Copyright © 2005, The European Society of Cardiology
Can we Justify the Cost of Echocardiography? Lessons from Outcomes Research
University of Queensland Department of Medicine, Princess Alexandra Hospital, Ipswich Road, Brisbane, Qld Q4102, Australia
Received 8 January 2005; accepted after revision 24 January 2005.
* Tel.: +61 7 3240 5340; fax: +61 7 3240 5399. E-mail: tmarwick@soms.uq.edu.au
 |
Abstract |
|---|
 Top Abstract 1. Cost2. Effectiveness3. Cost effectiveness4. ConclusionReferences |
|---|
The inexorable growth in the performance of cardiac imaging studies has led to a significant reduction in reimbursement for these services, which are already being reimbursed at levels that are quite similar to their cost. This paper seeks to apply the principles of analysis of cost-effectiveness to echocardiography. Opportunities are summarized for reducing cost and maximizing effectiveness, and different approaches to defining cost-effectiveness are examined for valvular, myocardial and ischemic indications.
Keywords: Echocardiography; Cost-effectiveness; Outcomes
The health care budget continues to expand, driven by the ageing population, the efficacy of treatment of cardiac disease, and the cost of new technology. In the face of constant revenues, as well as competing demands for funding from other medical and non-medical commitments, governments and insurance organisations are seeking to control expenditure. In most countries, the performance of imaging is increasing at a rate >10% per year, and as the technique is not obviously connected to health outcomes, the necessity of this expenditure is unclear to many payors.
In the context of this challenging milieu, it is important for echocardiographers and our professional groups to understand the relationships between cost and effectiveness, including how to measure each of these parameters. Excellent reviews of these topics have been written by experts in the field,1,2 so this paper will attempt to address how clinicians can apply these data to the practice of echocardiography, in the hope that the cost-effectiveness of this investigation will be better defined.
|
1. Cost |
|---|
|
TopAbstract1. Cost2. Effectiveness3. Cost effectiveness4. ConclusionReferences |
|---|
1.1 Definitions
One of the difficulties posed by this field is understanding the jargon. It is tempting to make decisions about the selection of imaging techniques based on cost minimisation, which is reasonable if the tests are comparable and the clinical outcomes are the same. Unfortunately, this is rarely the case – for example when a technique is being selected to assess left ventricular function, nuclear ventriculography and echocardiography might seem superficially similar, but differ with respect to reproducibility and ability to obtain supplementary information – for example regarding valvular disease. Cost-effectiveness analysis is the solution to comparing similar but non-identical tests with different costs and outcomes (e.g. selection of stress echocardiography versus SPECT imaging), and this will be attended to in detail below. A full cost benefit analysis measures all costs and outcomes in monetary terms, and this is an important prelude to a cost utility assessment, which is usually done by health planners when deciding to invest additional funds into comparable outcomes of a completely different nature, such as decisions to screen for cancer versus heart disease.
1.2 Analysis of cost
Measuring costs is difficult. The use of charges, although easily obtained, is unreliable because these are highly variable and determined by the market. Actual costs should be the desired metric for a complete cost-effectiveness analysis. This involves an assessment of fixed cost (including capital, depreciation and service cost), variable costs (which include consumables), stepped costs (e.g. salaries, which increase with the number of services), and induced costs, which reflects the additional cost that arises due to complications of testing. Interestingly, a full assessment of cost performed in Australia about a decade ago (Deloitte, data on file) found that the true cost of performing an echocardiogram was very close to the Medicare reimbursement for the test. Thus, government or insurance company reimbursements are a reasonable approximation to cost, presumably because these organisations have performed at least an approximate analysis of cost.
1.3 Implications
In most countries, reimbursement for echocardiography services continues to fall, as this is the only means of capping expenditure in the face of increasing volume. Given that the cost of the test and the reimbursement are now closely aligned, the echo community must consider ways of further reducing costs, or our work will become unsustainable. In fact, this process is already occurring in the angiography laboratory, where significant cost savings may be attained by focusing on efficiency and tracking procedural times. Two approaches are feasible in the echocardiography laboratory, based on technical developments.
