PoCUS: Just because we can doesn't mean we should.

Abstract

As the need for expedited and accurate diagnostic testing continues to increase across care settings, point of care ultrasound (PoCUS) has emerged as an indispensable modality. However, this commentary is designed to stimulate conversations and new research directions related to the indications for PoCUS and the potential harms associated with potentially indiscriminate use. PoCUS has found its place as an invaluable tool in a number of clinical scenarios: one, in helping guide the resuscitation of critically ill patients and determining underlying etiology; two, in supplementing clinical history and findings to help narrow a differential diagnosis; and three, to provide procedural guidance for common bedside procedures including vascular access, thoracentesis, lumbar puncture, and endotracheal intubation, among others.1, 2 PoCUS has also been shown to be cost-effective in the emergency department (ED) setting.3 The systematic review and meta-analysis published in Academic Emergency Medicine evaluated the accuracy of clinical signs and symptoms, computed tomography (CT) with angiography, and PoCUS for the diagnosis of ruptured abdominal aortic aneurysms (rAAA). Pooled sensitivities were 27.8%–61.7% for clinical symptoms, 30.9%–47.1% for examination findings, and 91.4% for CT with angiography; while PoCUS did not have the fidelity to detect rAAAs, its pooled sensitivity and specificity for AAAs in patients with suspected ruptures was 97.8% and 97.0%, respectively. The authors concluded: “CTA has reasonable accuracy, but misses some cases of rAAA. PoCUS is a valuable tool that can help guide the need for urgent transfer to a vascular center in patients suspected of having rAAA.”4 Where PoCUS is increasingly seen as an extension of the stethoscope in clinching critical diagnoses, it has important caveats. First, its diagnostic yield is operator-dependent; while it is commonly used at the bedside by non-echocardiographers and non-radiologists, it requires mastery of skills in image generation through appropriate transducer manipulation, accurate interpretation of findings, and integration into clinical decision-making.1 Second, widespread adoption of the tool has been challenged due to gaps in structured training across medical curricula and faculty development programs, lack of clear competency standards, and variable resource availability. Third, while it may improve diagnostic accuracy, its utility in impacting patient-important outcomes have not yet been clearly established.5 A recent systematic review of studies evaluating PoCUS as a diagnostic tool to determine the underlying cause of acute dyspnea found significant improvement in diagnostic sensitivities and significantly more correct diagnoses, but no clear benefit in terms of in-hospital mortality and length of hospital stay. Fourth, the tool has been recommended as an adjunct to standard testing for acute presentations like dyspnea, rather than as a replacement; thus, PoCUS has generally not been recommended as a stand-alone diagnostic tool.6 Let us consider a case. A 51-year-old male presents to the ED with sudden severe back pain developing gradually after getting out of his vehicle. He has hypertension and an extensive pack-year smoking history. His symptoms and physical examination suggest a musculoskeletal etiology. As you work through your differential, abdominal aortic aneurysm is considered as a potential cause for his acute back pain. You decide to use PoCUS to evaluate the aorta; it appears normal. You extend the ultrasound probe further to do a renal assessment and notice a cystic mass on the left kidney. CT confirms a well-circumscribed renal lesion. You consider this to be an unlikely explanation for the patient's pain. Nonetheless, the patient is admitted to hospital. Urology is consulted, and the patient is consented for left-sided nephrectomy. The narrative is not uncommon. Data from 15 million fee-for-service Medicare beneficiaries aged 65–85 years between 2010 and 2014 shows that 43% of patients received a chest or abdominal CT. Furthermore, those residing in regions with higher scanning rates were at increased risk of undergoing nephrectomies, likely attributable to an increased rate of incidental detection of renal masses.7 Incidentalomas may account for close to half of kidney cancer diagnoses8, 9; despite kidney cancer incidence rapidly growing since the advent of cross-sectional imaging, there has been no marked change in kidney cancer-related mortality.8, 10 Richard Lehman, an editorialist for the British Medical Journal, captured the phenomenon aptly: ”more CTs, fewer kidneys.”11 Responding to busy clinical settings and delays to formal diagnostic imaging, faster patient turnarounds, increased fear of litigation for missed diagnoses, and risk aversion, as well as a growing need to manage patient and public expectations, many physicians may naturally turn to PoCUS as part of first-line work-ups. Overtesting, however, may propagate a domino effect of unanticipated harms. Incidental findings are an inevitable occurrence with any diagnostic work-up; but unanticipated, and frequently insignificant, findings on low-yield imaging tests may drive further imaging investigations with