Background
The 2008 appropriate use criteria (AUC) for stress echocardiographic (SE) examinations were revised in 2011 to cover a wider range of scenarios of use. Data comparing the 2008 and 2011 AUC for SE are limited.
Methods
A retrospective chart review of SE studies performed at an academic medical center was conducted, and ordering of SE studies was assessed using the 2008 and 2011 AUC.
Results
A total of 252 consecutive SE studies performed in 2011 were reviewed. The 2008 AUC classified 126 SE studies (50%), and the 2011 AUC classified 221 SE studies (88%) ( P < .001). Of all SE studies, 106 (42%) were performed as part of evaluations for noncardiac solid-organ transplantation. The majority of these studies (79%) were classified as inappropriate using the 2011 AUC and were not classifiable using the 2008 AUC. Of the 146 SE studies performed for other reasons, 69% were appropriate and 22% were inappropriate using the 2011 AUC. Inappropriate SE studies for both the general nontransplant and noncardiac solid-organ transplantation populations were nearly exclusively ordered for perioperative assessment with normal functional capacity. The SE studies that remained unclassified by the 2011 AUC were related to the assessment of specialized cardiac conditions, including hypertrophic cardiomyopathy and follow-up after heart transplantation.
Conclusions
The 2011 AUC classify a significantly greater proportion of SE studies compared with the 2008 AUC. Most of the reclassified SE studies using the 2011 AUC were inappropriate and were ordered for perioperative assessments in patients with normal functional capacity.
Increasing use of noninvasive cardiac stress testing has been cited as one of the reasons for the increased costs of health care delivery. To facilitate the appropriate use of stress echocardiographic (SE) imaging, the American College of Cardiology Foundation (ACCF) and the American Society of Echocardiography along with other subspecialty societies developed appropriate use criteria (AUC). Awareness of and adherence to the AUC has become an important quality goal for many echocardiography laboratories, and tracking appropriateness is now required for accreditation.
Only a few studies have examined the performance of the 2008 AUC for SE imaging since they were published. Results from these studies have shown that the AUC for SE imaging classified 80% to 90% of studies and that the inappropriateness rate was approximately 20%. The AUC for SE imaging were subsequently revised in 2011 in an attempt to better classify SE studies by addressing common clinical scenarios not covered in the initial AUC and to reflect the most current evidence for noninvasive stress testing. The revised AUC expanded the number of SE indications from 49 to 87.
To assess the performance of the 2011 AUC for SE imaging, we applied both the 2008 and 2011 AUC to SE studies performed at an academic medical center. We hypothesized that the 2011 AUC would classify a greater number of SE studies compared with the 2008 AUC, leading to a more comprehensive determination of the rates of appropriate and inappropriate studies.
Methods
Study Population
A retrospective chart review was conducted on consecutive SE studies performed from July 1 to November 30, 2011, at the Massachusetts General Hospital Cardiac Ultrasound Laboratory. Exclusion criteria were (1) age < 18 years, (2) SE imaging ordered as part of a research study, (3) insufficient information in the electronic medical record (EMR) to make an accurate determination of appropriateness, and (4) inability to access the EMR because of privacy restrictions.
The study protocol was reviewed and approved by the Partners Healthcare institutional review board. The investigators had no contact with the patients or ordering health care providers, nor did they attempt to influence the care of the patients in any way.
Data Collection
Detailed review of the EMR was performed. Signs and symptoms and the reasons for SE studies were abstracted from the study orders and the EMR. Ordering of SE studies is performed electronically for inpatients and using paper orders for outpatients. There are no decision support tools to guide appropriate ordering of SE examinations embedded within the EMR or the paper ordering system. The EMR is a comprehensive database including physician notes, imaging results, and laboratory findings and captures the clinical activities of the health care network.
