Aeromedical Evacuations from Operation Iraqi Freedom: A Descriptive Study

Harman, Dale R

Objective: To describe health patterns in evacuated military members during Operation Iraqi Freedom (OIF) and utilize demographic, diagnostic, and pre- and postdeployment health information to understand the utility of data collected for aeromedical evacuations. Methods: We conducted a descriptive analysis of U.S. evacuees from 2003 using data from the U.S. Transportation Command's Regulating and Command & Control Evacuation System and the Defense Medical Surveillance System. Results: The typical patient was an Army male under the age of 29 requiring orthopedic or surgical care. Disease/ nonbattle injuries were six times as common as battle injuries and 94% were classified as routine evacuees. Eighty-six percent had health data available in the Defense Medical Surveillance System. Two thirds had pre- and/or postdeployment questionnaire data. Conclusions: Combining data sources increases our understanding of disease patterns in deployed troops. Targeted preventive interventions can then be implemented. Changes in the U.S. Transportation Command's Regulating and Command & Control Evacuation System database can improve its utility as an epidemiological tool.

Introduction

The enormous utility of health-related data that is thoroughly and systematically acquired is well documented.1-5 When automated health information relating to exposures, demographic attributes, illnesses, injuries, and other health outcomes is combined into an analyzable form, it offers the opportunity to identify and characterize potential health hazards and their consequences and to detect trends over time. More importantly, these data can lead to the design, implementation, and evaluation of preventive interventions and provide the basis for efficient resource allocation and appropriate policies to reduce future adverse health events.6

Advances in electronic databases and information management systems have made large amounts of health-related information more accessible. However, the ability to locate specific data sources can be challenging, since health information is frequently stored in a variety of disparate locations under the authority of different organizations. Moreover, data are collected and stored in nonstandardized formats and, thus, are difficult to collate, analyze, and compare. Overcoming the hurdles of variability in coding patterns and the different structure and architecture of automated databases still makes the formation and use of large-scale, integrated information systems difficult, but the rewards are worth the effort.7,8

Assembling epidemiological health information under combat conditions provides additional unique challenges. The pace of military medical operations in this type of setting is almost always frantic, and the need to treat, stabilize, and evacuate patients to higher echelons of care is fundamental to saving lives and minimizing disability. These situational factors combine to make complete and accurate documentation of injury and illness data extremely challenging. Add the stressors of a combat environment, including the constant threat of hostile fire, and it is easy to understand why only information deemed absolutely critical to the patient's immediate care and well-being is recorded. Unfortunately, vital communications are frequently verbal, temporary (handwritten directly on the injured patient), or on paper triage tags. Thus, a clear and comprehensive characterization of injury and illness patterns, at the population level, is inherently difficult to obtain under these conditions.

For most U.S. military combat operations, summary reports of disease and nonbattle injury (DNBI) have been the mainstay of medical surveillance.9-13 At best, these data provide only a general overview of injury and illness patterns not directly due to combat, within a framework of location and time. DNBI reports present aggregate numbers of illnesses and injuries in broad categories. As a result, it is not possible to document the circumstances of the illness or injury event or relevant exposures. Nor is it possible to link these data to other databases containing medical outcomes or other patient demographic data at the individual level. Denominator data used to calculate DNBI rates is also reported in broad categories, precluding the ability to conduct other important subgroup analyses. At best, DNBI data can only be used to explore possible associations using ecological study designs.

To make progress toward a comprehensive medical surveillance system for combat operations and the subsequent development and implementation of preventive interventions, existing health-related data sources should be systematically assessed. Operation Iraqi Freedom (OIF) provides us with the opportunity to evaluate our current ability to identiry illness and injury patterns and trends as well as to enhance how we currently collect and use health data in combat settings. As a first step, this report focuses on the U.S. Transportation Command's (TRANSCOM) Regulating and Command and Control Evacuation System (TRAC^sup 2^ES) data, which is collected for the purposes of tracking aeromedical evacuations.

