Improving Pedestrian and Motorist Safety Along Light Rail Alignments (2009)

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.

66 This chapter develops recommendations designed to facili- tate the compilation of accident data in a coordinated and con- sistent manner across LRT systems in the United States. The goal is to generate data that permits meaningful comparisons of accident rates across transportation modes, and more rigorous assessment of the effectiveness of safety measures, devices, and practices implemented on LRT alignments. The chapter pro- vides an overview of existing data collection experience and the various report forms used by transit agencies. Additional detail regarding the accident data collection process is provided in Appendix E. It is essential to have detailed and comprehensive informa- tion to accurately identify problems, rates, and trends in LRT accident data and to develop corrective actions and remedial measures. Accident data should be collected and stored in a consistent format that facilitates electronic data management and analysis. Diagrams of crossing conditions and character- istics depicting roadway/LRT geometry where collisions occur, and collision diagrams illustrating crash experience (“condition diagrams” and “collision diagrams,” respectively, in MUTCD terminology), can be useful in pinpointing prob- lems. The information contained in these diagrams is not, however, directly transferrable to the numeric format of elec- tronic databases and thus cannot be used in the statistical analysis of large sets of collisions. It is desirable to have consis- tent accident reporting formats across transit agencies. Transit Agency Data The research team reviewed the incident reporting forms used by 12 LRT transit agencies. The incident report forms used by these transit agencies are included in Appendix E. Agencies with Multiple Accident Report Forms Most agencies provided only one incident report form, but three agencies provided more than one. These agencies were the TTC, SEPTA, and SF MUNI. The TTC provided three separate incident reports: Occur- rence Report, Surface Supervisory Occurrence Report 185L, and Assessment and Summary Disposition Report. • The Occurrence Report and the Surface Supervisory Occur- rence Report contained generally very similar informa- tion. However the Occurrence Report focused on collecting information only available at the scene such as the type of impact (i.e., sideswipe), the precise location of injured pas- sengers, and whether or not each vehicle sounded its horn or had its headlights on before impact. The Surface Supervisory Report focused on reporting damage and factors contributing to the incident, such as detailed descriptions of the location and extent of damage, the specific actions of all drivers and pedestrians involved, classification of the severity of the incident and any injuries, and judgments (yes, no, or unknown) as to whether or not each environmental factor contributed to the occurrence. • The Assessment and Summary Disposition Report focused on identifying any evidence that could implicate the TTC operator or the other motorist/pedestrian in legal respon- sibility for the incident. SEPTA provided three separate incident report forms: Operator’s Accident Incident Report, the Supervisor’s Acci- dent Investigation Report, and the SEPTA Public & Opera- tional Safety Division Incident Report. • As with the TTC, the data collected on the Operator’s Acci- dent Form and the Supervisor’s Accident form were nearly identical. The Operator’s Accident Form included descrip- tions of the specific actions of all parties involved in the inci- dent, while the Supervisor’s Accident Investigation Form contained additional information classifying the incident and indicating the presence and types of safety features and traffic control devices at the scene. • The Public and Operational Safety Division Incident Report was a summary level report compiled using the previous two C H A P T E R 7 Improving the Accident Data Collection Process

accident reports in conjunction with other available infor- mation sources. A list of potential sources is provided with the report and includes: Operator Report, Supervisor Report, Interviews, Photographs, Vehicle Inspection, Event Recorder Log, CD-ROM, Field Notes, Sketch, Chain of Custody, Evidence, Control Centre Report, Police Report, D&A Report, Radio/Telephone Tapes, and Infrastructure Inspection. SF MUNI also provided three incident report forms: The Employee Form, The Supervisor Form, and The Safety Form. • The Employee Form focused on reporting information rel- evant to loss prevention and contained less detail regard- ing the environmental conditions or actions of persons involved. • The Supervisor Form focused on determining whether the condition of the driver contributed to the incident (includ- ing drugs/alcohol), what emergency services attended the scene, and recommendations as to what training might be necessary for the driver to prevent another occurrence in the future. • The Safety Form was clearly designed to facilitate reporting to the NTD. Many of