Background
The recent report that appropriately performed echocardiographic examinations result in active changes in management in only one third of patients has challenged the validity of current appropriate use criteria. Limited information exists about the clinical importance of transthoracic echocardiography (TTE) to guide management and rule out important alternative pathology.
Methods
The clinical impact of inpatient TTE performed at the Mayo Clinic over a 20-week period between October 14, 2013, and March 3, 2014, was investigated. Studies were included if they were ordered within 72 hours of admission, and treating physicians participated in a real-time survey regarding the clinical importance of TTE. Appropriate use was determined by two independent investigators, with differences adjudicated by a third investigator. Clinical impact was derived from physicians’ survey responses and independently confirmed by chart review.
Results
Of the 539 transthoracic echocardiographic examinations included in this study, 512 (95%) were appropriate, 16 (3%) may be appropriate and 11 (2%) rarely appropriate. Although only 48% of participating physicians actively changed management on the basis of findings on TTE, 97% responded that TTE answered their clinical questions, and 95% would still order TTE in similar clinical contexts.
Conclusions
Most early inpatient transthoracic echocardiographic studies at our institution were appropriate and answered specific clinical questions important for management decisions in the opinion of the treating physician. Confirming a plan of care already in place and ruling out alternative pathology may be as important clinically as uncovering new findings or changing management.
Highlights
- •
Most echocardiographic examinations performed within 72 hours of hospital admission are appropriate.
- •
Most physicians surveyed thought the echocardiographic studies answered their clinical questions.
- •
Only 48% of physicians changed management, but 95% would order additional TTE.
- •
Echocardiography uncovered unexpected cardiac abnormalities in 33% of cases surveyed.
Echocardiography is a versatile diagnostic modality, providing real-time structural and physiologic information with virtually no side effects. However, in an increasingly cost-conscious health care environment, the expanding use of diagnostic imaging has come under scrutiny. Increased access to imaging modalities in the early 2000’s led to exponentially increased clinical use, with a rate of growth greater than any other physician-ordered services for Medicare beneficiaries. As government and private insurers took steps to limit the number of imaging studies by limiting reimbursement or requiring testing preauthorization, professional organizations increased their efforts to document the value of various imaging modalities and create guidelines for meaningful clinical use.
The American College of Cardiology Foundation (ACCF) and American Society of Echocardiography (ASE) appropriate use (AU) criteria (AUC), for the most part, lack the evidence of randomized controlled trials and were developed mainly by expert consensus. The paradigm that an “appropriate” test is necessarily a clinically meaningful test has been challenged recently. A retrospective analysis concluded that although 90% of echocardiographic examinations performed within 1 month at a large university medical center were “appropriate,” only one third resulted in “active changes in management.” However, the assumption that only an active change in management is consistent with efficient use of resources has been criticized.
Although advanced imaging can enhance patient care by providing valuable insights that often cannot be obtained from clinical information alone, there is still an unmet need for a body of evidence linking noninvasive imaging, including echocardiography, with physician decision making and patient outcomes. In the absence of clear patient outcomes data, surrogates have been used, such as the capacity of a test to change patient management. With this assumption, tests that do not change management could be considered unnecessary health care expenses and be potentially eliminated from the therapeutic armamentarium. However, we believe stronger evidence is needed before concluding that less is indeed more in the case of appropriately performed echocardiography. Thus, we set forth to assess the clinical utility of echocardiography in a systematic and prospective manner, with direct input from the treating physicians involved in patient care.
