Background
Professional societies have made efforts to curb overuse of cardiac imaging and decrease practice variation by publishing appropriate use criteria. However, little is known about the impact of physician-level determinants such as liability concerns and risk aversion on decisions to order testing.
Methods
A web-based survey was administered to cardiologists and general practice physicians affiliated with two academic institutions. The survey consisted of four clinical scenarios in which appropriate use criteria rated echocardiography or stress testing as “may be appropriate.” Respondents’ degree of liability concerns and risk aversion were measured using validated tools. The primary outcome variable was tendency to order imaging, calculated as the average likelihood to order an imaging test across the clinical scenarios (1 = very unlikely, 6 = very likely). Linear regression models were used to evaluate the association between tendency to order imaging and physician characteristics.
Results
From 420 physicians invited to participate, 108 complete responses were obtained (26% response rate, 54% cardiologists). There was no difference in tendency to order imaging between cardiologists and general practice physicians (3.46 [95% CI, 3.12–3.81] vs 3.15 [95% CI, 2.79–3.51], P = .22). On multivariate analysis, a higher degree of liability concerns was the only significant predictor of decisions to order imaging (mean difference in tendency to order imaging, 0.36; 95% CI, 0.09–0.62; P = .01).
Conclusion
In clinical situations in which performance of cardiac imaging is rated as “may be appropriate” by appropriate use criteria, physicians with higher liability concerns ordered significantly more testing than physicians with lower concerns.
Highlights
- •
Physicians with higher liability concerns reported ordering more imaging tests.
- •
Practice specialty had no impact on ordering imaging of uncertain benefit.
- •
Practice experience had no impact on ordering imaging of uncertain benefit.
- •
Liability concerns may hinder efforts to improve the appropriate use of testing.
Although echocardiography and cardiac stress tests are critical for the diagnosis and management of patients with cardiovascular disease, rapid growth in the use of cardiac imaging has led to speculation that testing may be overused. Although some observational studies have suggested that patients who undergo echocardiography in the hospital setting have improved outcomes, a recent meta-analysis found that the rate of “rarely appropriate” transthoracic echocardiographic examinations is decreasing but remains significant at 15%. Cardiac imaging during stress testing provides excellent ability to detect ischemia in patients with suspected coronary artery disease, but there is a considerable degree of variation in the use of testing and high rates of testing in low-risk populations.
Reducing overuse of cardiac imaging is important not only to reduce health care costs but because such imaging could lead to worse patient outcomes from unnecessary radiation exposure or downstream effects such as subsequent invasive procedures. Physician specialty groups such as the American Society of Echocardiography and the American College of Cardiology have made efforts to reduce overuse of imaging by disseminating appropriate use criteria (AUC) for cardiac imaging procedures. However, the barriers to the effective application of AUC and other interventions (i.e., the American Board of Internal Medicine’s Choosing Wisely campaign) have been poorly characterized. In particular, little is known about the impact of physician-level determinants on decisions to order echocardiography and stress tests. Identification of physician characteristics associated with greater use of discretionary imaging could lead to the development of novel policies or more specific interventions to reduce overuse of testing. Therefore, we sought to determine the degree to which physician specialty, practice experience, risk aversion, and liability concerns influence decisions to order echocardiography and stress tests in clinical situations in which testing is of uncertain benefit.
Methods
Study Population
Our study population was composed of physicians who were affiliated with two large academic centers (the Hospital of the University of Pennsylvania and Massachusetts General Hospital). Prior studies have shown that the majority of nonurgent cardiac imaging tests are ordered by either general practitioners or cardiologists. Therefore, we restricted our study population to physicians with primary specialties in general internal medicine or family medicine (collectively referred to as general practice) or cardiovascular medicine. All physicians with primary specialties in one of these three areas who practiced at any site affiliated with the academic medical center (i.e., main teaching hospital, community-based outpatient practice, etc) were invited to participate. Respondents were recruited via e-mail distribution lists that were obtained from the offices of the respective division or department heads.
