Implementation Process: Appendix E

Issues Associated with Use of Safety Information in Highway Design Role of Safety in Decision Making

Safety is often not given adequate consideration in making road planning, design, operation and maintenance decisions, since safety-related data are not usually of sufficient quality because :

• Appropriate information can not be readily produced
• Appropriate analytical tools are not available to designers

Designers wish to properly address safety in their work. However, the problems they face in obtaining and using safety data result in safety information not being given proper weight in decision making. This is illustrated by Exhibit 1.

Designers tend to weigh those factors in which they have the greatest confidence. These include estimates of construction costs and quantified right-of-way impacts. Public opinions regarding alternative decisions also tend to carry weight in decision-making. The designer often considers information about crashes — their location, environmental factors, causes, etc., as incomplete, missing, or clearly erroneous. Typically, they have neither the time nor resources to correct the problems, and so must proceed with information that is of lesser quality than other factors being considered. As a result, designers tend to lack confidence in the ability to predict or quantify safety effects of alternatives. In this manner, safety inevitably becomes a lesser or secondary factor in decision-making. Specific issues are described in the sections below.

The difficulties that result from having poor data will affect more than design decisions. This material, however, focuses upon highway design. It is adapted from a report prepared for the NCHRP (Adapted from Ronald Pfefer, Timothy Neuman and Richard Raub, Improved Safety Information to Support Highway Design, Final Report, National Cooperative Highway Research Program Project 17-12, 1997.)

The purpose of this supplementary section is to outline the limitations that safety analysts have due to poor data, and to provide some direction for those who would wish to begin an effort to change this. This is not something that can be achieved in the short term, but it is something so important that the effort should be started immediately.

Safety Data Collection Shortcomings

Problems exist with respect to the police crash report, the type and quality of data maintained that describe the highway environment, and dynamic data such as traffic volume and speed. Data shortcomings can be attributed in part to lack of a focus, or priority, for safety, when the agencies responsible for data collection establish procedures and protocols.

The primary source of crash information is the police crash report (PCR). In many cases, such reports do not exist. Where they do, the information which the designer believes to be most important is missing, especially the location of the crash. Because the reporting police officer has primary responsibility to secure the scene, see to the treatment of the injured, and minimize resulting congestion, he or she will not likely take the time to properly locate the points of impact, or precisely record other roadway related elements.

With respect to the roadway, much data of critical importance to safety typically are not collected and maintained by highway agencies. For example, the character of the roadside is acknowledged to be among the most important factors affecting safety on two-lane rural highways, While most states maintain a video log of a portion of their system, few states maintain quantitative records of roadside features, or conditions, sufficient for use in decision making. Most agencies also fail to inventory geometric features such as horizontal alignment, vertical alignment, and intersections, all elements acknowledged by current thinking to be related to crash frequency and severity.

Similarly, although traffic data are maintained by most agencies, it is generally not of sufficient quality to support safety decisions. Traffic volume data for intersection movements generally are not available. Adequate information about vehicle speeds often is lacking. The frequency with which traffic volume data are updated may be sufficient for, pavement management systems, for example, but not for characterizing use in safety analyses.

The needs to improve aspects of data management include establishment of linkages between databases, making use of information not traditionally accessed by designers, and maintenance of history files.

Potentially-valuable data are maintained by various agencies for purposes other than those related to highway safety. These are often not accessible, and hence not utilized. Examples include maintenance records (guardrail replacement for example), pavement and bridge management systems, trauma records, and citizen complaints to police or highway agencies. Such data could offer insights to designers or safety analysts, but are typically not linked to safety data systems.

Most agencies do not maintain long-term records of the safety history of a highway. Most also do not maintain records of historical changes in geometric conditions, traffic volumes, or other related conditions for specific locations. Such data, maintained over a long time, and accessible to designers, would provide support to evaluations of the effectiveness of site-specific improvements, and thus would enhance decision making for future design decisions.

Furthermore, the analyst might not be able to generate information to make a connection between historical crash patterns and related conditions (for which data are available in the roadway inventory) because the location reference is often not precise enough, or is missing. Even if the linkage were possible, achieving it would be very time consuming, and records often do not contain adequate information about longitudinal geometrics. If the user were to evaluate a change in geometry at a site, a time-series analysis would probably be appropriate, requiring 10 years of data. However, a history of the roadway inventory might not be maintained, so the user cannot be sure if and when related changes were made to the highway. Finally, the designer may not have available traffic or other data that would be of value in considering design changes.

