October 9th, 2013
Defining the “Appropriate Use” of Transthoracic Echo
CardioExchange’s Harlan Krumholz interviews Susan Matulevicius about her study group’s analysis of the use of transthoracic echocardiography at a tertiary-care academic medical center. The article is published in JAMA Internal Medicine.
Researchers retrospectively reviewed medical records from 535 consecutive transthoracic echocardiograms (TTEs). Two cardiologists, blinded to clinical impact, classified the TTEs according to appropriate use criteria (AUC) from 2011. Then the TTEs were assessed for clinical impact by 2 cardiologists who were blinded to AUC.
Of the TTEs, 31.8% resulted in an active change in care, 46.9% in continuation of current care, and 21.3% in no change in care. According to the 2011 AUC, 91.8% of TTEs were appropriate, 4.3% inappropriate, and 3.9% uncertain. The percentage of appropriate versus inappropriate TTEs that led to an active change in care did not differ significantly (32.2% vs. 21.7%, respectively; P=0.29).
Krumholz: In your study, 92% of the TTEs were appropriate but only 31.8% resulted in an active change in care. Isn’t that a number needed to change of roughly 3? That seems very good, so why does your conclusion characterize the 31.8% figure as low? In my 1994 study of changes in management as a result of an echocardiogram, the rate was much lower.
Matulevicius: To optimize the use of a diagnostic test in providing high-value care, we first need to understand its current use. The initial purpose of our study was simply to ask whether ordering providers are using the TTE’s results in any way. We did not assess whether the change in care that occurred could have occurred without the TTE. The 1 in 3 TTEs that leads to active change included many that altered care in nonmeaningful ways. Our online supplement shows data from our exploratory analysis: TTEs that led to active change were rated on a scale of 1 to 5 (where 1 was “misused” and 5 was “very useful”) by the consensus of two independent cardiologists. Only 19% (32/170) of all active-change TTEs were rated as a 4 or 5, which equates to only 6% (32/535) of all TTEs in our study being useful or very useful, similar to the rate of meaningful change in your 1994 study. For a diagnostic test that is used frequently, I would argue that even 1 in 3, especially when only a minority are being used in a valuable manner, is too low.
The harder part to evaluate (and which I did not address in the paper thoroughly) is the societal importance of the 47% of TTEs classified as “continuation of care.” Sometimes, reassuring patients or obtaining echocardiographic information that informs (even if it does not actively change) management is equally valuable. However, there is a spectrum of necessity of testing. For example, sometimes TTEs are performed when the pretest probability of an abnormality is low, given all the clinical and examination data. Similarly, it sometimes doesn’t matter what the TTE shows because nothing more can be done for the patient. Those are cases where the incremental value of TTE may not be high. However, at other times, when the pretest probability is intermediate from the physical exam and clinical data, a normal TTE that reassures the physician and the patient is of great value and, likely, of equal value to the TTE that was “very useful” in actively changing care. As a professional community, we must acknowledge the spectrum of value in testing and try to refrain from ordering a TTE just because it is part of the “protocol” for evaluating a given condition, if it provides very little necessary information for delivering high-value care to an individual patient.
Krumholz: How should we improve the use of echocardiography?
Matulevicius: Patient selection is a central aspect of quality in cardiac imaging or any diagnostic procedure. Imaging must be used in the proper patient subset at the optimal time, and we must be able to act on the results. Given limited resources and rising health care costs, additional research into the necessity of TTE is needed, requiring collaboration among hospitals, administrators, politicians, economists, the government, and patients. Previous attempts to reduce reimbursement for — and utilization of — TTE have been unsuccessful; TTE volume has continued to increase. Efforts to value physicians’ time in communicating the treatment plan and expectations of care to the patient may reduce use of diagnostic testing while enhancing patient-centered care. Medical school and post–medical school training about cost and value, as well as testing and professional society endorsement of programs like “Choosing Wisely,” may increase our stewardship of health resources. Incorporating necessity into the appropriateness framework may refine the use of TTE. In addition, quantifying the impact of TTE-based reassurance on patient-centered outcomes, and specifically examining whether alternate strategies that do not involve TTE provide a similar benefit to patients at a lower cost, may offer the greatest potential to decrease utilization while maintaining high-quality care.
Krumholz: In many academic centers, more volume brings more revenue, training material, and prestige. What is the incentive to reduce the volume?
Matulevicius: Doing more means more revenue and clinical volume not only in academic medical centers, but also in private practice offices and non-academic centers. And that goes for all diagnostic testing, not just TTE. In our current environment, it is difficult to incentivize reducing volumes. The key is to restructure the incentives. Among the countries in the Organization of Economic Cooperation and Development (OECD), the U.S. has among the lowest rates of doctor visits per capita (3.9, vs. 6.5 in the OECD overall) but one of the highest rates of MRI scans per 1000 population (91.2, vs. 46.6 in the OECD overall). If prestige came with providing high-value, efficient care — and if we were reimbursed equitably for definitive evaluation and management — we would probably be more willing to order fewer tests and provide more direct patient care.
We must remember that by 2021, national health expenditures are expected to grow to by nearly 5 trillion dollars. A person who retires at age 65 in 2030 will have to set aside 52% of his or her salary each and every year to cover the costs of retirement savings, health insurance, social security, and Medicare and Medicaid payroll taxes. This is 5 times more than a 65-year-old who retired in 1960 needed, and over 30% of those costs are directly related to health care (see the work by Sylvester Schieber).
Krumholz: You had a single site. How generalizable is your study?
Matulevicius: We did this study at a U.S. tertiary care academic medical center, limiting generalizability to other settings, especially non-U.S. or private practice settings. Adherence to AUC — as well as the clinical impact of TTE by region, practice, type, practice size, clinician experience, and payor mix — may differ in ways that our single-center study cannot capture.
Krumholz: How hard was this assessment to do? Can other sites repeat what you did?
Matulevicius: It takes a lot of time to go through each chart (and to convince three other people to do the same thing) and to ensure that all chart abstractors interpret the evaluation criteria similarly. Nevertheless, it is completely reproducible, especially in a system that has an electronic medical record. The one issue will be to ensure that the definition of continuation of care — “lack of escalation or de-escalation of current care, but direct communication provided to patients and/or documentation by providers in the chart about the TTE results” — is applicable in a given institution. Some institutions have automated methods for informing patients of test results, so this category would need to be refined at those institutions (otherwise, the “no change” TTEs would be included in the “continuation of care” category). The overall process, which could be applied not only to TTE but also to other diagnostic testing modalities, may inform how differences in payor mix, practice type, and clinician experience factor into use of diagnostic testing.
JOIN THE DISCUSSION
How do the findings from Dr. Matulevicius’s study affect your perspective on the appropriateness of TTE use at its current volume?