Many veterans who were deployed to the Persian Gulf during the 1990-1991 Gulf War developed multiple unexplained symptoms such as pain, fatigue, and neurocognitive problems. This constellation of symptoms has been termed Gulf War Veterans' Illnesses (GWVI). Although there is no proven explanation for the cause of GWVI, one fairly widespread explanation is systemic Mycoplasma fermentans infection. The Antibiotic Treatment Trial of GWVI is a randomized placebo-controlled trial to determine whether a 1-year course of doxycycline treatment in deployed Gulf War veterans with GWVI and testing as Mycoplasma species positive will improve their overall functional status as measured by the Physical Component Summary of the SF-36V questionnaire. The study of a multisymptom illness such as GWVI is complicated by the nonspecific nature of the illness, the unknown etiology, and the lack of a widely accepted outcome measure. The presumption of mycoplasma infection raises concerns regarding the methodology for determination of mycoplasma infection, the choice of treatment, and the duration of treatment. However, such a presumption allows the formulation of a clear testable hypothesis that can be tested with treatments with known rates of adverse events and known activity against Mycoplasma species. This paper describes the major issues faced by the investigators during planning, the study design, the patient screening results, and the baseline characteristics of the study patients. There were 2712 patients screened for study entry at 26 Department of Veterans Affairs and two Department of Defense medical centers. Of these, 491 met all study entry criteria and were randomized to either 1 year of doxycycline (200 mg/day) or 1 year of placebo. All patients were seen monthly during treatment and at 6 months after the end of treatment. Study patients had a mean age of 41 years and were mostly male (86%), white (64%), married (68%), and employed full-time (71%).