the categories of data included in the form and their answers were taken directly from the NTD Reportable Incident Report Form. This form focused on including all the fields required by the NTD, such as classi- fication of incident, ROW type, intersection controls, spe- cific actions of drivers, etc. Incident versus Accident Report Forms Many of the report forms contained information pertaining to the collision of a transit vehicle with either another vehicle, a pedestrian, or a fixed object. Other report forms also included data related to other types of incidents, such as criminal activ- ity on transit property, passenger illness, etc. From the per- spective of safety analysis, it would be ideal to keep these types of incidents completely separate from collision reports. As indicated in the Collision Data Available, Requested, and Received section in Chapter 3, failure to do so often leads to incorrect reporting of incidents, resulting in the need to under- take significant data cleaning before databases can be used for analysis. Categories of Information Included in Accident Reports The accident report forms reviewed by the research team contained several categories of information that were common to many of the forms. These included: incident classification, location, weather, illumination, road/rail conditions, action of 67 driver(s), safety equipment, damage, injuries/fatalities, witness information, and emergency services present. In addition, many forms contained sections for collecting data pertaining to pedestrian/passenger incidents. As mentioned in the Loca- tion section, a few of the incident report forms also contained information regarding incidents that were not related to trans- portation safety. Table 34 shows the common categories of data reported in the hardcopy and electronic incident report forms supplied by the LRT agencies. The table also shows the data reporting for- mat used for each category of data. The letters used for the data reporting format (C, T, etc.) are explained (checkbox, text field, etc.) in the table’s footnote. Incident Classification Most incident report forms required the investigator to pro- vide a classification of the incident. Generally, incidents were classified based on type of incident and/or severity. Forms intended for use in the investigation of a wide variety of inci- dents included an extensive list of possible incidents, most of which were not collisions. Incident reporting forms generally classified collisions based on the object or individual that col- lided with the transit vehicle. Location The most common method of identifying the location of collisions was to provide a text field on the incident report form. The limited amount of space dedicated to location on most of the forms suggested that most agencies expect a min- imal description of the incident location. Some of the forms required both the street being travelled and the nearest cross- street to be reported. Most forms also included details per- taining to the transit agency such as the run number, route number, switch number, and line/branch number. Some of the forms included more specific details regarding the incident location. The SF MUNI Safety Form reported the exact latitude and longitude of the incident location. A few forms further classified the location using a series of check- boxes. For example, the LACMTA form required the location of the rail vehicle to be classified as being on the mainline, shop, yard, or other location. The location of the other vehi- cle could be identified from the location of the person involved in the incident, which included the categories ROW, grade crossing, tunnel, or yard. The TTC Supervisor’s Accident Investigation Form pro- vided the most detailed list of descriptors to classify the inci- dent location, including whether the location was at an intersection, midblock, loop, garage, terminal, near side stop, far side stop, island, and/or curb. If the incident occurred at a bay or stop, the investigator could further indicate whether the

Table 34. LRT incident report forms: categories of data included and data format. Agency/Report Name Incident Classification Location Weather Illumination Road/Rail Conditions Actions of Driver(s) Safety Equipment (Vehicle/ROW) Damage Injuries/ Fatalities Hardcopy Forms LACMTA C C/T T C C T A/T Santa Clara Valley TA C T C C C C/T C C RTD Denver T C T C/T Memphis Area TA T T C/T C/T/D C/T Portland Tri-Met C C/T C C C C/T/D C/T C/T/D C/T SEPTA Supervisor' s Accident Investigation Form C T C C C C/D C C/T C/T Operator Accident Incident Report C T C C C C/D T C/T St. Louis RT T T T T T Toronto Transit Commission Occurrence Report C T/D C C C C/T/D C T C/T Surface Supervisory Occurrence Report C C/T/D C C C C/T/D C C/T C/T Edmonton Transit System C T T T T T/D T T C/T City of Calgary Transi t C T C C C C/T/D T/D Electronic Forms SF MUNI Supervisor Form T/P T/P T/P P T T/P T/P T/P Safety Form C/T/P T/P P P T/P T/P P T/Tb Employee Form T/P T/P P P T/P T/P Utah TRAX T P P P P P P/T/D P/T/Tb Note: C – Checkboxes, T – Text field, D – Diagram, A – Alphanumeric Code, P – Pull-down Menus, Tb – Table. A blank indicates that the category was not included in the form examined. Source: review of all referenced forms and reports received from LRT agencies