Methods
Study Design and Recruitment
This study was approved by the Mayo Clinic Institutional Review Board and undertaken at the hospitals associated with the Mayo Clinic in Rochester, Minnesota. All adult inpatient transthoracic echocardiographic studies performed within 72 hours of hospital admission between October 14, 2013, and March 3, 2014, were screened for potential study enrollment on the day of the study requisition. The 72-hour time line allowed the inclusion of examinations requested by the hospital inpatient team for clinical evaluation of the presenting symptoms, rather than late reassessment during a prolonged hospital course. Transthoracic echocardiography (TTE) performed in the emergency department before admission or in the cardiac catheterization laboratory for procedure guidance, or ordered for left ventricular (LV) assist device follow-up, congenital heart disease, or research were excluded from this study. The ordering staff physicians received an electronic invitation to participate in a survey regarding the clinical impact of TTE for their patients. An informed-consent statement was embedded in the invitation e-mail, with a link to the survey; only those transthoracic echocardiographic examinations for which the survey was completed were included in the final analysis.
Echocardiography laboratory requisitions were screened every day for eligible studies. In the case of physicians ordering multiple studies in 1 day (which is often the case for cardiologists on the inpatient service), up to three surveys were sent for three different examinations in 1 day, depending on how fast physicians answered to the first request. The order in which surveys were sent to each physician was determined by the order in which TTE requisitions were received by the echocardiography laboratory. Attempts to capture all eligible studies were made on subsequent days. Unanswered surveys were resent to the same ordering physician up to three times over 3 consecutive days or to an alternatively designated “treating” physician identified from the medical record (new physician writing daily progress notes, cardiology fellow, or consulting staff member). As soon as physicians answered surveys, they were sent more surveys for subsequent examinations, if available, until all examinations were captured or there was no response.
Electronic Survey
A brief five-question survey was designed to capture the opinions of the treating physicians regarding the clinical utility of TTE in a real-time fashion, upon seeing echocardiographic results for each patient. Selected questions were focused on TTE’s ability to establish a primary cardiac diagnosis, the types of changes in management that followed, whether there were any unexpected findings on TTE, and whether the physician, having the benefit of this new information, would still order another study in a similar patient in the future. With feedback from our participating physicians after the first 3 weeks of enrollment, the survey was updated to include an additional question: “Did the echocardiogram answer the clinical question you ordered it for?” ( Supplemental Table 1 ).
Physician responses were automatically collected via the Research Electronic Data Capture software, a validated platform for electronic collection and management of research data. In the case of duplicate survey responses received for the same patient encounter from two different physicians (four cases [0.7%]), only the first received survey was included in the data analysis. Therefore, in the final analysis, each completed survey corresponded to a unique echocardiographic examination and a unique patient encounter.
Defining Appropriateness
Echocardiographic studies were classified as “appropriate,” “uncertain,” or “inappropriate” in accordance with the 2011 ACCF and ASE AUC at the time of the data analysis and subsequently renamed “appropriate,” “may be appropriate,” or “rarely appropriate” in accordance with the most recent terminology. Independently, two investigators (A.C. and A.K. or A.C. and E.J.D.) determined AU from detailed electronic medical record (EMR) review, without knowledge of the survey responses. When more than one AU category fit a particular scenario, the most clinically relevant indication was selected as the reason for TTE (i.e., for AU adjudication). A third experienced investigator (S.V.P.) adjudicated disagreements on AU category between the first two investigators by independently reviewing the EMR.
Independent EMR Review
To confirm the changes in care management reported on the survey, investigators reviewed in detail the EMRs of all patients included in the final analysis. This comprehensive review included hospital admission notes, physical examination on admission, prior imaging, relevant consults, hospital course, medication changes, and other cardiac diagnostic tests or procedure notes. Finally, the hospital summary was reviewed for the main diagnoses on discharge, relevant medication changes, and follow-up plans.
Echocardiographic Data
The main findings on index TTE were compared with the most recent previous echocardiographic study when a prior study had been performed. Abnormal findings were considered new when the patient had no previous echocardiogram on record or the results of index TTE were significantly different from those of the most recent earlier study. Significant changes were defined as new or worsening LV systolic dysfunction (e.g., new LV ejection fraction < 50%, LV ejection fraction decrease > 10% of previous, new regional wall motion abnormalities), new right ventricular systolic dysfunction, increase in pulmonary arterial pressure of >10 mm Hg, new or worsening valvular dysfunction (moderate degree or worse), new pericardial effusion, evidence of constrictive pericarditis, new or enlarging aortic aneurysm, or aortic dissection.
