Abstract
Background:
Emergency physicians (EP) perform critical actions while operating with diagnostic uncertainty. Point-of-care ultrasound (POCUS) is useful in evaluation of dyspneic patients. In prior studies, POCUS is often performed by ultrasound (US) teams without patient care responsibilities.
Objectives:
This study evaluates the effectiveness of POCUS in narrowing diagnostic uncertainty in dyspneic patients when performed by treating EPs versus separate US teams.
Methods:
This multicenter, prospective non-inferiority cohort study investigated the effect of POCUS performing team in patient encounters for dyspnea. Before and after surveys assessing medical decision making were administered to attending physicians. Primary outcome was change in most likely diagnosis after POCUS. This was assessed for non-inferiority between encounters where the primary or US team performed POCUS. Secondary outcomes included change in differential diagnosis, confidence in diagnosis, interventions considered, and image quality.
Results:
156 patient encounters were analyzed. In the primary team group, most likely diagnosis changed in 40% (95% CI, 28–52%) of encounters versus 32% (95% CI, 22–41%) in the US team group. This was non-inferior using an a priori specified margin of 20% (p < .0001). Post-POCUS differential decreased by mean 1.8 diagnoses and was equivalent within a margin of 0.5 diagnoses between performing teams (p = 0.034). Other outcomes were similar between groups.
Conclusion:
POCUS performed by primary teams was non-inferior to POCUS performed by US teams for changing the most likely diagnosis and equivalent when considering mean reduction in number of diagnoses. POCUS performed by treating EPs reduces cognitive burden in dyspneic patients.
Keywords: ultrasound, dyspnea, differential diagnosis, POCUS
Introduction
Shortness of breath is a common Emergency Department (ED) complaint that often requires prompt evaluation and treatment. Patients presenting with dyspnea are often in extremis and have a high risk of mortality.1 Decision-making can be complex, and physicians have to act in a short period of time with a high degree of diagnostic uncertainty. Physical examination may direct clinical suspicion, but it is often difficult to differentiate based on exam alone and diagnosis often relies on laboratory and radiographic investigation.2,3 Chest X-ray (CXR) is often the initial diagnostic study of choice, but poor sensitivity compared to other modalities such as ultrasound, may limit its ability to resolve diagnostic uncertainty.4–6 In addition, obtaining and reading a CXR takes valuable time and may contribute to a delay in diagnosis. Likewise, chest computed tomography (CT) is an alternative and effective option,7 but can be limited in both its availability and ease of use in critically ill patients.
Point-of-care ultrasound (POCUS) has become a widely used and accurate tool for the initial evaluation of the critically ill emergency department patient with well established guidelines for appropriate use8. POCUS has been shown to have excellent test characteristics for the diagnosis of common ED causes of dyspnea including pulmonary edema, pneumothorax, pneumonia, and pleural effusion.9–12 Previous studies have also suggested that initial POCUS evaluation may facilitate rapid diagnosis in the acutely dyspneic patient,9,10 suggesting that there may be a wider role for the routine use of POCUS in the dyspneic patient in the ED. However, these studies are often limited by performance of POCUS by Emergency Ultrasound fellowship trained clinicians and/or a dedicated ultrasound team without patient care responsibilities that is distinct from the primary treating team. Prior studies evaluating the use of POCUS in the medical decision-making process have had similar limitations.13 We evaluated the effectiveness of POCUS in narrowing diagnostic uncertainty and guiding initial management of acutely dyspneic patients when performed by the treating providers compared to POCUS performed by a separate team of ultrasound-trained clinicians. This study is the first to evaluate the utility of POCUS in the medical decision-making process when it is performed by the primary treating team of clinicians on acutely dyspneic patients.
