Utilisation of echocardiography and application of the appropriate use criteria at a large tertiary hospital in Queensland

Abstract

Background

Medical imaging, including echocardiogram, is a major driver of escalating health‐care costs.

Aims

To examine the utilisation of transthoracic echocardiograms (TTEs) at a large tertiary hospital in Australia (Gold Coast Hospital and Health Service).

Methods

We retrospectively looked at 500 transthoracic echocardiograms ordered sequentially. The stated reason for requesting the TTE was then classified into one of 98 indications constructed by the ACCF. These requests were then determined to be ‘appropriate’, ‘inappropriate’, or ‘uncertain’ as per the appropriate use criteria (AUC).

Results

Overall, 443 TTEs were ordered appropriately (88.60%), 50 inappropriate (10.00%), and 8 met the criteria for uncertain. TTEs ordered by the cardiology team comprised of 131 appropriate (91.61%), 6 inappropriate (4.20%), and 6 uncertain (4.20%) indications. There were a higher percentage of appropriate TTE (93.23%) when performed on inpatients then when compared to those performed as an outpatient (82.91%).

Conclusions

Overall, Gold Coast Hospital and Health Service shows a good adherence to the AUC in keeping with similar studies in large hospitals overseas.

Introduction

Medical imaging overuse is a major driver of escalating health‐care costs,1 with cardiac imaging in particular being identified as an important area subject to overuse.2 Inappropriate testing and overuse of diagnostic tools not only equates to increased health‐care costs but also associated patient harm in the form of subjecting patients to the risks of the test itself and then to subsequent additional tests and procedures.3 Medical imaging overuse has manifested itself in particular in the United States,4 with cardiac imaging contributing to the significant majority of inappropriate or non‐contributory testing.5 Specifically, the annual rate of growth for transthoracic echocardiography (TTE) compared with other services provided by the US public health system was found to be disproportionately high.6

Transthoracic echocardiogram prima facie is a benign investigation; with the test itself carrying very few risks. It is also widely available in many medical centres around Australia. The investigation therefore lends itself to the high potential for misuse both locally in Australia and overseas.

In ‘an effort to respond to the need for the rational use of imaging services in delivery of high‐quality care’, the American College of Cardiology Foundation (ACCF) published the ‘Appropriate Use Criteria (AUC) for echocardiography’.7 The report was initially published in 2007, then expanded upon and revised in 2011. The report consists of 98 possible indications for transthoracic echocardiogram and each is assigned a rating indicating whether that indication is either ‘appropriate’, ‘inappropriate’, or ‘uncertain’.7

An appropriate imaging study is defined by the ACCF as ‘one in which the expected incremental information, combined with clinical judgement, exceeds the expected negative consequence by a sufficiently wide margin for a specific indication that the procedure is generally considered acceptable care and a reasonable approach for the indication’.7

There have been several studies examining the application of the AUC to transthoracic echocardiograms at various medical centres in the USA. However, data and testing of the AUC to transthoracic echocardiograms in other countries are much more limited.

Our aim is to apply the ACCF appropriate use criteria to the use of transthoracic echocardiograms at a large tertiary hospital in Queensland.

Methods

Our study retrospectively examined 500 TTE performed sequentially from 1 January 2014 to 6 February 2014 at the Gold Coast Hospital and Health Service. This included all transthoracic echocardiograms performed at the Gold Coast University Hospital and Robina Hospital in both an inpatient and outpatient setting.

The clinical reasoning behind each TTE was determined by examining the patient's electronic medical records (EMR) and echocardiogram request form. The clinical reason for requesting the TTE was then classified into one of the 98 TTE indications and seven disease categories constructed by the ACCF. All requests were assigned an indication as per the ACCF AUC and then determined to be either appropriate, inappropriate, or uncertain. Also tabulated was which specialty ordered the TTE, whether the specialty was cardiology or non‐cardiology, and whether the TTE was performed as an inpatient or outpatient. These results were then tabulated on an excel spreadsheet from which our data could be extrapolated. This process was performed by only one reviewer for all 500 TTEs.

All sequential TTEs were included in our study. Transoesophageal and stress echocardiograms were excluded on the basis that the ACCF has separate indications for these and that our aim is to examine TTEs only.

