ORIGINAL RESEARCH CONTRIBUTION
Barriers to Computed Tomography Radiation Risk Communication in the Emergency Department: A Qualitative Analysis of Patient and Physician Perspectives Thomas E. Robey, MD, PhD, Kelly Edwards, PhD, and Mary K. Murphy, PhD
Abstract Objectives: This qualitative study aimed to characterize the barriers to informed discussions between patients and emergency physicians (EPs) about radiation risk from computed tomography (CT) and to identify future interventions to improve patient understanding of CT radiation risk. Methods: This study used a focus group approach to collect concepts about radiation risk exposure from a national sample of EPs and a local sample of emergency department (ED) patients. A directed content analysis used an a priori medical ethics framework to explore themes from the focus groups while a subsequent normative ethics analysis compared these results with existing perceptions about discussing CT radiation risk. Results: Focus groups (three each for a total of 19 EPs and 27 patients) identified concepts consistent with core medical ethics principles: patients emphasized autonomy and nonmaleficence more than physicians, while physicians emphasized beneficence. Subjects’ knowledge of radiation dose and risk were equivalent to previously published reports. When asked about whether they should talk about radiation with patients, 74% of EPs reported that radiation exposure should be discussed, but the study EPs self-reported doing so with only an average of 24% of patients. Patients reported wanting to hear about radiation from their physicians the next time they need CT scans and thought that a written handout would work better than any other method. When presented with options for how to discuss risk with patients, EPs reported needing easy access to risk information and preferred discussion over other communications approaches, but had mixed support of distributing patient handouts. Conclusions: The normative view that radiation from diagnostic CT should be discussed in the ED is shared by patients and physicians, but is challenged by the lack of a structured method to communicate CT radiation risk to ED patients. Our analysis identifies promising interest among physicians and patients to use information guides and electronic order prompts as potential informational tools to overcome this barrier. ACADEMIC EMERGENCY MEDICINE 2014; 21:122–129 © 2014 by the Society for Academic Emergency Medicine
From the Waterbury Hospital Emergency Department (TER), Waterbury, CT; the Yale University Department of Emergency Medicine (MKM, TER), New Haven, CT; and University of Washington Department of Bioethics and Humanities (KE), Seattle, WA. Received May 9, 2013; revisions received August 2 and September 4, 2013; accepted September 4, 2013. Presented at the Northeast Regional SAEM Meeting, Providence, RI, April 2013; and the Society for Academic Emergency Medicine Annual Meeting, Atlanta, GA, May 2013. Funded by the Greenwall Foundation to TR. The authors have no potential conflicts of interest to disclose. Supervising Editor: Daniel L. Theodoro, MD, MSCI. Address for correspondence and reprints: Thomas E. Robey, MD, PhD; e-mail:
[email protected]. A related commentary appears on page 211. Editor’s note: This paper discusses the topic of the 2016 Academic Emergency Medicine consensus conference, “Shared decisionmaking in the emergency department: development of a policy-relevant patient-centered research agenda.” The date and location of this consensus conference, which will be held in conjunction with the SAEM annual meeting, will be released shortly. The conference chair is Corita Grudzen; co-chairs are Chris Carpenter and Erik Hess.
122
ISSN 1069-6563 PII ISSN 1069-6563583
© 2014 by the Society for Academic Emergency Medicine doi: 10.1111/acem.12311
ACADEMIC EMERGENCY MEDICINE • February 2014, Vol. 21, No. 2 • www.aemj.org
D
iagnostic computed tomography (CT) is a powerful imaging modality that reduces the frequency of surgery and time to definitive diagnosis from the emergency department (ED).1 Its use has increased dramatically, from approximately 2 million scans in 1980 to about 70 million in 2010,1–3 with as many as a third of these studies ordered from the ED.1 Patients have come to expect CT as part of their emergency visits,2 so it is concerning that recent high-profile studies estimate lifetime malignancy risks from CT to be as high as 1%4 and that 2% of all cancers in the United States may now be attributable to radiation from diagnostic CT scans.4–6 Despite these data, risk of radiation exposure from CT scans is controversial. Discrepancies between calculations of malignancy risks3,4,7 raise concerns about risk accuracy,8 and only one study retrospectively associates tumorigenesis to CT scans.6 Only 15% of patients at academic centers are told about radiation risk, compared to 85% being informed of contrast dye allergy.9 Multiple studies over the past decade, including one of emergency physicians (EPs),10 have reported physician lack of CT radiation dose knowledge,10–14 but no efforts have been made to elucidate reasons for this lack of knowledge and other barriers to effective communication about radiation risk between physicians and patients in the ED. Focus groups with semistructured interviews can collect value-laden information that is often not available by quantitative methods,15–19 and normative ethics analyses can provide insight into expectations and motivations19 to explain behaviors. In this study we aimed to collect and compare EPs’ and patients’ perceptions of personal or systemic barriers to CT radiation risk communication in the ED. Specific goals of the investigation included ascertaining whether patient and physician perspectives about radiation risk align and identifying opportunities to improve physician–patient interaction to improve overall radiation risk communication. METHODS Study Design We developed an a priori framework based on major medical ethics themes to serve as the foundation for organizing focus group discussion questions and subsequent analyses. The focus group design used a normative ethics analysis19,20 incorporating directed content analysis21 and phenomenologic methods.16 The directed content analysis mapped subjects’ comments against the ethical principles and related concepts defined in Table 1. All patient and EP participants were offered gift cards and a meal for their time. The institutional review board at Yale University determined the research to be exempt from review. Study Setting and Population Emergency physicians (n = 19) participated in a total of three focus groups. They were unknown to the facilitators and recruited by an author (TER) by e-mail fliers at the Society for Academic Emergency Medicine annual meeting in Chicago, Illinois, in May 2012. Patients (n = 27) were recruited for three focus groups in July 2012 using fliers posted in local physi-