1. Methods
1.1. Study setting
This prospective, multi-center, non-inferiority cohort study was performed between January 2018 and February 2020 in two emergency departments: a university affiliated hospital with approximately 75,000 adult visits per year (university site) and a large community hospital (community site) with approximately 90,000 adult visits per year. Of note recruitment was suspended for a 7-month period during 2019 due to lack of available research staff. Both hospitals have full-time resident coverage and are core sites in the same Emergency Medicine residency program. However, the attending physicians at both sites are employed in different groups and have no overlap in practice locations or clinical guidelines. Patients are evaluated by an emergency physician (EP) or by a resident or advanced practice provider with EP supervision at both sites. More than 6,000 educational and diagnostic POCUS studies per year are performed at the university site and approximately 3,500 per year are performed at the community site. Both sites have weekday, daytime coverage by a separate diagnostic ultrasound team that performs POCUS studies that is focused on trainee education. The institutional review boards of both sites approved the study. Written informed consent was obtained from all attending EPs whose decision-making was studied.
1.2. Selection of Participants
This study evaluated emergency physicians’ decision-making when evaluating patients with undifferentiated dyspnea and each physician-dyspnea patient encounter was considered a single observation. A convenience sample of encounters between EPs and patients with undifferentiated dyspnea was considered for this study. Undifferentiated dyspnea was defined as patients greater than 18 years of age who did not have an identified cause of dyspnea at the time the pre-POCUS survey was administered and thus presented diagnostic uncertainty. The determination of undifferentiated dyspnea was made by the treating physician. Care was taken to minimize the time between completion of the pre-POCUS survey, performance of POCUS and completion of the post-POCUS survey however there was no blinding to other new clinical data. All eligible attending physicians had completed an emergency medicine residency and were employed by their institutions. EPs completed pre- and post-POCUS surveys at both locations based on the presence of a recruiter, the presence of the ultrasound team, or physician preference. Patient encounters were excluded if the patient had already received a POCUS study (either by the primary treating team or the ultrasound team) and thus no patient encounters included in this study had POCUS performed by both groups.
1.3. Study Design and Measurements
Enrolled physicians initially evaluated patients with undifferentiated dyspnea with detection of vital signs, a medical history, and physical examination. They then ordered all diagnostic evaluations deemed necessary such as CXR, CT, blood sampling, or blood gas analysis. The treating attending physician then filled out a standardized form (Supplemental Figure S1) specifying which diagnoses among congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), pneumothorax (PTX), pulmonary embolism (PE), pneumonia (PNA), pericardial effusion, pleural effusion, acute coronary syndrome (ACS), acute respiratory distress syndrome/acute lung injury (ARDS/ALI) and other diagnosis formed their differential diagnosis as well as which was the most likely diagnosis. The attending physician then indicated which initial interventions they would initiate among IV fluids, breathing treatments, nitroglycerin, chest tube, IV tissue plasminogen activator (TPA), non-invasive positive pressure ventilation (NIPPV), and intubation. Finally, the treating physician indicated their overall confidence in their diagnosis.
After completion of the form, patients underwent ultrasonographic evaluation by the treating physician and team (primary team) or by a separate diagnostic ultrasound team (US team) led by ultrasound faculty. Enrollment into one of these two groups was based on the presence and availability of the ultrasound team which was generally Monday through Friday during daytime hours. Both groups underwent a similar patient-specific, focused ultrasonographic evaluation. For the ultrasound team, all ultrasound scans were either performed or supervised by faculty. For the treating physician group, training of the ultrasound operator varied from ultrasound fellowship trained attending physicians to junior residents with less than one month of ultrasound specific training. Ultrasonographic evaluation consisted of targeted scans consisting of one or more of lung, cardiac (echocardiogram), inferior vena cava (IVC), or other focused study based on patient presentation, history and physical exam. POCUS was performed with multiple ultrasound systems based on site: M9 (Mindray Inc. Mahwah, NJ) at the university site and TE7 (Mindray Inc.) or M-Turbo (Sonosite Inc. Bothell, WA) at the community site.
