Combined Analysis of the VA and DoD Gulf War Clinical Evaluation Programs

As the previous description indicates, in the USA there have been three distinct Gulf War clinical evaluation programs:

  1. Initial VA Registry begun in 1992
  2. Revised VA Registry initiated in 1995
  3. DoD’s CCEP begun in 1994

These three clinical evaluation programs were not typical "registries" as this term is generally used in medical care. Clinical registries customarily enroll patients with a specific disease in order to better understand the natural progression of that disease or how well patients respond to different types of treatment. In contrast, the three Gulf War clinical programs accepted all veterans, regardless of their symptoms or even whether they had a health complaint. Through a multifaceted outreach effort, all Gulf War veterans were encouraged to participate in the clinical programs and undergo an examination if they had any health concerns or questions, even if they did not have a current illness. As a consequence, it was possible in these clinical evaluation programs to assist a large number of Gulf War veterans and to assess a wide spectrum of health problems.

There are difficulties comparing data from the three registries because of the differences in how the information was collected and coded. Nevertheless, a combined study of these clinical databases provides more useful information than individual analysis for several reasons. First, combining the registries affords a larger number of evaluated Gulf War veterans for study, which increases the chance of detecting a rare or less clinically obvious abnormality. Combined analysis also provides greater numbers and increased statistical power to conduct subgroup analysis by gender, age, location, and time of Gulf War service. Another value of combined analysis is that it allows for comparison of health problems between different populations of veterans: those who remained on active duty after the war and veterans who left active military service or entered the inactive Reserves/National Guard and became eligible for VA health care. Finally, combined analysis permits longitudinal assessment of illnesses over a multi-year period among veterans who participated in more than one registry.

Combined analysis of registry data is possible because the characteristics of these three clinical evaluation programs are similar. The primary goal of the clinical registries was to arrive at a definitive diagnosis and to code diagnoses accurately using ICD-9-CM. Therefore, veterans in the registries received comparable clinical work-ups and diagnoses for any apparent abnormality or complaint, even though the VA and DoD registries used slightly different diagnostics guidelines. In addition, the revised VA registry and the DoD registry were patterned closely on each other, utilizing a similar two-stage diagnostic work-up that targeted veterans with unexplained symptoms and then referred the most difficult to diagnose and treat patients to specialized hospital centers for more extensive evaluation. Importantly, similar self- reported exposure data was obtained from veterans in the three registries.

There are, however, consequential differences in how symptoms and diagnostic data were coded in these three clinical registries (Table 1). The VA Gulf War Registry Health Examination Program had two separate data formats as a result of a major revision and expansion of the original codesheet in 1995. On the original codesheet, up to three symptoms and three diagnoses could be recorded, but for the revised VA codesheet, up to 10 symptoms and 10 diagnoses were captured. In contrast, the DoD's CCEP database lists, as text, the chief complaint and up to 6 other symptoms, and codes a primary diagnosis and no more than 6 secondary diagnoses. In addition to these differences, the VA registry specifically codes for CFS and fibromyalgia, unlike the CCEP. Lastly, coding of registry participants found to be healthy, without a significant health problem, was done differently in the VA and DoD registries (Table 1).

Even though the three databases contain comparable information, the major limitation of all the registries, whether analyzed together or separately, is that participants are self-selected: veterans voluntarily requested and participated in the clinical evaluations. Because of this selection bias, the patterns of illnesses and participation rates in the registries are not necessarily representative of the entire population of Gulf War veterans. It is suspected that participation in the VA registry is influenced by many factors, including: financial need and lack of health insurance; proximity to a VA medical facility; notification by VA outreach efforts; over 100,000 letters sent by DoD to veterans who were potentially exposed to nerve agents from the Khamisiyah demolitions; and, popular media coverage. Participation in the CCEP may have been both discouraged by career concerns during a period of military downsizing and bolstered because the CCEP provided more rapid access to the military health care system.

In a prior study of the initial 74,653 VA and DoD registry patients,168 participation in the registries was associated with:

  1. service in the Army
  2. service in the National Guard
  3. service in the Gulf during wartime hostilities (Operation Desert Storm)
  4. older age
  5. being a construction worker
  6. female sex
  7. having been hospitalized during the year before the Gulf War

Registry participation was also greatly influenced by media attention. This finding was demonstrated by increased enrollment in the registries during periods of high media interest in Gulf War health issues (Figure 3).168

Another problem related to the unique aspects of this population is that for purposes of external comparison, there is no similar, large, control group of outpatients who: 1) have been as intensively evaluated as veterans in these three registries, 2) have survived a life-threatening experience, and 3) have been the subject of prolonged, media reporting. The systematic clinical examination of a distinct group of over 100,000 individuals is without parallel.

In conclusion, it is possible to conduct a combined analysis of the three Gulf War clinical registries because of similar methodology and ICD-9-CM coding of diagnoses. However, direct integration of the three computer databases is not possible for most analyses because of procedural differences among the registries.

The following combined analysis allowed for the assessment of clinical data from comprehensive clinical examinations of over 100,000 veterans, which represents more than 14% of all U.S. forces deployed in the Gulf War. This analysis also provides clinical data on the spouses and children of Gulf War veterans. Due to the self-referred nature of the registry population and the absence of a comparable control group, statistical analysis of this database has to be done with caution. Nevertheless, because such a large proportion of Gulf War veterans have been systematically evaluated over an extended eight-year period, the findings of this combined analysis adds to our understanding of the health problems experienced by this population.

Overall Aims of Study

Although not designed as research studies, the VA Gulf War Registry Health Examination Program and the CCEP provide valuable clinical information about the health of Gulf War veterans. This study had the following aims:

  1. Describe the patterns of illnesses among Gulf War veterans evaluated in the three clinical evaluation programs;
  2. Assess changes in health status of veterans as individuals and as a group over time;
  3. Describe the patterns of illnesses among the family members of Gulf War veterans; and,
  4. Establish a linked database of the three clinical programs, which can be utilized in future research efforts to help understand military and veteran health problems.


This study utilizes data already collected in three health care programs. All data analyses and reporting of findings was done anonymously. This study was approved by two institutional review boards:

  1. Department of Veterans Affairs: Research and Development Human Studies Subcommittee, VA Medical Center, Washington, DC
  2. Department of Defense: Uniformed Services University of the Health Sciences Institutional Review Board, Bethesda, MD

| First Page | Prev Page | Next Page |