ideal choice. As a result, having alternatives to choose from provides flexibility and acknowledges that all candidate locations in a system are not the same. Additionally, including alternative countermeasures offers an opportunity to incorporate a few higher-cost countermeasures proven to significantly reduce crashes. Agencies should use these higher-cost countermeasures in a limited number of circumstances and only in situations that might not be effectively addressed by the low-cost countermeasures in the short list. The high-priority locations identified in the second step (Screen and Prioritize Candidate Locations) also might influence the final countermeasure selection, which is why it is important to identify a range of countermeasures appropriate for the range of high-priority sites selected for project development. The final selection represents the highest-priority countermeasures that agencies will maintain for safety investment, based on expected crash reductions and estimated implementation costs. The MnDOT Case Study provides an illustration of how to select countermeasures for deployment.

Outcomes

An outcome of this step is documentation of the countermeasures, which will be used in future program evaluation. The primary outcome of this step is a short list composed mainly of effective, low-cost countermeasures and a few higher-cost countermeasures for each focus crash type. These countermeasures then become the focus of the safety project development efforts that are described in the next step of the systemic safety planning process.

Prioritize Projects

Objective

The objective of the fourth and final step in the systemic safety planning process is to identify and develop a list of high-priority safety improvement projects for implementation. This process considers the prioritized at-risk locations identified in the second step (Screen and Prioritize Candidate Locations) and applies the most appropriate countermeasures from the list developed in the third step (Select Countermeasures) to develop a list of high-priority safety improvement projects. For the purposes of the Systemic Tool, the term projects includes dedicated safety projects as well as safety improvements implemented as one component of a traditional construction or maintenance project (e.g., resurfacing) or as part of routine maintenance efforts.

Data Needs

The information needed to support the project prioritization process includes a basic understanding of an agency's priorities, practices, and policies as they relate to project and program development. This understanding is especially important to ensure the countermeasures identified in this step are only included in projects for areas where they fit with local practices for installation and maintenance. In addition, defining specific safety projects requires current information about countermeasure implementation costs and estimated effectiveness.

Process

Prioritizing projects involves developing a decision process so that selected countermeasures can be consistently assigned to projects for the prioritized focus facility elements identified in the second step. This fourth and final step results in a prioritized implementation order for safety projects. Figure 7 shows the three basic tasks conducted to prioritize systemic safety projects.

FIGURE 7. Systemic Safety Planning Process: Tasks to Prioritize Projects

Task 1: Create Decision Process for Countermeasure Selection

Creating a decision process for each focus crash type provides a means to consistently assign countermeasures to focus facility locations. A decision process is simply a set of criteria that considers issues such as volume, environment, adjacent land use, or cross-section, and uses them to identify the appropriate countermeasure for high-priority locations. An important distinction of the systemic safety planning process is that the decision-making process does not just identify the most appropriate countermeasure for each location (like when addressing hot spots), it also considers multiple locations with similar crash and risk characteristics to select preferred one or more countermeasures suitable and affordable for widespread implementation. Alternative countermeasures, and the criteria for deploying each countermeasure, might consider typical variations in site conditions.

For example, urban intersection and rural road departure systemic safety programs address uniquely different problems. Therefore, each program needs countermeasures specific to the problem and different decision processes for selecting the final countermeasure. Additionally, programs addressing similar issues on different facility types (e.g., road departure on rural divided expressways versus two-lane highways) might have some of the same countermeasures, but the criteria and thresholds for selecting the countermeasures might be completely different. An agency can develop a process recognizing that no single countermeasure is likely to be best for every location due to the variety of features and characteristics that are encountered along a roadway system. The MnDOT Case Study illustrates a decision tree created to select countermeasures.

Task 2: Develop Safety Projects

The next task in the project prioritization process is to apply the decision process to identify one or more specific countermeasure for each candidate locations selected for safety investment. Safety projects are developed by providing a detailed site description (e.g., route number, mile point, intersecting roadway, and segment termini), identifying the specific countermeasure selected, estimating the implementation cost, and summarizing how the site scored with the risk factors. After the countermeasures for safety investments are selected, agencies can then decide how to most efficiently bundle projects into a design package for contract letting. The MnDOT Case Study shows an example of the documentation produced during the process to develop a safety project.

Task 3: Prioritize Safety Project Implementation

Lastly, projects are prioritized for implementation by considering factors such as funding, other programmed projects, time to develop project plans, expected crash reduction, amount of public outreach needed, and environmental and right-of-way constraints. The prioritized list of locations infers a particular order based on the number of risk factors present at a given location. However, other factors often influence the ability to let a specific project. For example, a project might need to be included in an established capital improvement program or coordinated with other projects/programs. The next chapter presents a case study that illustrates how MnDOT applied the systemic safety planning process to develop prioritized projects for deployment on the county roadway network throughout the state.