There was no mandated POCUS protocol and studies performed were consistent with routine care at either site. Generally, a combination of lung ultrasound, focused transthoracic echocardiography, and inferior vena cava (IVC) evaluation was performed. Lung ultrasound was performed to evaluate for pulmonary edema, pneumonia, and pneumothorax as previously described.14,15 Evaluation of left ventricular ejection fraction, acute right heart strain and pericardial effusion were performed also as described elsewhere.16–21 If applicable, analysis of IVC variation with respiratory cycle was performed.22,23 Interpretation of any POCUS findings was at the discretion of the treating attending physician and no protocolized interventions or additional diagnostic actions were mandated. In the case of primary team performed POCUS, the treating attending physician interpreted the images and integrated findings into their medical decision-making directly. When the US team performed the study, the results were reported to the treating attending physician who could also view the images if desired. In both cases, the treating physician then filled out a post-ultrasound form indicating the most likely diagnosis, differential diagnosis, interventions and physician confidence.
1.4. Bias
Neither the treating physicians or US team were blinded to other diagnostic evaluations that resulted during the time taken to perform the ultrasound and discuss results. While care was taken to minimize the time elapsed between completion of pre- and post-ultrasound forms, there remains a degree of maturation bias as this time interval was not measured. This was thought to represent a realistic clinical scenario given that POCUS is generally not performed in isolation from other diagnostic evaluations and is further discussed in the limitations section below. At both sites, routine practice in POCUS is to document radiographic findings rather than provide specific diagnoses to minimize editorializing findings. Evaluating this difference between US team and primary team performed POCUS is the central aim of this study.
1.5. Exposure Variables, Outcomes and Secondary Data
The primary exposure for each dyspnea patient encounter was considered performance of the POCUS study by either the primary team or the US team. The primary outcome was the change in the most likely (primary) diagnosis as reported by the treating EP after receiving the results of POCUS. Secondary outcomes included change in differential diagnosis, change in confidence in diagnosis assessed on a Likert-type scale, and change in interventions.
At both sites, the majority of POCUS studies are stored in Qpath (Telexy Inc. Maple Ridge, BC, Canada) with matching patient identifiers for the purposes of billing and/or quality assurance. However, a minority of POCUS studies are stored without patient data entered at the bedside due to time constraints. For studies with matching patient identifiers, the patient record was reviewed after patient discharge to determine presenting vital signs, oxygen support, demographics, and ED disposition. All data was abstracted by one of three reviewers using single data entry (AB, VL, BF). These data were objective and not subject to interpretation by reviewers who were either house officers or senior medical students.
1.6. Post-hoc image quality analysis
Images from the first 12 consecutive POCUS studies from both the primary team performed and US team performed studies with linked patient data were extracted from the Qpath image archive and deidentified. These 24 studies were interpreted by 4 Emergency Ultrasound fellowship trained physicians. All image reviewers were blinded to the study findings and performing team. Image quality was rated on a four-point scale using a rubric (Supplemental Figure S2). Reviewers were also asked to identify the most likely of the study diagnoses based only on images and a short clinical vignette that included chief complaint, age and triage vital signs.
1.7. Statistical Analysis
Descriptive statistics were performed on all outcomes of interest. Differences between patient encounter characteristics were assessed with Welch’s t-test or Fisher’s exact test where appropriate. Non-inferiority was defined as a change in primary diagnosis with a non-inferiority margin of 20% below that of the US team for POCUS completed by the primary team. Statistical significance for non-inferiority was assessed using the Farrington-Manning method24. Differences in means of number of diagnoses or management options were assessed using Welch’s t-test. Equivalence between groups was assessed using the two one-sided test method25 and an equivalence margin of plus or minus 0.5 diagnoses, intervention options or levels on the Likert-type confidence in diagnosis scale. In our analyses, a p-value < 0.05 indicated statistically significant non-inferiority or equivalence. The non-inferiority margin of 20% was determined a priori by consensus of the study investigators as the limit of acceptable difference for change in diagnosis when POCUS is performed by the primary team compared to the US team. The equivalence margin of 0.5 diagnoses, intervention options or levels on the Likert-type confidence scale was determined post hoc but thought to be clinically meaningful and represents a higher statistical bar than applying a superiority test. For image review, a two-way random, consistency single measure intraclass correlation coefficient was used with image quality assessment and Fleiss’ kappa statistic was calculated to assess interrater reliability of diagnoses.26 All statistical analyses were completed using SAS 9.4 (SAS Institute Inc. Cary, NC).
