Safety Data Initiative

Overview

While the U.S. Department of Transportation (DOT) has reduced accident rates in every major category of transportation, diminishing gains suggest that many programs might be reaching their performance limits. Transportation fatalities still rank third as the cause of lost years of life in the U.S. (behind heart disease and cancer).

The goal of the Safety Data Initiative (SDI) is to improve the quality of transportation data such that U.S. travel risk factors can be identified, quantified and minimized. Based on a series of workshops and the Safety in Numbers Conference with stakeholders, the Bureau of Transportation Statistics (BTS) developed (and the DOT Safety Council approved) the Safety Data Action Plan, which includes ten research projects to improve data quality and fill data gaps.

Work in Progress

BTS has been established as the lead agency charged with improving safety data. Four working groups have been established to accomplish this goal. Teams include members from all transportation modes (air, rail, highway, water, and pipelines), other federal agencies, as well as academia. The purpose of the working groups is to develop implementation plans for the assigned research projects. These workgroups are designated as A, B, C, and D and include the following research projects:

• Research Project #1: Reengineer DOT data programs
• Research Project #2: Develop common criteria for reporting injuries and deaths
• Research Project #3: Develop common denominators for safety measures
• Research Project #4: Advance the timeliness of safety data
• Research Project #5: Develop common data on accident circumstances
• Research Project #6: Develop better data on accident precursors or leading indicators
• Research Project #7: Expand the collection of "near-miss" data to all modes
• Research Project #8: Link safety data with other data [This research project is being addressed by the Intermodal Transportation Data Base (ITDB) group]
• Research Project #9: Explore options for using technology in data collection
• Research Project #10: Expand, improve, and coordinate safety data analysis

Working Group A

Research Project #1: Reengineer DOT data programs

DOT maintains in excess of 40 programs that capture either safety data or crucial related information, such as measures of exposure. But a recent data quality review requested by Congress suggested that quality improvements can be made to better serve the DOT mission. The first step toward improvement will be to assess the quality of the major data systems, including a full quality audit of all major safety data systems in FY 2001.

With improved data, DOT's safety programs will become not only more effective, but more cost-effective as well. With better data, DOT can develop more targeted inspection, education, regulatory, and research programs to address its strategic goal of improving safety. The objective of this project is to conduct an assessment of major data collection systems and work closely with the modal data collection organizations and data suppliers to identify areas for possible improvement. Process improvements will then be developed and implemented. The final report for Research Project #1 is due to be completed by December 2001.

Research Project #4: Advance the timeliness of safety data

Much of our safety data is reported only on an annual basis, and for most of our safety data there are reporting lags of up to several months after an accident or end of a reporting period. The DOT community needs to have safety data at least on a monthly basis and with no more than a 30-day lag to identify trends and take corrective actions earlier, thus reducing transportation-related injuries and deaths.

The goal of this project is to provide information to advance the timeliness of transportation safety data. BTS proposes to develop plans to increase the reporting frequency of five databases that support DOT safety performance measures from annually or quarterly to monthly. The final report for Research Project #4 is due to be completed by November 2001.

Working Group B

Research Project #6: Develop better data on accident precursors or leading indicators

DOT has focused strongly on safety outcomes but often only subjectively assessed the relationship between contributing factors and the resulting accidents, deaths, and injuries. Such a focus limits the department's ability to intervene early in the process and may result in program designs that are not as targeted or as cost-effective as they could be.

The goal of the Safety Data Action Plan is to improve safety data and data analysis across transportation modes. Collecting or reporting more accurate data on counts of transportation-related injuries is important for presenting an accurate picture of the present condition. Although accurate data on transportation-related injuries does not contribute to the reduction of actual adverse events or accidents, "accurate" information on risk factors or precursors of adverse events can theoretically lead to development of intervention measures, policies, and work practices that can potentially prevent injury or decrease the likelihood that an adverse event will occur. Better information on accident precursors can help us forecast trends, and we can use that information to help set more realistic program goals.

The main goal of this project is to identify leading indicators or accident precursors for further research that have been shown to predict a safety outcome based on demonstrated correlations. The results of this research can also be used to target or redirect programs for greater effectiveness and fewer fatalities and injuries. The final report for Research Project #6 is due to be completed by December 2001.

Research Project #7: Expand the collection of "near-miss" data to all modes

Research project #7 can be viewed as a subset of research project #6, because its purpose is to extend the collection of near-miss data, a particular type of accident or event precursor data, to all modes. Virtually all transportation accidents are preceded by a chain of events or circumstances, any one of which might have prevented the accident if it had gone another way. In a large number of cases, operators are aware of these "close calls" or "near misses" and may have information that could prevent future accidents. Each mode has a different operating environment and its own set of issues with the industry. The Federal Aviation Administration (FAA) has had a near-miss reporting system in place for several years, including confidentiality protections for those reporting. Other modes such as the United States Coast Guard (USCG) and Federal Motor Carrier Safety Administration (FMCSA) are in the process of developing similar near-miss reporting systems.

The goals of this project are to study existing near-miss systems, identify potential benefits and problems, explore the transferability of near-miss reporting from aviation and maritime modes to other modes, and propose a coordinated effort across DOT for implementing such systems.

Near-miss reporting systems can provide information that complements accident investigations and may provide key data that would prevent low probability/high consequence accidents in particular. Under certain conditions they also provide information on what might work to break the accident chain before an accident occurs. With better data on near-miss cases, policy makers can identify better prevention measures and transportation planners can develop more focused prevention efforts. In addition, better analysis of near-miss information can enhance safety researchers' ability to mitigate accidents. The final report for Research Project #7 is due to be completed by November 2001.

