Clinical utility and management impact of M-mode echocardiography

Clinical Utility and Management Impact of M-Mode Echocardiography

LEE GOLDMAN, M.D. PETER F. COHN, M.D.*

GILBERT H. MUDGE, Jr., M.D. BEVERLY

HASHIMOTO,

HERBERT

SHERMAN,

JOSHUA

WYNNE,

MARGARET

M.D.7 D.E.E.

M.D.

FLATLEY,

B.A.

Boston, Massachusetts

From the Cardiovascular Division and the Division of General Medicine, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston. Massachusetts. This study was supported in part by a grant (18-P975 17/1-O 1) from the Health Care Financing Administration. Dr. Goldman is a Teaching and Research Scholar of the American College of Physicians, and a Henry J. Kaiser Family Foundation Faculty Scholar in General Internal Medicine. Dr. Wynne is a Young Investigator of the National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland. Requests for reprints should be addressed to Dr. Lee Goldman, Department of Medicine, Brigham and Women’s Hospital, 75 Francis Street. Boston, Massachusetts 02 115. Manuscript accepted September 9, 1982. * Present address: Cardiology Division, Department of Medicine, S.U.N.Y. Health Science Center, Stony Brook, New York. t Present address: Department of Radiology, University of California, San Francisco, California.

To determine the clinical utility and management impact of M-mode echocardiography, 182 echocardiograms were analyzed at a university teaching hospital. The physicians who ordered the echocardiograms said that 12 percent provided crucial information that was not available from other tests and that 28 percent resulted in a change in patient management. According to two independent board-certified cardiologist-reviewers, 88 percent of echocardiograms were appropriately ordered, but only 15 echocardiograms (8 percent) were actually needed for a change to a new and appropriate management. According to the reviewers, the 77 Group I M-mode echocardiograms (those ordered to evaluate left ventricular function, lefl atrial size, potential cardiac sources of emboli, or the possibility of bacterial endocarditis, or those ordered in patients who, according to the ordering physician, had undergone or would undergo catheterization regardless of the results of echocardiography) were less likely than the 105 Group II M-mode echocardiograms (those ordered to evaluate possible mitral valve prolapse, hypertrophic cardiomyopathy, valvular function, or the pericardium) to be ordered appropriately, to provide helpful information, or to provide crucial results. Group I echocardiograms had reviewer-assessed appropriate management impact in onty one case (1 percent) compared with a 13 percent rate of management impact for Group II M-mode echocardiograms (p
July 1993

The American Journal of Medicine

Volume 75

49

IMPACT OF ECHOCARDIOGRAPHY-GOLDMAN

ET AL

actual clinical practice if it is not always ordered or used appropriately. To evaluate the clinical utility and management impact of echocardiography, we surveyed 182 echocardiograms that were ordered at Brigham and Women’s Hospital as part of routine clinical practice during a six-week period. Each of these echocardiograms was subjected to a detailed analysis that included an interview with the physician who ordered the test as well as independent reviews by two board-certified cardiologists. METHODS This analysis is based on 196 consecutive M-mode echocardiograms that were ordered by a variety of physicians and performed in the Noninvasive Cardiac Laboratory of the Brigham and Women’s Hospital. At the time of the study, two-dimensional echocardiograms were available on an investigational basis but could not be routinely ordered by physicians. When each patient arrived in the Noninvasive Cardiac Laboratory, the reason for which the echocardiogram was ordered was obtained from the requisition form, and a standardized brief history and physical examination form was completed by the cardiology fellow who was to perform or supervise the procedure. Echocardiography was then routinely performed and the results were evaluated as part of the standard clinical practice at our hospital. Because it was not feasible for two reviewers to interview each ordering physician, and because we feared that duplicate interviews would not be tolerated by busy clinicians, the following protocol was developed. As soon as possible after the echocardiogram was evaluated and the results were available, each ordering physician was contacted by a physician-researcher who asked the ordering physician these questions: For what clinical reasons (to make a diagnosis, to plan or monitor therapy, to determine the extent of a known disease or estimate prognosis, to reassure yourself, to reassure the patient, to reassure another physician) did you order the echocardiogram? What anatomic structures were you most interested in? Did the echocardiogram provide useful results? Did the echocardiogram lead to a change in management plans? Were any tests avoided because of the echocardiograms? How do you plan to manage the patient? Do you wish you had ordered another test rather than echocardiography? We attempted to interview physicians before they knew the echocardiographic results. Although the interview usually occurred before the physicians received the formal echocardiographic evaluation, we could not prevent clinicians from obtaining preliminary results for their own patients. The interview was conducted in a nonjudgmental fashion by a physician who had no specialized cardiology training and who did not participate in the subsequent review process. Physicians’ responses were recorded verbatim. Two independent reviewers were provided with the ordering physician’s responses to the interview, the cardiology fellow’s history and physical examination, and the final interpretation of each echocardiogram. Each of the echocar-

