loose stool, incomplete rectal evacuation, and abdominal pain were more prevalent during and after the war in deployed than in nondeployed veterans from the same unit (Sostek et al., 1996). The results were based on a 64-item questionnaire administered after the war. Subjects reported that their gastrointestinal complaints began in the Gulf and persisted after return to the United States.
Over the last 5 months of 1991, hospitalizations for testicular cancer were slightly elevated in a large study of active duty deployed versus nondeployed veterans (Gray et al., 1996). In a follow-up study, the investigators extended their analysis through 1996. They replicated their earlier finding, but also found that by 4 years after the war, the cumulative risk of testicular cancer was similar for the two groups of veterans (Knoke et al., 1998). They attributed the transient increase in testicular cancer immediately after the war to regression to the mean because of the healthy-servicemen selection effect and to deferring care during deployment (during which time they would not have had the opportunity for diagnosis and treatment).
The epidemiologic studies of Gulf War veterans summarized above have contributed greatly to our understanding of veterans’ symptoms, but they are beset by limitations commonly encountered with epidemiologic studies. A major limitation is representativeness; most studies focus on groups that are not representative of all Gulf War veterans, by virtue of either their military duties and location during deployment; their military status during the war (active duty, reserves or National Guard); their military status after the war (active duty, reserves, discharged); their branch of service (Army, Navy, Air Force, Marines); or ease of ascertainment (IOM, 1999a). The Iowa study, with its population-based design, had the broadest coverage of U.S. Gulf War veterans. Although it is considered the most representative, the cohort contained few members of racial and ethnic minorities (Iowa Persian Gulf Study Group, 1997). The findings from population-based studies from Canada (Goss Gilroy, 1998) and the United Kingdom (Unwin et al., 1999) are generally consistent with U.S. studies.
Other limitations of epidemiologic studies include small sample size, low participation rates that could result in selection bias in some studies, and recall bias.24 The potential for recall bias is particularly important because most studies rely on self-reporting of symptoms and exposures years after the event, rather than on biological measures (Joellenbeck et al., 1998). Additionally, studies may be too narrow in their assessment of health status. The measurement instruments may have been too insensitive to have detected abnormalities affecting deployed
Selection bias would occur if Gulf War veterans who are symptomatic choose to participate in a study more frequently than those who are not symptomatic. Recall bias would occur if Gulf War veterans who are symptomatic tend to overestimate their previous exposures in comparison with veterans who are not symptomatic (see Chapter 3).