These are just a few of the difficulties faced when an attempt is made to use highway safety information to support highway design, and other highway safety activities. However, many agencies are, or already have, overcome many of these problems. There are various technological and non-technological steps to be taken to improve the quality of safety data.

Potential Resolution of Safety Information Issues

Advances in technology in both data collection and management offer many cost-effective opportunities to overcome current issues related to the collection and management of safety data. Sensitivity to designers’ safety-related data needs in data-collection programs is part of the solution. Organizational strategies are recommended, designed to change from "business as usual," and upgrade the importance of safety information.

Despite recent technological advances, few agencies employ anywhere near the full range of technologies that could optimize their use of safety data. Technologies can be applied to data collection, management, and communications, including sharing databases, user interfaces, analytical tools, and decision support. With the advent of computer-aided engineering (CAE), highway agencies now have the ability to build and maintain detailed inventories of their system by direct reference to computer files of as-designed roadways. Many states now use portable computers, global positioning equipment, instrumented vehicles, and geographic information system-based data as their means of assisting with data collection. Some are using optical reading of forms, and have created distributed databases. Linkage of crash reports and medical data and use of modeling tools or decision support systems are also done on a limited basis. Overall, agencies still are not taking full advantage of the powerful computer-aided tools now available.

Application of technologies offers the opportunity for agencies to improve the quality of data with less effort or cost than currently being expended. Traffic volume data gathering, for example, can take advantage of computerized signal systems in place that already monitor intersection turning movements, vehicle classification, and speeds. Additionally, current devices for measurement of these data have become low cost. Such data, captured and made accessible to a safety information system, could result in agencies no longer having to conduct expensive manual or other counting programs that yield less data.

Need For a Design Decision-support System

Given improved data quality and data management through strategies such as effective use of the latest technologies, the current highway design process still would lack a comprehensive means for incorporating safety information. Improved information may only leave the designer with a larger dilemma of how to deal with it. The results of NCHRP Project 17-12 included recommendations for the development of a design decision-support system for safety (DDSS), which has the general objective of assisting the designer in making design decisions at each point in the process.

The recommended framework for a DDSS provides for several elements including user interface, decision-support module, information presentation module, and design analysis module. This is shown diagrammatically in Exhibit 2.

The user interface should be as intuitive as possible for both its end-users and managers. End-users (designers) primarily interact with the system through the decision-support module. This module provides an "intelligent" core to the system. The information-presentation module contains the basic software needed to display information that is derived either directly from the data stores, or generated from the design-analysis module. Finally, the design-analysis module will provide an arsenal of tools for analysis of existing and proposed designs. It would be able to draw upon the best analytical devices available at any given time.

The types of analytical tools discussed to this point are designed to facilitate identification of the underlying factors contributing to highway crashes. The tools are intended to allow the designer to investigate the data to the level of detail desired, through performing a series of inquiries, each of which can "drill down" further into the details available in the data set. In addition to these traditional analytical tools, the module will also deliver to the designer results of analyses using methods potentially based upon safety audit procedures. Safety audits may be used as a vehicle for summarizing and assessing the results from other design-analysis tools available in this set. There also will be a series of analytical devices to test and assess a proposed design. Finally, analytical tools are being developed, such as those being created by FHWA for the Interactive Highway Safety Design Model (IHSDM; http://www.tfhrc.gov/safety/ihsdm/vdmweb.htm). IHSDM tools will include, but not be limited to, vehicle-driver dynamics simulation, human factors-based design analysis procedures, models for predicting operational and safety outcomes of alternative designs, and economic analysis tools. In addition there is a new initiative of the Transportation Research Board, to produce a Highway Safety Manual, which will include methods for estimating the safety effects of various elements of the road.

The DDSS is structured to assist the designer at the various stages of the process. It will be capable of generating information at various levels of detail from the most aggregate to the most disaggregate. The report demonstrates the "drill down" concept by diagramming and describing processes used in identifying locations considered especially hazardous. Implementation of the proposed system over time should produce substantial benefits to the highway design community and others. Examples are given in Exhibit 3.