vehicle was entering, exiting, or dwelling at the bay or stop. Providing additional details regarding the location of the inci- dent can give individuals unfamiliar with the location informa- tion pertinent to the determination of whether or not location contributed to the collision. This level of detail and categoriza- tion would be especially useful in the analysis of records at a national level. Weather Most of the incident forms required the investigator to report the weather conditions at the time of the incident. There was a high degree of consistency in the format of the weather reporting. Most agencies provided the investigator with a series of checkboxes from which he/she could select the appropriate response. As would be expected, the responses available were determined by the climate of each location. Northern locations, for example, tended to provide detailed responses for winter weather conditions. Illumination Most of the incident report forms included information about lighting conditions at the time of the incident. In almost all cases, the incident form provided the investigator with a series of checkboxes from which to select the appropriate response. Although virtually all agencies contained identical responses for environmental lighting conditions (i.e., daylight, dark, dawn/dusk), certain agencies also allowed the investiga- tor to indicate whether glare or street lighting were present. These added details provide a more comprehensive picture of the lighting conditions at the time of the incident, and can help the analyst more accurately determine whether illumina- tion was a factor contributing to the incident. Road/Rail Conditions The condition of the roadway or railway at the time of the incident was also included in most incident report forms. Almost all the forms that recorded road/rail conditions pro- vided a series of checkboxes from which the investigator could select the appropriate response. Responses focused on envi- ronmental conditions that might reduce surface friction and contribute to a collision (i.e., leaves, water, ice, etc.). The Toronto Transit Commission Occurrence Reports also required the investigator to indicate whether any of the fol- lowing road conditions were applicable: asphalt or concrete, gravel or other, upgrade, downgrade, construction, straight, curve. The inclusion of this information on the incident reporting form provides a more comprehensive picture of how design and environmental road/rail conditions might contribute to collisions. 69 Action of Driver(s) The actions of the drivers involved in the incident were reported to various degrees on the incident report forms. The most commonly reported driver actions were: direction and speed of travel; use of head lights, tail lights, or horns; and maneuvers being executed at time of impact. In many of the incident forms, diagrams were an important source of infor- mation regarding driver action. The level of detail provided was not consistent among transit agencies. For example, some of the incident forms only required the direction of travel of the vehicles, with potential for additional explanation in the statements and/or diagrams. In contrast, forms such as the TTC Occurrence Report reported speed prior to impact, type of impact (i.e., sideswipe, head-on, etc.), distance travelled after impact, headlights (on/off), and horn sounded (yes/no), in addition to providing room for diagrams and statements. An important point to note is that the source of information was not identified on many of the incident forms. This can be critical information when considering the validity of informa- tion about driver actions. Safety Equipment The presence of safety equipment was inconsistently reported across transit agencies, and often omitted entirely. The safety equipment information contained on the incident report forms included traffic controls at intersection; aspect of signals at time of crash; type and condition of switch; visi- bility and functionality of traffic signs/signals; presence, visi- bility and functionality of grade crossing devices; type of traffic lines; and presence, type, and indication of transit sig- nal. Most incident forms included only a few of the above listed factors. Damage The accurate estimation of damage at the scene of a colli- sion can be a difficult task. The incident report forms gen- erally used one or more of the following three methods to indicate the extent of damage: diagrams, classification of dam- age, and description of damage. There were significant varia- tions among transit agencies in the methods employed to report damage. UTA staff noted during the workshop that if the fault lies with the operator of the motor vehicle and not the LRT operator (which is almost always the case), the agency never receives formal reports of total damage on the auto or on the LRV. This made estimating damage a difficult and imprecise task. Diagrams typically required the investigator to shade or mark the areas of damage on the transit vehicle and/or other vehicle. Classification of damage was based either on a qual-