50

July 1993

The American Journal of Mediclne

Volume 75

diograms was reviewed by the director of clinical cardiology at our hospital; for half the patients, the second review was performed by his successor, and for the other half of the patients it was performed by another board-certified faculty cardiologist. Each of the reviewers was considered to be up-to-date in terms of the uses of echocardiography, but none was primarily involved in echocardiographic research nor had primary clinical responsibilities in the Noninvasive Cardiac Laboratory. Without knowing the identities of the patient or of the ordering physician, the reviewers answered 10 questions for each of the echocardiograms: (1) Did the ordering physician have a clear clinical purpose for obtaining the echocardiogram? (2) Did the reviewer share the questions or uncertainties for which the echocardiogram was obtained? (3) Was echocardiography the best test to use in terms of the stated reasons for ordering it at that particular time in the evaluation of the patient? (4) Based upon available scientific data, did the echocardiogram have documented or at least presumed efficacy for answering the question posed by the ordering physician? (5) Did the ordering physician understand the meaning of the report of the echocardiographic results? (6) Did the ordering physician appropriately integrate the echocardiographic result with other available data? (7) Was the echocardiographic result useful in answering the questions that prompted the physician to order the test? (8) Assuming a proper interpretation for all data, how useful would the echocardiographic result have been to someone trying to answer the questions originally posed by the ordering physician? (9) Assuming a proper interpretation for all available clinical and laboratory data, should the echocardiographic result have led to a change in patient management? (10) Was the patient appropriately managed? For each of these questions, a reviewer’s opinion could be positive, neutral, or negative. Because of the wide variety of clinical situations, the reviewers used implicit criteria derived from their familiarity with the clinical value of echocardiography and their knowledge of how these procedures compared with competing or complementary techniques at the time the procedure was obtained. Our reviewers and our review grading system were chosen to ensure that reasonable variations in practice patterns would not be judged negatively. The review system graded an overall review on each question as positive if one reviewer gave a positive response (regardless of the other reviewer’s response) or if both reviewers were neutral; the overall review on a question was considered negative if both reviewers gave a negative response or if one reviewer gave a negative response and the other reviewer was neutral. In Question 8, positive responses were further subdivided: the reviewers termed a study crucial if it supplied conclusive information not previously available, contributory if it provided important data for a patient in whom no test ordered thus far could be diagnostic, or confirmatory if it substantiated a conclusion that was already reasonably well established but not proved conclusively. In Question 8, the usefulness of the study was classified according to the highest degree of usefulness indicated by either of the two reviewers. Of the 196 consecutive echocardiograms that were eligible for this study, 14 were eliminated because the ordering