Survey Development
Clinical Scenarios
We designed a web-based survey instrument to assess physicians’ decision making to order echocardiography or imaging stress tests in four common clinical scenarios. Two scenarios assessed respondents’ likelihood of ordering resting transthoracic echocardiography, and two scenarios assessed their likelihood of ordering an imaging stress test (stress echocardiography or nuclear single-photon emission computed tomography). The clinical scenarios were deliberately constructed to have a high degree of uncertainty with regard to whether the performance of an imaging test would be of clinical benefit. Therefore, we chose scenarios in which performance of imaging is rated as “may be appropriate” by AUC for echocardiography and stable ischemic heart disease. Our goal was to maximize variation among responses, thereby increasing our ability to detect a difference between provider-level determinants of ordering an imaging test.
Questions were developed on the basis of the AUC for echocardiography and stable ischemic heart disease and revised on the basis of feedback from coauthors knowledgeable in the field of appropriate use. Specifically, changes were made to clarify time period distinctions and make symptom burden and examination findings straightforward. All scenarios contained information about the age and sex of the patient, as well as relevant signs and symptoms, to ensure sufficient information to make a reasonably informed decision without being excessively detailed to reduce respondent burden.
The four clinical scenarios were as follows: (1) routine stress imaging in a patient who underwent coronary artery bypass grafting >5 years ago with no assessment for ischemia since surgery but with no new symptoms, (2) preoperative stress imaging before intermediate-risk surgery in a patient with known coronary artery disease but with unknown functional capacity, (3) routine surveillance of left ventricular ejection fraction in a patient with stable heart failure who last underwent echocardiography >1 year ago, and (4) surveillance of mild mitral regurgitation in a patient with stable symptoms and normal left ventricular ejection fraction who last underwent echocardiography >3 years ago. We also included one negative control question in which performance of imaging was rated as “rarely appropriate” (repeat assessment of left ventricular ejection fraction within 1 year in a patient with stable symptoms). Respondents were asked how likely they would be to order an imaging stress test (scenarios 1 and 2) or nonurgent transthoracic echocardiography (scenarios 3 and 4). Responses were chosen from a six-point Likert-type scale ranging from “very unlikely” to “very likely,” without a neutral option. The complete survey is provided in the Supplemental Appendix .
Liability Concerns and Risk Aversion Tools
Respondents’ degree of liability concerns was assessed using a previously validated survey instrument. Five specific questions were asked, including whether respondents have had to change practice patterns because of previous lawsuits, whether they were concerned about being involved in future lawsuits, whether they ordered tests to avoid the appearance of malpractice, and whether they believed that relying on technology rather than clinical judgment reduces the risk for being sued. Each response was provided on a six-point Likert-type scale (1 = strongly disagree, 6 = strongly agree).
Risk aversion was also assessed using a previously validated tool. Respondents were asked six questions pertaining to their risk-taking behaviors in general life (not necessarily related to patient care). Questions included whether respondents enjoyed taking risks (i.e., “thrill-seeking”), whether they preferred to avoid uncertainty, and whether a sense of security was important to daily life. Each response was provided on a six-point Likert-type scale (1 = strongly disagree, 6 = strongly agree).
Finally, respondents were asked about their familiarity with AUC for echocardiography and stress testing for stable ischemic heart disease, years spent as attending physicians in their primary specialties, and demographic information regarding their practices, including location and setting.
Outcome Variables
The primary outcome variable was tendency to order imaging (TOI), defined as each respondent’s average likelihood of ordering an imaging test across the four scenarios. TOI was calculated by taking the mean numeric value of responses on the six-point Likert-type scale that each respondent provided for the clinical scenarios. TOI could range from a value of 1 (very unlikely to order imaging for each of the four clinical scenarios) to 6 (very likely to order imaging for each of the four scenarios).
Covariates included physician specialization in cardiovascular medicine, degree of liability concerns (calculated as the mean numeric value of responses on the six-point Likert-type scale; 1 = minimal liability concerns, 6 = high liability concerns), degree of risk aversion (calculated as the mean numeric value of responses on the six-point Likert-type scale; 1 = minimal risk aversion, 6 = high risk aversion), practice experience (<5, 5–10, 11–20, or >20 years), familiarity with AUC (determined by the numeric value indicated on the six-point Likert-type scale; 1 = not at all familiar, 6 = very familiar), academic institution, and practice setting (teaching vs nonteaching).
