As discussed in our article,1 and in the article by Ashkin and Pitchenik, the chest roentgenogram is not an infallible tool in the diagnosis of tuberculosis. However, the attention that has been given to the "atypical" presentation of tuberculosis may lead some clinicians to minimize the value of the chest roentgenogram. As our study showed, this would be unfortunate. Physician judgment was far from infallible, ie, 58% of the infected patients were not isolated on admission.
Ashkin and Pitchenik expound the alternative: that total reliance on the chest roentgenogram to the exclusion of physician judgment would result in under-isolation of some patients who are infectious. This is quite correct. Indeed, our results show that 14% of the patients with tuberculosis had chest roentgenograms that, while abnormal, were not typical for tuberculosis. Thus, the results of the chest roentgenogram should not override a physician's decision to isolate a patient.