If a patient was under the care of a particular treating time, but the request was as per the advice of another team (e.g. a general medicine team requests a transthoracic echocardiogram based on cardiology advice), this request is to be taken as ordered by the advising team and not the treating team.

There was strictly no questioning of any clinical judgements and all requests and medical records were taken on their face value. For example, if a request form states ‘widespread ST elevation ?pericarditis’, the ECG was not checked for accuracy.

If an echocardiogram was ordered as an inpatient but performed as an outpatient, that is an outpatient TTE for the purposes of our study.

Lastly, often the request for TTE will fit in more than one of the 98 ACCF indications. In this case, it will be classified under the most appropriate indication. This is in fact how the ACCF intended the indication process to work.

Results

Who orders TTEs?

Out of our 500 TTEs, 143 (28.6%) were ordered by cardiologists, 136 (27.2%) by the general medicine teams, 49 (9.8%) by the neurology teams, 37 (7.4%) by oncology teams, 28 (5.6%) by respiratory physicians, 22 (4.4%) by the anaesthetic department, 21 (4.2%) by surgical teams, 18 (3.6%) by intensivists, and the last 46 (9.2%) ordered by other remaining medical and surgical specialities such as rheumatology or orthopaedics (Figure 1).

Figure 1.

Number of TTEs ordered by specialty.

This equates to 143 (28.6%) of TTEs ordered by cardiologists, and 357 (71.4%) ordered by non‐cardiologists.

Of these 500 TTEs, 266 (53%) were performed as an inpatient, with 69 (25.94%) ordered by cardiologists, and 197 (74.06%) ordered by non‐cardiology teams. Two hundred and thirty‐four (47%) TTEs were performed as an outpatient, with 74 (31.62%) of these being ordered by cardiologists and 160 (68.38%) being ordered by non‐cardiology teams.

Indications for TTE

Allocating the indications for our 500 echocardiograms into the seven disease categories developed by the ACCF, we found that 194 (38.8%) were ordered for ‘general evaluation of cardiac structure and function’, 109 (21.8%) for ‘evaluation of hypertension, heart failure of cardiomyopathy’, 71 (14.2%) for ‘evaluation of valvular function’, 60 (12%) for ‘cardiovascular evaluation in an acute setting’, 59 (11.8%) for ‘evaluation of intracardiac and extracardiac structures and chambers’, 4 (0.8%) for ‘evaluation of aortic disease’, and 3 (0.6%) for ‘evaluation of adult congenital heart disease’ (Figure 2).

Figure 2.

The seven different disease categories.

Classification of studies as per the appropriate use criteria

We found 443 (88.6%) TTEs were ordered for an appropriate indication as per the AUC, 50 (10%) for indications deemed inappropriate, and 8 (1.6%) uncertain (Figure 3).

Figure 3.

Comparison of AUC adherence by speciality.

In regard to the 143 TTEs ordered by cardiology teams, 131 (91.61%) indications were classified appropriate, 6 (4.20%) were classified as inappropriate, and the remaining 6 (4.20%) classified as uncertain. Of the 357 TTEs ordered by non‐cardiology teams, 311 (87.11%) indications were classified as appropriate, 44 (12.32%) classified as inappropriate, and 2 (0.56%) classified as uncertain (Figure 3).

Of the 266 TTEs performed on inpatients, 248 (93.23%) were for appropriate indications, 18 (6.77%) were for inappropriate indications, and none classified as uncertain. Of the 234 TTEs performed in an outpatient setting, 194 (82.91%) were for appropriate indications, 32 (13.68%) for inappropriate indications, and 8 (3.43%) classified as uncertain (Figure 4).

Figure 4.

AUC adherence by patient setting.

Inappropriate indications

The most common inappropriate indication for TTE was for ‘routine perioperative evaluation of ventricular function with no symptoms or signs of cardiovascular disease’, with 19 TTEs falling under this indication. Following this was ‘initial evaluation (of a murmur of click) when there are no other symptoms or signs of valvular or structure heart disease’, then ‘transient fever without evidence of bacteraemia or a new murmur’ (Figure 5).

Figure 5.

Inappropriate TTE categories.