Exhibit 2. Generalized Concept for a Design Decision-Support System

Exhibit 3. Examples of Impact on Safety Analyses for a Design Project with Implementation of the DDSS Concept
Current Practice	Becomes	Improved Practice Using the New System
Extensive manual assembly of safety data tabulations from different offices within and outside the design agency, requiring many hundreds of staff hours		Direct access to a comprehensive data warehouse, within a CADD/GIS environment
Manual inspection of hard copy of crash reports from microfilm record, to correct coding errors		Pre-screened data minimizes errors including proper location of crash; plus immediate access to digital images of crash reports within CADD/GIS environment
Labor-intensive translation of tabulations of key data into calculated fields and graphics using independent spreadsheets and graphics software into which data must be manually keyed		Integrated statistical routines, business graphics and advanced data-visualization software; including intelligent collision diagrams plotted on actual geometry and highway attribute time lines
Lack of historical traffic data and history of site geometry		Data warehouse with records of all available traffic counts, and a continual, up-to-date, history of physical attributes of the site
Lack of safety models and limited flexibility of data systems to provide support for positions taken regarding the hazards at the site and the effectiveness of proposed improvements		A suite of design analysis models to assess existing and proposed designs from a variety of perspectives, making it feasible to produce a documented and defendable estimate of safety impacts of alternatives.
Support for decisions regarding data to use, analysis of it, conclusions to be drawn, and arriving at recommended improvements, comes from staff or specialists.		A central decision-support function within the system to provide the user with guidance at every step of the analysis.
Decision tools which are applied are done manually		A suite of decision tools using the latest developments in decision science, are directly accessible with information generated by the system, based upon the previous work of the designer.

• Addresses site-specific improvements based on site-specific needs, not general programs like vehicle design improvements, occupant restraints, etc.
• Uses state-of-the-art technology to advance the state of the practice
• Should be comprehensive – including all stages of the safety management process
• Should be rigorous enough to have scientific merit, yet flexible enough to fit into diverse highway agency operating environments
• Should draw upon knowledge and experience from other ongoing initiatives
• Uses PC-based software to help agencies move away from legacy mainframe systems for safety management

Further details on the planned system may be found in a paper on the subject, which is placed in Appendix F.

Organizational and Institutional Considerations

Even when technological tools are available, organizational and institutional issues may limit their usefulness. Key among these issues are lack of integration of data files, dependence on sources of data which may derive from organizations that are not oriented to the data needs of the design community, lack of integration of computer tools, and inadequate designer computer literacy.

Recommended organizational and institutional strategies include:

• Policy Actions—Giving safety data the status of "strategic enterprise data," establishing safety-performance measures, requiring state and local agencies to participate in safety-data collection oversight, and allowing access to high precision high level, remote imaging for use in obtaining data for documenting the physical highway environment.
• Organizational Actions—Reorganizing agencies to facilitate the efficient and effective collection, management, and use of safety data, establishing entities to provide specialized data and decision-support services, establishing entities dedicated to quality assurance of data, and coalescing state and local agencies into a cooperative safety data collection and management function.
• Legislative Actions—Establishing and authorizing data collection and management functions with oversight over enterprise-strategic data for users within the state, and enabling, through legislation, the implementation of real-time, on-board vehicle dynamics data collection from vehicles involved in crashes.
• Funding Actions—Funding technological advancements, comprehensive research, and training in the use of safety data, establishing mechanisms for cooperative funding of strategic enterprise safety data, and establishing performance-based funding for highway projects.
• Training Actions—Offering training for designers in principles of highway safety related to design; principles, practices and tools of safety analyses; and training to users on new systems for using safety information.
• Research Activities—Continuing support of tools such as FHWA’s Interactive Highway Safety Design Model, development and use of additional tools and methods to support highway design, and committing to develop improved design guides and standards based on improved safety information.

To Improved safety information to support highway design, the design community should:

• Recognize the importance of the role of safety in making design decisions, the inadequacy of present methods for incorporating safety, and the need to improve and enhance methods for analyzing safety.
• Take an active role as part of the broader safety community to help define designers’ requirements for safety data, and work with other users of safety data, in cooperation with collectors and managers of the data, to plan, design, finance and finally implement systems which are useful for design purposes.
• Begin to work on developing new ways of delivering information, using all the appropriate technology.
• Better prescribe and promote proper approaches to analysis and decision making.
• Embark on an ambitious program to demonstrate and develop the decision-support systems that are needed, once improved information is available that justifies their use.
• Re-examine their institutional and organizational structure and environment to identify where barriers to positive change exist, and to begin to anticipate how and in what manner organizational change must take place to maintain effective operations.