The practice of flexible bronchoscopy is not standardized. Current guidelines are concerned primarily with safety aspects of the procedure. In view of this, and the authors' own observations of individual variation in preparation and technique, a national survey of bronchoscopic procedure was performed to assess physicians' methods. A structured questionnaire was mailed to 547 consultant physicians in adult respiratory medicine. Physicians' routines of patient preparation, drug therapy, sampling methods, and experience of complications with the flexible bronchoscope were assessed. A 60% response (328 physicians) was obtained. Patient consent was obtained by a junior doctor in 31% of replies. 205 (63%) physicians gave benzodiazepine sedation, 46 (14%) used opioid, and 38 (12%) administered both. Ninety-four (29%) physicians prescribed an antimuscarinic agent, and 235 (74%) gave antibiotics to patients with mechanical heart valves. Only 22% of physicians monitored electrocardiogram and 10% monitored blood pressure during all procedures. Marked variance was noted in sampling routines of suspected lung tumours. Physicians who used fluoroscopic guidance for transbronchial lung biopsy reported a significantly lower incidence of pneumothorax requiring drain insertion over the previous 12 months compared to those who did not (2.68 of 1000 versus 9.17 of 1000, (p<0.03)), but no difference in the total incidence of pneumothorax. Only 87 (27%) of responders had performed transbronchial needle aspiration sampling over the previous 12 months. The preparation and practice of flexible bronchoscopy varies greatly for each physician. Use of radiographical screening for performing transbronchial lung biopsy was associated with a lower likelihood of pneumothorax requiring chest tube drainage.