TAIHOD: The Total Army Injury and Health Outcomes Database
The impact of injuries on the mission, readiness, and budget of the Armed Forces is dramatic. To uncover the complete spectrum of injury morbidity and mortality among Army Soldiers, the U.S. Army Research Institute of Environmental Medicine (USARIEM) developed a research database, the Total Army Injury and Health Outcomes Database (TAIHOD).
The TAIHOD is a versatile system that joins multiple personnel and health datasets from various Defense Department agencies. Each agency, at the request of USARIEM, created a dataset that included only active duty Army Soldiers. All records are linked by social security numbers at the level of the individual Soldier. Social security numbers are scrambled to protect the anonymity of the Soldier. The TAIHOD now links four general categories of data: demographics, health outcomes (hospitalizations, outpatient visits, lost-time injuries, permanent disabilities, and fatalities); self-reported health habits and risk-taking behaviors from surveys; and chemical exposures from the Defense Occupational Health Readiness System (DOHRS).
Research epidemiologists at USARIEM use the databases to directly link Army personnel records, self-reported health habits, and chemical and noise exposures to specific health outcomes, and to trace the interrelationship of these outcomes and exposures over time. Analysis of broad categories of data from multiple sources over long periods of time (currently spanning 1970 to 1999) will give researchers an improved understanding of where to optimally focus injury and illness prevention resources. Research inquiry is essential to improve the efficiency and power of surveillance programs. By improving data quality, research efforts provide valuable feedback to data collection agencies and ultimately increase the accuracy and efficiency of surveillance and research programs alike.
The Defense Manpower Data Center (DMDC) dataset, the core of the TAIHOD, includes 4.6 million Soldiers, active duty or former active duty, and contains demographic and exposure variables used to identify subpopulations for further study. Key variables include gender, race/ethnicity, age, military rank, education level, time in service, occupation, marital status, and service discharge codes. The DMDC also has files indicating people who were on reserve status.
The Patient Administration System and Biostatistics Activity (PASBA) is the source of over 3.3 million hospitalization records in the form of a Standard Inpatient Data Record (SIDR), covering all Army personnel admitted to military medical treatment facilities (MTFs) and civilian hospitals. Although the PASBA inpatient data system was not specifically implemented for the purpose of injury surveillance or prevention, its comprehensive nature and highly standardized and complete record system make it an especially useful tool for injury and health outcomes research. The combination of an extensive cause of injury coding system (NATO STANAG codes) and the ability to track readmission give this data exceptional power. We have recently added information from the Standard Ambulatory Data Record (SADR), which contains information on over 4 million outpatient visits from 1997 forward. Variables include ICD-9-CM codes for the nature of the condition, procedure codes, nature and date of visit, disposition, and Medical Treatment Facility.
A dataset obtained from the US Army Safety Center (USASC) contains detailed cause and activity data on over 130,000 ground and aviation "accidents" involving equipment, weapons systems, and vehicles involved in crashes. Additional details on many hospital and fatality cases are also available. This database contains many cases not serious enough to require hospitalization and, therefore, provides useful information on outpatient injuries as well. Variables include date and time of accident, safety devices used (helmets, gloves, goggles, seatbelt), cause and type of injury, body part affected, and environmental conditions present. Extensive information on potential risk or mediating factors may also be culled from narratives that accompany the accident reports.
The Army Disability dataset contains records on 150,000 disability board cases with functional disability ratings according to the Veterans Administration System for Rating Disabilities (VASRD). Relationships between disabilities, hospital ICD-9-CM codes, and occupational exposures can be explored when the disability data is linked to the other TAIHOD components. Variables include medical review board date, percentage of disability, and result/disposition of case.
The Army Casualty Information Processing System (ACIPS) contains information on the cause, time, and place of death on over 11,000 Army Soldiers. By linking casualty data to safety reports and hospitalization data, very elaborate fatality studies can be accomplished. The use of self-reported health habits and risk-taking behavior will also allow many risk factors for injury fatalities to be evaluated. Variables include date and time of death, and manner of death (e.g., accidental, self-inflicted, combat).
The Health Risk Appraisal (HRA) dataset includes information collected in more than 500,000 surveys administered to active duty Army Soldiers. These files include self-reported health habits such as diet, exercise, tobacco and alcohol use, stress, job-satisfaction, risk-taking behavior, and health care utilization. By analyzing this information against other TAIHOD files for the same Army sub-populations, TAIHOD investigators can evaluate the relationship between health habits and the incidence of injury and illness.
The Health Enrollment Assessment Review (HEAR) dataset presently contains over 60,000 surveys administered to active duty Army Soldiers, former active duty Soldiers, and their dependents. These data were obtained recently and are being cleaned and matched for integration into the TAIHOD. Like the HRA, these files also include self-reported health habits such as diet, exercise, tobacco and alcohol use, stress, job-satisfaction, risk-taking behavior, and health care utilization. The HEAR is now the survey instrument of choice in the Army, as well as other branches of the military and the DoD, as the HRA is being phased out.
The Defense Occupational Health Readiness System (DOHRS) data was recently obtained and is being cleaned and matched for integration into the TAIHOD. This data will allow assessment of cumulative exposures for specific chemicals, for general classes of chemicals, and for noise. Data can be evaluated at the level of the individual Soldier, by occupation, or by specific worksite. This dataset contains over two million observations by field industrial hygienists on approximately 1 million individuals, thus for most individuals there are multiple exposure measures over time. This allows evaluation of the impact of cumulative exposures and changes in exposure on various health outcomes. Variables include location (Army base and room number), time of day, type of chemical exposure, and air quality measurements.
The Gulf War Datasets contain demographic information on over 600,000 Army active duty and reserve personnel deployed to the Persian Gulf. These data include information about deployment status, such as dates of deployment and re-deployment.
The Comprehensive Clinical Evaluation Program (CCEP) is a database of health evaluations of more than 50,000 Gulf War veterans describing their health complaints after serving in the Gulf War. These data can be linked to the Gulf War Datasets, DMDC, HRA, PASBA (SIDR and SADR) data to identify illnesses and injuries common among Gulf War Veterans.
The Airborne Datasets contain demographic information on 40,000 people who were in Airborne School in Ft. Benning, GA from 1995-1998. These data can be linked to DMDC, Safety, and PASBA data to evaluate injuries and illnesses in this unique group of people.
Access is strictly controlled through a combination of a locked security door with pass-key access, and a secure intranet (with no access to the Internet).
All database inquiries require review and approval, and are subject to restrictions on data use as defined by IRB/HURC approval protocol. All analytic efforts utilizing TAIHOD data must comply with a protocol approved by theUSARIEM HURC committee, and often other Human Use requirements specified by particular funding institutional guidelines (e.g., NIH).
All researchers sign confidentiality agreements.
Only aggregate information is published, no individual identifiers are ever published. Individual identifiers are used only for initial data linkage purposes. Identifiers are encrypted in the working datasets.