123
Table 1 Concepts From Medical Ethics Used in Directed Content Analysis Principles of medical ethics Beneficence Activity promoting the good of other persons Autonomy Capacity to act independent of external forces Nonmaleficence “Do no harm” Justice Fair distribution of resources Related concepts Paternalism Medical decisions are best left to physicians The Golden Rule “Do unto others as you would have them do unto you” Medico-Legal Concern that a decision will incur future litigation
cians’ offices, on community bulletin boards, and in waiting areas for two urban EDs serving New Haven, Connecticut. Inclusion criteria were self-report of at least one CT scan ordered from the ED in the past 2 years. Patients unknown to the facilitators were prescreened via telephone (TER) to ensure these requirements. Both participant populations knew in advance that the discussions would be about CT scans, but not radiation or ethics. Study Protocol A complete description of this study’s methods following the standard of the consolidated criteria for reporting qualitative research (COREQ) is included in Data Supplement S1 (available as supporting information in the online version of this paper). All focus groups were conducted in closed conference rooms. Focus group sessions followed a semistructured script (Data Supplement S2, available as supporting information in the online version of this paper) of open-ended questions specifically matched between the patient and physician groups, using terminology familiar to each population. Questions were designed to encourage reflection about conversations surrounding recent CT scans ordered or received. When needed, probe questions were used to clarify or provide elaboration. In an effort to unmask participant reactions in a setting of uncertainty, one question offered a hypothetical situation of ethical uncertainty16 to better understand latent motivations and underlying beliefs about the role of risk in diagnostic CT. The session was facilitated (MKM), participants’ comments were documented by a scribe (TER), and digital audio was transcribed by Fox Transcribe (San Francisco, CA). Physician and patient groups were approximately equal in length (70 to 80 minutes). Demographic forms were completed by physician and patient participants before the focus groups began. Patients and physicians also completed intake surveys to compare individual knowledge and experience discussing CT radiation with previous reports.10,11,13 Physician surveys (Data Supplement S3, available as supporting information in the online version of this paper) underwent focused design and sequential external pilot testing,15 while patients completed a written survey (Data Supplement S4, available as supporting
124
Robey et al. • PATIENT AND PHYSICIAN PERSPECTIVES OF CT RADIATION RISK
information in the online version of this paper) that was adapted from Baumann et al.2 Data Analysis Participants’ concepts and phrasing were organized into categories (“coded”) and analyzed both by hand and using ATLAS.ti (version 7, ATLAS.ti Scientific Software Development GmbH, Berlin, Germany). Sample sizes were determined by confirming diminishment of new concepts raised in subsequent groups (“theme saturation”).15,18 Saturation was assessed both in real time (TER, MKM) and in transcription analysis (TER). Real-time saturation was measured using summative content analysis, in which coded concepts were counted and compared immediately after each session. Concepts extracted from meeting notes were plotted against an a priori codebook created for this project (TER, KE) and expanded with each additional session. After transcription, saturation was further verified using summative analysis of participant phrasing and word selection. One major theme (out of 13) was raised in the final physician group, while no new major theme (out of 10) was raised in the final patient group. A directed content analysis used the a priori theme structure to guide questions and prompts during interviews. Comments were categorized and compared by rank order according to codes based on those themes as well as de novo codes based on text analysis.21 A phenomenologic analysis of text compared patient and physician experiences discussing radiation risk. Descriptive statistics as well as medians and interquartile ranges (IQRs) for the demographic and survey data were calculated by hand. RESULTS Demographics of physician and patient focus group participants are described in Table 2. EP age ranged from 32 to 67 years with a median of 37 years, while patient age ranged from 19 to 70 years with a median of 41 years. Patients self-reported a median of 20 (IQR = 7 to 25) lifetime ED visits and a median of 3 (IQR = 1 to 4) lifetime CT scans. On a four-position scale, eight patients reported fair health, 15 reported good health, and four thought their health was excellent. Patients and physicians alike were found to lack knowledge of precise radiation risk estimates for common CT scans, as shown in Table 2. With regard to the importance of discussing risk, 74% of physicians reported that radiation exposure should be discussed more than half of the time, but the EPs self-report talking about radiation risks with an average of 24% of their patients. Most patients (70%) preferred physicians tell patients about every CT risk that had side effects that lasted a day or more. Table 3 illustrates patient and physician comments within the a priori ethics framework. Patients emphasized major themes of autonomy and nonmaleficence more than physicians, and the majority of physicians emphasized beneficence. Emergency physicians remarked that they were driven primarily to do the best thing for their patients and
Table 2 Study participant demographics and baseline knowledge of radiation exposure. Characteristics Physicians (n=19) Sex (male) Race/ethnicity African American White Hispanic Asian Years in practice
