Thank you for granting us the opportunity to reply to the letter by Dr Iwata regarding our article "Evaluation of Clinical Parameters to Predict Positive Cultures for Mycobacterium tuberculosis in Inpatients," recently published in the ARCHIVES.1
Unrecognized TB among individuals who present with atypical clinical features is a well-documented problem. Unfortunately, although those patients are a potential source for nosocomial transmission of TB, their identification is very difficult. The clinical prediction rule was not developed to help physicians identify all TB cases, but rather to assist in determining whether an individual case has a high enough probability of TB so that the patient should be placed on respiratory isolation ward. Given that the prediction rule was developed for future use in patients already suspected of having TB, the inclusion of patients with unrecognized TB in the derivation process would have biased the estimators of the true relationship between the clinical predictors and the presence of a positive culture for TB in patients with suspected TB.
Wisnivesky JP, McGinn T, Crystal RG. Are Fewer Patient Isolations Justified for Active Tuberculosis?—Reply. Arch Intern Med. 2001;161(8):1116-1117. doi: