Pneumonia due to infection with Candida spp. is extremely rare even though these yeasts are commonly cultured from respiratory secretions. The diagnosis of pneumonia due to Candida spp. should be made only by demonstrating tissue invasion of a biopsy specimen. Physicians might misinterpret the presence of Candida spp. in respiratory secretions as being the etiological agent of pneumonia. This study describes the practice of limiting identification (ID) of rapidly growing yeasts (i.e., Candida spp.) in respiratory secretions and its impact on patients. Before November 2001, rapidly growing yeasts found in respiratory secretions were identified to the species level. After November, rapidly growing yeasts were reported as "yeasts, not Cryptococcus." The group of patients with respiratory secretions processed before November 2001 is called the full ID group (n = 267); the group with samples processed after that date is called the limited ID group (n = 77). Full ID patients had an average length of hospital stay of 12.1 days/patient; that of limited ID patients was 10.1 days/patient, a decrease of 2 days/patient (P = 0.02). The full ID patients had an average cost of 9,407 dollars/patient; that of limited ID patients was 6,973 dollars/patient, a decrease of 2,434 dollars/patient (P = 0.03). Antifungal medications were used in 103 of 267 (39%) full ID patients and in 16 of 77 (21%) limited ID patients, a decrease of 18% (P = 0.004). Limited ID patients had a mortality rate of 14.3%; that of full ID patients was 18.7%, a decrease of 4.4% (P = 0.37). This policy of limiting yeast ID did not impair the diagnosis of pneumonia. Rather, decreases in lengths of stay, costs, and administration of unnecessary antifungal therapy were observed after instituting this policy.