IMPACT OF ECHOCARDIOGRAPHY-GOLDMAN

physician was unavailable or uncooperative. The remaining 182 echocardiograms represent the basis for this report. This sample size was believed to be sufficiently large to provide a reasonable assessment of contemporary practice patterns at our hospital. Our detailed review process was very timeconsuming and made a larger study impractical. Statistical Methods. Univariate comparisons were performed using the Fisher exact test. Statistical significance was defined as p CO.05 RESULTS is often shared in a teaching hospital, as best as could be determined, the type of physician most responsible for ordering the echocardiogram was an intern or resident in 108 cases (59 percent), a cardiology fellow in 24 cases (13 percent), a full-time faculty internist or surgeon in 17 cases (9 percent), a full-time faculty cardiologist in 26 cases (14 percent), and a private internist or cardiologist in seven cases (4 percent). Only five of 182 echocardiograms were ordered by physicians whose official duties included echocardiographic evaluation in our hospital. Interview with the Ordering Physician. When requesting the echocardiogram, the ordering physician often had more than one clinical purpose in mind, including to make a diagnosis in 129 cases (7 1 percent), to estimate prognosis or to define the extent of a known disease in 111 cases (61 percent), to guide or monitor management in 107 cases (59 percent), to reassure the ordering physician in 5 1 cases (28 percent), to reassure another physician in 31 cases (17 percent), and to reassure the patient in 27 cases (15 percent). From an anatomic standpoint, physicians sometimes were interested in evaluating more than one cardiac structure, but in each case they were able to identify a principal anatomic reason for ordering the echocardiogram (Table I). In 30 cases (16 percent), cardiac catheterization either already had been performed or, according to the ordering physician, was scheduled and would be performed regardless of the result of echocardiography. According to the interview with the ordering physician, the echocardiogram provided crucial information that was not available from other tests in 22 cases (12 percent), provided contributory information when no single test could be diagnostic in 79 cases (43 percent), provided confirmatory information when the diagnosis was already reasonably well established in 56 cases (3 1 percent), and was not useful in 25 cases (14 percent). In 13 cases (7 percent), the ordering physician indicated that, in retrospect, he or she wished that a different test had been ordered. In 48 cases (26 percent), the ordering physician indicated that management had been changed in some way on the basis of the echocardiographic result. Although decision-making

TABLE I

ET AL

Principal Anatomic Reasons for Ordering Echocardiograms Numberof Echocardiagrams (Percent)

Evaluate left ventricular function Evaluate question of mitral valve prolapse or hypertrophic cardiomyopathy Evaluate other valvular function Evaluate pericardium or pericardial fluid Evaluate left atrial size, question of cardiac source of emboli, or definite or suspected bacterial endocarditis

28 (15) 20 (11)

a5 (47) 25 (14) 24 (13)

The reviewers believed that 181 of the 182 echocardiograms were ordered for clinical reasons; in 168 of these 181 cases, one or both reviewers shared the physician’s reasons for ordering the echocardiogram (Figure 1). In 165 cases, reviewers judged that the echocardiography was the best test at the time ordered, and in 182 cases, the reviewers believed that echocardiography could be efficacious for answering the physician’s clinical question. All four of these criteria were met in 157 cases (86 percent), and these 157 echocardiograms were therefore considered to be ordered appropriately. If all available clinical data were appropriately interpreted, the reviewers-judged that 166 of the 182 echocardiograms provided useful information for the ordering physician. Of these 166 potentially useful echocardiograms, at least one of the reviewers believed that 22 provided crucial information, another 73 provided contributory information when no single test could be diagnostic, and another 45 provided important confirmatory information. According to the reviewers, 175 (96 percent) of the 182 echocardiograms were correctly interpreted and integrated with other clinical data by the ordering physician. The reviewers believed that 142 echocardiograms (78 percent) met all of these criteria: thus, in these cases, the echocardiograms were considered to have been both appropriately ordered and properly used by the ordering physician (Figure 1). Changes in management: Although the ordering physicians indicated that management had been changed after evaluation of 48 (26 percent) of the echocardiograms, the reviewers believed that just 15 of these echocardiograms (8 percent) were actually needed for a change to a new and appropriate management (Table II). According to reviewers, the types of appropriate management changes that were based on the echocardiographic results included: changes regarding cardiac catheterization in five cases, pericardial procedures in four cases, and noninvasive Reviewers’ Assessments.