Echocardiographic Studies Providing New Information versus Normal or “Unchanged” Findings
In a subanalysis of the main study, echocardiograms were classified on the basis of the presence or absence of significant new findings as defined above. Changes in management were recorded through detailed review of the medical record between the time of TTE and the time of hospital discharge.
Statistical Analysis
Because of the high prevalence of appropriate echocardiographic examinations and the known limitations of the κ statistic in this case, agreement between reviewers in appropriateness grading was assessed with two measures. First, overall agreement was calculated as the proportion of echocardiograms in agreement divided by the total number. Second, the percentage of agreement within each AU category was calculated and presented. These measures have been previously suggested to overcome limitations of the κ statistic when data are skewed toward one category (i.e., appropriate studies).
Differences in the overall percentage of appropriate studies ordered by each medical specialty were tested with the Fisher exact test. P values < .05 were considered to indicate statistically significant differences in appropriateness ordering pattern among the different specialties. All statistical analyses were performed with commercially available software (JMP version 10.0; SAS Institute Inc, Cary, NC).
Results
All adult TTE requisitions received by the Mayo Clinic Echocardiography Laboratory during the recruitment period of this study were screened for potential enrollment ( Figure 1 ); 1,055 electronic surveys were sent to ordering physicians for early performed inpatient TTE within 72 hours of admission, and 543 physician responses were received (a 51% response rate). Four surveys were excluded because they represented duplicate responses (0.7% of survey responses); therefore, a total of 539 physician surveys were collected and included in the final analysis, corresponding to 539 unique patient encounters and transthoracic echocardiographic examinations.
Inpatient TTE was ordered by all medical and surgical specialists ( Figure 2 ). Cardiologists ordered the largest number of inpatient examinations (635 [50% of eligible studies]) and accounted for the largest absolute number of survey responders (212 [39% of survey responders]), with a study enrollment rate of 33% from their eligible examinations. Internists ordered a smaller number of studies (226 [18% of eligible examinations]) and answered 165 surveys, resulting in a study enrollment rate of 73% from their eligible examinations. Critical care physicians and anesthesiologists also had high participation rates (53% and 62%, respectively).
Of the 539 physicians who participated in our survey, 212 (39%) were cardiologists, 165 (31%) were internists, 34 (6%) were medical subspecialists, 76 (14%) were pulmonary and critical care physicians, 31 (6%) were anesthesiologists, and 21 (4%) were surgeons.
Survey Validation
We received 88 physician responses from the initial five-question survey and 451 physician responses from the six-question survey, for a total of 539 unique survey responses corresponding to 539 early-performed inpatient echocardiographic examinations. The two survey groups were initially analyzed independently for the five common questions, and the results were found to be similar. Thus, the surveys were combined in the final analysis.
Clinical Indications for TTE and AU Category
Classification of examinations as appropriate, may be appropriate, or rarely appropriate, and the main clinical indications for TTE, are summarized in Table 1 . The first two investigators agreed on the AU category designation in 510 of the 539 studies (94.6% overall agreement). Of the 512 appropriate studies, the first two investigators agreed in 500 cases (97.6% agreement). There was a lower rate of agreement for studies that may be appropriate under certain circumstances (44% agreement for 16 studies) and rarely appropriate studies (36.4% agreement for 11 studies). Of the 539 patients, 236 (44%) were women (mean age, 70 ± 16.4 years), and 303 (56%) were men (mean age, 66 ± 15 years). Clinical indications for TTE in our cohort of hospitalized patients, classified by AU category, are found in Supplemental Tables 2 to 4 .
