Study objectives: Computed tomography (CT) scan use has become increasingly common in US emergency departments (EDs). Knowledge of trends in CT scan use has implications for ED design and patient flow. We undertake a study to investigate the change in use of ED CT scan use and compare it with the change in use of other tests, such as chest radiographs or blood tests. We also investigate the relationship between CT scan use and changes in patient acuity.

Methods: Data were obtained from the 1996 to 2000 National Hospital Ambulatory Medical Care Survey (NHAMCS), a national probability sample survey conducted by the National Center for Health Statistics, Centers for Disease Control and Prevention. All visits that resulted in a CT scan being obtained were identified. Similar data were collected for the use of chest radiographs and blood tests. Indicators of patient severity (triage category, hospital admission, and ICU admission rates) were also collected, and demographic data were also obtained. Continuous variables were summarized by mean and SDs. Proportions were compared with a χ² test; an a priori significance level of P value less than .05 was selected. Descriptive analyses were performed with statistical software (Microsoft Excel 2000 version 9.0; Microsoft Corporation, Redmond, WA). The data used for this study are openly accessible to the public and excluded all patient identifiers. The study was therefore exempt from institutional board review approval.

Results: We analyzed 115,011 visits that represented 496.4 million visits to US EDs during the 5-year study period. Data were weighted in accordance with NHAMCS guidelines for statistical analysis. The total number of CT scans performed increased from 2.9 million to 5.7 million (χ²=99.2, P=.001). The percentage of ED patients undergoing CT scanning in the study period increased from 3.2% to 5.3%. The use of the CBC count also increased moderately from 25.8% to 26.1% (χ²=8.1, P=.01). The rate of chest radiographs remained constant throughout the study period (16.5%, χ²=0.23, P=NS). The hospital admission rate increased from 10.6% (9,620,213 admissions in 1996) to 12.4% (13,364,017 admissions in 2000; χ²=29.1, P=.001). The percentage of patients admitted to the ICU (1.7% in 1996, 1.6% in 2000) decreased moderately but significantly (χ²=9.4, P=.01). There was a decrease in the triage acuity of patients presenting to the ED: the percentage categorized as level 1 (to be seen within ≤15 minutes) decreased significantly from 21% to 15.7% (χ²=159, P=.001). Overall, the growth in CT scanning was higher than for the growth in the use of a CBC count, chest radiographs, or hospital admission rate, which suggests that despite a lack of increased patient acuity, the rate of CT scan use is increasing. Projections based on 5 years of data suggest that by 2010, more than 10% of all ED patients will undergo a CT scan while in the ED.

Conclusion: The use of CT scanning is increasing at a faster rate than other common tests. This increase cannot be explained by increasing patient acuity. These trends need to be factored into ED design and may have implications for the training of emergency medicine residents. These conclusions also suggest the need for the investigation of the rational use of CT scanning in patient subpopulations.