Some problems in the determination of the false positive and false negative rates of tuberculin tests.