The sensitivities and specificities of the indirect microimmunofluorescent antibody (IFA) and Weil-Felix (OXK) tests for scrub typhus were established for a range of titers using groups of diseased and control (other febrile illnesses) patients diagnosed by other methods. At a cut-off point of greater than or equal to 1:400, the IFA test was 0.96 specific, and at greater than or equal to 1:320, the OXK was 0.97 specific. Using either these highly specific levels of antibody or other rigorous diagnostic criteria (isolation or 4-fold rising titers), the prevalence of scrub typhus infection was determined to be 0.22 in an unselected population of febrile patients in a rural Malaysian hospital. Probability values (Pr) for the correct diagnosis of scrub typhus were then calculated from the specificity, sensitivity and prevalence determination for a range of titers. The Pr for an OXK titer of greater than or equal to 1:320 was 0.79, and the Pr for an IFA titer of greater than or equal to 1:400 was 0.78. When both these titers were present in a single specimen, the Pr increased to 0.96.