First, the digital handling, review and storage of echocardiograms represents a readily achievable source of cost savings, partly by reduction of reading time, and partly by the reduction of storage costs. A number of studies3–6 have compared the results of digital and video interpretation (Table 1a), and the only study which has shown an incremental benefit of videotape above digital review was a study performed using stress echocardiography, where sampling of a single beat carries the potential for sampling error, but can now be addressed using digital techniques involving multiple cycle capture. Similarly, a number of studies have addressed the adequacy of image quality with digital compression (Table 1b) in order to evaluate whether such a change would have a detrimental effect on patient care.7–11 There is no evidence that standard approaches to compression lose meaningful amounts of data, The results of these studies firmly indicate that changing the laboratory to a digital format loses no measurable information, but has a material impact on cost. Moreover, with the transition to DICOM file structure, there is no need to incur vast capital expenditure in the process of doing this. Simple DICOM viewers can be loaded onto generic work stations for a few thousand dollars, and the images burnt to DVD, obviating the need for expensive on-line storage.
|
|
The second possible approach to reducing cost would be to reduce scan acquisition time by three-dimensional acquisition. The goal here would be obtain a number of three- dimensional data-sets, with and without colour, as well as the relevant Doppler measurements, in the space of maybe 10–15 minutes rather than the usual hour examination time. This strategy would free up an echo machine, which is the major capital cost in the laboratory, while the sonographer who performed that study obtains the conventional imaging planes and measurements from the three-dimensional data set. Unfortunately, while this remains the target for the future, a recent comparative study between two- and three-dimensional acquisitions demonstrated that in 13% of patients, there were important missed findings using a pure three-dimensional echo approach,12 and further technical developments will be needed before this dream can be realised.
|
2. Effectiveness |
|---|
|
TopAbstract1. Cost2. Effectiveness3. Cost effectiveness4. ConclusionReferences |
|---|
2.1 Definitions
The measurement of effectiveness is crucial - there can be no cost effectiveness without effectiveness! Moreover, even when cost-effectiveness is not perceived as an important topic (e.g. in publicly-funded health systems, where workload and revenue are often dissociated and resources are limited), understanding effectiveness addresses the importance of appropriate use of limited resources. Thus, an assessment of the effectiveness of our efforts should be an interest of all echocardiographers.
The measurement of effectiveness by echocardiography is not intuitive – unlike measuring a response to drug therapy, it is the decisions that arise from the test rather than the test itself that provides favourable health effects. The first step in demonstrating effectiveness is to show that the test results correlate with outcome, and the second (and more difficult) step is to show that this outcome can be altered by changes in management. The most meaningful measure of prediction of outcome is mortality, which requires large and longitudinal studies – in the field of echocardiography, these have generally been obtained with stress echocardiography. An alternative is to examine composite cardiac events. Quality of life is an important parameter, particularly in decision making about valve disease, and the optimal approach is to combine both survival and quality of life into a metric called quality of life year saved (QALY), a measure of a number of life years saved at full quality of life. An important distinction to keep in mind is the differentiation between efficacy (the ability to obtain an outcome) and effectiveness (the ability to attain an outcome in routine practice).
2.2 Implications
The indications for echocardiography in our laboratory over a period of several months are illustrated in Fig. 1. The following sections illustrate the assessment of effectiveness by examining the evidence of benefit from echocardiography in three situations: evaluation of left ventricular function, echocardiography for infective endocarditis, and the effectiveness of investigations for cardiac source of embolism.
|
2.2.1 Follow-up assessment of left ventricular function
The assessment of left ventricular function is the most common indication for echocardiography. Table 2 summarizes the indications associated with previous echocardiography. It might be expected that a high proportion of patients with valvular disease have sequential echocardiograms, but it is very striking that a quarter of the patients undergoing left ventricular assessment have had a previous echocardiogram. Given the preponderance of this indication, left ventricular assessment is the commonest cause of a repeat study.
|
Left ventricular function is an important determinant of outcome. However, the prognostic implications of these findings in populations are often applied to individuals, and a further extrapolation occurs when the test is used sequentially. In fact, the latter is difficult to justify with two-dimensional echocardiography; the variability of sequential left ventricular ejection fraction measurement is about 15%, 12% of which reflects test/retest variability. The latter reflects the performance of repeat echocardiograms from slightly different windows and in slightly different axes from one visit to the next. This phenomenon limits the smallest change of ejection fraction that can be detected with 95% confidence to 11%.13 Similarly, the smallest change of mass that can be detected with 95% confidence is 59 g, which compares to a 20–40 g change of mass in the course of a year of anti-hypertensive therapy.14 Clearly, the performance of sequential two-dimensional measurements to follow patients may be inappropriate with current technology.
