Analysis of sensitivity estimates by lesion location (peripheral or central) and size (<2 cm, 2-5 cm, or >5 cm) revealed the same basic findings summarized in Table 4; all 3 groups of respondents (all physicians as a group, pulmonologists, and PAs) underestimated the median reported sensitivities for all of the diagnostic tests, with the greatest underestimations reported for sputum cytology and TFNAB.
However, in contrast to the physician respondents, PAs selected sputum cytology most frequently as their first diagnostic test selection for half of the case simulations (eg, simulation 1, Figure 1, and simulation 10, Figure 2).
Ten percent to 20% of physicians selected sputum cytology first for peripheral lesions, as did 20% to 40% of PAs, even though they estimated it had the lowest sensitivity for these lesions.
Physician assistant respondents frequently selected sputum cytology as their first diagnostic test, despite their estimate of its low sensitivity for central lesions.
The findings described here suggest that clinicians' perceptions of diagnostic test sensitivities for the tests examined are lower than actual values, and that they underestimate sensitivities for sputum cytology and TFNAB to the greatest degree.
On the whole, PA respondents chose sputum cytology first more frequently than physician respondents.
Physician assistant respondents did choose sputum cytology as their first choice of diagnostic test for half of the case simulations, but they also chose bronchoscopy first for the other half.
However, the consistent underestimation of test sensitivity reported by each of the respondent groups for all sizes and locations of lesions examined, as well as the consistent underestimation of TFNAB and sputum cytology to the greatest degree by all groups, supports the validity of our results.
who order or perform Sputum cytology 12 (22) Bronchoscopy 35 (65) TFNAB * 13 (24) Thoracoscopy 6 (11) Thoracotomy 7 (13) * TFNAB indicates transthoracic fine-needle aspiration biopsy.
The importance of sputum cytology in the diagnosis of lung cancer: a cost-effectiveness analysis.