A significant number of 1991 Gulf War veterans experienced debilitating symptoms that could not be linked to a specific disease. In 1995 Congress gave VA the authority to compensate certain veterans with undiagnosed illness through passage of PL 103-446. The purpose of this epidemiologic database study is to understand better the characteristics of the veterans who have been compensated for undiagnosed illness and to determine the effect of this compensation decision on the veteran's use of VA healthcare services. We propose a retrospective cohort study of subgroups of N=697,000 veterans who served in the Gulf War between August 2, 1990 and July 31, 1991. The approach is highly quantitative in that we shall use existing databases from VA and Department of Defense (DOD) and adopt multivariate statistical methods to compare demographic, military characteristics, and health care utilization of subgroups who 1) receive compensation for undiagnosed illness, 2) receive compensation for other conditions, 3) were denied compensation for undiagnosed illness, and 4) responded to the National Health Survey of Gulf War Era Veterans.
Using epidemiologic methods, we address the primary research question: Do Gulf War veterans who have been granted compensation for undiagnosed illness have a unique profile of demographic, military, in-theater environmental exposures, and psychosocial morbidity that is distinct from that of all 697,000 veterans of the 1991 Gulf War. Secondly, we evaluate the effect of being approved for compensation for undiagnosed illness in terms of utilization of out-patient and in-patient health care services at VA facilities.
Information is obtained by linking five electronic databases: (1) The VBA Compensation and Pension Master Record File, which is the source of data on compensation granted or denied, diagnostic codes, and psychosocial variables; (2) The Army Center for Health Promotion and Preventive Medicine, which will provide data on exposures; (3) the VHA Medical Outpatient Datasets; (4) the VHA Inpatient Datasets; and (5) The National Health Survey of 30,000 Gulf War-Era Veterans. We evaluate data quality of specific variables from these databases; cluster analyses is used to determine whether the veterans compensated for undiagnosed illnesses form clusters of patients based upon their characteristics. Characteristics include demographic, military, environmental exposures, and psychosocial variables. We assess frequency of health care utilization during the 2 year period before and the 2 year period after the compensation decision. Multivariate analyses include multiple linear or logistic regression analysis, use of propensity score to reduce the number of independent variables, log-linear modeling, and discriminant analysis. No new data is collected on veterans.
Among the total population of 696,661 Veterans who served in the Persian Gulf during the 1991 Gulf War, a subset of 8,247 unique Veterans were identified as having a claim approved by VBA for "undiagnosed illness (UDX) during the period Jan 1995 through Dec 2006, with median time 1999, 3rd qtr. The Definition of "undiagnosed illness" as in Public Law (103-466) 1995, includes all VBA Diagnostic Codes which correspond to UDX 8800-8899 series and, in addition, any of the 3 conditions Fibromyalgia, Chronic Fatigue Syndrome, or Irritable Bowel Syndrome (irritable colon). Thus, the conditions defined in VBA data handbook prefixed by "UDX" form the collection of acceptance criteria. The linking with the VHA Medical Outpatient Datasets dated January 1997 through December 2006 showed that 7,909 unique Veterans presented at VA facilities during the 9 year period for a cumulative total of 548,508 visits. The remaining 338 Veterans [8,247 - 7,909] did not match and it is likely that they would link with the VHA Medical Inpatient Datasets. These hospitalizations are not reported upon during this annual report. It is also possible that these SSNs would be found among the non-Gulf Veteran database, a reflection of the misclassification problem of deployment status inherent in DMDC files. This will be explored later.
This report considered the 7,909 unique veterans for whom 548,508 multiple visits were documented in the outpatient datasets during the interval January 1997 through December 2006. Within-person paired analyses were performed in which the number of visits made for health care during the 24 months prior to the approved claim were compared with the number of visits made during the subsequent 24 months. Veterans use of VA health care services did not decrease over baseline use, despite the potential loss of the UDX benefit. Specifically, the frequency of health care visits during the 2-yr baseline period for 7,909 Veterans averaged 12.5 visits, while the frequency of visits during the 2-yr period following approval of claim for UDX benefits averaged 15.9 visits. This average increase of 3.4 outpatient visits was significantly greater than zero, giving evidence for rejection of the hypothesis of no change in frequency of VAMC outpatient visits, p<.001.
The results of the analysis answer in the affirmative the question whether receipt of benefits for "undiagnosed illness" tend to decrease, increase, or have no effect on frequency of visits to VA facilities for healthcare during the initial 2-year period subsequent to approval of the claim. The analyses indicate an average increase of 3.4 visits during the 2-yr period following the approved claim for benefits for "undiagnosed illness".
None at this time.
Health Systems, Military and Environmental Exposures
Gulf War I, Research method, Utilization patterns