Emerg Radiol DOI 10.1007/s10140-014-1237-x
ORIGINAL ARTICLE
Patterns in computed tomography utilization among emergency physicians in an urban, academic emergency department Jonathan Kirschner & Kaushal Shah & Daniel Runde & David Newman & Brandon Godbout & Dan Wiener & Jarone Lee
Received: 20 March 2014 / Accepted: 6 May 2014 # Am Soc Emergency Radiol 2014
Abstract We sought to determine if CT utilization rates varied by characteristics of the physician. A chart review was performed at an urban academic emergency department (ED) to identify all the CT scans ordered and patients seen for subjects 21 years of age and older by physicians between January 2001 and December 2008. “Years of experience” was defined as years of practice after residency. Various experience cutoffs were determined a priori. Physicians were labeled “academic” if they had reduced clinical hours for academic duties and “clinical” if they were physicians without “protected time.” We categorized physicians as “high users” (top quartiles) and “low users” (bottom quartiles), and compared utilization rates from 2001 to 2003 to utilization rates from 2005 to 2007. There were 280 physician-years of J. Kirschner Department of Emergency Medicine, Indiana University School of Medicine, Indianapolis, IN, USA K. Shah : D. Newman Department of Emergency Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA D. Runde (*) Department of Emergency Medicine, Harbor-UCLA Medical Center, David Geffen School of Medicine UCLA, Box 21, 1000 W Carson St., Torrance, CA 90509-2910, USA e-mail: [email protected] B. Godbout Department of Emergency Medicine, Lenox Hill Hospital, New York, NY, USA D. Wiener Department of Emergency Medicine, Mount Sinai St. Luke’s and Roosevelt, Mt. Sinai School of Medicine, New York, NY, USA J. Lee Division of Trauma, Emergency Surgery and Surgical Critical Care, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
practice, with an average experience of 6.1 years. When comparing groups of physicians with more or less than 3, 5, 10, and 15 years of experience, there were no statistically significant differences between the number of CT scans per 1,000 visits (p=0.85; p=0.21; p=0.57; p=0.08, respectively). Comparison between clinical and academic physicians yielded no differences (clinical=98.4, academic=104.2, p= 0.10). Low users ordered 78 CT scans per 1,000 patient visits (95 % CI 76.6–78.5), as compared to the high users that ordered 135 CT scans per 1,000 patient visits (95 % CI 131.8–139.0). We found that all of physicians stayed within their quartiles except one. While there was substantial variation among CT utilization rates by physicians at this urban emergency department, our data shows no differences between physicians with more or less clinical experience and no change in individual utilization patterns during the study period. Keywords Computed tomography . Utilization . Emergency medicine
Introduction The use of computed tomography (CT) scanning as a diagnostic modality has rapidly increased since its inception in the 1970s. An estimated 62 million CT scans are performed in the USA each year compared to 8 million in 1980 [1]. CT utilization in the emergency department (ED) is also steadily increasing [2–4]. For example, the rate of CTs performed at one center more than doubled in a 7-year period, and the use of advanced imaging increased nearly threefold for injuryrelated ED visits from 1998 to 2007 [3, 5]. The increased use of CT technology means increased costs, increased ED length of stay, and a significant radiation exposure leading to increased cancer risk [1, 6, 7]. There have been numerous
Emerg Radiol
efforts to reduce the number of unnecessary CT scans, for example, with clinical decision rules such as the Canadian CT Head Rule [8]. We sought to determine if the characteristics of emergency physicians (EP) with regard to years of experience or having academic responsibilities were correlated with the number of CT scans obtained per ED patient visit. Additionally, we aimed to determine if CT utilization patterns for individual providers changed over 8 years, specifically whether providers moved from higher to lower utilizing groups or vice versa.
