Conclusions
As healthcare increasingly shifts its focus toward value-based care, echocardiography has been a prime target for refinement as it remains a frontline tool for the diagnosis and management of various cardiovascular diseases.18 Given its wide availability in all types of clinical settings as well as its safety profile, echocardiography is at risk for being misused. Common outpatient scenarios in which rarely appropriate echocardiograms are ordered include routine surveillance for heart failure with recent imaging less than one year prior, repeat assessment of left ventricular ejection fraction in patients with coronary artery disease, monitoring trivial or small pericardial effusions, and reassessment of valvular heart disease at an interval more frequent than recommended by professional society guidelines.10 Common inpatient scenarios resulting in rarely appropriate echocardiogram orders include fevers without bacteremia or new murmur, lightheadedness or presyncope without other evidence of cardiovascular disease, evaluation for pulmonary embolism, and surveillance of left ventricular ejection fraction in patients with coronary disease or patients with prior normal ventricular function and no change in clinical status.11,19 It is especially important to target these rarely appropriate clinical scenarios as the volume of echocardiography testing continues to increase.
In our search through existing literature, studies evaluating the effectiveness of AUC-based QI interventions have demonstrated mixed results. However, most of these studies have been single-center studies with only the Echo WISELY trial being a multicenter trial.12 The results of our meta-analysis help to further clarify this topic, demonstrating that AUC QI interventions are associated with a successful reduction in rarely appropriate echocardiography testing. While some small studies showed that the effects of QI interventions resulted in an initial significant change in target behavior with eventual loss of effect over time, our results suggest that the effects of AUC-based QI interventions are persistent over long-term follow-up with a trend towards further reduction in rarely appropriate testing.
It must be noted that when the initial professional societies AUC documents were published, echocardiography studies were characterized as “appropriate”, “uncertain”, or “inappropriate”. However, this was subsequently revised to the categories of “appropriate”, “may be appropriate”, and “rarely appropriate”, to acknowledge that studies previously characterized as “inappropriate” may be appropriate in certain specific, infrequently encountered clinical settings. In our review of existing literature, we included both sets of nomenclature based on time of publication. However, to be consistent with the most updated AUC recommendations, published results utilizing the outdated terminology of “inappropriate” were included as “rarely appropriate” in our analysis. It is important to emphasize that for each specific indication the inappropriate and rarely appropriate criteria are not exchangeable, however, since we didn’t look at specific indications, but rather investigated the effect of an intervention of clinicians’ behavior, including both terminologies in our analysis is perfectly reasonable.
One issue that has been called into question is whether the effectiveness of AUC QI interventions demonstrated in single-center studies, is generalizable. Of particular interest is differences in adaptation of QI interventions by physicians in training and more senior physicians. It is expected to see greater degree of behavioral changes in practice among physicians in training, who are more receptive to changes in practice behavior and feedback. Single center studies that are focused on trainees only, can therefore demonstrate greater degree of effectiveness of such AUC QI interventions.10,11The advantage of our meta-analysis is that it encompasses providers of different specialties as well as providers of differing levels of training, and the results are therefore more generalizable.
Another important question is whether educational QI tools are effective in changing provider behavior even when not combined with a feedback tool. Our study demonstrate that the use of a feedback tool is not necessary for educational tools to be effective and that the addition of feedback tools to educational or decision support tools did not further enhance QI intervention. Several studies have suggested that the effectiveness of educational tools when used alone stems from the fact that many providers are simply not aware of the existence of appropriate use criteria for ordering echocardiograms.13,15 These authors suggest that by simply teaching providers about AUC indications for echocardiograms via modalities including lectures and reference cards, there is a significant change in ordering behavior.
While we demonstrated in this meta-analysis that the addition of feedback tools was not necessary when combined with other QI modalities, there is limited available data to draw conclusions regarding the effectiveness of feedback tools when used as the solitary QI intervention. It has been seen in QI studies aimed at reducing rarely appropriate testing done using other types of radiological imaging that feedback tools are effective when used alone as an intervention.20–22 To our knowledge, there have not been any similar studies conducted for AUC-based QI interventions targeted towards rarely appropriate echocardiogram testing. Furthermore, it is possible that the format for feedback delivery is also important. In the study by Bhatia et al, it was observed that when feedback was given in both email format as well as performance reports, many providers ignored email feedback but did access their performance reports.12 Further studies are needed to better clarify both the role of feedback tools and format for feedback in AUC-based QI intervention.