July 1993

The American Journal of Medicine

Volume 75

51

IMPACT

OF

ECHOCARDIOGRAPHY-GOLDMAN

ET

AL

1182

Were

there

order

the

Are

the

by

this

Is

the

to

be

clinical

reasons

to

the

the

for

ordering

reviewers? best

test

to

echocardiogram efficacious

for

order

at

present?

presumed

or

known

the

physician’s

indications? Is

the

echocardiogram

useful Did

to the

the

the it

the

needed in

result

ordering

ordering

interpret integrate Does

potentially

physician?

physician

correctly

echocardiogram with

other

result clinical

echocardiogram contribution

patient

result to

a

correct

and data?

make

a

change

management?

evaluation, therapy, or instructions

8% appropriate management

impact

in six cases (Table

III). Correlates

of the yield of echocardiography: According to the reviewers, M-mode echocardiograms that were ordered principally to evaluate left ventricular function, left atrial size, a cardiac source of emboli, or the question of endocarditis had management impact in only one (2 percent) of 52 cases. This one patient, who had the new onset of catheterization-documented aortic regurgitation and pulmonary edema three months previously, was suspected to have endocarditis despite 12 negative blood culture results; his echocardiogram

TABLE II

Changes in Management Based on Echocardiographic Results: Comparison of the Ordering Physicians and the Reviewers Number of Echocardiagrams (Percent)

No change in management according to ordering physician or reviewers Change in management reported by ordering physicians Reviewers agreed with management but said echocardiogram was not needed to determine appropriate management Reviewers disagreed with new management One or both reviewers said echocardiographic result definitely (seven cases) or possibly (eight cases) needed to change management and that actual final management was appropriate

52

studies\

echocardiogram? indications

shared Is

true

clinically-requested

July 1983

134 (74) 46 (26) 29 (16)

3 (2) 15 (6)

The American Journal of Medicine

Volume 75

Figure 1. Summary of the reviewers’ opinions on all 182 echocardiograms. In this flow chart, the percentages in the left column reflect the echocardiograms that were approved for all preceding criteria and for the criterion on that line. The percentages in the right column indicate the echocardiograms that were approved for all preceding criteria but were not approved for the criterion on that line.

demonstrated a flail aortic leaflet that had no obvious vegetations but that nevertheless suggested endocarditis and influenced his physicians to begin antibiotics before the operative specimen eventually grew enterococci. By comparison, M-mode echocardiograms ordered for other anatomic purposes (valvular function, the pericardium, hypertrophic cardiomyopathy, mitral valve prolapse) had management impact 12 percent of the time (14 of 130 cases; p = 0.08). Similarly, M-mode echocardiograms that were obtained in patients who, according to the ordering physician, already had or were scheduled to have cardiac catheterization regardless of the echocardiographic result were less likely to receive favorable reviews. Of 30 such echocardiograms, the only one with appropriate management impact according to the reviewers was in the same patient with aortic regurgitation just described. By means of a combination of these two factors-the principal anatomic reason for ordering the echocardiogram and the concomitant availability or pending availability of cardiac catheterization results regardless of the result of M-mode echocardiography-echocardiograms were divided into two groups (Table IV). Group I included 77 M-mode echocardiograms (42 percent of our sample) that were significantly less likely to be ordered appropriately, to be useful, to provide crucial information, or to have management impact than were the 105 M-mode echocardiograms in Group II. Although Group I M-mode echocardiograms comprised 42 percent of our sample, only one (7 percent) of the 15 echocardiograms with reviewer-assessed appropriate management impact was in Group I. Conversely, the usefulness and management impact of M-mode echocardiograms were not

