Factors associated with the treatment of latent tuberculosis infection among health-care workers at a midwestern teaching hospital

Study objective: To assess factors associated with initiating therapy and compliance with treatment for latent tuberculosis infection among health-care workers with positive tuberculin skin test results. Design: Prospective cohort study. Setting: An urban midwestern teaching hospital in St. Louis, MO. Study population: Health-care workers with positive tuberculin skin test results. Measurements: (1) Rates of initiating therapy for latent tuberculosis infection among all health-care workers with positive tuberculin skin test results, and (2) compliance rates with therapy for latent tuberculosis infection among health-care workers with recent tuberculin skin test conversion. Results: A total of 440 tuberculin skin test-positive health-care workers were evaluated from January 1, 1994, to May 1, 2000. Of those evaluated, 1 health-care worker had presumed active tuberculosis, 1 had no record of being evaluated, 1 had missing records, and 33 were not recommended isoniazid therapy, leaving 404 workers for analysis. Overall, 396 of 404 health-care workers (98%) with positive tuberculin skin test results initiated isoniazid therapy. In univariate analysis, bacille Calmette-Guérin (BCG) vaccination (p = 0.02) and foreign birth (p = 0.03) were significantly associated with not initiating isoniazid therapy. Compliance data were available for 388 of 404 health-care workers (96%). Of these, 318 of 388 health-care workers (82%) were compliant with 6 months of therapy. BCG vaccination (odds ratio [OR], 3.5; 95% confidence interval [CI], 1.8 to 7.1) and symptoms while receiving therapy (OR, 4.5; 95% CI, 2.0 to 10.1) were significantly associated with noncompliance in multivariate analysis. Among new converters, Asian race (p = 0.006), foreign birth (p = 0.01), BCG vaccination (p = 0.006), and symptoms while receiving therapy (p < 0.001) were significantly associated with noncompliance in univariate analysis. Conclusion: This hospital had a high rate of initiating isoniazid therapy for tuberculosis infection among their health-care workers, and a high rate of compliance with therapy. These rates of initiation and completion of isoniazid therapy were much higher than those previously reported in the literature. This may be largely due to a focused program, which includes active follow-up of health-care workers with positive tuberculin skin test results, consisting of physician counseling and monthly phone consultations by nurses, along with free services and medications provided on-site.