Discussion

In our study, only 10% of TTEs ordered by the Gold Coast Hospital and Health services were for an indication deemed inappropriate by the AUC. We can see that this number decreases when a TTE was ordered by a cardiology team as opposed to a non‐cardiology team (Figure 3). However, it must be noted that the non‐cardiology teams include a conglomerate of both medical and surgical teams. TTEs were also more likely to be categories as appropriate if performed on an inpatient (Figure 4).

Comparisons with the USA

Similar studies have been performed in many centres in the United States of America. Three major USA studies using the 2007 AUC all found between 87% and 89% of TTE requests to be appropriate and 11–12% of TTE requests of be inappropriate, with only a minor increase in the number of appropriate requests when ordered by a cardiologist.8, 9, 10 These results are consistent with our study.

Due to similar methods, sample size and hospital size, we directly compared our results to two major studies in the USA. The first study by Aggarwal et al. was a chart review at a large tertiary centre in Minnesota reviewing 299 TTEs using the old 2007 AUC.6 Their study found that 84.50% of TTEs were ordered for appropriate indications, 6.40% ordered inappropriately and 9.10% of indications did not fit into any of the AUC indications.6 Their inability to classify some of the indications for TTEs was likely due to use of the old 2007 AUC, which did not contain as many TTE indications as the 2011 update. The second study for comparison was by Patil et al., which used the revised 2011 AUC.11 They reviewed 1825 patients who underwent TTE within the Saint Luke's Health System in Missouri and found that 82% of TTEs were ordered for an appropriate indication and 12.30% ordered for an inappropriate indication.11 Again, these results are not dissimilar to ours.

Comparisons outside the USA

Studies examining the AUC are much more limited outside the USA. A large study in Wales by Gurzun and Ionescu reviewed 1070 TTE requests from 14 different hospitals.12 Using the AUC, they found 86% of their TTEs were ordered for an appropriate indication, 11% for an inappropriate indication, and 3% classified as uncertain.12 Like the USA studies, these results are again similar to our study and it was also noted in their study that inpatient requests were more likely to be appropriate than outpatient requests.12

A study by Al‐Kaisey et al. looked at the use of TTE in a regional centre in Victoria, Australia. Using the AUC, they reviewed 1000 TTEs and found 77.20% to be appropriate, 20.30% inappropriate, and 2.50% classified as uncertain.13 As this is a marked reduction in the number of appropriately ordered TTEs when compared to our study, this could lead to the possibility that a Victoria regional setting has an impact on the adherence to the AUC when compared to a large tertiary centre in Queensland (Figure 6).

Figure 6.

A direct comparison of AUC adherence in our study with an interstate Australian regional centre.

Feasibility of the AUC

The appropriate indications for TTE determined by the ACCF are very broad. For example, appropriate indications include ‘symptoms or conditions potentially related to suspected cardiac aetiology including but not limited to chest pain, shortness of breath, palpitations, TIA, stroke, or peripheral embolic event’, ‘prior testing that is concerning for heart disease or structural abnormality including but not limited to chest X‐ray, baseline scout images for stress echocardiogram, ECG, or cardiac biomarkers’ and ‘sustained or non‐sustained atrial fibrillation, SVT, or VT’.7 An appropriate request for an echocardiogram may therefore include something as simple as ‘AF’ or ‘Shortness of breath of possible cardiac origin’. In contrast, the inappropriate indications are quite narrow, for example ‘routine perioperative evaluation of ventricular function with no signs or symptoms of heart disease’.7 This effectively means that TTE requests are much more likely to fit into appropriate criteria. This was also noted in the study by Aggarwal et al. who felt that criterion 1 in particular – symptoms potentially due to suspected cardiac aetiology, including but not limited to, shortness of breath, light‐headedness, syncope, TIA, cerebrovascular events – was far too broad.6 However, the ACCF does acknowledge that the appropriateness criteria are ‘purposefully broad to cover an array of cardiovascular signs and symptoms as well as the ordering physician's best judgement as to the presence of cardiovascular abnormalities’.7

Another comment on our study is that some requests for TTE can be difficult to fit into one of the 98 indications. For example, one request form in our study contained the information ‘past history of pericarditis, presented with chest pain, likely MSK, TTE to rule out pericarditis’. This was classified under the indication ‘suspected pericardial conditions’, even though it was not strictly suspected. Interestingly, in their study, Aggarwal et al. used two different reviewers to categorise their TTEs into one of the AUC criterion, and disagreement between observers was present almost one third of the time.6 This may make reproducibility of our study difficult.

