Some Ministries of Health in Africa plan to make blood slide microscopy available in peripheral health centers to improve malaria diagnosis over the current practice, which relies solely on clinical findings. To assess whether microscopy improves the management of febrile persons in health centers, we prospectively reviewed medical records of all outpatients visiting six health centers with laboratories in Zambia during a 2-3-day period. Staff interviews and a blinded review of a series of blood slides from each facility by two expert microscopists were also conducted. Of 1,442 outpatients, 655 (45%) reported fevers or had a temperature > or = 37.5 degrees C. Blood slide microscopy was ordered in 28-93% of patients with fever (mean = 46%). Eighty-eight (35%) patients without parasitemia were prescribed an antimalarial drug. Antimalarial drugs were prescribed with equal frequency to those who were referred for a blood slide (56%) and those not referred (58%). The sensitivity of microscopy was 88% and the specificity was 91%. Use of malaria microscopy varied widely, indicating that clinicians are not using standard criteria for ordering this test. Although diagnosis by microscopy was generally accurate, it appeared to have had little impact on the treatment of persons with fever. Guidelines for using blood slide microscopy are needed and prescription of antimalarial drugs should be discouraged when slide results are negative.