Outcome

The outcome for this step is development of a safety project for each identified at-risk candidate location along a roadway system. The roll-up of the prioritized projects produces the systemic safety program for the city, county, district, region, or state. This final step completes the systemic safety planning process, which is Element 1 of the Systemic Safety Project Selection Tool.

Case Study: Minnesota Department of Transportation's Application of the Systemic Safety Planning Process

The following case study illustrates how the Minnesota DOT applied the systemic safety planning process as part of a statewide effort to develop county road safety plans for each of their 87 counties. The focus crash types were severe lane departure crashes on rural segments and curves, severe right-angle crashes at rural thru-stop and urban signalized intersections, and severe pedestrian crashes in urban areas. The FHWA's HSIP Noteworthy Practice Series provides information about this effort (FHWA, Office of Safety). The FHWA's Developing Safety Plans: A Manual for Local Rural Road Owners provides guidance for developing local road safety plans (FHWA, 2012). The case study provides illustrations of each successive task of the systemic safety planning process, using the same data set throughout. Each illustration describes the purpose, provides a description, and typically shows example data and analysis results.

Identify Focus Crash Types and Risk Factors – Task 1: Select Focus Crash Types – Crash Data for All Public Roads

Note: Numbers are not additive, as one crash may involve a young driver at an intersection. The numbers represent severe crashes (Fatal and A-type Injury crashes).

Source: Department of Public Safety Crash Data Records, 2006 to 2010

Purpose

Identify the high-priority emphasis area – categories of severe crashes that represent the greatest opportunities for reduction.

Description

• Demonstrates disaggregation of Minnesota's statewide crash data into AASHTO's designated emphasis areas.
• In this case, the data suggests Minnesota's high-priority emphasis areas (focus crash types) include:
  • Drivers
    • Seatbelt usage
    • Impaired driving
    • Young drivers
  • Highways
    • Intersections
    • Road departure
• The data also supports the importance of actively including the county highway system in statewide safety planning efforts.

Identify Focus Crash Types and Risk Factors – Task 2: Select Focus Facilities – Crash Tree for All Public Roads

Purpose

Identify where crash types most frequently occur.

Description

• This is a crash tree analysis focusing on the county highway system in Minnesota.
• Analysis shows there are more severe crashes on the county system than the state system, and the majority of these crashes are rural, involve road departure, and occur on curves more than 50 percent of the time.
• Rural intersection crashes primarily occur at thru-stop controlled intersections. The most common crash type is a right-angle crash.
• Urban intersection crashes primarily occur at signalized intersections. The most common crash type is a right-angle crash.
• This chart provided a reasonable preview of the likely outcome of the safety analysis – the safety focus of Minnesota's County Roadway Safety Plans has been rural, road departure, curves, and angle crashes at intersections.

Identify Focus Crash Types and Risk Factors – Task 3: Identify and Evaluate Risk Factors – Risk Factor Identification

Road Features

• Shoulder Width/Type
• Horizontal Curvature
• Access Density
• Roadside Rating
• Intersection Skew

Traffic Volume

• Average Daily Traffic (ADT)

Other Features

• Presence of Commercial Development
• Proximity to Rail Crossing
• Distance from Previous Stop
• Operating Speed

Data screened by:

• Agency Experience
• Availability

Purpose

Identify roadway characteristics to use as an initial set of potential risk factors to be further evaluated for use in systemic network screening.

Description

• MnDOT reviewed published research to identify roadway features strongly related to crash experience. MnDOT compared these findings to available data and identified potential crash risk factors to apply to:
  • Two-lane rural county highways
  • Rural horizontal curves
  • Rural thru-stop intersections
  • Urban signalized intersections – focus on angle and pedestrian crashes
• The potential risk factors for each facility type selected by MnDOT are shown in the bottom table.

Identify Focus Crash Types and Risk Factors – Task 3: Identify and Evaluate Risk Factors – Evaluation of Segment Traffic Volume as a Potential Risk Factor

Road Departure Crashes

Purpose

Use descriptive statistics to evaluate Average Daily Traffic (ADT) as a potential risk factor.

Description

600-1,200 ADT was selected as a risk factor because severe crashes were overrepresented on roadways with ADTs in this range.

Identify Focus Crash Types and Risk Factors – Task 3: Identify and Evaluate Risk Factors – Evaluation of Access Density as a Potential Risk Factor

Crash Rate for Minnesota State Highways Based on Access Density

Crash Rate for Rural Minnesota County Roadways Based on Access Density

Purpose

Evaluate access density as a potential risk factor.

