Hospitalizations for tuberculosis in New York City: how many could be avoided?

To determine which factors were associated with hospitalization for tuberculosis (TB) in New York City (NYC), United States, and to estimate the proportion of potentially avoidable admissions.

Patients diagnosed with TB from April to June 2003 were included. Records of hospitalized patients were reviewed to determine whether hospitalization was appropriate. Hospitalization was considered appropriate if patients met ≥1 of the NYC health department hospitalization criteria and/or needed hospitalization per study physicians' judgment. The association of patient characteristics with hospitalization and with having an inappropriate hospitalization was evaluated using multivariate analyses. TB cases from 2008 were also evaluated to determine whether more recent cases had similar associations with hospitalization.

Of 315 patients diagnosed with TB during the study, 226 (72%) were hospitalized. Hospitalized patients were more likely to have a cavitary chest radiograph (adjusted odds ratio [aOR] 8.11, 95%CI 1.82-36.20), abuse alcohol/drugs (aOR 6.53, 95%CI 2.06-20.67), be Black non-Hispanic (aOR 3.05, 95%CI 1.00-9.38), have unknown human immunodeficiency virus status (aOR 2.67, 95%CI 1.24-5.76), and to have been first evaluated by a private medical provider (aOR 2.37, 95%CI 1.11-5.08). Eighty-seven (38%) of the hospitalizations may have been inappropriate; foreign-born (aOR 3.16, 95%CI 1.39-7.14) and acid-fast bacilli sputum smear-positive (aOR 2.49, 95%CI 1.18-5.23) patients were more likely to be hospitalized inappropriately.

Many TB hospitalizations in NYC may be avoidable. Existing guidelines for diagnosing and managing cases as out-patients need to be put into practice.