Trends in thoracic radiology over a decade at a large academic medical center.

Abstract

To investigate thoracic radiology usage over and above the secular trends associated with hospital-wide changes in the number of patients over a decade. We retrospectively reviewed administrative data from our 905-bed tertiary-care hospital between January 1, 1992, to December 31, 2001. Three points of entry to the radiology department were identified: inpatient (IP), outpatient (OP), and the emergency room (ER). The total numbers of patients, imaging studies, chest radiographs, chest CTs, CTs for pulmonary embolism, pulmonary angiograms, ventilation/perfusion scintigrams (V/Qs), lung biopsies, cardiac and chest MRIs, and FDG-PET scans for lung nodules and masses were collected. The significance of trends using linear regression analysis was evaluated. IP and OP numbers have significantly increased over a decade (P = 0.04 and P = 0.01 respectively); ER patient numbers have not. There has been an increase in the ratio of chest radiographs per patient arising from the ER area (P = 0.0002). All 3 areas demonstrated an increase in the ratio of chest CTs per patient: IP (P = 0.0002), OP (P = <0.0001), and ER (P = <0.0001). IP and ER areas demonstrated an increase in the ratio of CTs for pulmonary embolism per patient (P = 0.006, P = 0.04 respectively). There was a decrease in the ratios of pulmonary angiograms and V/Qs per IP (P = 0.02 & P = 0.0003 respectively). Cardiac MRIs per patient demonstrated an increase (IP P = 0.01, OP P = 0.02). FDG-PET for lung nodules and masses per patient demonstrated an increase in IP (P = 0.03) and OP (P = 0.003) areas. The total number of chest imaging studies divided by the total number of imaging studies demonstrated an increase in IP and ER areas (P = 0.02 and P = 0.02 respectively). There has been an increase in thoracic radiology usage above secular trends, particularly in the regions of chest CT and FDG-PET. CT is replacing more traditional techniques to diagnose pulmonary embolism for inpatients.