IMPACT OF ECHOCARDIOGRAPHY-GOLDMAN

TABLE III

Cases in Which the Ordering Physician and at Least One Reviewer Agreed the Echocardiogram Appropriately Contributed to a Change In Management

Changesregardingcardiac catheterization(n = 5) Catheterizationavoided(suspectedaortic stenosisin two patients,mitral stenosisin one patient) Catheterizationperformed(suspectedaortic stenosisin one patient, lett atrial myxoma in one patient) Changesregardingpericardialprocedures(n = 4) Pericardialaspirationavoidedin a patientwith lymphomaand enlargingheart after radiationand chemotherapy Pericardialprocedureperformed(one pericardiocentesisin a patient with renal failure. open pericardial biopsy performed

3 2 1 3

in two patients with suspected tuberculous pericarditis) Changes in noninvasive evaluation, therapy, or instructions (n = 6) Radionuclide shunt study ordered for suspected atrial septal defect Antibiotics (begun in one patient whose massive aortic regurgitation was ascribed to endocarditis before the operative specimen eventually grew enterococci. and prescribed prophylactically for procedures in two patients with abnormal aortic valve echocardiograms) Patient instructions (palpitations ascribed to mitral valve prolapse in one patient, and one patient with questionable asymmetric septal hypertrophy identified and instructed to tell future physicians to avoid certain medications)

correlated with the ordering physician’s specialty or degree of echocardiographic expertise or with whether the ordering physician was a house officer, fellow, or attending physician. COMMENTS In a study of 1,000 patients, Merkiewicz and colleagues [8] indicated that echocardiograms established a diagnosis in 23 percent of patients who had evidence of heart disease but no firm clinical diagnosis before echocardiography. They also said that echocardiography demonstrated totally unexpected findings in 10

percent of the cases in whom a presumptive clinical diagnosis had been made. Pollick [9] noted that the introduction of echocardiography in his general hospital appeared to improve diagnostic accuracy, but he did not directly analyze the impact on management. By comparison, Grimmer et al [lo] found that M-mode echocardiograms usually merely confirmed the diagnosis of the ordering physician. They concluded that echocardiograms rarely revealed totally unsuspected informa-

TABLE IV

ET AL

1 3

2

tion and were most useful for assessing the severity of known cardiac disease. Although our ordering physicians said that 26 percent of echocardiograms had management impact, our reviewers reported that echocardiograms were necessary for an appropriate change in management in only 8 percent of cases. As suggested by Grimmer and colleagues [IO], these latter cases tended to be patients in whom echocardiography was ordered to assess the severity of known disease or to answer very specific questions. We recognize that no review process can be totally unbiased, but our review process was designed to accept reasonable variations in practice patterns, and our criteria for a positive review always gave the benefit of the doubt to the ordering physician. Furthermore, our grading scheme was so liberal that even a change in patient instructions regarding potential future medications was considered to represent management impact. Thus, we believe that overall our reviewers were more accurate in their assessment of the impact of the echocardiograms than were the ordering physicians.

Correlates of Reviewers’ Assessments

Circumstances of Echocardlcgram Group I: catheterization scheduled or already performed, or echocardiogram ordered to evaluate left ventricle function, left atrium size, source of emboli, or bacterial endocarditis Group II: catheterization neither scheduled nor already performed, and echocardiogram ordered to evaluate mitral valve prolapse, hypertrophic cardiomyopathy, valvular function, or the pericardium

Numberof Appropriately EchocardIograms Ordered

Properly Useful

Provided Crucial lnlormation

Contributed to Appropriate Changein Management

77

59 (77%)

47(61%)

105

96 (93%)

93 (69%)

19 (18%)

14 (13%)

p
p
p
p
July 1963

3 (4%)

The American Journal of Medicine

l(l%)

