Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
47 As acknowledged previously, the steps in this guidebook are not always followed consecu- tively, and there is ample room for reflection and returning to earlier steps in an iterative way. This may be particularly true for Step 6, as the process of refining a treatment plan may require revisiting the actions, assumptions, and decisions made in prior steps. In general, the process to refine and implement a systemic treatment plan will involve the following activities. They are as follows: â¢ Consider additional community priorities. â¢ Perform additional diagnostics. â¢ Perform economic assessments. â¢ Allocate funding and implement projects. The method of execution of these activities can vary widely based on procedures developed for project selection, inter-agency coordination, and others. There are ample resources available describing how to perform these activities in general, and links to key resources are provided at the end of this section. Thus, the following guidance will focus on key systemic safety consid- erations to make during these activities, which may differ in some ways from traditional safety management practices. It is important to note, however, that at the time this guidebook was produced, few agencies had completed a systemic pedestrian safety process to this extent. Best practices surrounding prioritization and implementation of systemic pedestrian safety projects are still in the early stages. As more agencies go through the process, it will be essential to evaluate systemic projects and systemic programs in terms of both processes and outcomes. The next step in this guidebook outlines evaluation considerations and activities. Consider Additional Community Priorities Step 4 provided guidance on identifying, and to some degree prioritizing, sites for treatment based on risk, using several different methods. The results from that step are dependent on the quality and completeness of the risk data available, including how well pedestrian exposure to traffic is accounted for and the robustness of the analysis and screening and ranking procedures used. During this step in the process, there should be an opportunity to reflect on the results from Step 4 and consider additional information from outside the systemic process, as well as engage key stakeholders. For example, it may be important to consider additional community priorities or opportunities, such as geographic balance, equity, installation, and maintenance issues, or other considerations, especially if resource constraints limit the opportunity to treat all similarly ranked and economically viable locations. C H A P T E R 7 Step 6: Refine and Implement Treatment Plan
48 Systemic Pedestrian Safety Analysis NCHRP Research Report 803: Pedestrian and Bicycle Transportation Along Existing Roads, and with the TRB Webinar titled ActiveTrans Priority Tool (Lagerwey et al. 2015), may be useful resources to think through to complete the final prioritization steps in a systemic process. The tool could be used to supplement the initial risk-based ranking and to consider additional risks or roadway constraints, as well as factors such as equity, compliance with existing guidance, or others. Once these determinations are made, weights may be assigned to factors to complete the prioritization. The result will be a well-documented process that may also be used iteratively and help identify areas where data improvements or other changes in the process are desirable. Perform Additional Diagnostics It may also be important to gather additional data needed to make final decisions. No general safety management process can cover all situations that may be encountered, including risks that remain unmeasured and unidentified in prior studies or in analyses using local data. Also, no guide is a substitute for experience and a thorough diagnosis of conditions before finalizing a treatment package. Thus, it remains important to conduct diagnoses and field visits such as road safety audits to identify potentially unobserved safety issues that should be considered, to verify those risks that seem indicated by the data, and to help ensure that the systemic counter- measures considered for application are appropriate and will not have unintended adverse con- sequences on other types of crashes or users. Perform Economic Assessments All agencies are likely to be interested in maximizing cost-effectiveness of systemic projects. Since systemic treatment sites are identified because they have high risk characteristics but may not yet have experienced crashes, this may make them less suited for a traditional benefitâcost analysis, which relies heavily on crash estimates, unless SPF data are available. Since safety perfor- mance functions aim to predict expected crashes in the absence of treatment, they provide a useful tool and metric in the absence of crash data or to complement crash data. To determine expected benefits from treatments, it is also necessary to have quantifiable information on expected crash effects of treatments, or crash modification factors, to apply to expected or predicted crashes. Since systemic countermeasures are typically lower cost, expected crash benefits per site may be lower and still achieve cost-effectiveness overall, especially if a longer time frame is considered. In a traditional benefitâcost analysis, values are typically assigned to crashes, lives, or injuries as well as to the treatments to perform the assessment. The data types needed include costs of countermeasures, estimates of expected crashes without treatment, crash reductions expected from the countermeasures to be applied, and crash or injury cost estimates (many states have their own), which may include various estimates by crash severity of the economic value of a statistical life (Hauer 2011). Other considerations in an economic assessment include the number of treated locations, expected life span of each treatment, and determinations of the Noteworthy Practice Equity was an important consideration for the City of Seattle. The City chose to stratify their initial list of locations by geographic area and then prioritize locations for treatment within each area, to ensure there was strong geographic distribution of their systemic treatments.