The Department of Defense (DoD) has developed a superior aeromedical evacuation network. Capable of global patient movement, TRANSCOM has undoubtedly saved numerous lives with its rapid and efficient response to medical transportation requirements, particularly in combat environments. When aeromedical transportation is requested. TRAC^sup 2^ES coordinates and transmits all information to the patient's originating and receiving locations and to the transportation provider. Thus, TRAC^sup 2^ES is a real-time, single information interface for the entire patient movement process dedicated to protecting the health and welfare of our deployed troops.14

In addition to supporting this vital transportation function, TRAC^sup 2^ES contains health-related information at the individual level, and it has the potential to provide useful data for surveillance purposes as well as for epidemiological studies. The Army Medical Surveillance Activity (AMSA) recently published a report on OIF aeromedical evacuations of injured Army personnel using TRAC^sup 2^ES data.15 Our assessment of this transportation database focuses on its use as an epidemiological tool with the potential to identify important force health protection issues and to provide the basis for enhanced medical planning and resource allocation and eventually the development and evaluation of appropriate preventive measures. The medical issues pertaining to OIF deserve such analyses, and our goal is to apply the lessons learned to reduce American and allied losses in future operations.

Methods

This is a descriptive study of injuries and illnesses among OIF deployed forces using an automated transportation database of aeromedical evacuations. The study population consisted of all military patients who were aeromedically evacuated from the Central Command (CENTCOM) area of operations during the calendar year 2003 (Figure 1, map of CENTCOM).

Data Sources

USTRANSCOM is comprised of elements from the Air Force Air Mobility Command, the Navy Military Sea Lift Command, and the Army Military Traffic Management Command and is responsible for meeting airlift needs, including aeromedical evacuations for members of the DoD. TRAC^sup 2^ES combines logistic information, transportation requirements, as well as clinical decision-support elements to provide complete visibility of all patient movement and worldwide responsiveness to the DoD's medical transportation needs.17

AMSA maintains the Defense Medical Surveillance System (DMSS), which includes automated, individual level data on hospitalizations, ambulatory visits, reportable diseases, acute respiratory diseases, health-risk appraisals, as well as longitudinal personnel and deployment data for all branches of the U.S. military. AMSA publishes monthly summaries of reportable diseases, illness and injury trends, disease outbreaks, and case reports.18

Construction of the Analytic Data Set

The Office of the Secretary of Defense for Health Affairs provided access to data from TRAC^sup 2^ES on medical evacuation flights originating in CENTCOM between January 1 and December 31, 2003. The data fields within TRAC^sup 2^ES include patient name, social security number, cite number (specific for a particular patient and flight), age, service branch, unit name, movement classification (ambulatory, litter, etc.), precedence (routine, priority, or urgent), originating and destination facility, conflict name (e.g., OIF), medical specialty (MEDSPEC) code (identifies the primary medical service involved in care), injury type (battle vs. nonbattle), a primary and two secondary International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes, clinical diagnoses, and a short text field for an abbreviated medical history and other information, as well as a large number of logistical fields necessary for transportation mission success.

The total number of lines of data from TRAC^sup 2^ES for aeromedical evacuations from OIF, originating in CENTCOM in the calendar year 2003, was 28,404. Each line containing an invalid or incomplete social security number was excluded (n = 476). Because each leg of an aeromedical evacuation is entered into TRAC^sup 2^ES as a new line of data, in many cases there are multiple lines of data for an individual patient. Thus, all but the first record per patient was excluded (n = 11,650), when more than one flight was required to reach the desired destination. Each duplicate record was carefully reviewed to ensure that it was temporally linked to the first record and not a separate evacuation because of a second injury or illness. This left 16,278 remaining records.

After invalid or duplicate social security numbers had been excluded, AMSA linked these data from TRAC^sup 2^ES (n = 16,278) with DMSS data elements, including ICD-9-CM diagnostic codes from pre- and postdeployment inpatient and outpatient encounters and selected portions of the most current pre- and postdeployment health surveys. Once data linking was completed and before analysis, individual identifying information was removed to protect patient confidentiality. A large number of individuals (n = 4,534) in the linked data set were identified as deployed in support of Operation Enduring Freedom or were not specifically designated as OIF participants in the appropriate data field. To validate deployment information, we scanned alternate data fields and identified an additional 88 OIF deployed personnel. Data pertaining to patients not deployed to OIF (n = 4,446), as well as all nonmilitary personnel (n = 649), were subsequently excluded. The remaining 11,183 cases constituted our final analytic data set. Figure 2 illustrates this process.

For purposes of this initial assessment of OIF illnesses and injuries, we examined demographic information, ICD-9-CM diagnostic codes, and other health-related information from TRAC2ES to characterize aeromedical evacuations. Our linked data set, containing pre- and postdeployment survey information, as well as pre- and postdeployment inpatient and outpatient encounter data, will be used in future analyses. Descriptive statistics were generated using SPSS 11.5 for Windows (SPSS Inc., Chicago, Illinois) and Microsoft Excel 2000 (Microsoft Corporation, Redmond, Washington).