Description

• Illustrates macro analysis showing that as access density increases, crash rates increase.
• Shows significant relationship between density and crash rate – risk factor selected for systemic network screening.

Identify Focus Crash Types and Risk Factors – Task 3: Identify and Evaluate Risk Factors – Evaluation of Edge Risk Assessment as a Potential Risk Factor

1 – Good Shoulder, Good Clear Zone

2 – Good Shoulder, No Clear Zone

2 – No Shoulder, Good Clear Zone

3 – No Shoulder, No Clear Zone

Identify Focus Crash Types and Risk Factors – Task 3: Identify and Evaluate Risk Factors – Evaluation of Edge Risk Assessment as a Potential Risk Factor

Purpose

Use descriptive statistics to evaluate edge risk assessment as a potential risk factor.

Description

• Severe road departure density is the number of crashes per mile per year, which illustrates the frequency of these crashes.
• The crash rate is the number of crashes normalized by vehicle miles traveled, indicating the relative risk based on exposure.
• Shows that a roadside condition of no usable shoulder and the presence of fixed objects (Risk Rating = 3) generates a higher severe road departure rate.

Screen and Prioritize Candidate Locations – Task 1: Identify Network Elements to Analyze

This table represents 19 of the 77 segments.

Purpose

Identify the network elements from the focus facility types which represent the locations where the focus crash types tend to occur. The elements are for use in network screening.

Description

• For the application of the Systemic Safety Planning Process on corridors, the Minnesota process split corridors into segments with consistent design features (e.g., cross section or volume) to minimize the variation of risk factors within a corridor and so that the same countermeasure could be applied to the entire segment.
• The end points of the segments also considered practical issues about how projects could be deployed. Segments typically ended at the edge of cities since the preferred countermeasure, shoulder and edgeline rumble strips, could not be deployed within cities.
• The table on this page illustrates the results of segmentation of Minnesota 2-lane county roads.

Screen and Prioritize Candidate Locations – Task 2: Conduct Risk Assessment

This table represents 19 of the 77 segments.

Purpose

Evaluate the risk factors of the systems and locations selected for analysis using roadway and traffic characteristics in order to rank/prioritize at-risk locations.

Description

• Risk factors are not weighted.
• A star indicates the corresponding risk factor is present.
• More stars identify locations as higher priority candidate for safety investment.

Screen and Prioritize Candidate Locations – Task 2: Conduct Risk Assessment – Determine Feasibility of Selected Risk Factors

Greater Minnesota Segment Risk Factors

Purpose

Assess selected risk factors to determine their feasibility for differentiating between elements (curves, segments, and intersections) on the focus facility type.

Description

• The line representing the percent of focus facility locations with the risk factor present indicates that the risk factors generally appear in 20 to 30 percent of focus facility type segments.
• MnDOT concluded these percentages are high enough to distinguish between segments, but not so high that most or all segments had the risk factor present (making it difficult to distinguish between segments for prioritization purposes.)
• Additionally, the bars representing segments with the risk factors present show a higher severe lane departure crash density than segments without the risk factor.

Screen and Prioritize Candidate Locations – Task 3: Prioritize Focus Facility Elements – Threshold for Selecting Candidate Locations

Purpose

Summarize the locations based on the number of risk factors present to establish a threshold for selecting candidate locations.

Description

The bars indicate the percent of the system that has the number of risk factors present (i.e., stars received). Additionally the chart shows that the severe crash density and severe lane departure crash density increases as more risk factors are present. MnDOT used this chart to determine that those locations with three or more stars present, approximately 30 percent of the focus facility type, would be candidate locations for systemic safety projects.

Select Countermeasures

Purpose

Review NCHRP 500 Reports, Highway Safety Manual, CMF Clearinghouse, FHWA Office of Safety website, and state Strategic Highway Safety Plans to identify a comprehensive list of potential countermeasures for each focus crash type.

Description

• The key is to review the latest research and other available information to identify those countermeasures with the greatest potential to address the focus crash types.
• The countermeasures will be screened using:
  • Crash data,
  • Effectiveness,
  • Cost, and
  • Agency policies, procedures, and experience.
• The end result will comprise a short list of countermeasures for each focus crash type to use in development projects.

Select Countermeasures – Task 1: Assemble Comprehensive List of Countermeasures – List of Potential Countermeasures

Source: NCHRP 500 Volume 6 (2003)

Purpose

Assemble an initial comprehensive list of countermeasures associated with focus crash type.

Description

Initial list of countermeasures MnDOT identified to reduce road departure crashes using the NCHRP 500 Volume 6 report.