Step 6: Refine and Implement Treatment Plan 49 numbers of years to be used in estimating potential benefits. The CMF Clearinghouse guidance also describes how to apply costs and normalize them to the current or investment year in a traditional benefitâcost analysis. As mentioned in Chapter 3, there is a need to be able to estimate where crashes may occur to perform an economic assessment and to determine how crashes might change with treat- ment. This presents a challenge in a systemic process that might identify many locations with no observed prior crashes. To address this need, agencies have used two basic approaches: â¢ Bundled high risk sites with similar high crash sites for benefitâcost or cost-effectiveness assessment and treatment (see Case Example 3); or â¢ Metrics derived from SPFs to estimate potential crash savings (see Case Example 2). An advantage of developing SPFs to determine risk factors is that the models can also be used to generate SPF-predicted crashes and empirical Bayes estimated crashes. SPF-predicted crashes represent the average risk for locations having the set of factors in the model. Empirical Bayes estimates are weighted by crash history, which helps to account for crash factors that may not have been included in the model or that are specific to a given site. In the absence of crashes, SPF-predicted crashes may be the best estimate of crash potential for a given location. The higher of empirical Bayes estimated or SPF-predicted crashes could be used in prioritization. Considering the difficulty in assigning costs to human lives and injuries and value judgments implied (Hauer 2011), some agencies such as those with Vision Zero programs may prefer a cost-effectiveness approach. The cost-effectiveness approach does not assign values to lives and injuries but simply attempts to estimate which projects will provide the largest expected reduc- tion in crashes or injuries per expenditure. Noteworthy Practice A lack of crash frequency often excludes segments and corridors with few prior crashes, but high crash potential, from being selected using a crash-based benefitâcost analysis. To account for this, after corridors for potential treatment were identified for improvements through an initial risk-based screening and ranking process, Oregon DOT predicted combined vehicleâpedestrian and vehicleâbike crashes using the available safety performance functions and the predictive method in Part C of the HSM. For prioritization, the number of predicted crashes were compared with the number of observed prior crashes, and the higher value was used in the cost-effectiveness analysis. See Case Example 2 for details. Definition: Cost-Effectiveness Index The cost-effectiveness index (CEI) estimates the cost to reduce one vehicleâpedestrian crash. It is calculated using the equation below: CEI project cost expectedreductioninpedestrian crashes =
50 Systemic Pedestrian Safety Analysis Table 18 offers a hypothetical example for how to apply a cost-effectiveness index to a selec- tion of 26 potential treatment sites identified in Step 4, using model-derived SPFs and established CMF and countermeasure cost data. Site 1 represents the site with the highest predicted crash risk based on the SPF model, as indicated by the relatively high number of predicted crashes (A); inversely, Site 3 represents the lowest risk site. The values in the second column are derived from an SPF model created for the City of Seattle to predict crashes involving pedestrians and motor vehicles traveling straight at midblock locations and are provided here for illustration purposes only. These SPF values have been multiplied by the expected countermeasure life span (in this case assumed to be 10 years). The countermeasure options (B) show three countermeasures rec- ommended for midblock crossing locations that have established CMFs (C) and costs (D) (see table footnotes). The values for (E) are the expected number of crashes that would be prevented over the assumed 10-year period if the site was treated with the countermeasure. Dividing the cost of the treatment (D) by this number produces the cost-effectiveness index or the cost (in thousands) to reduce one crash. The CEI value can be used to explore the cost-effectiveness of different treatment sce- narios. For example, there is high cost-effectiveness (as demonstrated by the low CEI value Site Predicted No. of Crashes (A) Countermeasure Options (B) CMF (C)a Countermeasure Cost per Site (D)b Expected Crash Reduction if Treated (E) = A â (A x C) Cost- Effectiveness Index (in thousands) (F) = D/E 1 3.6 High visibility crosswalk 0.63 $2,540 1.33 2 Median island 0.69 $13,520 1.12 12 High visibility crosswalk and median island 0.44c $16,060 2.02 8 PHB 0.53 $57,680 1.69 34 2 1.36 High visibility crosswalk 0.63 $2,540 0.50 5 Median island 0.69 $13,520 0.42 32 High visibility crosswalk and median island 0.44 $16,060 0.76 21 PHB 0.53 $57,680 0.64 90 3 0.45 High visibility crosswalk 0.63 $2,540 0.17 15 Median island 0.69 $13,520 0.14 97 High visibility crosswalk and median island 0.44 $16,060 0.25 64 PHB 0.53 $57,680 0.21 273 Other 23 sites 18.51 High visibility crosswalk 0.63 $2,540 6.85 9 Median island 0.69 $13,520 5.74 54 High visibility crosswalk and median island 0.44 $16,060 10.37 36 PHB 0.53 $57,680 8.70 152 All 26 sites 23.98 High visibility crosswalk 0.63 $2,540 8.87 7 Median island 0.69 $13,520 7.43 47 High visibility crosswalk and median island 0.44 $16,060 13.43 31 PHB 0.53 $57,680 11.27 133 aCMFs provided here are the highest (most conservative) value from Table 17 in Step 5; see the technical report for details. bAverage cost estimates are from Costs for Pedestrian and Bicyclist Infrastructure Improvements, Appendix D table on pages 42â44, available at http://www.pedbikeinfo.org/cms/downloads/Countermeasure%20Costs_Report_Nov2013.pdf. cAssuming multiplicative effects on crashes (0.63 x 0.69) for high visibility crosswalk and MI (multiply CMFs). Table 18. Example cost-effectiveness analysis for different treatment scenarios.