Three study investigators (R.S.B., V.K., and R.B.W.) were involved in the review of the EMR to determine indications for the SE studies. Investigators classified the clinical indication for each SE examination using both the 2008 and 2011 AUC, allowing the determination of an appropriateness rating (appropriate, uncertain, or inappropriate). If the reason for the SE study did not have a corresponding indication in the 2008 or 2011 AUC, it was considered unclassifiable. Specifically, two investigators independently reviewed each SE order blinded to the other reviewer and assigned the clinical indication number and appropriateness rating. Agreement was required by the two reviewers. For SE examinations for which there was no initial agreement on appropriateness rating, review of the study orders and EMR was conducted by a third investigator to achieve consensus on both the appropriateness rating and the clinical indication number. If there was initial agreement on appropriateness rating but not on clinical indication number, the two reviewers reconciled the clinical indication. Review by a third investigator was required for 10 SE studies. If a patient underwent multiple SE studies during the study time period, each study was considered separately and included in the analysis.
Clinical data collected included patient demographics (age, sex, and cardiac risk factors), previous cardiac procedures (coronary artery bypass grafting, valve surgery, and percutaneous coronary intervention), comorbidities (angina, heart failure, prior myocardial infarction, atrial fibrillation, peripheral arterial disease, chronic obstructive pulmonary disease, chronic kidney disease, connective tissue disease, dementia, and cancer), and hospitalization within the past 6 months.
We also assessed the impact of the SE results on patient management by a review of the EMR. We defined a change in patient management as an SE finding that resulted in cardiac catheterization or coronary revascularization, a change in medications, a cancellation of a preplanned surgical procedure, or a hospitalization.
SE Imaging
All SE examinations were performed at the Massachusetts General Hospital Cardiac Ultrasound Laboratory using either supine bicycle exercise or intravenous dobutamine infusion. Adjunctive atropine was administered in the dobutamine studies as clinically needed to achieve target heart rate. All SE studies were performed using a Philips iE33 ultrasound machine (Philips Medical Systems, Andover, MA). All SE studies were performed in accordance with the recommendations of the American Society of Echocardiography and were interpreted by level III and/or board-certified echocardiographers using visual assessment of left ventricular wall thickening and motion. The use of intravenous contrast administration to improve left ventricular opacification was determined at the discretion of the echocardiographer who supervised the study. In our study, intravenous contrast was used in 24 SE examinations (9.5%). For the purpose of this study, SE results were determined from the final clinical report in the EMR.
Statistical Analysis
Categorical variables for ordering characteristics, patient demographics, and appropriateness ratings using the 2008 and 2011 AUC were compared using χ 2 or Fisher’s exact tests as required. Continuous variables were compared using analysis of variance and are reported as mean ± SD. Statistical significance was indicated by a two-tailed P value < .05.
Results
Study Population
Two hundred eighty-eight consecutive SE examinations performed at the Massachusetts General Hospital Cardiac Ultrasound Laboratory were reviewed. Of these, 33 were excluded because of inability to access the EMR because of privacy issues, and three were excluded because of insufficient data to allow accurate classification. Therefore, 252 SE studies were included in the final analysis. These studies were ordered by 83 different health care providers.
Patient characteristics are highlighted in Table 1 . The patient population was relatively young, more commonly male, and had histories of cardiac risk factors, including diabetes mellitus (26%), hypertension (65%), hyperlipidemia (48%), and smoking (46%). In addition, 34% of patients had chronic kidney disease, attributable to the large number of pretransplantation assessments.
| Characteristic | Value |
|---|---|
| Age (y) (mean ± SD) | 58.1 ± 12.2 |
| Men | 145 (58%) |
| Previous MI | 32 (13%) |
| Previous PCI | 21 (8%) |
| History of valve surgery | 3 (1%) |
| History of CABG | 8 (3%) |
| Angina | 88 (35%) |
| Heart failure | 26 (10%) |
| Hospitalization in past 6 mo | 31 (12%) |
| Diabetes mellitus | 65 (26%) |
| Hypertension | 165 (65%) |
| Hyperlipidemia | 122 (48%) |
| History of smoking | 115 (46%) |
| Present smoking | 19 (8%) |
| Atrial fibrillation | 28 (11%) |
| Chronic kidney disease | 85 (34%) |
| Cancer | 45 (18%) |
| COPD | 9 (4%) |
| Peripheral arterial disease | 14 (6%) |
The types of ordering providers for SE studies are detailed in Table 2 . Cardiologists (50%) were the most common ordering providers, and 63% of patients had been seen at least once by a cardiologist in the preceding 6 months. The next largest ordering provider groups were hepatologists (19%) and nephrologists (16%), who ordered SE studies as part of pretransplantation assessments.