Volume 75

53

IMPACT OF ECHOCARDIOGRAPHY-GOLDMAN

ET AL

Because we could not prevent some ordering physicians from knowing the echocardiographic result before our interview, we cannot be certain that all their responses were totally unbiased. However, our protocol also included the reasons for ordering as outlined on the echocardiogram requisition form as well as the data on history and physical examination supplied by the cardiology fellow. When the various data sources were compared, we were unable to identify any obvious contradictions in the physician interviews that would bias our study results. Overall, only 8 percent of our echocardiograms were judged by experienced reviewers to have contributed to an appropriate change in patient management. Most notably, solely on the basis of the anatomic reasons for ordering the echocardiogram and on whether the patient had already undergone or was scheduled to undergo cardiac catheterization regardless of the echocardiographic results, we were able to group patients so that the 77 (42 percent) Group I echocardiograms had virtually no chance of having an important impact on management. Even among the 105 Group II echocardiograms, the reviewers believed that the rate of appropriate management impact was just 14 percent. Although our groupings were constructed retrospectively and not prospectively, they were based on prior hypotheses that were responsible for our interest in conducting this study. The low yield of M-mode and two-dimensional echocardiography for the assessment of patients with cerebrovascular accidents or focal cerebral ischemic episodes has been reported previously by Greenland and colleagues [ 161 and by Lovett et al [ 171. We also found that M-mode echocardiograms that were ordered to evaluate left ventricular function were unlikely to have management impact, presumably in part because of their uncertain reliability in patients whose coronary disease might cause regional contraction abnormalities. Two-dimensional echocardiography would be more accurate for the assessment of ventricular function, but we cannot comment on how often this increased accuracy would result in a change in management strategies. The role of M-mode echocardiography in patients with definite or suspected bacterial endocarditis is controversial. Our data suggest that M-mode echocardiography will rarely demonstrate definitive vegetations in patients with negative blood culture results, although an echocardiogram was helpful in one of our patients whose history was consistent with culture-negative endocarditis with resulting aortic regurgitation. We thus suggest that echocardiography may be helpful in the diagnosis of endocarditis in occasional cases [ 181, but that echocardiograms should not be ordered indiscriminantly in patients with fever, systolic murmurs, and

54

July 1963

The American Journal of Medicine

Volume 75

negative blood culture results. In patients with the established diagnosis of endocarditis, early data suggested that echocardiography provided important prognostic information [7]. Later data [ 19,201, however, were not so impressive regarding the independent value of echocardiography for predicting complications in patients who were known to have bacterial endocarditis. Nevertheless, echocardiograms can be helpful in assessing the severity of aortic regurgitation [21] or ventricular compromise [22] in patients with endocarditis. The potential value of echocardiography in patients with valvular heart disease has been emphasized by St. John Sutton et al [23] and by Alpert and colleagues [24]. During the time when our study was conducted, echocadiography was often useful for the evaluation of patients with suspicious murmurs; however, in some patients, echocardiography was a low-yield test because catheterization had been performed or, according to the ordering physician, would be performed regardless of the echocardiographic result. If in the future, echocardiography is more frequently ordered and used appropriately to avoid cardiac catheterization and proceed directly to surgery [23,24], the management impact of echocardiography in valvular heart disease could be higher than we have reported. However, if echocardiography leads to the inappropriate avoidance of cardiac catheterization, it could actually have a negative impact on management. Although many patients with mitral valve prolapse will have classic physical findings that might obviate the need for echocardiography, the documentation that a murmur is caused by mitral valve prolapse and is not simply a flow murmur may lead to the institution of prophylactic antibiotics in patients who otherwise might not receive them [25]. In echocardiography laboratories that are not hospital-based or that study young ambulatory patients, mitral valve prolapse may be a more common reason for ordering echocardiograms than it was in our series, and further data regarding the clinical utility of the test in such patients is needed. Our study concentrated on the value of M-mode echocardiography, whereas two-dimensional echocardiography is presently available in many hospitals and is believed to be preferable to M-mode echocardiography for the evaluation of left ventricular function and intracardiac masses [ 11,26,27]. Despite the improved accuracy of two-dimensional echocardiography for such conditions, it is not at all clear how such improved accuracy will translate into increased management impact. We suggest that an analysis of the impact of two-dimensional echocardiography will be a natural supplement to our present report. We have previously [ 121 reported that ordering physicians believed that 28 percent of cardiac nuclear