The recent development of three-dimensional echocardiography allows us to avoid the variations that arise from different two-dimensional measurement planes, reducing the problem of test/retest variation.15 A number of validation studies have confirmed that although two-dimensional echo underestimates left ventricular volumes, three-dimensional echo provides volumes that are closely analogous to those measured with magnetic resonance imaging.
2.2.2 Evaluation for infective endocarditis
The evaluation of suspected infective endocarditis accounts for about 6% of echo requests in our laboratory. As nearly 80% of patients undergoing echo and transoesophageal echo for this purpose have a normal study, the question of effectiveness is a pertinent one. Part of this problem relates to the role of echocardiography in standard clinical criteria for the diagnosis of endocarditis,16 which do not consider the underlying probability of the disease when the performance of echocardiography is recommended. In a study of 500 inpatients undergoing transthoracic echo for possible bacterial endocarditis, an echo diagnosis of endocarditis was confirmed in 8.6% of patients.17 The independent predictors of these echo findings are illustrated in Table 3, and include a history of embolism, central venous access, intravenous drug abuse, prosthetic valves, and positive blood cultures. No individual who lacked any of these parameters actually turned out to have a positive echocardiogram. Thus, the first step in the avoidance of unnecessary echocardiograms for this diagnosis is to consider the risk of the disease.
|
A second question that warrants consideration in patients with endocarditis is whether transthoracic or transoesophageal echo should be performed. This has been addressed in a decision analytic model of six strategies for the management of patients with suspected endocarditis and positive blood cultures, including two weeks treatment for bacteremia, the use of transthoracic or transoesophageal echo alone, combinations of the two, or universal treatment for endocarditis in all patients.18 The published sensitivity and specificity of transthoracic and transoesophageal echo were then used to identify the likely outcomes of treatment, including an expectation of 15% mortality, 2.5% endocarditis relapse, and the risk of non-treatment from the literature, as well as identifying the costs of procedures, treatment and impaired quality of life. All of these were modelled according to the pre-test probability of the condition. The authors' findings (Table 4) identified transoesophageal echocardiography as the most cost efficient approach to identifying and treating this condition, minimising the numbers of missed cases of endocarditis and also avoiding over-diagnosis (with its incumbent costs). In contrast, the use of transthoracic echocardiography led more patients to have a missed diagnosis, with a consequent minor reduction of life expectancy as well as greater treatment costs incurred because of these missed diagnoses. Combined approaches led to excess treatment which invoked high levels of cost, and inappropriate treatment decisions involving no echocardiography led either to excess treatment or missed diagnoses with consequent high levels of cost. In a sensitivity analysis where the authors altered the probability of the condition, transthoracic echo was only more cost effective when there was a very low likelihood of endocarditis.
|
The third area of potential cost control in echocardiography for endocarditis is the consideration of repeat echocardiograms. Vieira19 examined transthoracic and transoesophageal studies in 266 episodes of endocarditis among 262 hospital patients, 55% of whom had prosthetic valve endocarditis, and 26% of whom had culture negative endocarditis. Transthoracic echocardiograms were repeated in 72% of patients and transoesophageal studies were repeated in 18% of patients. The results indicated that the return from repeat echocardiography was limited, and in particular after the third study, no additional findings were identified.
2.2.3 Evaluation of cardiac source of embolism
Referrals for echocardiography with suspected cardiac source of embolism account for 4% of the workload in our laboratory. While the 26% of these patients who have atrial fibrillation could have a transthoracic echocardiogram justified on the basis of assessment of the left ventricle and valves, the return from the majority of these patients in sinus rhythm was very limited (Fig. 2). Indeed, only a single patient with an aortic arch atheroma was identified, and we would have to conclude that the cost of this activity would be difficult to justify.
|
This audit of indications for echocardiography illustrates how better selection of patients may allow us to focus on indications where echocardiography is more likely to be effective. Unfortunately, the unquestioned performance of echocardiograms for weak indications consumes resources and limits timely access by other patients.
|
3. Cost effectiveness |
|---|
|
TopAbstract1. Cost2. Effectiveness3. Cost effectiveness4. ConclusionReferences |
|---|
In situations where the indications for echocardiography can be shown to be effective, and when the costs of testing have been minimised, the next step is to consider cost-effectiveness analysis. The simplest strategy is to show that the cost effectiveness of echocardiography makes this technique preferable to alternative investigations, while a more challenging approach would be to prove that decision-making from echocardiography (compared to clinical evaluation alone) improves outcome and/or reduces downstream costs. Three models of cost-effectiveness analysis will be reviewed. First, computer models can be used to document the diagnostic value of the technique, based on accuracy and costs derived from the literature, or preferably informed by observational studies or meta-analyses. Second, retrospective studies may be compared, with appropriate corrections for selection processes to identify superior strategies for identifying outcomes and thereby cost effectiveness analysis. Finally, and most desirably, cost effectiveness analyses can be incorporated into new controlled prospective studies.