1.8. Sample Size Estimation
A sample size of 146 subjects was estimated based on an expected change in primary diagnosis 40% of the time in both the US team and primary team groups and a non-inferiority margin of 20% (β = 0.2, α = 0.05). The estimate of 40% change in primary diagnosis was based on results from the initial 20 subjects.
2. Results
2.1. Demographics of participating attending physicians
Descriptive characteristics of the cohort of participating attending physicians at both sites are listed in Table 1. 60% of the attending physicians were faculty at the university affiliated ED, 35% were female, and the mean post-graduate years was 16.3. There were 8 total attending physicians who were designated ultrasound faculty by their departments and of these, 4 were fellowship trained. Individual attending physicians were not linked to specific POCUS studies per our study protocol to maintain confidentiality for both attending physicians and patients.
Table 1. Attending Physician Characteristics.
Characteristics of attending physicians participating in the study. Individual participants were not linked to POCUS studies performed per study protocol due to concern for confidentiality of medical decision making. No significant differences seen in site demographics. PGY = postgraduate year(s), US = Ultrasound.
| University Site | Community Site | Total | |
|---|---|---|---|
| Participating attendings, n (%) | 39 (60%) | 26 (40%) | 65 (100%) |
| Female, n (%) | 11 (28%) | 12 (46%) | 23 (35%) |
| PGY, mean years (SD) | 16.0 (10.1) | 16.6 (7.3) | 16.3 (9.1) |
| PGY, median years (IQR) | 15 (8 – 21) | 15 (10 – 20) | 15 (9 – 20) |
| Ultrasound Faculty, n (%) | 5 (12.8%) | 3 (11.5%) | 8 (12.3%) |
| US Fellowship Trained, n (%) | 3 (7.8%) | 1 (3.8%) | 4 (6.2%) |
2.2. Description of POCUS studies performed and patient encounter characteristics
156 total encounters were studied. Of these, 91 had POCUS performed by the US team, and 65 had POCUS performed by the primary team and their breakdown by study site are reported in Figure 1. On retrospective review of the Qpath system at each site, 119 studies (76%) contained direct identifiers from which the patient record could be linked to study data (Table 2). For these 119 patient encounters with available clinical data, demographics, presenting characteristics and vital signs are summarized in Table 2. Patient encounter characteristics were not significantly different for primary and US team performed studies. Out of 156 completed surveys, 154 indicated which POCUS components were performed. Primary team performed POCUS studies were less likely to include all three of echocardiography, IVC and lung views (59% vs 84%, p = 0.0004).
Figure 1. Patient Encounter Flow Diagram.
156 total dyspnea patient encounters were deemed by the treating attending physician to have diagnostic uncertainty and POCUS was used in the evaluation of these patients. For these encounters, Qpath image archive was searched for matching images and thus corresponding linked patient data.
Table 2. Patient Encounter Characteristics.
Demographics of patients and ultrasound studies included in the study. Patient data was obtained retrospectively from ultrasound studies where patient identifiers were saved to the Qpath system. No significant differences were seen in patient demographics between primary team and US team. Primary respiratory chief complaint was defined as a reported chief complaint at triage including shortness of breath, dyspnea, or respiratory distress. Secondary respiratory chief complaint was defined as any other chief complaint with the patient later reporting dyspnea to the primary provider. Admission included both inpatient and observation admission status. SD = standard deviation, ESI = Emergency Severity Index (1–5), BPM = beats or breaths per minute, NIV = non-invasive ventilation, ECHO = echocardiogram, IVC = visualization of the inferior vena cava.