itative description (minor/moderate/major) or quantitative estimate (greater or less than a dollar value). Description of damage relied on the investigator to provide a meaningful description of the damage to each vehicle. The methods of reporting damage were generally inconsis- tent across transit agencies. Although there were some agencies that combined the above techniques in an effort to create a more comprehensive view of the damage, many agencies relied solely on a rough dollar value estimate or arbitrary classifica- tion of damage. In addition, the incident reporting forms gen- erally did not provide any guidance as to how the user should decide between various classifications of damage. Finally, dam- age reported in either diagrams or text descriptions would likely be very difficult to translate into an electronic database in a format suitable for analysis. Injuries/Fatalities In general, the incident report forms focused on reporting the number of injuries, the classification of each injured per- son (e.g., transit agency employee, passenger, pedestrian, etc.), the extent of each injury, and how each injured person was transported from the scene. A combination of checkboxes and text fields was usually used to record the relevant infor- mation. The incident report forms also included sections where the contact information of each injured individual could be reported. The method of reporting the number and severity of injuries used on the Utah TRAX Supervisor’s Accident/ Incident Report Form seemed particularly useful. While the majority of forms relied solely on a description of injuries provided by the investigator in a text field, the TRAX from also provided a table with three rows of injury classification: Class A (bruising, abrasions, minor to moderate bleeding, sprains, and strains), Class B (unconsciousness, fractures, severe bleeding), and Class C (death, paralysis, and dismem- berment). Each vehicle involved in the incident was assigned a column in the table, and the user was required to indicate the number of individuals in each vehicle whose injuries fell under each category. This method of reporting injuries appears useful to concisely convey most of the information relevant to the transit agency, while providing the user with some concrete guidance on how to report the extent of injuries. Contact Information All the incident report forms included sections where the contact information of all individuals involved in the incident was reported. Almost all the forms collected contact informa- tion for all drivers, vehicle owners, witnesses, and emergency personnel present at the scene. The contact information of all 70 employees involved in the investigation or reporting process was also collected. Existing Accident Reporting Standards The desire to promote uniformity and comparability of accident data and statistics across agencies and levels of gov- ernment has led to the publication of a number of accident reporting guidelines that include: • American National Standard Manual on Classification of Motor Vehicle Traffic Accidents (ANSI D16) (7th ed., Amer- ican National Standards Institute, ANSI D16.1-2007, 2007) • Data Element Dictionary for Traffic Record Systems (ANSI D20) (American Association of Motor Vehicle Adminis- trators, ANSI D20-2003, April 2003) • Model Minimum Uniform Crash Criteria (MMUCC) Guide- line (3rd ed., 2008 http://www.mmucc.us/, accessed Aug 28, 2008) Supervisory Agency Data As discussed in Chapter 3, there are other agencies at the state and national levels that receive and compile data reported by local LRT agencies, and the reporting process is outlined below. SSO Agency Data SSOs fill multiple roles. They collect data to forward to the FTA, but their larger role is to oversee accident investigations and to undertake corrective action. All SSOs follow the same basic reporting process following an incident. Transit agencies are required to notify their SSO of an incident (which may or may not be a crash) over a certain severity threshold within two hours of the incident occurring. The SSO then proceeds with a more formal safety review. The SSO may conduct an investiga- tion directly, or the transit agency may conduct the investiga- tion and then report it to the SSO. If warranted, the SSO formulates a corrective action plan. The SSO submits all data to the FTA in an annual report. All SSO agencies interviewed for this study expressed interest in a consistent standard for accident data collection. FTA/NTD Data The National Transit Database is “the Federal Transit Administration’s primary national database for statistics on the transit industry” (National Transit Database Federal Tran- sit Administration 2008 Safety and Security Reporting Manual). Transit agencies are required to report all safety and security incidents to the NTD using two forms. The Safety and Secu-

