Variation in Head Computed Tomography Use for Emergency Department Trauma Patients and Physician Risk Tolerance

Physician risk intolerance and malpractice fear have been implicated as leading to defensive behaviors and increased use of health care resources.¹ In the emergency department (ED), physician risk aversion has been associated with higher rates of testing and hospitalization for low-risk patients with chest pain,^2,3 greater use of diagnostic imaging for patients with abdominal pain,⁴ and an increased likelihood of performing head computed tomography (CT) in scenarios involving pediatric patients with minor head injury.⁵

In this study, we explore predictors of head CT use for ED trauma patients, including assessing the impact of physician risk intolerance and malpractice fear. Patients with trauma were chosen because (1) the most common indication for head CT in the ED is trauma,⁶ (2) head CT can definitively exclude life-threatening injuries, and (3) missed diagnoses can have dire consequences for both patients and physicians. Furthermore, because emergency physicians overestimate the likelihood of clinically significant findings on head CT and the medicolegal risk of not obtaining diagnostic imaging,⁷ head CT use may be associated with defensive practice. We hypothesized that variation exists in head CT use and is predicted by physician risk intolerance and malpractice fear.

This prospective study included a physician survey and review of administrative data at an urban academic level 1 trauma center with 60 000 annual visits. We first surveyed attending emergency physicians to quantify their risk tolerance and malpractice fear using 3 previously validated risk scales: the Risk Taking Scale (RTS), Stress from Uncertainty Scale (SUS), and Malpractice Fear Scale (MFS).^2-4 Physician demographic characteristics, including age, sex, years in practice, clinical hours worked, 3- or 4-year residency training, and prior lawsuit involvement, were also captured. We reviewed all patient visits to the ED during calendar year 2009. Visits were eligible for inclusion if they included patients 16 years or older and their medical charts included International Classification of Diseases, Ninth Revision, Clinical Modification codes for either “head trauma” or “trauma.” For each patient encounter we recorded patient age, sex, triage acuity as measured by the Emergency Severity Index, time of day at presentation, and emergency attending physician of record. The primary outcome was whether head CT was performed during the ED visit. All statistical analyses were performed using JMP software (version 8.0.2; SAS Institute Inc, Cary, North Carolina). A multivariate logistic regression model was used to assess the impact of patient and physician factors on use of head CT.

Thirty-seven of 39 physicians (95%) completed the survey. Physician risk tolerance scores were reasonably normally distributed and differed significantly among ED physicians for each of the RTS, SUS, and MFS scales (analysis of variance; P < .001). A total of 55 281 patient encounters during the study period were reviewed; of these, 7905 patient encounters were eligible for inclusion, including 329 with a diagnosis of “head trauma” and 7576 with a diagnosis of “trauma.” A head CT was performed in 1145 of these encounters (14.5%), yielding a median of 28.5 head CT scans performed per ED physician. There was wide variation in head CT use by physician, ranging from 7.2% to 24.5% of patient encounters, with a single outlier of 41.7%.

Head CT use was associated with increasing patient age (P < .001), male sex (P = .001), higher triage acuity (P < .001), and time of presentation (overnight greater than evening or day) (P < .008) (Table). Physician demographic characteristics and risk taking (P = .32), stress from uncertainty (P = .30), and malpractice fear (P = .26) were not predictive of head CT use. Even after controlling for the patient factors, physician identity was still strongly associated with head CT use (P < .001), suggesting that significant variation in head CT use exists among physicians.

We found significant variation in head CT use for ED trauma patients that was not explained by patient factors and was not associated with physician risk tolerance or malpractice fear. Variation in head CT use by ED physicians has been demonstrated previously,^8,9 but not specifically for ED trauma patients. Because well-validated decision rules are available to guide imaging for patients with mild traumatic brain injury, this suggests that an opportunity for quality improvement exists.

The lack of association between physician risk tolerance and head CT use contrasts with prior findings of physician risk aversion being associated with increased use of diagnostic imaging and health care resources.^3-5,7 One reason for this discrepancy may be the shared decision making involved in the care of ED patients at our academic center, which, at a minimum, includes either a resident physician or physician assistant in addition to the attending physician. For patients with trauma team activation, the surgical trauma service and possibly other consulting services may be involved. Because this study measured only the risk tolerance of ED attending physicians, its impact on use may have been diluted by the influence of other decision makers. Initiatives to reduce variation and inappropriate imaging for ED trauma patients should consider all decision makers involved in patient care and might ultimately require institution-level support for clinical decision rules or appropriateness measures.

Correspondence: Dr Andruchow, Brigham and Women's Hospital Center for Evidence-Based Imaging, 20 Kent St, Second Floor, Brookline, MA 02445 (jandruchow@partners.org).

Published Online: March 26, 2012. doi:10.1001/archinternmed.2011.2243

Author Contributions:Study concept and design: Andruchow, Raja, Zane, and Khorasani. Acquisition of data: Andruchow, Raja, and Prevedello. Analysis and interpretation of data: Andruchow, Raja, Zane, and Khorasani. Drafting of the manuscript: Andruchow, Raja, Zane, and Khorasani. Critical revision of the manuscript for important intellectual content: Andruchow, Raja, Prevedello, and Khorasani. Statistical analysis: Andruchow, Raja, and Prevedello. Obtained funding: Raja and Khorasani. Administrative, technical, and material support: Khorasani. Study supervision: Raja, Zane, and Khorasani.