Materials and methods Study design A retrospective electronic chart review was performed to identify all the CT scans obtained and total patients seen by each attending physician between January 2001 and July 2008. A minimum age cutoff of 21 years was applied to the CT scan review and the calculation of total patients seen. The hospital’s institutional review board approved the research protocol, and it qualified for “exempt status.” Setting and selection of participants The study population included all attending emergency physicians working in the adult emergency department at our urban academic center with an annual adult emergency department census of over 110,000 patients and a 3-year residency program. To collect the raw data, departmental informatics specialists queried our electronic medical record database using a structured search designed to detect all patients over the age of 21 that received a CT scan during the study period. Specific data elements (e.g., date, type of CT scan, and ordering physician) were obtained. The raw data was given to us as a spreadsheet document. Physicians who ordered fewer than 15 CT scans or evaluated fewer than 300 patients per year were excluded from the study (n=5 physicians). Methods of measurement Nearly all CT scans ordered in our emergency department will appear in our electronic medical record, which also acts as our computerized ordering system. As a result, any CT ordered by the physician is automatically linked to the patient’s chart. On rare occasion, our electronic medical record and computerized order entry system goes “offline” for a brief period, either as a preplanned maintenance event or as an unplanned system failure event. Based on internal data from our informatics personnel, we estimate our total “downtime” during an annual period typically averages less than 1 %. We presume that we did miss CT scans ordered during “downtimes” in our dataset,
but we would also miss a proportional number of patient visits. The number of CTs obtained per ED visit should therefore not be dramatically impacted. As a result, we did not pursue these missed data points for our study. Data collection and processing Each physician included in the study was made anonymous and assigned a number. All data extraction and analysis were performed after the data was made anonymous. Two physician abstractors extracted data with formatted data sheets. Both were emergency medicine residents trained by the principle investigator in training sessions designed for the protocol. Abstractors and data extraction were monitored by the principal investigator for adherence to protocol, and abstractors met regularly with the research team. Specific elements on the form included attending number, number of patients visits seen by the physician, type of CT procedure (e.g., head, neck), and date the CT was completed. Number of CTs completed was grouped by attending, year, and anatomic region, including head, neck, chest, abdomenpelvis, facial bones, and others. For example, CTs of the head included those without contrast, with contrast, CT angiographies, and CTs with and without contrast. Definitions Physicians were labeled “academic” if they had reduced clinical hours for specific academic duties and “clinical” if they were full-time physicians without protected time. “Years of experience” was defined as years of clinical practice after residency. Various experience cutoffs were determined prior to data analysis. Physicians were grouped into “high users” (top quartile) and “low users” (bottom quartile) and analyzed to see if they stayed within their assigned quartiles. We only included physicians if they had data throughout our study period and compared their utilization rates from 2001 to 2003 against their utilization rates from 2005 to 2007. Additionally, we also examined their rates of CT utilization by anatomic region, including head, neck, chest, abdomen-pelvis, facial bones, and others. Outcome measures Our primary outcome measure was CT utilization rate, defined as the rate of CTs ordered by individual attendings per 1,000 patient visits. Our secondary outcomes were CT utilization rates for each of the following anatomic areas: head, neck, chest, abdomen-pelvis, and facial bones, defined as the rate of CTs performed per 1,000 patient visits.
Emerg Radiol
Data analysis Standard descriptive statistical methods including Student’s t test and Pearson’s chi-squared analysis were performed using Microsoft Excel 2008 (Microsoft Corp., Redmond, WA, USA) and SPSS 13.0 (SPSS, Inc., Chicago, IL, USA). Statistical significance was set a 0.05 and confidence intervals (CIs) were set at 95 %.