IMPACT OF ECHOCARDIOGRAPHY-GOLDMAN ET AL

medicine procedures contributed to a change in management. Other investigators [ 13- 151 have reported that ordering physicians believed various radiographic procedures contributed to a change in management in 17 to 23 percent of cases. When such reports of ordering physicians have been subjected to impartial review, however, upper gastrointestinal series contributed to a change in management in just 7 percent of cases [ 131 and cardiac nuclear medicine studies contributed to a change in management in 12 percent of cases [ 121. Thus, our 8 percent reviewer-assessed rate of appropriate management impact for echocardiography was virtually identical to the rates for other noninvasive diagnostic techniques. Echocardiograms were appropriately ordered in 86 percent of cases in the present study, whereas, in our earlier report [ 121, cardiac nuclear medicine procedures were appropriately ordered just 65 percent of the time (p
in clinical efficacy that may not be emphasized sufficiently in traditional medical training. Although our data document that a substantial proportion of noninvasive diagnostic tests may be obtained for purposes that can be predicted at the time of ordering to have little or no potential management impact, we do not intend to imply that potential management impact is the only justification for ordering a test. For example, echocardiography may be valuable as a baseline study in a patient with a condition such as a newly implanted prosthetic valve or as a rule-out procedure in a patient with unexplained congestive heart failure, even though such M-mode echocardiograms are unlikely to have immediate management impact. In addition, Sox [28] has shown that some patients receive substantial psychologic benefit from tests that might not be medically necessary. One would hope, however, that appropriate education of physicians and patients could reduce the psychologic [28] and sociologic [29] stimuli for tests that are not truly needed. Most published studies of the utility of diagnostic tests report their sensitivity and specificity on selected subgroups of patients. We believe it mandatory that all such procedures be tested on a wide variety of patients so that physicians can be made aware of the procedure’s usefulness and management impact in the kinds of patients in whom the test is actually ordered. If such information is available, accurate and safe noninvasive procedures such as echocardiography are more likely to be used to improve patient management in a costeffective manner. ACKNOWLEDGMENT We would like to thank the referring physicians and the cardiology fellows of the Brigham and Women’s Hospital for their assistance.

REFERENCES 1. 2. 3. 4.

5.

6.

Feigenbaum H: Echocardiography, 3rd ed. Philadelphia: Lea & Febiger; 1981. Popp RL: Echocardiographic assessment of cardiac disease. Circulation 1976; 54: 538-552. Feigenbaum H, Waldhausen JA, Hyde LP: Ultrasonic diagnosis of pericardial effusion JAMA 1965; 191: 711-714. Henry WL, Clark CE, Epstein SE: Asymmetric septal hypertrophy. Echocardiographic identification of the pathognomanic anatomic abnormality of IHSS. Circulation 1973; 47: 225-233. MacDonald IG, Feigenbaum H, Chang S: Analysis of left ventricular wall motion by reflected ultrasound. Application to assessment of myocardial function. Circulation 1972; 46: 14-25. Popp RL: Current concepts: echocardiographic evaluation of left ventricular function. N Engl J Med 1977; 296: 856858.

7.

8.

9. 10. 11.

12.

