Background: Despite increased awareness of tuberculosis, delays in management are common.
Objective: To investigate management delays among hospitalized patients with tuberculosis.
Design: Retrospective cohort study.
Setting: The Barnes-Jewish-Christian Health System, a network of eight community and tertiary-care facilities serving the St. Louis, Missouri, metropolitan area.
Patients: All 203 patients with tuberculosis hospitalized in the Barnes-Jewish-Christian Health System from 1988 to 1996.
Measurements: Time from admission to first consideration of the diagnosis (suspicion interval), first consideration and treatment initiation (treatment interval), and admission and treatment initiation (overall management interval) were determined. Delays were defined as intervals longer than 24 hours.
Results: The overall management interval (median, 6 days [5th and 95th percentiles, 1 and 52 days]) exceeded 24 hours in 152 patients (74.9% [95% CI, 68.9% to 80.9%]). The suspicion interval (median, 1 day [5th and 95th percentiles, 0 and 16 days]) exceeded 24 hours in 54 patients (26.6% [CI, 20.5% to 32.7%]), and the treatment interval (median, 3 days [5th and 95th percentiles, 0 and 51 days]) was prolonged in 130 patients (64.0% [CI, 57.4% to 70.6%]). Overall management delays of more than 10 and 25 days occurred in 33.5% (CI, 27.0% to 40.0%) and 18.7% (CI, 13.3% to 24.1%) of patients, respectively. The 55 patients with smears that were positive for acid-fast bacilli had a median treatment interval of 3 days (5th and 95th percentiles, 0 and 33 days); in 58.2% of patients (CI, 45.2% to 71.2%), this interval exceeded 24 hours.
Conclusions: Delays in initiation of treatment were more common than delays in the initial suspicion of tuberculosis. Both types of delays were common even in patients with disease that was confirmed by a positive smear. These data illustrate a need for improved education of physicians about the benefits of early initiation of therapy for tuberculosis.