exposure to other risks such as those associated with contrast dye exposure, more diagnostic labels, and increased medical resource utilization and care burden. This may drive a rising incidence of diagnoses with no change in patient-important outcomes.12 A similar case in point from the emergency medicine world, applicable to overdiagnosis and overtreatment, is that of subsegmental pulmonary embolism (PE) diagnosed by CT. A previous study has shown that CT scans identify PEs in 5% more cases than ventilation-perfusion (VQ) lung scans in patients considered likely to have acute PE based on Wells scores and D-dimer assays; these patients subsequently carry a lifelong medical label and are treated with months of anticoagulation, but have been shown to have no significant difference in patient-important outcomes over intermediate-term follow-up.12, 13 Most clinicians understand that overdetection and overmedicalization is harmful, not helpful. Most are not invested in imaging or surgeries that cause more harm than benefit. Nevertheless, a desire for early detection, rapid assessment, diagnostic ”FOMO” (fear of missing out), and public and systemic perceptions may drive the juggernaut forward as can financial incentives in some jurisdictions. While incidental findings are an inevitable component of justifiable investigation, the liberal approach to PoCUS application can result in net harm in the pursuit of benefit. Was this avoidable? PoCUS has been shown to have higher sensitivity and specificity than other radiologic counterparts and clinical evaluations for AAAs.4 With regard to population level screening for AAA, pooled results of four large population-based randomized trials with over 130,000 patients have shown a mortality benefit and reduced risk of ruptures and need for emergent procedures; the Canadian Task Force on Preventive Health Care therefore updated its guidance in 2017 to recommend screening men aged 65–80 years; but not those younger. The Task Force acknowledged that, although AAA identification may result in elective procedures for some AAAs that may never have ruptured or caused morbidity, the relatively benefits greatly outweighed the harms associated with overdiagnosis in this setting. In other words, had the patient in our scenario been a 70 year old male, PoCUS could be considered far more justifiable. PoCUS may be most valuable in patients with a reasonably high pre-test probability or likelihood of having a AAA. The relative benefits in these patients, weighed against the risk of incidental findings and downstream consequences of overdiagnosis and medical excess, are likely most favorable. Routine screening for asymptomatic AAA requires substantial resources in a resource-limited ED setting, and offers a relatively low success rate.14 However, PoCUS should not replace the physical examination in this setting, which has moderate sensitivity in detecting AAAs large enough to be referred for surgery.15 Where palpation cannot be relied on to exclude AAA and pre-test probability is sufficiently high, PoCUS may be most strongly indicated.15 It is important to consider PoCUS-related limitations as well; where it may be useful to identify AAAs in cases of suspected ruptures, its ability to identify rAAAs reliably is more limited. This is well-acknowledged by the recent systematic review by Fernando and colleagues.4 Overdiagnosis is now a new medical subject heading (MeSH) term. In the year 2000, 52 articles were indexed on PubMed with the term “overdiagnosis”; by comparison, in 2021, 460 articles were indexed with the same. Multiple international efforts have moved the needle on recognizing the importance of reducing overdiagnosis, overtreatment, and medical excess. Choosing Wisely has emerged as an international campaign to reduce low-value tests, treatments, and procedures. Preventing overdiagnosis has also emerged as an international not-for-profit movement focused on winding back the harms of too much medicine, combating widening disease definitions that overmedicalize and over-label health as disease, and garnering an appreciation for unanticipated harms of excessive testing. Preventing Overdiagnosis has its next annual meeting in June 2022 (Calgary, Alberta) to further the dialogue on this area, and to guide further action through policy and research. The Lown Institute Right Care Alliance has similarly grown from a vision of matching the right evidence-based care to the right patient at the right time, with a parallel focus on reducing both overuse and underuse. Where these international efforts have frequently operated in silos, there appears to be a growing need to build bridges to place overdiagnosis on the agenda for system leaders, health care workers and patients. The onus of preventing overdiagnosis is relevant to all of us and is relevant to PoCUS. In the hands of frontline clinicians, selective approaches and caution are warranted. Informed decision-making around risks and benefits of this technology, with due consideration of the consequences associated with overdetection are warranted. In the same vein, PoCUS research efforts should also go beyond the realm of diagnostic test accuracy studies and should evaluate reporting of incidental findings and the outcomes associated with these.