[Study on the prediction of smear negative pulmonary tuberculosis with classification trees].

OBJECTIVE

To improve the respiratory isolation policy for patients with suspected pulmonary tuberculosis (TB).

METHODS

All consecutive patients with suspicion of having pulmonary TB when seeking health care at the County TB dispensary of the Center of Disease Control and Prevention received face to face interview.

RESULTS

A Classification model was constructed with a sensitivity of 90.9% and specificity of 90.2%, while predictive factors of culture-proven pulmonary TB among smear negativecases were soakage in Chest X-ray exam (77.0% vs. 4.4%; P<0.0001), bilateral lung's abnormal (1.6% vs. 19.4%; P<0.0001) and reaction of tuberculin skin testing (0.0% vs. 2.6%; P=0.014).

CONCLUSION

Soakage, bilateral lung's abnormal and positive reaction of tuberculin skin testing were important predictors to prognosticate culture positive diagnosis. The model had been proved to have promising sensitivity and specificity in the rural population covered by NTP-DOTs.