Our protocol was reviewed and approved by the Institutional Review Board at the Uniformed Services University of the Health Sciences.

Results

The demographic distribution of aeromedically evacuated military personnel is shown in Table I. The Army comprised the majority of patients (85.5%), with the Marines, Air Force, and Navy constituting nearly all of the remaining patients (14.4%). Most evacuees were male (81.4%) and nearly one-third (31.9%) were between the ages of 20 and 24 years old. The proportion of evacuees in older age categories steadily decreased.

Table II illustrates the number of aeromedical evacuations by quarter during the calendar year 2003. Most (40%) occurred during the second quarter and the least (9.7%) during the first quarter. The numbers declined in the third (34%) and fourth (16.3%) quarters. The most common reason for evacuation was DNBI (86.5%). The remainder (13.5%) was for injuries directly related to combat. Only a few evacuations were classified as urgent (1.4%), i.e., requiring transport immediately to save life or limb or to help prevent serious complications. Slightly more were classified as priority evacuations (4.6%), requiring transport within 24 hours with minimal delay en route. The vast majority of evacuations were classified as routine (94%), indicating that evacuation could safely take place within 72 hours.

Figure 3 depicts the distribution of DNBI and combat injury by month. Battle injuries peaked in April (n = 419) and DNBI peaked in July (n = 1,506). Figure 4 represents the proportion of battle injuries versus DNBI across age categories. As age increased, the proportion of battle injuries declined.

For purposes of future epidemiological analyses using linked TRAC^sup 2^ES and DMSS data, we determined that pre- and postdeployment questionnaire information was available for about two-thirds of the patients (61.6% and 65.2%, respectively). Additionally, 86.2% of evacuees had one or more outpatient encounters and 33% were hospitalized at least once for any cause before or following their deployment to OIF.

Table III provides the age distribution of aeromedical evacuees by gender and service branch. The "typical" patient is a young adult male in the Army. Data on gender for 411 patients (3.7%) and age for 330 patients (2.9%) were missing.

Of the 59 MEDSPEC codes included in the TRAC^sup 2^ES data, the 10 most common are listed in Table IV. Orthopedic (21.5%) and general surgery (13.3%) patients were the most common. The next most common MEDSPEC code was psychiatry, which accounted for nearly 7% of all cases.

Table V lists the primary ICD-9-CM codes assigned to evacuated patients within each of the 17 major diagnostic categories. Injuries and musculoskeletal conditions were the two largest categories, and, when combined, they constituted 40.8% of all cases. Diseases of the digestive system (9.1%) followed by symptoms, signs, and ill-defined conditions (8.3%) were the next largest groups. Nervous system disorders (6.4%), mental disorders (6.1%), and disorders of the genitourinary system (6.1%) comprised the remaining major categories. All other categories each made up less than 3% of the total number of patients.

Discussion

We have provided a broad overview of all military patients who were aeromedically evacuated from the OIF theater in the calendar year 2003. Not only do we gain useful information about the medical operations in OIF, but we also acquire insight into the potential utility of a transportation database as an epidemiological tool.

ICD-9-CM codes and MEDSPEC categories together provide useful information on resource needs. As expected, MEDSPEC codes indicated that orthopedic and general surgery capabilities were in high demand. However, psychiatric care was also a major requirement for patients evacuated from theater. Not surprisingly, injuries and musculoskeletal conditions were by far the most common ICD-9-CM diagnostic categories for evacuated patients. Less common, but each constituting between 5 and 10% of diagnostic categories were diseases of the digestive system; signs, symptoms, and ill-defined conditions; nervous system disorders; mental disorders; and disorders of the genitourinary system. These medical events warrant further investigation into their nature and circumstances.

Given the young Army male's predominant role in OIF, it follows that this group comprised the vast majority of aeromedically evacuated patients. Reliable denominator data would allow us to calculate rates and conduct subgroup analyses but are difficult to obtain because of constant troop movement, the high operational tempo, and the combat environment, as well as for security reasons. Reliable information on unusual patterns of illness or injury rates within the service branches and trends by age, gender, and occupation would assist medical planners and decision makers.