| Primary team (n=45) | US team (n=74) | Total (n=119)* | |
|---|---|---|---|
| Age, mean years (95% CI) | 68.7 (64.5 – 72.9) | 70 (67.2 – 72.8) | 69.5 (66.5 – 72.5) |
| Female, n (%, 95% CI) | 19 (42%, 29 – 57%) | 39 (53%, 41 – 64%) | 58 (49%, 40 – 58%) |
| Triage ESI, mean (95% CI) | 2.0 (1.9 – 2.2) | 2.1 (2.0 – 2.2) | 2.1 (2.0 – 2.2) |
| Admission, n (%, 95% CI) | 42 (93%, 82 – 98%) | 70 (95%, 87 – 98%) | 112 (94%, 88 – 97%) |
| Triage Vital Signs, mean (95% CI) | |||
| Temperature (°C) | 36.9 (36.7 – 37.1) | 36.7 (36.6 – 36.8) | 36.8 (36.7 – 36.9) |
| Heart Rate (BPM) | 96 (89 – 102) | 92 (88 – 96) | 93 (90 – 97) |
| Systolic Blood Pressure (mmHg) | 135 (127 – 143) | 136 (129 – 143) | 136 (130 – 141) |
| Diastolic Blood Pressure (mmHg) | 74 (70 – 78) | 77 (72 – 81) | 76 (73 – 79) |
| SpO2 (%) | 94 (92 – 96) | 94 (93 – 95) | 94 (93 – 95) |
| Respiratory Rate (BPM) | 23 (21 – 25) | 23 (21 – 24) | 23 (22 – 24) |
| Oxygen treatment at arrival, n (% total) | |||
| None | 18 (40%) | 28 (38%) | 46 (39%) |
| Nasal Cannula < 15 L | 21 (47%) | 40 (54%) | 61 (51%) |
| Nasal Cannula >= 15 L | 2 (4%) | 4 (5%) | 6 (5%) |
| NIV | 4 (9%) | 2 (3%) | 6 (5%) |
| Views obtained, n (%) | Primary team (n=65) | US team (n=89) | Total (n=154)** |
| ECHO only | 3 (5%) | 1 (1%) | 4 (3%) |
| Lung only | 6 (9%) | 2 (2%) | 8 (5%) |
| ECHO + Lung | 13 (20%) | 8 (9%) | 21 (14%) |
| ECHO + IVC | 5 (8%) | 3 (3%) | 8 (5%) |
| ECHO + Lung + IVC | 38 (59%)# | 75 (84%) | 113 (73%) |
119 out of 156 studies had linked patient data for which patient demographics could be reported.
154 out of 156 encounters reported views obtained for the study
Primary team performed POCUS was less likely to include all of ECHO, Lung and IVC views (p = 0.0004).
2.3. Changes in diagnosis
The attending physician changed their primary diagnosis in 32% (95% CI 22.2 – 41.4%) of cases when POCUS was performed by the US team and in 40% (95% CI 28.1 – 51.9%) of cases when a member of the primary team performed POCUS (Table 3). Primary team performed POCUS was non-inferior to US team performed studies when assessed with a non-inferiority margin of 20% (p < 0.0001). The mean number of diagnoses considered was decreased by 1.8 after POCUS from 4.1 diagnoses to 2.3 (p < 0.0001). This result was consistent regardless of the team performing POCUS (Table 3a) and also similar across both sites (3.9 diagnoses decreased to 2.2 at the university affiliated site, compared to 4.6 diagnoses decreased to 2.5 at the community site, Table 3b). Using an equivalence margin of plus or minus 0.5 diagnoses, the change in mean number of diagnoses considered was statistically equivalent between the primary and US team performed studies while the difference between the university and community sites was not statistically equivalent. Similarly, the mean confidence level of attending physicians in their primary diagnosis increased post-POCUS by 0.7 points on a 5-point Likert-type scale (p < 0.0001). Using an equivalence margin of plus or minus 0.5 points on this scale, this increase in confidence was equivalent between primary and US team performed studies (Table 3a) as well as between the community and university sites (Table 3b).
Table 3a.