July 1983

Wann LS, Dillon JD, Weyman AE, Feigenbaum H: Echocardiography in bacterial endocarditis. N Engl J Med 1976; 295: 135-139. Markiewicz W, Peled B, Hammerman H, Greif 2, Hir J, Riss E: Contribution of M-mode echocardiography to cardiac diagnosis. An assessment in 1,000 successive patients. Am J Med 1978; 65: 803-807. Pollick C: Echocardiography in a district general hospital. Postgrad Med J 1978; 54: 297-301. Grimmer SFM, Tindall H, Hill JD: Diagnostic contribution of echocardioaraphv. Lancet 1982: I: 440-441. Brown AK, Aiders& V: The use of cross-sectional echocardiography in a general hospital. Postgrad Med J 1980; 56: 221-228. Goldman L, Feinstein AR, Batsford WP, Cohen LS, Got&chalk A, Zaret BL: Ordering patterns and clinical impact of cardiovascular nuclear medicine procedures. Circulation 1980;

The American

Journal of Medicine

Volume 75

55

IMPACT OF ECHOCARDIOGRAPHY-GOLDMAN ET AL

13.

14.

15. 16.

17.

16.

19.

20.

56

62: 660-667. Marton KI, Sox HC, Wasson J, Duisenberg CE: The clinical value of the upper gastrointestinal tract roentgenogram series. Arch intern Med 1960; 140: 191-195. Fineberg HV, Bauman R, Sosman M: Computerized cranial tomography: effect on diagnostic and therapeutic plans. JAMA 1977; 236: 224-227. Wittenberg J, Fineberg HV, Black EB, et al: Clinical efficacy of computed body tomography. AJR 1976; 131: 5-14. Greenland P, Knopman DS, Mikell FL, Asinger RW, Anderson DC, Good DC: Echocardiiraphy in diagnostic assessment of stroke. Ann Intern Med 1961; 95: 51-53. Lovett JL, Sandok BA, Giuliani ER, Nasser FN: Two-dimensional echocardiography in patients with focal cerebral ischemia. Ann Intern Med 1961; 95: l-4. Rubenson DS. Tucker CR, Stinson EB, et al: The use of echocardiography in diagnosing culture negative endocarditis. Circulation 1961; 64: 641-646. Wann LS, Hallam CC, Dillon JC, Weyman AE, Feigenbaum H: Comparison of M-mode and cross-sectional echocardiography in infective endocarditis. Circulation 1979; 60: 726-733. Stewart JA, Silimperi D, Harris P, Wise NK, Fraker TD, Kisslo JA: Echocardiographic documentation of vegetative lesions in infective endocarditis: clinical implications. Circulation 1980; 61: 374-380.

July 1963

The American Journal of Medicine

Volume 75

21.

22.

23.

24.

25. 26.

27. 28. 29.

Morganroth J, Perloff JK, Zeldis SM: Acute severe aortic regurgitation: pathophysiology, clinical recognition, and management. Ann Intern Med 1977; 87: 223-232. Henry WL, Bonow RO, Rosing DR, Epstein SE: Observations on the optimum time for operative intervention for aortic regurgitation. II. Serial echocardiographic evaluation of asymptomatic patients. Circulation 1980; 61: 484492. St. John Sutton MG, St. John Sutton M, Oldershaw P, et al: Valve replacement without preoperative cardiac catheterization. N Engl J Med 1981; 305: 1233-1238. Alpert JS, Sloss LJ, Cohn PF, Grossman W: The diagnostic accuracy of combined clinical and noninvasive cardiac evaluation: comparison with findings at cardiac catheterization. Cathet Cardiovasc Diagn 1980; 6: 359-370. Kaplan EL, Anthony BF, Bisno A, et al: Prevention of bacterial endocarditis. Circulation 1977; 56: 139A-143A. Popp RL, Fortuin NJ, Johnson ML, Kisslo J: Optimal resources for ultrasonic examination of the heart. Circulation 1982; 65: 423A-431A. Abbasi AS: Echocardiography in the differential diagnosis of the large heart. Am J Med 1976; 60: 677-686. Sox HC, Margulies I, Sox CH: Psychologically mediated effects of diagnostic tests. Ann Intern Med 1981; 95: 680-685. Eisenberg JM: Sociologic influences on decision-making by clinicians. Ann Intern Med 1979: 90: 957-964.