Step 6: Refine and Implement Treatment Plan 51 of 2) in treating Site 1 with a high visibility crosswalk but similar cost-effectiveness (CEI value of 7) in treating all 26 sites with high visibility crosswalk markings. Since there is still some uncertainty about which sites will actually experience future crashes (in the absence of treat- ments), the more sites that can be treated the more opportunity to reduce crashes that may occur at some sites. On the other hand, there is much lower cost-effectiveness in treating only the lowest predicted risk site (Site 3) in this group with a more expensive PHB treatment. Similarly, it may be more cost-effective to treat more higher-risk sites with multiple counter- measures (to address more types of risk) than to treat lower ranking sites. This is illustrated by the estimates for combining high visibility crosswalk markings and median islands. Other considerations (as described in Chapter 6) may certainly come into play, but this example illustrates one method in current use by a state DOT to take cost-effectiveness into account (see Case Example 2). The CEI example is one way to approach cost-effectiveness without monetizing the cost of a crash, injury, or fatality. It can easily be extended to a benefitâcost analysis by applying crash cost estimates to the predicted prevented crashes and comparing these with the costs of treatment. For an extended example of a more traditional benefitâcost analysis approach, see pages 41â50 of Arizona DOTâs 2017 Pedestrian Safety Action Plan, section on benefitâ cost evaluation, available at http://www.azbikeped.org/downloads/ADOT-Pedestrian-Safety- Action-Plan.pdf. Regardless of which method is used, it is important in a systemic approach to perform the economic assessment for all locations anticipated to be treated with a similar package of treat- ments. The sum of âpredictedâ crashes or combined expected crashes and risk estimates across many locations can help to justify the treatment cost and to determine the balance of invest- ment and number of sites for different types of systemic projects, as well as with high crashâ high cost treatments. CMF Caveats in Cost-EffectivenessâBenefit Analyses A unique challenge for pedestrian safety is that there are fewer pedestrian CMFs available for use in estimating benefits of treatments. In an ideal world, high-quality CMFs would be available for each pedestrian treatment and would be calibrated using local data to pro- vide precise estimates of potential effectiveness. However, there may be insufficient data to calibrate crash effect estimates of pedestrian crash type countermeasures in a given jurisdic- tion. Furthermore, there is little research on synergistic or multiplicative effects of applying multiple countermeasures to a site. Thus, the estimates available in the guidebook (Table 17 in Chapter 6) and from the CMF Clearinghouse and other sources may serve as the best avail- able estimates. Other Effectiveness Assessments Since crash effect estimates may be lacking or not appropriate for the locations considered, agencies could also consider estimates of behavioral or operational effects such as documented speed reductions to estimate crash effects. For example, the HSM provides CMF estimates for changes in baseline speeds on total fatal or injury crashes; additional assumptions are that effects on pedestrian crashes may be similar. These crash effect estimates, adjusted for local experience or conditions, can aid agencies that wish to perform a benefitâcost analysis. Other resources that detail methods and metrics for performing economic assessments with different data limitations are listed in the Additional Resources section.