Post-POCUS change in primary diagnosis, # diagnoses in differential, management, and attending confidence in diagnosis by team that performed the POCUS study; Primary Team vs dedicated US Team. For non-inferiority, the null hypothesis was primary team change in primary diagnosis greater than 20% below that of the US team. For equivalence, the null hypothesis was primary team change in mean # of diagnoses, mean confidence level, or mean # of management options greater than 0.5 diagnoses, confidence levels, or interventions. Frequency of change in the primary diagnosis when the primary treating team performed the POCUS was non-inferior to when POCUS was performed by the US team within a margin of 20%. Likewise, equivalence between the two performing teams was declared for other metrics within a margin of +/− 0.5. POCUS=Point of Care Ultrasound. A p-value < 0.05 indicates statistically significant non-inferiority or equivalence.
| Primary Team | US Team | p-value | |
|---|---|---|---|
| Number POCUS Studies Performed, n | 65 | 91 | |
| Change in Primary Diagnosis, n (%, 95% CI) | 26 (40%, 28 – 52%) | 29 (32%, 22 – 41%) | < .0001* |
| Any Change in Management, n (%, 95% CI) | 22 (34%, 22 – 45%) | 29 (32%, 22 – 41%) | 0.0014* |
| Pre-POCUS Confidence, Mean 5-point scale | 3.2 | 3.2 | |
| Post-POCUS Confidence, Mean 5-point scale | 3.9 | 3.9 | |
| Difference in mean confidence (95% CI) | 0.7 (0.5 – 1.0) | 0.6 (0.5 – 0.8) | 0.0059** |
| Pre-POCUS Differential, mean # diagnoses | 4.1 | 4.1 | |
| Post-POCUS Differential, mean # diagnoses | 2.3 | 2.3 | |
| Difference in mean # of diagnoses post-POCUS (95% CI) | −1.8 (−2.2 – −1.4) | −1.8 (−2.1 – −1.4) | 0.034** |
| Pre-POCUS management, mean # interventions | 1.0 | 0.8 | |
| Post-POCUS management, mean # interventions | 1.0 | 0.8 | |
| Difference in mean # of interventions post-POCUS (95% CI) | 0.0 (−0.2 – 0.1) | −0.1 (−0.2 – 0.1) | < .0001** |
p-value for non-inferiority with −20% margin
p-value for equivalence with ± 0.5 diagnosis, ± 0.5 intervention or ± 0.5 level of confidence difference
Table 3b.
Post-POCUS change in primary diagnosis, # diagnoses in differential, management, and attending confidence in diagnosis by team that performed the POCUS study; Community Site vs. University Site. For non-inferiority, the null hypothesis was primary team change in primary diagnosis greater than 20% below that of the university site. For equivalence, the null hypothesis was community site change in mean # of diagnoses, mean confidence level, or mean # of management options greater than 0.5 diagnoses, confidence levels, or interventions. Frequency of change in the primary diagnosis when POCUS performed at the community site was non-inferior to when POCUS was performed at the university site within a margin of 20%. Equivalence between the two sites was declared for other metrics within a margin of +/− 0.5. A p-value < 0.05 indicates statistically significant non-inferiority or equivalence.
| Community Site | University Site | p-value | |
|---|---|---|---|
| Number POCUS Studies Performed, n | 36 | 120 | |
| Change in Primary Diagnosis, n (%) | 14 (39%, 23 – 55%) | 41 (34%, 26 – 43%) | 0.0013* |
| Any Change in Management, n (%, 95% CI) | 8 (22%, 9 – 36%) | 43 (36%, 27 – 44%) | 0.21* |
| Pre-POCUS Confidence, Mean 5-point scale | 3.1 | 3.3 | |
| Post-POCUS Confidence, Mean 5-point scale | 3.8 | 3.9 | |
| Difference in mean confidence (95% CI) | 0.6 (0.3 – 1.0) | 0.7(0.5 – 0.9) | 0.01** |
| Pre-POCUS Differential, mean # diagnoses | 4.6 | 3.9 | |
| Post-POCUS Differential, mean # diagnoses | 2.5 | 2.2 | |
| Difference in mean # of diagnoses post-POCUS (95% CI) | −2.1 (−2.7 – −1.59) | −1.7 (−2.0 – −1.4) | 0.43** |
| Pre-POCUS management, mean # interventions | 1.1 | 0.8 | |
| Post-POCUS Management, mean # interventions | 1.3 | 0.7 | |