rity Monthly Summary Form (S&S-50) is a monthly summary of the number of safety and security events that resulted in an arrest/citation but did not meet the criteria of a “reportable incident.” Transit agencies are also required to submit data pertaining to a reportable incident using the Reportable Inci- dent Report Form (S&S-40). This form must be submitted within 30 days of the incident occurrence. Redundancies in Data Collection As noted in the preceding section, agencies are responsible for collecting and reporting data twice. They collect and send data to the SSO (which reviews the data and may develop cor- rective action plans), and they collect and send data to the NTD through NTD reporting procedures. Agencies that operate on a shared alignment with heavy rail also send in-depth internal records and/or report incidents to the FRA. Local LRT agen- cies are tasked with the role of reporting the same information to multiple agencies, but all the local agencies visited reported that they rarely receive any feedback or results from the data they report. A Potential National Standard Accident Data Collection Procedure This section outlines a potential method to facilitate the compilation of transit accident data in a coordinated and consistent manner across the United States, with a focus on collecting data that can be used to assess the impact of safety treatments. Two sources provide the data required for safety analysis: 1. The local LRT agency investigation of a collision, coupled with an associated report from police, necessary for assess- ment of causal factors. 2. The local authority responsible for traffic data manage- ment, which could provide exposure data such as road and pedestrian traffic volumes, necessary to establish rates. Since both of these sources are at the local level, all data col- lection and entry should take place locally and electronically to minimize data manipulation and transcription problems later. The database format and structure should be designed to easily permit all data users (for whatever purpose) to extract the information they need with a minimum of manual effort. This will increase efficiency by reducing the redundancy of reporting differently to the various agencies involved at local, state, and national levels. It will also increase accuracy by avoiding subjective decisions and second-guessing of evidence (e.g., during transcription of one report into another sys- tem) that weakens the resulting data and any conclusions drawn later. 71 Standardized Electronic LRT Incident Reporting To ensure that data relevant to assessing the safety of transit systems and impact of safety treatments are available, it is critical that these data should be consistently collected during the incident reporting process. Since incident report forms are the primary source of information used by transit agencies in incident investigation, site-specific conditions such as the right- of-way classification and the presence or absence of traffic con- trols, signage, and other safety devices should be consistently reported in the incident report forms used by transit agencies. Develop and implement a standardized, comprehensive, electronic LRT incident reporting form. This form should be developed and implemented nationally to ensure that the same data are collected in all cases and to ensure that the same data are accessible to all who need them. The form should pro- vide useful prompts and other support for the user entering data to improve clarity and reduce effort to a minimum. Structure of Reporting Forms The structure and format of incident report forms is also critical to ensuring that all relevant data are collected, easily transferred into an electronic database, and easily analyzed. Checkboxes and pull-down menus should be used for data fields, and the number of possible or relevant responses should be limited, avoiding free-form input where practical. The responses can be supplemented with text descriptions that can be used to report unusual circumstances. The text descriptions can also be used to verify data accuracy. For example, report forms may include a comprehensive list of safety treatments and require the investigator to indicate if each is present or absent, followed by a text field where the investigator can describe any relevant features not included on the list. This for- mat will not only provide the investigator with a sense of what information is relevant to collect, but will also result in data that is more useful for analysis. The responses in the checkboxes and pull-down menus should conform to national or industry standards where applicable (e.g., right-of-way classification by alignment type as developed as part of TCRP Report 17 and described in Chapter 2). Sources such as the Model Minimum Uniform Crash Criteria (MMUCC) can be extremely useful in provid- ing transit agencies with guidance as to what categories of information are critical to include in incident reporting. The use of data reporting methods such as hand-drawn sketches that cannot be used for queries in electronic databases should be minimized, and additional categorized data fields should be included to ensure that relevant information in sketches is also provided in a more usable format.