52 Systemic Pedestrian Safety Analysis Allocate Funding State and local agencies may need to change or adapt existing policies to allocate funds to systemic safety projects. For some states, this may require an adjustment in how projects are prioritized (see section on considering additional community priorities) to allocate safety funding to sites based on risk if, for example, crash prediction estimates are unavailable and only crash histories are available to estimate future crashes. Additionally, agencies may wish to balance expenditures across high crash or hot spot treat- ment projects and systemic safety implementation projects, particularly if data to drive confi- dence in one process over another are lacking. Compared with the traditional crash hot spot approach, in which economic decisions may be made on a per-site basis until funding limits are reached, systemic applications may be decided based on expected crash savings across a group of sites, including those with low or no prior crashes but potential crashes based on crash predictions or estimates or other risk assessment. While expected crash savings may be low at many individually treated systemic sites, investment risk is also typically low (for lower cost, but effective treatments). By treating many locations with similar risk, a positive benefitâcost ratio can be achieved. Conversely, treating a few more sites with high cost treatments may work against a positive benefitâcost ratio as the numbers of expected crashes or injuries decrease on a per-site basis. Figure 5, reproduced from Figure 12 in Reliability of Safety Management Methods: Systemic Safety Programs, illustrates a framework for how agencies might optimize the balance of fund- ing between the high crash and systemic treatment approaches (Gross et al. 2016b). A simi- lar approach may be used in assessing allocation of funding by enforcement and educational options if funds may be transferred among program types. (See the highlighted example from Caltrans.) Some assumptions may have to be made when applying this approach; for example, if there is uncertainty about crash effects of treatments. The FHWAâs guide (Gross et al. 2016b) also mentions that there is a need to test this framework for allocating funds under real-world conditions and scenarios. Figure 5. Optimizing investment in high-cost and low-cost (systemic) improvements (From Reliability of Safety Management Methods: Systemic Safety Programs, Figure 12, Gross et al. 2016b.)
Step 6: Refine and Implement Treatment Plan 53 Additional Resources Additional implementation considerations are provided in Chapter 5, Section 5.3, and Chapter 6 of the technical report. The following resources offer guidance on ways to prioritize treatments and evaluate different scenarios. FHWAâs Systemic Safety Project Selection Tool https://safety.fhwa.dot.gov/systemic/fhwasa13019/ NCHRPâs ActiveTrans Prioritization Tool and NCHRP Report 803: Pedestrian and Bicycle Transportation Along Existing RoadsâActiveTrans Priority Tool Guidebook http://www.pedbikeinfo.org/planning/tools_apt.cfm http://www.pedbikeinfo.org/pdf/PlanDesign_Tools_APT_ Guidebook.pdf FHWAâs CMFs in Practice, Introduction to Safety Performance Functions https://safety.fhwa.dot.gov/tools/crf/resources/cmfs/ pullsheet_spf.cfm FHWAâs CMFs in Practice, Quantifying Safety in the Roadway Safety Management Process https://safety.fhwa.dot.gov/tools/crf/resources/cmfs/ AASHTOâs Highway Safety Manual http://www.highwaysafetymanual.org/Pages/default.aspx FHWAâs Reliability of Safety Management Methods: https://safety.fhwa.dot.gov/rsdp/downloads/ Systemic Safety Programs fhwasa16041.pdf FHWAâs Guidebook for Developing Pedestrian and Bicycle Performance Measures https://www.fhwa.dot.gov/environment/bicycle_pedestrian/ publications/performance_measures_guidebook/ Resource Link Noteworthy Practice There are several examples of how states and cities are allocating funding toward systemic safety programs. Caltrans set aside $10 million from the Highway Safety Improvement Program (HSIP) and established the Systemic Safety Analysis Report program. The state funding for the Systemic Safety Analysis Report program was made available by exchanging the locally available HSIP funds for state highway account funds. Caltrans has issued two rounds of calls for all types of systemic safety project funding applications by local agencies. Oregon DOTâs All Roads Transportation Safety program splits its available funding (primarily HSIP funds) evenly between two project prioritization types: hot spot and systemic analyses. Funding for systemic projects is further disseminated into three emphasis areas identified by Oregonâs Strategic Highway Safety Plan, which include roadway departure, intersection, and pedestrianâbicycle safety improvements. Together, these three emphasis areas account for approximately 90% of the fatal and injury crashes in Oregon. Seattle is pursuing HSIP funds based on crash-based screening and empirical Bayes estimates. According to Seattle DOT staff, Washington State DOT is allowing prediction models to be used to justify treating sites that have not yet experienced crashes. The Seattle crash prediction models were based on crash data from previous years, but a recent review of more current crash data showed that some sites predicted to have crashes based on the model did in fact experience a pedestrian crash later (Chris Svolopoulus, Seattle DOT, personal communication, 2018).