Implementation of an emergency department triage procedure for the detection and isolation of patients with active pulmonary tuberculosis

Study objectives: To investigate the ability of an emergency department scr eening protocol to initiate respiratory isolation of patients with pulmonar y tuberculosis at ED triage before chest radiography. Methods: We conducted a prospective cohort study with retrospective medical record review of adult patients who presented for care to an urban, univer sity-affiliated hospital in Los Angeles County over a 4-month period. Ambul atory patients were administered a triage screening protocol that used pati ent-reported tuberculosis risk factors and symptoms in combination with sel ective chest radiography to screen patients at ED triage for active pulmona ry tuberculosis. Results: A total of 10,674 patients were screened; 2,218 were isolated at t riage and underwent chest radiography, and 378 were kept in isolation in th e ED. The respiratory isolation of pulmonary tuberculosis (RIPT) protocol d etected 17 of 27 visits made by patients with unsuspected pulmonary tubercu losis, yielding a sensitivity of 63% (95% confidence interval [CI] 42% to 8 1%). The estimated specificity was 78%. For each patient with tuberculosis who was detected by the RIFT protocol, 624 patients were screened at triage , 130 chest radiographs were taken, and 22 patients were placed in respirat ory isolation in the ED. Patients with undetected pulmonary tuberculosis mo re commonly had nonpulmonary chief complaints (76% versus 20%; odds ratio [ OR] 13, 95% CI 2.1 to 78.3), and only 60% (95% CI 26% to 88%) were ultimate ly isolated in the hospital. Among RIFT screen-positive patients, radiograp hic findings predictive of pulmonary tuberculosis were cavitary lesions (OR 84.3, 95% CI 22.6 to 315), upper robe infiltrates (OR 24.2, 95% CI 9.1 to 64.4), pleural effusions (OR 8.9, 95% CI 2.5 to 31.8), diffuse/interstitial infiltrates (OR 5.7, 95% CI 1.8 to 17.9), and non-upper lobe infiltrates ( OR 3.1, 95% CI 1.0 to 9.5). Conclusion: The RIFT screening protocol was only moderately sensitive for i solating patients with pulmonary tuberculosis at ED triage. Future studies should evaluate modified and abridged screening protocols, as well as the c ost-effectiveness of triage screening.