| Table | Indication classification | A | M | R | No. of transthoracic studies |
|---|---|---|---|---|---|
| 1 | General evaluation of cardiac structure and function | 114 | 4 | 7 | 125 |
| 2 | Cardiovascular evaluation in an acute setting | 225 | 7 | 1 | 233 |
| 3 | Evaluation of valvular function | 26 | 1 | 3 | 30 |
| 4 | Evaluation of intracardiac and extracardiac structures | 41 | 0 | 0 | 41 |
| 5 | Evaluation of aortic disease | 4 | 0 | 0 | 4 |
| 6 | Evaluation of hypertension, heart failure, or cardiomyopathy | 102 | 4 | 0 | 106 |
| Total | 512 | 16 | 11 | 539 |
AU and Medical Specialty of Ordering Physicians
The proportion of appropriate TTE in our cohort varied according to the medical specialty of ordering physicians ( P = .006), with cardiologists ordering a higher percentage of appropriate studies (98.5%) compared with internists (93%), critical care physicians (93%), anesthesiologists (90%), and other medical subspecialists (88%). Medical subspecialists had the highest percentage of may be appropriate studies (11%). Surgeons ordered only a small number of studies in certain high-acuity clinical situations; 95% of these had appropriate indications.
Clinical Impact of TTE
According to the physician survey, 259 studies (48%) led to changes in management. Changes in management occurred more often in appropriate studies: 50% of the appropriate examinations led to changes in management, whereas only 25% of the may be appropriate and 18% of the rarely appropriate studies led to changes in management ( P = .016). TTE uncovered primary cardiac etiologies for the presenting symptoms in 44% of cases, according to the survey. In 23% of cases, normal results on TTE within a specific clinical context helped direct attention toward noncardiac causes. The types of therapeutic changes are summarized in Table 2 .
| Question | No. of transthoracic studies | % of total |
|---|---|---|
| Did TTE uncover a primary cardiac etiology for the current acute presentation? | 237 | 44 |
| Did it help direct attention to a noncardiac etiology? | 122 | 23 |
| Did TTE lead to a change in your management? | 259 | 48 |
| 1. New cardiovascular consult | 65 | 12 |
| 2. New other consult (ID, CV surgery, EP) | 35 | 6 |
| 3. Other echocardiogram (TEE, stress echocardiography) | 16 | 3 |
| 4. Cardiac CT/MRI | 13 | 2 |
| 5. Cardiac catheterization | 37 | 7 |
| 6. New medication changes | ||
| a. β-blockers | 51 | 9 |
| b. ACE inhibitors/ARBs | 37 | 7 |
| c. Diuretics | 62 | 12 |
| d. IV fluids | 26 | 5 |
| 7. New procedures (cardioversion, PPM/ICD, CRT, ablation, pericardiocentesis, etc) | 59 | 11 |
| Total number of surveys | 539 | 100 |
After seeing the results of TTE, 95% of physicians (511 of 539) responded that they would still order TTE for similar patients in the future, and 97% of physicians (437 of 451) responded that TTE had answered their clinical questions ( Table 3 ).
| Question | Yes | % of total | No. of surveys |
|---|---|---|---|
| Were there any unexpected findings on TTE compared with your initial clinical assessment? | 180 | 33 | 539 |
| Knowing what you know now, would you choose to order additional TTE for a similar patient with similar admission problems? | 511 | 95 | 539 |
| Did TTE answer your clinical question? ∗ | 437 | 97 | 451 |
∗ This question was added at the second iteration of the survey and was answered by 451 physicians.
Among the 28 indications for which interviewed physicians would not order additional TTE in the future (5% of cases), 25 studies were appropriate, one was uncertain, and two were inappropriate. In four cases with appropriate indications for TTE, physicians went on to order other imaging studies (transesophageal echocardiography, stress echocardiography, or nuclear perfusion study) that answered the clinical questions.