Results Over the study period, there were 60 physicians included, with a mean experience of 6.1 years accounting for 280 physicianyears of practice (see Table 1). There were 325,426 patient visits and 32,706 CT scans performed for a CT rate of 101 per 1,000 visits. The academic group ordered 104.2 CT scans per 1,000 patients and the clinical group ordered 98.4 CT scans per 1,000 patients; the clinical group overall utilized less CT scans, and the results approached but did not achieve statistical significance (see Table 2). There was no difference in CT utilization if the physicians were compared by the following experience cutoffs: 3, 5, 10, and 15 years. There was a trend toward decreased utilization among physicians with greater than 10 years of experience, but it was not significant (see Table 2). When comparing the bottom quartile of CT utilizers (low users) to the top quartile (high users), there were several notable differences between the groups that existed independent of physician experience. Low users (n=15 physicians) ordered an average of 78 CT scans per 1,000 visits (95 % CI 76.6–78.5), compared to high users (n=15 physicians) who ordered an average of 135 CT scans per 1,000 patient visits (95 % CI 131.8–139.0). High users ordered 2 per 1,000 visits more neck CT scans (p
ORIGINAL ARTICLE
Patterns in computed tomography utilization among emergency physicians in an urban, academic emergency department Jonathan Kirschner & Kaushal Shah & Daniel Runde & David Newman & Brandon Godbout & Dan Wiener & Jarone Lee
Received: 20 March 2014 / Accepted: 6 May 2014 # Am Soc Emergency Radiol 2014
Abstract We sought to determine if CT utilization rates varied by characteristics of the physician. A chart review was performed at an urban academic emergency department (ED) to identify all the CT scans ordered and patients seen for subjects 21 years of age and older by physicians between January 2001 and December 2008. “Years of experience” was defined as years of practice after residency. Various experience cutoffs were determined a priori. Physicians were labeled “academic” if they had reduced clinical hours for academic duties and “clinical” if they were physicians without “protected time.” We categorized physicians as “high users” (top quartiles) and “low users” (bottom quartiles), and compared utilization rates from 2001 to 2003 to utilization rates from 2005 to 2007. There were 280 physician-years of J. Kirschner Department of Emergency Medicine, Indiana University School of Medicine, Indianapolis, IN, USA K. Shah : D. Newman Department of Emergency Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA D. Runde (*) Department of Emergency Medicine, Harbor-UCLA Medical Center, David Geffen School of Medicine UCLA, Box 21, 1000 W Carson St., Torrance, CA 90509-2910, USA e-mail: [email protected] B. Godbout Department of Emergency Medicine, Lenox Hill Hospital, New York, NY, USA D. Wiener Department of Emergency Medicine, Mount Sinai St. Luke’s and Roosevelt, Mt. Sinai School of Medicine, New York, NY, USA J. Lee Division of Trauma, Emergency Surgery and Surgical Critical Care, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
practice, with an average experience of 6.1 years. When comparing groups of physicians with more or less than 3, 5, 10, and 15 years of experience, there were no statistically significant differences between the number of CT scans per 1,000 visits (p=0.85; p=0.21; p=0.57; p=0.08, respectively). Comparison between clinical and academic physicians yielded no differences (clinical=98.4, academic=104.2, p= 0.10). Low users ordered 78 CT scans per 1,000 patient visits (95 % CI 76.6–78.5), as compared to the high users that ordered 135 CT scans per 1,000 patient visits (95 % CI 131.8–139.0). We found that all of physicians stayed within their quartiles except one. While there was substantial variation among CT utilization rates by physicians at this urban emergency department, our data shows no differences between physicians with more or less clinical experience and no change in individual utilization patterns during the study period. Keywords Computed tomography . Utilization . Emergency medicine
Introduction The use of computed tomography (CT) scanning as a diagnostic modality has rapidly increased since its inception in the 1970s. An estimated 62 million CT scans are performed in the USA each year compared to 8 million in 1980 [1]. CT utilization in the emergency department (ED) is also steadily increasing [2–4]. For example, the rate of CTs performed at one center more than doubled in a 7-year period, and the use of advanced imaging increased nearly threefold for injuryrelated ED visits from 1998 to 2007 [3, 5]. The increased use of CT technology means increased costs, increased ED length of stay, and a significant radiation exposure leading to increased cancer risk [1, 6, 7]. There have been numerous
Emerg Radiol
efforts to reduce the number of unnecessary CT scans, for example, with clinical decision rules such as the Canadian CT Head Rule [8]. We sought to determine if the characteristics of emergency physicians (EP) with regard to years of experience or having academic responsibilities were correlated with the number of CT scans obtained per ED patient visit. Additionally, we aimed to determine if CT utilization patterns for individual providers changed over 8 years, specifically whether providers moved from higher to lower utilizing groups or vice versa.