More than one-half of those evacuated from the OIF theater were between 18 and 29 years of age, and we would expect the distribution of battle injuries to be higher in this age group due to their combat roles. Correspondingly, we expect a larger proportion of individuals in the older age groups to require evacuation for diseases and nonbattle injuries. In fact, this is what we found (Fig. 4).

The observed pattern in the number of aeromedical evacuations by quarter in 2003 is consistent with operational tempo. The pre-war buildup of troops and supplies for OIF began in late December 2002, and the period from March 19 to late April 2003 defined the major combat phase of OIF. As expected, the dramatic increase in the number of aeromedical evacuations during the second quarter of 2003 coincided with major combat operations and is reflected in the spike in battle-related injuries seen in Figure 3. One possible explanation for the July peak in other than combat-related evacuations is continued troop buildup. The decline in numbers of aeromedical evacuations during the remainder of 2003 may be explained by fewer combat casualties during the later portion of the year, declining rates of noncombat-related illnesses and injuries as service members adjusted to the environment, and/or improved local medical capabilities obviating the need for some evacuations.

Considering that OIF is a major armed conflict, it is remarkable that so few patients required urgent transport. The vast majority of evacuations were considered routine, defined as requiring transport within 72 hours. The high quality of medical care provided to patients in theater likely contributed to this finding. Once critically injured or ill patients are stabilized, they no longer require urgent evacuation. The realities of war also dictate that some patients who are critically ill or injured do not survive long enough to allow movement from the theater.

It is important to note that this data set does not account for all injuries and illnesses sustained by 0IF deployed forces. Movement to a level I treatment facility uses vehicles attached to the military units which do not have medical capabilities beyond buddy aid. Transport to a level II facility (where medical care is physician-directed and includes resuscitative care) generally uses Service-specific ambulances, ships, and aircraft. TRAC^sup 2^ES is used when transporting patients to level III (fixed inpatient) facilities and beyond. Other medical transportation means may have been used before the patient's entry into TRAC^sup 2^ES, and those patients not requiring transport to a level HI facility are not accounted for in our study.

Understandably, only the most vital information required for purposes of patient transport was consistently entered into the TRAC^sup 2^ES database. Therefore, we found a substantial number of fields with missing data. For example, in our data set, 15.6% of the data for gender and 4.8% for age were initially missing and had to be abstracted from free text fields before analysis. Additionally, 1.7% of the initial 28,404 lines of data had to be excluded from the data set because of invalid or incomplete social security numbers. Since the military status or nationality of these patients could not be verified, information on these patients was lost to analysis. Ensuring completion of a minimum number of specific data fields at the point of origin would be optimal. As an absolute minimum, the system should require a unique patient identifier that can subsequently be used to capture missing information from other data sources. This would also allow data linking with health and personnel data for epidemiological studies. Future technological advances may at some point allow automated data entry and tracking of patient transport in a manner similar to that currently used by global shipping companies.

Our efforts to construct a linked data set yielded pre- and postdeployment questionnaire information from DMSS for about two-thirds of the patients at the time of the data request, even thought the completion of this form is required for all deploying personnel. However, updates to the DMSS database are continuous, and this rich source of health information can enhance or validate other data sources, such as TRAC^sup 2^ES. In addition to pre- and postdeployment surveys, DMSS includes entrance and periodic physical examination data, service and deployment information, and vaccination history. From an epidemiological perspective, the ability to link TRAC^sup 2^ES and DMSS offers great potential for descriptive and analytic studies to aid in the development of targeted preventive measures to reduce future evacuations out of theater, to mitigate illness and injury complications, and to inform policy on postdeployment health care needs.

In-theater ICD-9-CM codes represent preliminary diagnoses at the time of aeromedical evacuation. Linking with medical encounter information from the DMSS will allow us to compare initial diagnoses with final hospital discharge diagnoses and to describe follow-up care, disability, and separation from military service in the postdeployment period. In addition, examination of predeployment ICD-9-CM codes for medical encounters may identify predisposing conditions or circumstances, not captured at the time of predeployment screening, that increase the risk of medical evacuation during deployment.

This study corroborates the AMSA report on TRAC^sup 2^ES as a valuable source of health data. This real-time data source can fulfill some medical surveillance needs as well as establish a dynamic cohort for contingency operations that can be studied for modifiable risk factors and short- and long-term health outcomes as part of the DoD's force health protection efforts. Future detailed subgroup analyses are planned and will include analyses of battle and DNBI injuries and ICD-9-CM codes by gender, occupation, service component, and time in theater. Results of our TRAC^sup 2^ES study will be combined with a systematic assessment of other data sources for in-theater medical events to complete the picture. These steps will ensure progress toward instituting and evaluating preventive measures to minimize morbidity, mortality, and disability, and to maximize quality of life among our fighting forces.

