Abstract
BACKGROUND: There are no clinical guidelines on best practices for the use of bronchoscopy and esophagoscopy in diagnosing head and neck cancer. This retrospective cohort study examined variation in the use of bronchoscopy and esophagoscopy across hospitals in Michigan. METHODS: A total of 17,828 patients were identified with head and neck cancer in the 2006 to 2010 Michigan State Ambulatory Surgery Databases. A hierarchical, mixed-effect logistic regression was used to examine whether a hospital's risk-adjusted rate of concurrent bronchoscopy or esophagoscopy was associated with its case volume (< 100, 100-999, or ≥ 1000 cases per hospital) for those undergoing diagnostic laryngoscopy. RESULTS: Of 9218 patients undergoing diagnostic laryngoscopy, 1191 (12.9%) received concurrent bronchoscopy and 1675 (18.2%) underwent concurrent esophagoscopy. The median hospital rate of bronchoscopy was 2.7% (range, 0%-61.1%), and low-volume (odds ratio [OR] = 27.1; 95% confidence interval [CI] = 1.9, 390.7) and medium-volume (OR = 28.1; 95% CI = 2.0, 399.0) hospitals were more likely to perform concurrent bronchoscopy compared to high-volume hospitals. The median hospital rate of esophagoscopy was 5.1% (range, 0%-47.1%), and low-volume (OR = 9.8; 95% CI = 1.5, 63.7) and medium-volume (OR = 8.5; 95% CI = 1.3, 55.0) hospitals were significantly more likely to perform concurrent esophagoscopy relative to high-volume hospitals. CONCLUSIONS: Patients with head and neck cancer who are undergoing diagnostic laryngoscopy are much more likely to undergo concurrent bronchoscopy and esophagoscopy at low- and medium-volume hospitals than at high-volume hospitals. Whether this represents overuse of concurrent procedures or appropriate care that leads to earlier diagnosis and better outcomes merits further investigation.