3.1 Computer models
This strategy is well-exemplified by a computer model for selecting the most cost-effective approach between exercise ECG, exercise echocardiography, myocardial perfusion scintigraphy (SPECT) and coronary angiography.20 The first step is to use assumptions about accuracy of testing, derived from the literature, and costs (in this case derived from Medicare reimbursement at the time of the analysis). In this study, the sensitivity, specificity and cost of exercise testing (70%, 70%, $117) were compared with exercise echocardiography (80%, 85%, $281), exercise SPECT (90%, 70%, $617), and angiography (99%, 99%, $2510). Table 5a summarizes the net costs of testing – comprising the cost of the original test, the cost of progressing to angiography in patients with positive tests, and the costs of complications and mortality for both stress testing (0.05% and 0.005%) and angiography (2% and 0.15%). The most important assumption was that the correct diagnosis of coronary disease gave an additional 3 years of life at full quality – the implications of this arising from true positive diagnoses is summarized in Table 5b. The assumption by the model that lives are saved by correct diagnosis of disease means that the certain diagnosis of disease with angiography drives the greatest improvements of QALY. Finally, the costs and effects of testing are summarized in Table 5c. These results show that the greatest value is obtained by using stress echocardiography in patients at low to intermediate risk, SPECT for patients at intermediate to high risk, and angiography for those at the highest risk. This exercise carries an important message about the interaction between pretest risk, test accuracy, outcomes and cost, but it has the disadvantage that the selected numbers can appear arbitrary and unreal.
|
|
|
A more sophisticated approach to computer modelling has been published by Kuntz et al.21 The authors based a decision analytic model on real accuracy data from meta-analyses, but still assumed the outcomes of subsequent interventions, based on revascularization strategies predicated on the extent of coronary disease. In that study, patients with typical angina had a life expectancy of 10.4 years, and this is improved most by the performance of angiography (because no patients with disease were missed). In contrast, in patients with atypical chest pain, where the performance of angiography is unduly expensive, the most appropriate approach is either the performance of exercise ECG or exercise echocardiography, the latter producing slightly greater benefit in terms of effectiveness.
There are several important limitations when models are based on diagnostic accuracy data with assumed rather than observed outcomes. First, "real world" angiography rates are well below that noted in the decision models, where all abnormal results provoke angiography. Second, the underlying risk of a diagnostic testing population is less than a revascularization population. Third, revascularization of all identified disease is not performed in clinical practice, and in any case, such interventions may lack the effectiveness that they are purported to show in various trials for a number of reasons. Thus, while these results are a guide to the cost-effectiveness of non-invasive testing, they are heavily dependent on the underlying assumptions, and application of cost-effectiveness analysis to studies with recorded outcomes is a preferable approach.
3.2 Controlled retrospective studies
Stress echocardiography now has a huge evidence base identifying the low risk associated with negative test, while the risk associated with a positive test is related to the nature and extent of abnormality. Moreover, in circumstances where stress echo is used in conjunction with standard stress testing, the greatest incremental information is obtained in the intermediate risk group,22 just as it is with nuclear imaging,23 keeping in mind that this group accounts for about one half of patients undergoing testing. This might then pose the question as to whether the exercise ECG should indeed be used as the initial test, especially if the response of clinicians to a positive stress ECG may not be as confident as to that of a positive stress echo. Of course, the replacement of exercise ECG by exercise echo carries significant immediate cost implications, but these may be dwarfed by changes of downstream costs related to subsequent testing, management and complications.