| Characteristic | n (%) |
|---|---|
| Ordering service | |
| Cardiology | 125 (50) |
| Hepatology | 49 (19) |
| Nephrology | 40 (16) |
| General internal medicine | 22 (9) |
| Surgery | 12 (5) |
| Other | 4 (2) |
| Seen by cardiologist within the past 6 months | 159 (63) |
| Previous TTE in the past 6 mo | 39 (15) |
| Other cardiac imaging (CT or SPECT) in the past year | 27 (11) |
SE Results and Further Management
The overall rate of abnormal SE results in our study was low. Only 7% of studies had findings positive for ischemia. Another 5% of SE studies were positive for new or worsening valvular disease or worsening left ventricular outflow tract obstruction. In total, 10% of patients underwent cardiac catheterization as a consequence of the SE results. Of those patients who underwent cardiac catheterization, 48% had no evidence of significant (≥70%) epicardial coronary stenosis, 12% had significant left anterior descending coronary artery stenosis, 8% had significant right coronary artery stenosis, 8% had significant left circumflex coronary artery stenosis, and 24% had multivessel coronary artery disease (CAD). The percentages of inappropriate and appropriate SE studies that affected clinical management were not significantly different (7% vs 14%, P = .21).
SE Study Classification
Results of SE study classification using the 2008 and 2011 AUC are shown in detail in Table 3 . The 2011 AUC classified a significantly higher number of SE studies than the 2008 AUC (88% vs 50%, P < .001). The rate of appropriate studies (of those classifiable) was significantly lower using 2011 AUC compared with the 2008 AUC (48% vs 83%, P = .006). The rate of inappropriate studies was significantly higher using the 2011 AUC (48% vs 11%, P = .004). The rate of uncertain studies was low and similar using each version of the AUC. Perioperative risk assessment before noncardiac solid-organ transplantation represented 106 (42%) of the total number of SE examinations in our study. When these studies were excluded, of the remaining 146 SE studies, the appropriate, inappropriate, and uncertain rates were 69%, 22%, and 9%, respectively.
| 2008 | 2011 | P | |
|---|---|---|---|
| Classifiable | 126 (50%) | 221 (88%) | <.001 |
| Appropriate of those classifiable | 104 (83%) | 105 (48%) | .006 |
| Inappropriate of those classifiable | 14 (11%) | 105 (48%) | .004 |
| Uncertain of those classifiable | 8 (6%) | 11 (4%) | .82 |
The most common appropriate indications for SE in our study were indications 116 (evaluation of ischemic equivalent with intermediate pretest probability of CAD with an interpretable electrocardiogram and ability to exercise), 161 (vascular [high-risk] surgery with one or more clinical risk factors and poor or unknown functional capacity), and 118 (evaluation of ischemic equivalent with high pretest probability of CAD), which accounted for 41%, 25%, and 16% of all appropriate SE studies, respectively. Common inappropriate indications for SE studies were indication 159 (vascular [high-risk] surgery with moderate to good functional capacity) and 155 (intermediate-risk surgery with moderate to good functional capacity), which represented 80% and 10% of all inappropriate SE studies, respectively.
Details of the reclassification of SE examinations using the 2011 AUC compared with the 2008 AUC are reported in Table 4 . In total, 100 SE studies (40%) changed classification. Of those, 95 that did not have corresponding clinical indications in the 2008 AUC were newly classified with the 2011 AUC. The majority of these previously unclassified studies ( n = 92) were now classified as inappropriate using the 2011 AUC, with the remainder ( n = 3) reclassified as appropriate. Of the SE studies reclassified as inappropriate, almost all (90%) were for indication 159 or 155. The remainder of the reclassified SE studies were for perioperative assessment for low risk-surgery (indication 154), evaluation of infrequent premature ventricular contractions (indication 131), and the assessment of symptomatic severe aortic stenosis (indication 192). Newly appropriate studies included assessment of moderate mitral regurgitation (indication 195) and newly diagnosed heart failure or left ventricular dysfunction (indication 128), though the numbers of these studies were small. Four SE examinations went from either inappropriate or appropriate to uncertain, with the most common indication being 157 (intermediate-risk surgery with poor or unknown functional capacity).