Acknowledgments

We thank Drs. Michael Kilpatrick, Ken Cox, and John Gardner from the Deployment Health Support Directorate, Office of the Assistant secretary of Defense, Health Affairs, the staff of the Army Medical Surveillance Activity, Ms Cara Olsen from the Department of Preventive Medicine and Biometrics, and Ms Kim Bellis from the Center for Force Health Protection Studies at the Uniformed Services University of the Health Sciences.

References

1. Thacker SB, Koplan JP, Taylor WR, et al: Assessing prevention effectiveness using data to drive program decisions. Public Health Rep 1994; 109: 187-94.

2. Thacker SB, Stroup DF: Future directions for comprehensive public health surveillance and health information systems in the United States. Am J Epidemiol 1994; 140: 383-97.

3. Thacker SB, Stroup DF, Rothenberg RB, Brownson RC: Public health surveillance for chronic conditions: a scientific basis for decisions. Stat Med 1995; 14: 629-41.

4. Yasnoff WA, O'Carroll PW, Koo D, Linkins RW, Kilbourne EM: Public health informatics: Improving and transforming public health in the information age. J Public Health Manag Pract 2000; 6: 67-75.

5. Virnig BA, McBean M: Administrative data for public health surveillance and planning. Annu Rev Public Health 2001; 22: 213-30.

6. Rubertone MV, Brundage JF: The Defense Medical Surveillance System and the Department of Defense Serum Repository: glimpses of the future of public health surveillance. Am J Public Health 2002; 92: 1900-4.

7. Morris G, Snider D, Katz M: Integrating public health information and surveillance systems. J Public Health Manag Pract 1996; 2: 24-7.

8. Burwen DR, Seawright MF: Challenges in electronic importing of health data. J Public Health Manag Pract 2000; 6: 87-94.

9. Blood CG, Jolly R: Comparisons of disease and nonbattle injury incidence across various military operations. Milit Med 1995; 160: 258-63.

10. McKee KT, Kortepeter MG, Ljaamo SK: Disease and nonbattle injury among United States soldiers deployed in Bosnia-Herzegovina during 1997: Summary primary care statistics for Operation Joint Guard. Milit Med 1998; 163): 733-42.

11. Writer JV, DeFraites RF, Keep LW: Non-battle injury casualties during the Persian Gulf War and other deployments. Am J Prev Med 2000; 18: 64-70.

12. Taylor SF, Lutz RH, Millward JA: Disease and nonbaltle injury related to peacekeeping operations in South America: summary patient care statistics for CABANAS 2000. Milit Med 2001; 166: 1059-61.

13. Bohnker BK, Sherman SS, McGinnis JA: Disease and nonbattle injury patterns: afloat data from the U.S. Fifth Fleet (2000-2001). Milit Med 2003; 168: 131-4.

14. United States Transportation Command, Office of Public Affairs. USTRANSCOM News Service. Available at http://www.transcom.mil/pa/body.cfm?relnumber= 021113-1: accessed December 7, 2003.

15. Army Medical Surveillance Activity: Frequencies and characteristics of medical evaluations of soldiers by air (with emphasis on non-battle injuries). Operation Iraqi Freedom (OEF/OIF). January-November 2003. Medical Surveillance Monthly Report 2004; 10: 8-12.

16. United States Central Command. Map of CENTCOM Area of Responsibility. Available at www.centcom.mil/aboutus/aor.htm: accessed March 9, 2005.

17. United States Transportation Command. USTRANSCOM History. Available at http://www.transcom.mil; accessed December 7, 2003.

18. Army Medical Surveillance Activity, DMSS Overview. Available at http://amsa. army.mil/AMSA/amsa_home.htm: accessed December 7, 2003.

Guarantor: LCDR Dale R. Harman, MC (FS) USNR

Contributors: LCDR Dale R. Harman, MC (FS) USNR; Tomoko I. Hooper, MD MPH[dagger]; Col Gary D. Gackstetter, USAF BSC[double dagger]

Copyright Association of Military Surgeons of the United States Jun 2005
Provided by ProQuest Information and Learning Company. All rights Reserved