Statistical Analysis
Continuous variables are expressed as mean ± SD or as medians with interquartile ranges depending on normality. Cohort characteristics are described according to medical specialty (cardiologists vs general practice physicians). Student’s t tests were used to compare continuous variables, and χ 2 tests were used to compare categorical variables between cardiologists and general practice physicians. Linear regression models were used to evaluate the association between TOI and the physician-level determinants of testing. Parsimonious models were created using backward stepwise regression, with a threshold of P < .20 for inclusion in the final model. Analyses were performed using Stata version 13.1 (StataCorp LP, College Station, TX). The study protocol was reviewed and granted exemptions by the institutional review boards at the University of Pennsylvania and Harvard Medical School.
Results
Study Population
The survey was e-mailed to 420 physicians, and 108 complete responses were obtained (26% response rate). More responses were obtained from cardiologists than general practitioners (54% vs 46% of respondents). Four physicians (4%) chose to order echocardiography for the negative control scenario in which testing was rated as “rarely appropriate,” and their responses were excluded from the analysis. Characteristics of the study cohort are shown in Table 1 . Compared with cardiologists, general practice physicians were more likely to be women (15% vs 58%, P < .001), but there were no significant differences in institutional affiliation, practice experience, or practice setting between the two groups. Nearly half of respondents (48%) had >20 years of practice experience, and nearly all respondents (98%) identified their practice settings as primarily teaching facilities.
| Cardiologists ( n = 56) | General practice physicians ( n = 48) | P | |
|---|---|---|---|
| Institution (%) | .42 | ||
| A | 53 | 45 | |
| B | 47 | 55 | |
| Women (%) | 15 | 58 | <.001 |
| Practice experience (%) | .92 | ||
| <5 y | 19 | 12 | |
| 5–10 y | 15 | 9 | |
| 11–20 y | 19 | 31 | |
| >20 y | 47 | 48 | |
| Practice setting (%) | .99 | ||
| Teaching | 98 | 98 | |
| Nonteaching | 2 | 2 |
TOI
There was broad variation in physician responses to the clinical scenarios ( Figure 1 ). Among the four scenarios, physicians were most likely to order an imaging stress test in the patient undergoing intermediate-risk surgery with a history of coronary artery disease but unknown functional capacity (mean TOI, 4.83; 95% CI, 4.49–5.12). Physicians were least likely to order an imaging stress test in the patient with a history of coronary artery bypass grafting 6 years ago with no assessment for ischemia since surgery but with no new symptoms (median TOI, 2.49; 95% CI, 2.12–2.85).
Differences between Cardiologists and General Practitioners
Differences in physician-level determinants of ordering imaging tests between cardiologists and general practitioners are shown in Figure 2 . There was no difference in mean TOI between cardiologists and general practitioners (3.46 [95% CI, 3.12–3.81] vs 3.15 [95% CI, 2.79–3.51], P = .22). There was also no difference between cardiologists and general practitioners in liability concerns scores (2.72 [95% CI, 2.44–2.99] vs 2.68 [95% CI, 2.34–3.02], P = .86) or risk aversion scores (3.03 [95% CI, 2.89–3.18] vs 3.00 [95% CI, 2.89–3.12], P = .75). However, cardiologists rated themselves as being more familiar with the AUC for echocardiography and evaluation of stable ischemic heart disease (3.38 [95% CI, 3.08–3.67] vs 2.87 [95% CI, 2.60–3.14], P = .01).
Predictors of Decisions to Order Imaging
Our final linear regression model to assess the associations of physician-level determinants with TOI included the following independent variables: physician specialty, practice experience > 20 years, risk aversion score, liability concerns score, and familiarity with AUC. Results of the model are shown in Table 2 . A higher degree of liability concerns was the only significant predictor of decisions to order imaging (mean difference in TOI, 0.36; 95% CI, 0.09–0.62, P = .01).
| Mean difference in TOI (95% CI) | P | |
|---|---|---|
| Covariate | ||
| Liability concerns | 0.36 (0.09–0.62) | .01 |
| Cardiology specialty | 0.32 (−0.18 to 0.81) | .22 |
| Risk aversion | 0.43 (−0.17 to 1.02) | .16 |
| Practice experience (>20 y) | 0.47 (−0.25 to 1.18) | .20 |
| Familiarity with AUC | −0.15 (−0.42 to 0.12) | .28 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