Select Countermeasures – Task 2: Evaluate/Screen Countermeasures – Countermeasure Effectiveness

Proven

• Graduated drivers licensing
• Safety belt enforcement campaigns
• DWI checkpoints
• Street lights at rural intersections
• Access management
• Roadside safety initiatives
• Pave/widen shoulders
• Roundabouts
• Exclusive left turn signal phasing
• Shoulder rumble strips
• Improved roadway alignment
• Cable median barrier
• Removing unwarranted traffic signals
• Removing trees in hazardous locations
• Pedestrian crosswalks, sidewalk, and refuge islands
• Left turn lanes on urban arterials

Tried

• Rumble strips (on the approach to intersections)
• Neighborhood traffic control (traffic calming)
• Overhead red/yellow flashers
• Increased levels of intersection traffic control
• Indirect left turn treatments
• Restricting turning maneuvers
• Pedestrian signals
• Improve traffic control devices on minor intersection approaches

Experimental

• Turn and bypass lanes at rural intersections
• Dynamic warning devices at horizontal curves
• Static/dynamic gap assistance devices
• Delineating trees in hazardous locations
• Marked pedestrian crosswalks at unsignalized intersections

Source: NCHRP 500 Reports

Purpose

Evaluate and screen the initial list of countermeasures based on documented effectiveness of reducing the focus crash types and consistency with the agency's policies, practices, and experiences.

Description

MnDOT used the NCHRP 500 reports to determine an initial estimate of effectiveness with their classification of each countermeasure as Proven (effective), Tried, or Experimental to narrow down the initial list of countermeasures.

Select Countermeasures – Task 3: Select Countermeasures for Deployment – Countermeasures Selected for Deployment

Source: NCHRP 500 Volume 6 (2003)

Purpose

Identify and select a few countermeasures for each focus crash type based on the evaluation of the countermeasures and consideration of agency priorities, practices, and policies.

Description

MnDOT paired down the list of countermeasures identified in Task 1 to address road departure crashes based on the evaluation. These countermeasures were selected to develop safety projects.

Prioritize Projects – Task 1: Create a Decision Process for Countermeasure Selection – Decision Tree

Purpose

Develop a decision process to facilitate consistency in the selection of countermeasures.

Description

• Adapted from Minnesota County Road Safety Plan Project to represent a typical decision process for selecting roadway departure countermeasures on county roads.
• Utilizes Average Daily Traffic (ADT), lane width, and shoulder surface type to determine project type for rural 2-lane segments.

Key Point

Provides county engineer with opportunity to identify treatment preferences while selecting projects that MnDOT is comfortable funding.

Prioritize Projects – Task 2: Develop Safety Projects – Safety Projects Selected for Development

County Highway 26 from State Highway 23 to County Highway 15

Purpose

Apply decision process to develop specific safety projects for each candidate site selected for safety investment.

Description

• County Highway 26 Segment Project Form (left), provides information on:
  • Roadway Data – Average Daily Traffic (ADT), length, lane width, shoulder width/type
  • Crash Data – total and road departure crashes, density, rate
  • Deficiencies – risk ranking
  • Countermeasures Considered
  • Selected Countermeasure
  • Project Cost
• The table on this page provides regional summaries of the projects identified using the systemic process.

Prioritize Projects – Task 3: Prioritize Safety Project Implementation – Project Prioritization Process

Purpose

Identify the order in which projects will be implemented.

Description

Prioritizing the projects for implementation will start with the number of risk factors present at a given location. Several other factors may play a role in the selection process, such as funding availability, other programmed projects, time to develop project plans amount of public outreach needed, etc. For example, a lower priority project may be selected over a higher priority project if the location is in an established capital improvement program.

Summary

Element 1 of the Systemic Safety Project Selection Tool presents a general process for conducting systemic safety planning efforts, which is intended to complement an agency's site-specific analysis process—not replace it.

This systemic safety planning process involves four steps:

1. Identifying the problem by looking at systemwide data;
2. Screening and prioritizing candidate locations across the network;
3. Selecting relevant mitigating countermeasures most appropriate for broad implementation across those locations; and
4. Prioritizing projects for implementation.

The systemic safety planning process is a data-driven process that identifies the types of crashes which represent the greatest opportunity for substantially reducing the number of severe crashes, the locations that have the greatest potential for severe crashes but do not have a high crash history, and a short list of highly effective countermeasures. The systemic safety planning process is flexible. Any agency can adapt the tool to fit their needs, and this flexibility allows agencies to develop a process that leads to implementation within the constraints of their data availability, program requirements, staff capabilities, and funding. While there can be variation in the detail and time invested or specific actions conducted for each task, the ultimate goal is to direct low-cost, effective countermeasures where the greatest opportunities exist to prevent severe crashes and improve safety.