Provide a reporting form structure that can be easily trans- ferred into a searchable electronic database. Utilize check- boxes and pull-down menus with a limited number of relevant responses, supplemented by text fields where necessary. Record Exposure Data The information required to calculate and conduct mean- ingful comparisons of LRT collision rates should also be reported. Such information includes traffic exposure informa- tion such as the number of vehicles crossing at a level crossing, the number of passengers boarding/alighting at a specific platform/station, and the number of pedestrians crossing the tracks per segment per unit time. As this information will not be available to the investigator at the scene of the collision, the information will have to be obtained from the relevant source within the local agency and will need to be linked to the inci- dent report data after the on-scene data collection. Ensure that LRT collision reports include fields for storing geometric and traffic exposure measures that apply to the specific incident site. These data may be sourced from local municipalities or traffic agencies, and the database system storing the data should record the data source because it will not likely be the same as that of the rest of the collision report. LRT Crossing Database As LRT collisions are relatively rare events, there may never be enough LRT-related collisions at one location to develop a statistically significant before-and-after study. To have enough collisions to determine the safety effect of a treatment, researchers could compare the number of incidents at physi- cally similar locations throughout one or more LRT systems. To undertake such analyses, researchers need to have geomet- ric and traffic control data for each crossing and information about any special safety measures or treatments that have been applied. It is not enough to have this information only for locations where there are incidents. An analysis database should include all crossing locations to determine the impacts of different measures over the “do nothing” option in each of the alignment types. The FRA maintains a similar database for heavy rail cross- ings, but the database for LRT needs to be significantly larger as it needs to include information about crossings and segments between crossings where a vehicle or pedestrian may be able to enter the alignment. As compiling and updating the database is a significant undertaking and beyond the resources of most LRT agencies, a national program appears to be necessary. To support analyses of LRT safety, an LRT crossing data- base could be created. This database could include, for each 72 alignment location (crossing or segment), details of geom- etry, control devices, and traffic exposure. The database should be updated as alignments and traffic volumes change (perhaps on a five-year cycle and after major construction). Summary of Information to Be Included in the LRT Collision Database and Reporting Form Based on the analysis of the local, NTD, and SSO incident databases, the project team compiled a list of potential fields for an LRT collision reporting form and accompanying data- base. Consider including the following fields on the LRT collision reporting form when it is developed: • Incident classification: – Clear collision/incident distinction – Type of object or individual involved – Type of impact • Location: exact location (cross referenced with separate LRT crossing database) • Location detail: entering, exiting or dwelling in crossing or stop • Weather conditions • Illumination: daylight, dark, dawn/dusk, glare, streetlight- ing (and condition of repair) • Road/rail conditions: dry, leaves, water, ice, asphalt/ concrete/gravel, upgrade/downgrade/level, construction, straight/curve • Action of driver(s): – Direction and speed of travel – Use of head lights, tail lights, horn – Maneuver being executed – Source of information: driver, investigator, police, witnesses • Safety equipment: – Crossing type – Traffic controls – Transit signal presence, type, indication of transit signal – Aspect of signals – Type and condition of switch – Visibility and functionality of traffic signs/signals – Visibility and functionality of grade crossing devices (i.e., gates) – Presence and type of other safety treatments • Property damage: a repair cost estimate for all vehicles involved, but with an option to provide a simpler classi- fication of damage with clear definitions (i.e., repairable/ irreparable damage to LRV, repairable/irreparable damage to private vehicle)

73 • Injuries/fatalities: – Number of injuries of each class (predetermined classes of severity) in each vehicle. • Contact information: for drivers, vehicle owners, witnes- ses, emergency personnel, agency employees involved in investigation • Exposure (to be completed following contact with the appropriate agency): – Annual average daily traffic volume, number of pedes- trians boarding/alighting, number of pedestrians per segment per unit time – Source of exposure data