Materials and methods Study design A retrospective electronic chart review was performed to identify all the CT scans obtained and total patients seen by each attending physician between January 2001 and July 2008. A minimum age cutoff of 21 years was applied to the CT scan review and the calculation of total patients seen. The hospital’s institutional review board approved the research protocol, and it qualified for “exempt status.” Setting and selection of participants The study population included all attending emergency physicians working in the adult emergency department at our urban academic center with an annual adult emergency department census of over 110,000 patients and a 3-year residency program. To collect the raw data, departmental informatics specialists queried our electronic medical record database using a structured search designed to detect all patients over the age of 21 that received a CT scan during the study period. Specific data elements (e.g., date, type of CT scan, and ordering physician) were obtained. The raw data was given to us as a spreadsheet document. Physicians who ordered fewer than 15 CT scans or evaluated fewer than 300 patients per year were excluded from the study (n=5 physicians). Methods of measurement Nearly all CT scans ordered in our emergency department will appear in our electronic medical record, which also acts as our computerized ordering system. As a result, any CT ordered by the physician is automatically linked to the patient’s chart. On rare occasion, our electronic medical record and computerized order entry system goes “offline” for a brief period, either as a preplanned maintenance event or as an unplanned system failure event. Based on internal data from our informatics personnel, we estimate our total “downtime” during an annual period typically averages less than 1 %. We presume that we did miss CT scans ordered during “downtimes” in our dataset,
but we would also miss a proportional number of patient visits. The number of CTs obtained per ED visit should therefore not be dramatically impacted. As a result, we did not pursue these missed data points for our study. Data collection and processing Each physician included in the study was made anonymous and assigned a number. All data extraction and analysis were performed after the data was made anonymous. Two physician abstractors extracted data with formatted data sheets. Both were emergency medicine residents trained by the principle investigator in training sessions designed for the protocol. Abstractors and data extraction were monitored by the principal investigator for adherence to protocol, and abstractors met regularly with the research team. Specific elements on the form included attending number, number of patients visits seen by the physician, type of CT procedure (e.g., head, neck), and date the CT was completed. Number of CTs completed was grouped by attending, year, and anatomic region, including head, neck, chest, abdomenpelvis, facial bones, and others. For example, CTs of the head included those without contrast, with contrast, CT angiographies, and CTs with and without contrast. Definitions Physicians were labeled “academic” if they had reduced clinical hours for specific academic duties and “clinical” if they were full-time physicians without protected time. “Years of experience” was defined as years of clinical practice after residency. Various experience cutoffs were determined prior to data analysis. Physicians were grouped into “high users” (top quartile) and “low users” (bottom quartile) and analyzed to see if they stayed within their assigned quartiles. We only included physicians if they had data throughout our study period and compared their utilization rates from 2001 to 2003 against their utilization rates from 2005 to 2007. Additionally, we also examined their rates of CT utilization by anatomic region, including head, neck, chest, abdomen-pelvis, facial bones, and others. Outcome measures Our primary outcome measure was CT utilization rate, defined as the rate of CTs ordered by individual attendings per 1,000 patient visits. Our secondary outcomes were CT utilization rates for each of the following anatomic areas: head, neck, chest, abdomen-pelvis, and facial bones, defined as the rate of CTs performed per 1,000 patient visits.
Emerg Radiol
Data analysis Standard descriptive statistical methods including Student’s t test and Pearson’s chi-squared analysis were performed using Microsoft Excel 2008 (Microsoft Corp., Redmond, WA, USA) and SPSS 13.0 (SPSS, Inc., Chicago, IL, USA). Statistical significance was set a 0.05 and confidence intervals (CIs) were set at 95 %.
Results Over the study period, there were 60 physicians included, with a mean experience of 6.1 years accounting for 280 physicianyears of practice (see Table 1). There were 325,426 patient visits and 32,706 CT scans performed for a CT rate of 101 per 1,000 visits. The academic group ordered 104.2 CT scans per 1,000 patients and the clinical group ordered 98.4 CT scans per 1,000 patients; the clinical group overall utilized less CT scans, and the results approached but did not achieve statistical significance (see Table 2). There was no difference in CT utilization if the physicians were compared by the following experience cutoffs: 3, 5, 10, and 15 years. There was a trend toward decreased utilization among physicians with greater than 10 years of experience, but it was not significant (see Table 2). When comparing the bottom quartile of CT utilizers (low users) to the top quartile (high users), there were several notable differences between the groups that existed independent of physician experience. Low users (n=15 physicians) ordered an average of 78 CT scans per 1,000 visits (95 % CI 76.6–78.5), compared to high users (n=15 physicians) who ordered an average of 135 CT scans per 1,000 patient visits (95 % CI 131.8–139.0). High users ordered 2 per 1,000 visits more neck CT scans (p