Another difficulty faced was the idea of ‘routine surveillance’. Often, if an echocardiogram was performed for routine surveillance of a cardiac condition within a time frame [e.g. ‘routine surveillance (<1 yr) of heart failure when there is no change in clinical status or cardiac exam’], it was classified as an inappropriate indication under the AUC.7 However, if a TTE was for re‐evaluation of cardiac disease ‘with a change in clinical status or cardiac exam or to guide therapy’, this would make the indication for the TTE appropriate, even if ordered within the same time frame.7 This is a very hard differential to determine. The ACCF gives some guidance by stating that ‘Indications that describe routine or surveillance echocardiograms imply the test is being considered for a periodic evaluation since a certain period of time has elapsed. The test is not being ordered due to the anticipation of changing clinical decision making or guiding therapy’.7 In an outpatient setting, one could always argue that a test was performed for the purposes of guiding therapy, even if ordered simply because a certain time period has elapsed. The reviewer had to use his best judgement based on the TTE request form and patient EMRs as to whether a TTE was routine or not.

It is of this author's opinion that it is not appropriate to compare specialties in regard to adherence to the AUC. This is because specialties will tend to be directed towards certain indications when ordering a TTE. For example, a neurology team will almost always order a TTE to look for a cardiac thrombus as part of their stroke work‐up. This will fall into the indication ‘suspected cardiovascular source of embolus’ which is an appropriate indication.7 In fact, in our study all 49 TTEs ordered by the neurology team were appropriate because they were all ordered for this indication. In contrast, the general medical teams will often be faced with diagnostic dilemmas, meaning that they will be required to use TTEs for a broader range of indications.

Risk reduction for inappropriate testing

There may be strategies to reduce the number of inappropriate testing in both inpatient and outpatient settings. A prospective study by Bhatia et al. applied an educational tool to the Massachusetts General Hospital inpatient general medical team.14 This included lectures to house staff, a pocket card applying the AUC to common clinical scenarios, and a bi‐weekly feedback regarding ordering behaviour.14 A similar randomised prospective study was also performed with the cardiology and general medicine physicians‐in‐training ordering outpatient TTEs at the same medical centre.15 Both studies saw a reduction in the number of inappropriately ordered TTEs, offering as high as a 62% reduction rate.14, 15 Given that in our study almost 80% of the inappropriate TTEs ordered fall into only four categories, and given that electronic tracking applying the AUC to TTEs at point‐of‐order has been shown to be both feasible and accurate,16 a study examining strategies to achieve a reduction in inappropriate TTEs would be an interesting follow up. It is this author's opinion that reduction of inappropriate testing based on the AUC criteria would require the AUC to be seen as a gold standard and an internationally accepted criterion for the ordering of TTEs.

In our study, whether a TTE is ordered appropriately or inappropriately is guided by the appropriate use criteria as determined by the ACCF. This does not mean that we can say the TTE was ordered appropriately or inappropriately per se, but rather that it was ordered appropriately/inappropriately as per the appropriate use criteria determined by the ACCF. Interestingly, Ghatak et al. looked at outcomes of repeat echocardiograms and whether they yielded new clinical findings, and observed that the AUC did not appear to be a good predictor for this.17 However, the AUC is currently the best objective measure we have for categorising whether a TTE is appropriate or not. This can be evidenced by the incorporation of the AUC by the Intersocietal Accreditation Committee in Maryland USA, which requires tracking and reporting of appropriateness according to the AUC as a mandatory requirement for cardiovascular imaging as part of their accreditation process.18 Whether the AUC will be further applied outside the borders of the USA remains to be seen.

Conclusion

When ordering transthoracic echocardiograms, the Gold Coast Hospital and Health Service demonstrates good adherence to the AUC. There is only a slight increase in the number of appropriate TTEs when ordered by a cardiologist, and TTEs are most likely to be appropriate when performed on an inpatient.

Our results are in line with those of our overseas colleagues, and there is a closer adherence to the AUC at our large tertiary centre in Queensland when compared to a smaller Victorian regional centre.

Although the AUC is not a perfect measure of whether an echocardiogram is ordered appropriately, it is the most widely recognised guideline we currently have available.