One approach to this question may be achieved by comparison of large survival databases. The steps involved in such an analysis are matching of pre-test risk of each group, the comparison of outcomes within categories of risk, and then the inclusion of management and cost data. In such a study,24 the clinical nature of patients in the low, intermediate and high risk groups in the two studies was similar, as were the results of standard stress testing, and outcomes. However, the responses to the results of the stress ECG and the stress echocardiogram were quite different – while the use of angiography increased appropriately with increasing levels of risk by stress echo, angiography use was quite uniform across risk levels defined by the stress ECG, perhaps reflecting the clinicians' disquiet about the reliability of the stress ECG results. As a consequence of more angiography, more patients underwent revascularisation, increasing downstream costs, so that when the cost for prediction of events was compared with exercise ECG and exercise echo, it was significantly higher with the exercise ECG (Fig. 3). Thus, although the diagnostic costs for the two techniques were similar over long term follow-up, the follow-up costs were very much greater with the standard stress test.
|
3.3 Prospective controlled studies
The optimal means of gathering cost effectiveness data is to design this into studies of echocardiography from the outset. One such example is the ACUTE trial of strategies for reduction of embolic risk at the time of cardioversion, which compared a strategy of immediate cardioversion after exclusion of thrombus using transoesophageal echocardiography, with a conventional strategy of three weeks anticoagulation followed by cardioversion. The results of these strategies showed similar frequencies of embolic events, haemorrhagic events, total mortality, and reversion to sinus rhythm.25 The authors performed two approaches to cost effectiveness, one based on observational data, and the second based on an analytic model informed by the results of the trial. There was no difference in costs between the two strategies using either analysis. Sensitivity analysis demonstrated that the efficacy of transoesophageal echocardiography was driven by the likelihood of haemorrhagic risk, which carried the main source of cost in these individuals.26
|
4. Conclusion |
|---|
|
TopAbstract1. Cost2. Effectiveness3. Cost effectiveness4. ConclusionReferences |
|---|
Cost-effectiveness is easier to define with a treatment intervention than a diagnostic test, and this has been one of the challenges for echocardiography. Nonetheless, by consideration of optimizing cost, maximising effectiveness by controlling referrals, and identifying situations where echo provides proven outcomes benefit, or the same outcome as a more expensive technique, echocardiography may develop a new and stronger evidence base that justifies the cost of this procedure, even in our challenging funding environment.
|
Acknowledgement |
|---|
The author gratefully acknowledges the input and advice of Dr Leslee Shaw, whose seminal work has informed much of this discussion.
|
References |
|---|
|
TopAbstract1. Cost2. Effectiveness3. Cost effectiveness4. ConclusionReferences |
|---|
- Shaw L.J., Eisenstein E.L., Hachamovitch R. A primer of biostatistic and economic methods for diagnostic and prognostic modeling in nuclear cardiology: Part II. J Nucl Cardiol (1997) 4:52–60.[CrossRef][Web of Science][Medline]
- Shaw L.J., Hachamovitch R., Eisenstein E.L. A primer of biostatistic and economic methods for diagnostic and prognostic modeling in nuclear cardiology: Part I. J Nucl Cardiol (1996) 3:538–545.[CrossRef][Web of Science][Medline]
- Haluska B., Wahi S., Mayer-Sabik E. Accuracy and cost- and time-effectiveness of digital clip versus videotape interpretation of echocardiograms in patients with valvular disease. J Am Soc Echocardiogr (2001) 14:292–298.[CrossRef][Web of Science][Medline]
- Shah D.J., Diluzio S., Ambardekar A.V. Evaluation of valvular regurgitation severity using digital acquisition of echocardiographic images. J Am Soc Echocardiogr (2002) 15:241–246.[CrossRef][Web of Science][Medline]
- Segar D.S., Skolnick D., Sawada S.G. A comparison of the interpretation of digitized and videotape recorded echocardiograms. J Am Soc Echocardiogr (1999) 12:714–719.[CrossRef][Web of Science][Medline]
- Attenhofer C.H., Pellikka P.A., Oh J.K. Is review of videotape necessary after review of digitized cine- loop images in stress echocardiography? A prospective study in 306 patients. J Am Soc Echocardiogr (1997) 10:179–184.[CrossRef][Web of Science][Medline]