Research Project #9: Explore options for using technology in data collection

Much of DOT's safety data collection continues to follow old patterns of paper reporting (sometimes still using paper) and sometimes an automated facsimile of paper reporting. This is manually intensive, costly, and prone to a variety of errors. Several areas seem particularly suited to using technology to improve data collection and reporting. For example:

• Highway vehicle-miles traveled data are currently reported on an annual basis, with significant variation in data-collection methods and significant error in measurement.
• Most highway accidents are reported through paper reports filed by police officers on the scene.

Better use of technology could greatly facilitate more timely data collection and improve data quality, since the likelihood of human error will decrease, and it may also be more cost-effective. The objective of this project is to explore options for using new technology to improve data collection and reporting. There is a great amount of overlap between this project and project #4 (Advance the timeliness of safety data). In project #4 we'll explore options for improving timeliness (and implicitly, quality) of safety data by improving the process as well as the technology used for data collection and reporting. The final report for Research Project #9 is due to be completed by January 2002.

Working Group C

Research Project #2: Develop common criteria for reporting injuries and deaths

Transportation-related deaths and injuries are key measures of interest in the department's strategic plan. Currently, definitions and reporting criteria for injuries and deaths are inconsistent across the modes. The lack of consistent definitions poses a problem for long term planning, policymaking, and applying best practices from one environment to another. In this project, we will identify and inventory inconsistencies and develop common principles for describing the circumstances of death/injury. Then we will explore options for crosswalking the data or expanding the level of detail to permit such crosswalks. Finally, we will explore the feasibility of consistent, national definitions and reporting.

Common criteria for injury will provide a better sense of relative risk in different operating environments. DOT can address its strategic goal of improving safety in transportation, and BTS can fulfill its legislative mandate to improve the quality of data on accidents (49 U.S.C. 111(i)). In addition, the outcome of this project will provide a foundation of information for discussing an evolution toward a DOT-wide standard for key data coding. The objective of this project is to develop recommendations for BTS and the DOT modes to consider a standard for coding common data elements and injuries across databases. The group's recommendations shall provide the best value to DOT for coding common data while meeting the operational and technical requirements identified in the Safety Data Action Plan. The final report for Research Project #2 is due to be completed by November 2001.

Research Project #3: Develop common denominators for safety measures

Each of the modes uses a different set of denominators for evaluating changes in safety risk. This variety makes aggregation or comparison unworkable and limits researchers' ability to perform comparative risk analysis across modes. The goal of this project is to define a set of denominators that can be used to characterize transportation safety in a comparable way for comparable circumstances, for example, to allow the risk of recreational boating to be compared to the risk of recreational flying or recreational driving.

In this project, BTS, in cooperation with all modes, will develop a variety of common denominators for safety measures, develop sources for estimates of exposure where there are currently gaps (e.g., recreational boating, bicycling, pedestrians, general aviation), and develop sampling techniques for collecting exposure data. The final report for Research Project #3 is due to be completed by November 2001.

Research Project #5: Develop common data on accident circumstances

Over time, each mode has developed its own taxonomy for describing the circumstances surrounding accidents. Consequently there is no consistency in how or what type of data describing an accident is being collected across modes. Though an effort has been made to define "causes" of accidents and develop "causal" data, such efforts are usually limited to a single factor, such as "operator error" which is not very informative. Systematic data collection on underlying factors contributing to an accident is limited. Perhaps more important is a description of the chain of events and the operating environment, including characteristics of the operator, the vehicle, the weather, and the system.

The objective of this project is to expand our understanding of the type of data needed to help identify the causal chain for individual accident investigations and to facilitate and support statistical analysis of data across a wide variety of accidents, even in different modes. To accomplish this we will explore using research on classifying accident circumstances and consider ways to generate consistent accident information across various environments with technology like event recorders. The final report for Research Project #5 is due to be completed by December 2001.

Working Group D

Research Project #10: Expand, improve, and coordinate safety data analysis

The purpose of collecting safety data is to use analysis to understand the causes and circumstances of transportation-related deaths and injuries and then help reduce them. Good analysis supports virtually every successful program. It can help identify transportation problems, as well as help develop policy options, interpret performance, and assess program effectiveness. However, analysis methods and proficiency levels vary widely across modes, and currently there is no good forum for sharing research findings or best practices in research methods. The level of resources for analysis/evaluation is also widely believed to be too low, after two decades of decline in its prominence. As a result of these shortcomings, the effectiveness of our analyses, and therefore the effectiveness of DOT's programs, is limited. The final report for Research Project #10 is due to be completed by November 2001.

Related BTS Web Sites

TranStats
To improve the quality of transportation safety data, it is necessary to gather all transportation safety-related data into one database, hence the creation of the TranStats. TranStats is an Internet-based repository for transportation-related data. It includes data from many separate databases that have been captured by the U.S. DOT. For a list of these source databases, visit the TranStats website. TranStats users will be able to query the data using dynamic tools and/or select from numerous canned queries. The results of the queries can be viewed in tables, graphed or mapped within the TranStats website. Users will also have the option of downloading the data sets to their computers for their own use.

Contact Us

Please contact the Safety Data Initiative at safetydata@bts.gov or call 1-800-853-1351 if you have questions or would like to comment on this website.