What is CODES?
CODES is a collaborative approach to obtain medical and financial outcome information related to motor vehicle crashes for highway safety and injury control decision making. It evolved as the result of the Intermodal Surface Transportation Efficiency Act (ISTEA) of 1991 which provided funds to the National Traffic Safety Administration (NHTSA) to report to Congress about the benefits of safety belts and motorcycle helmets for persons involved in motor vehicle crashes. To measure benefits in terms of reducing death, disability, and medical costs, NHTSA determined that statewide data were needed that included all persons involved in police-reported crashes -- those who were injured or who died as well as those who were not injured. In this manner, comparisons between those using and not using safety belts or motorcycle helmets could be made by identifying and contrasting the characteristics of the injured and uninjured persons within each of the restraint use groups. The CODES Report on the Benefits of Safety Belts and Helmets was delivered to Congress in February, 1996, and is available for download as a pdf document.
In 1992, seven states received a grant from the U.S Department of Transportation (DOT), National Highway Traffic Safety Administration (NHTSA) to participate in the development of a Crash Outcome Data Evaluation System. The number of states involved in the national CODES project has increased to include:
Data linkage and what a model CODES includes
A model CODES consists of linked statewide crash and injury data that match vehicle, crash, and human behavior characteristics to their specific medical and financial outcomes. Data related to motor vehicle crashes are located in multiple sources: crash data collected by police at the scene; EMS data collected by EMTs who provide treatment at the scene and enroute; medical data collected by physicians, nurses and others who provide treatment at the emergency department, in the hospital, or outpatient setting. Additional state data such as driver licensing, vehicle registration, citation/conviction records, insurance claims, HMO/managed care/etc. outpatient record data are also useful for linkage to meet State objectives.
As a minimum, states must have computerized statewide crash, hospital, and either EMS or emergency department data that have sufficient information to discriminate among the crash events and persons involved in each event.
State data are linked using probabilistic linkage technology which facilitates the linkage of state data files that are frequently imperfect. This technology identifies valid matches without requiring exact matches among the attribute variables and is capable of processing a phenomenal amount of data in a short amount of time. Data linkage has the side benefit of highlighting missing and inaccurate data which when corrected make state data more credible.
A model CODES depends upon an Advisory Committee consisting of the owners and major users of the state data to collaboratively resolve issues related to data access, patient confidentiality, management and release of the linked data and institutionalization of CODES.
The SC CODES Project
In 1998, South Carolina received a cooperative agreement from the U.S. Department of Transportation (DOT), National Highway Traffic Safety Administration to support the development of a Crash Outcome Data Evaluation System (CODES). CODES is a collaborative approach to facilitating roadway safety and injury control decision making through the use of medical and financial outcome information related to motor vehicle crashes.
The South Carolina CODES project is a multi-agency effort which includes the Department of Public Safety, Office of Highway Grants Management, The Department of Transportation, the Department of Health and Environmental Control, Emergency Medical Services and the South Carolina Budget and Control Board, Office of Research and Statistics. The purpose of the project is the linkage of databases containing information about individual persons collected from police crash reports, emergency medical transports, emergency room visits and inpatient hospitalization records. Using probabilistic linkage methods, South Carolina linked these databases for 1998 and 1999. To assist policymakers, Safe Communities, and other interested parties in analyzing crash patterns, the CODES project developed a mapping system to locate crashes based on such characteristics as crash severity, volume of crashes, age of driver and involvement of alcohol and/or drugs. Lastly, a CODES Internet site was developed to facilitate the dissemination of information from the project.
Statewide data was collected on all persons involved in police-reported crashes, transported by emergency medical services due to crashes, emergency room visits due to injuries and inpatient hospitalization due to injuries. Analyses of these data are then used to measure the impact of crashes by communities in terms of reducing injuries, deaths, and medical costs. Because data was collected from all crash cases, regardless of injury level, comparisons can be made between the characteristics of those using and not using safety belts, helmets, and other restraints. Property damage only crashes were not included in the linkage or analysis.
Data Linkage
Crash data alone do not indicate the injury problem in terms of the medical and financial consequences. By linking crash, vehicle, and behavior characteristics to their specific medical and financial outcomes, we can identify prevention factors. Here's how the process works: Crash and medical data are collected at the crash scene, en route to the emergency department, in the emergency department and in the hospital. The types of injuries, their severity, and the costs incurred by persons injured in motor vehicle crashes are described and computerized. These statewide data are linked and evaluated.
Data linkage has proven to be an efficient, cost-effective way to create an injury surveillance stem from existing data sources designed for other purposes. As with all data systems, a linked data file has limitations. Some arise from the linkage methodology, and others are inherent in the primary data sources.
Probabilistic linkage is based on statistical methods to assist in the linking of multiple records for the same person. The results of this methodology are a linked database and records that did not link for a variety of reasons. When trying to determine the completeness of the linkage of these data, two major factors impacting the linkage of data from these multiple systems must be addressed.
- The crash reports may indicate that no one was injured, but when compared with emergency department visits and inpatient hospitalization, persons in these crashes received medical treatment.
- Crashes occurring in the border areas, Aiken, Edgefield, Lancaster, York, and Jasper counties may have injuries transported to hospitals in North Carolina or Georgia. Consequently, they will not be reflected in the hospital data.
To illustrate the impact of these factors on data linkage Car Crashes reports show that from total 99,819 crashes there were 23,960 crashes with injuries. The linked database shows that 32,861 crashes with injuries were identified with the limitation that persons treated in other states were not included.
The linked data set provides a representative sample of the crashes in South Carolina on a statewide basis. From the linked data, an analysis of forty-three variables ranging from driver characteristics to crash characteristics to types of injuries shows that there is no systematic bias in the records that linked versus the records that did not link with the exception of location of crashes. As expected, boarder counties such as Aiken were lower due to the treatment of patients in Georgia. Some selected areas, crash reports and emergency medical reports systematically did not provide enough information to link records. County level analyses were done including these same forty-three variables and it was determined that types of crashes, drivers and injuries reported at the county level did not have any systematic biases. The summary tables both state and county level report the percentage of crashes that were linked.
For more information on the SC CODES Project, or to obtain data or reports, please contact Tracy J. Smith at (803) 898-9948 or tracy.joyce@ors.sc.gov.
Partners
- South Carolina Department of Public Safety
- South Carolina Department of Transportation
- South Carolina Department of Health and Environmental Control - EMS
- South Carolina Office of Research and Statistics
Links
- National Highway Traffic Safety Administration - CODES Home
- Other CODES Sites
- National Highway Traffic Safety Administration
- CDC Web-based Injury Statistics Query and Reporting System (WISQARS)
- United States Dept. of Transportation
- South Carolina State Government
- South Carolina Budget and Control Board