- Harris K.M., Schum K.R., Knickelbine T. Comparison of diagnostic quality of motion picture experts group-2 digital video with super VHS videotape for echocardiographic imaging. J Am Soc Echocardiogr (2003) 16:880–883.[CrossRef][Web of Science][Medline]
- Garcia M.J., Thomas J.D., Greenberg N. Comparison of MPEG-1 digital videotape with digitized sVHS videotape for quantitative echocardiographic measurements. J Am Soc Echocardiogr (2001) 14:114–121.[CrossRef][Web of Science][Medline]
- Spencer K., Solomon L., Mor-Avi V. Effects of MPEG compression on the quality and diagnostic accuracy of digital echocardiography studies. J Am Soc Echocardiogr (2000) 13:51–57.[Web of Science][Medline]
- Soble J.S., Yurow G., Brar R. Comparison of MPEG digital video with super VHS tape for diagnostic echocardiographic readings. J Am Soc Echocardiogr (1998) 11:819–825.[CrossRef][Web of Science][Medline]
- Karson T.H., Zepp R.C., Chandra S. Digital storage of echocardiograms offers superior image quality to analog storage, even with 20:1 digital compression: results of the Digital Echo Record Access Study. J Am Soc Echocardiogr (1996) 9:769–778.[CrossRef][Medline]
- Chan K.L., Liu X., Ascah K.J. Comparison of real-time 3-dimensional echocardiography with conventional 2-dimensional echocardiography in the assessment of structural heart disease. J Am Soc Echocardiogr (2004) 17:976–980.[CrossRef][Web of Science][Medline]
- Otterstad J.E., Froeland G., St John S.M. Accuracy and reproducibility of biplane two-dimensional echocardiographic measurements of left ventricular dimensions and function. Eur Heart J (1997) 18:507–513.
[Abstract/Free Full Text] - Gottdiener J.S., Livengood S.V., Meyer P.S. Should echocardiography be performed to assess effects of antihypertensive therapy? Test-retest reliability of echocardiography for measurement of left ventricular mass and function. J Am Coll Cardiol (1995) 25:424–430.[Abstract]
- Jenkins C., Bricknell K., Hanekom L. Reproducibility and accuracy of echocardiographic measurements of left ventricular parameters using real-time three-dimensional echocardiography. J Am Coll Cardiol (2004) 44:878–886.
[Abstract/Free Full Text] - Bansal R.C. Infective endocarditis. Med Clin North Am (1995) 79:1205–1240.[Web of Science][Medline]
- Greaves K., Mou D., Patel A. Clinical criteria and the appropriate use of transthoracic echocardiography for the exclusion of infective endocarditis. Heart (2003) 89:273–275.
[Abstract/Free Full Text] - Heidenreich P.A., Masoudi F.A., Maini B. Echocardiography in patients with suspected endocarditis: a cost-effectiveness analysis. Am J Med (1999) 107:198–208.[CrossRef][Web of Science][Medline]
- Vieira M.L., Grinberg M., Pomerantzeff P.M. Repeated echocardiographic examinations of patients with suspected infective endocarditis. Heart (2004) 90:1020–1024.
[Abstract/Free Full Text] - Patterson R.E., Eisner R.L. Cost analysis of non-invasive testing. In: Cardiac stress testing and imaging—Marwick T.H., ed. (1998) New York: Churchill Livingstone.
- Kuntz KM, Fleischmann KE, Hunink MG, et al. Cost-effectiveness of diagnostic strategies for patients with chest pain. Ann Intern Med 99 A.D.;130:709–18.
- Marwick T.H., Case C., Vasey C. Prediction of mortality by exercise echocardiography: a strategy for combination with the Duke treadmill score. Circulation (2001) 103:2566–2571.
[Abstract/Free Full Text] - Berman D.S., Hachamovitch R., Kiat H. Incremental value of prognostic testing in patients with known or suspected ischemic heart disease: a basis for optimal utilization of exercise technetium-99m sestamibi myocardial perfusion single-photon emission computed tomography. J Am College Cardiol (1995) 26:639–647.[Abstract]
- Marwick T.H., Shaw L., Case C. Clinical and economic impact of exercise electrocardiography and exercise echocardiography in clinical practice. Eur Heart J (2003) 24:1153–1163.
[Abstract/Free Full Text] - Klein A.L., Grimm R.A., Murray R.D. Use of transesophageal echocardiography to guide cardioversion in patients with atrial fibrillation. N Engl J Med (2001) 344:1411–1420.
[Abstract/Free Full Text] - Klein A.L., Murray R.D., Becker E.R. Economic analysis of a transesophageal echocardiography-guided approach to cardioversion of patients with atrial fibrillation: the ACUTE economic data at eight weeks. J Am Coll Cardiol (2004) 43:1217–1224.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||