Unexpected Findings
According to physicians, 180 of the 539 studies (33%) showed at least one unexpected finding compared with their initial clinical assessment. EMR review of these cases revealed a number of new echocardiographic abnormalities, which are classified in Table 4 .
| Unexpected finding | No. of transthoracic studies (%) ∗ † |
|---|---|
| New LV dysfunction (EF decrease, new RWMA) | 106 (59) |
| New RV dysfunction | 52 (29) |
| New or worsening valvular dysfunction (moderate or greater) | 35 (19) |
| Increase in PAP >10 mm Hg | 19 (11) |
| Thrombus, vegetation, or mass | 12 (7) |
| New pericardial effusion (moderate or greater) | 10 (6) |
| New or worsening aortic aneurysm | 7 (4) |
| Aortic dissection | 2 (1) |
| Constrictive pericarditis | 1 (1) |
∗ Unexpected results were found in 180 cases.
† Percentages do not total 100% because more than one unexpected result was found in some cases.
Physician’s Input in the Presence or Absence of a Recent Previous Echocardiographic Study
In a separate analysis, among the 539 patients, 153 (28%) had undergone prior TTE within the past year, and 386 (72%) had not. Overall changes in management occurred in 69 of 153 of patients (45%) with recent TTE and in 190 of 386 of patients (48%) without recent TTE, according to the physicians’ survey responses.
For patients with recent TTE, the physician questionnaire revealed that the new TTE supported active cardiovascular changes in 36% of cases, led to no active changes but supported therapeutic plans already in place or ruled out important alternative pathology in 56% of cases, and was of no help in 8% of cases. For the 386 patients without recent TTE over the previous year, the physician questionnaire revealed that TTE supported active cardiovascular changes in 39% of cases, led to no active changes but supported therapeutic plans already in place or ruled out important alternative pathology in 58% of cases, and was of no help in 3% of cases ( P = .07). In both categories of TTE, physicians endorsed similar rates of further specialty consults, further cardiac testing, and cardiovascular medication changes.
Independent EMR Confirmation of Changes in Management
Objective review of medical records confirmed the changes in management noted by the treating physicians on the survey in terms of new cardiac medication changes, new consults, and new diagnostic tests and procedures. The number of management changes noted in the EMR was greater than the number listed on the survey, which was often completed as soon as the results of TTE had become available to physicians. For example, there were new cardiovascular medication changes at the time of hospital discharge for 58% of patients; according to the survey, cardiac medication changes were made for 35% of patients.
Echocardiograms Providing New Information versus Normal or “Unchanged” Findings
Among the 539 transthoracic echocardiographic studies enrolled in our study, 242 (45%) provided new data (either worsening transthoracic echocardiographic abnormalities compared with the prior study or abnormal results when there was no previous study on record), and 297 (55%) did not show any new abnormalities (either the results were abnormal but identical to those of a previous examination on record, or the results were normal).
For the 242 examinations with new information, there were 170 further tests and procedures documented in the medical record; among these there were 20 transesophageal echocardiographic examinations, 10 stress tests, one repeat transthoracic echocardiographic examination for bubble study, 71 catheterizations (three additional catheterizations were recommended but declined or deferred), 33 chest computed tomographic or cardiac magnetic resonance imaging studies, 19 electrophysiologic procedures, three Holter monitors or tilt tests, three pericardiocenteses, and eight cardiovascular surgical procedures (including coronary artery bypass grafting, right ventricular biopsy, and new LV assist device implantation). LifeVests were offered for two patients, with the possibility of eventually placing implantable cardioverter-defibrillators. For the 297 “unchanged” transthoracic echocardiographic studies, there were 139 further tests and procedures; among these there were 18 transesophageal echocardiographic studies, five stress tests, 37 catheterizations, 40 chest computed tomographic or cardiac magnetic resonance imaging studies, 19 electrophysiologic procedures, eight Holter monitors or tilt tests, three pericardiocenteses, and nine cardiac surgical procedures. The most significant difference in further testing between transthoracic echocardiographic studies with new abnormalities and “unchanged” studies was the pursuit of cardiac catheterization: 71 of 242 (29%) versus 39 of 297 (12%), respectively ( P < .0001).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