| Difference in mean # of interventions post-POCUS (95% CI) | 0.1 (0.0 – 0.3) | −0.1 (−0.2 – 0.0) | 0.015** |
p-value for non-inferiority with −20% margin
p-value for equivalence with ± 0.5 diagnosis, ± 0.5 intervention or ± 0.5 level of confidence difference
When assessed after POCUS, the differential diagnoses were more often narrower than prior to POCUS. The most common diagnoses removed were tamponade, 89% (removed from the differential in 57 of 64 subjects where tamponade was part of the initial differential); pneumothorax, 95% (removed from the differential in 18 of 19 subjects where it was part of the initial differential); PE, 52% (removed from the differential in 51 of 98 subjects where it was part of the initial differential); pleural effusion, 61% (removed from the differential in 46 of 75 subjects where it was part of the initial differential). The diagnoses most frequently added to the differential were pleural effusion and pulmonary embolism. Pleural effusion was added in 6 of 156 subjects, of which it was not part of the initial differential in 81 and remained in the differential in 29 of 156 subjects. Pulmonary embolism was added in 7 of 156 subjects, of which it was not part of the initial differential in 51 and remained in the differential in 47 of 156 subjects. Notably, cardiac tamponade, which is a rare but critical diagnosis requiring significantly different management, was added to the differential in 3 subjects. Changes to individual diagnoses showed similar patterns of addition, removal and remaining in the differential when either the primary team performed the POCUS or when it was performed by the US team. Changes to individual diagnoses are presented in Figure 1.
When CHF was the most likely pre-test primary diagnosis, it was changed in 36% of cases (23 subjects out of 64) and was most often changed to pneumonia (9 subjects out of 64). When COPD was the most likely pre-test primary diagnosis, it was changed in 46% of cases (16 subjects out of 35) and was most often changed to CHF (8 subjects out of 35). Pneumonia was the most likely pre-test primary diagnosis in 33 subjects, and it was changed to CHF in 3 subjects and to pulmonary embolism in 3 subjects. These and other changes for individual diagnoses from pre-test to post-test are detailed in Table 4.
Table 4. Change to individual primary diagnosis.
The pre-POCUS primary diagnosis, as determined by treating attending physician, and total number within the cohort are reported. The subsequent post-POCUS diagnosis and percentage of pre-POCUS diagnosis are also reported.
| Pre-POCUS Primary Diagnosis (N) | Post-POCUS Primary Diagnosis | N (%) |
|---|---|---|
| CHF (64) | CHF | 41 (64%) |
| Pneumonia | 9 (14%) | |
| COPD/Asthma | 6 (9%) | |
| Pulmonary Embolism | 4 (6%) | |
| Other | 2 (3%) | |
| Cardiac Tamponade | 1 (2%) | |
| Acute Coronary Syndrome | 1 (2%) | |
| COPD/Asthma (35) | COPD/Asthma | 19 (54%) |
| CHF | 8 (23%) | |
| Pneumonia | 4 (11%) | |
| Cardiac Tamponade | 2 (6%) | |
| Pleural Effusion | 1 (3%) | |
| Other | 1 (3%) | |
| Pneumonia (33) | Pneumonia | 25 (76%) |
| Pulmonary Embolism | 3 (9%) | |
| CHF | 3 (9%) | |
| COPD/Asthma | 1 (3%) | |
| Acute Coronary Syndrome | 1 (3%) | |
| Pulmonary Embolism (8) | Pulmonary Embolism | 5 (63%) |
| Pneumonia | 1 (13%) | |
| CHF | 1 (13%) | |
| Acute Coronary Syndrome | 1 (13%) | |
| Other (6) | Other | 6 (100%) |
| Pleural Effusion (4) | Pleural Effusion | 3 (75%) |
| Pneumonia | 1 (25%) | |
| Acute Coronary Syndrome (3) | CHF | 2 (67%) |
| Acute Coronary Syndrome | 1 (33%) | |
| ARDS (2) | Pleural Effusion | 1 (50%) |
| ARDS | 1 (50%) | |
| Cardiac Tamponade (1) | Pulmonary Embolism | 1 (100%) |
2.4. Changes in management
Changes to management were less frequent overall than changes to differential diagnoses and the net change in number of management options considered pre- and post-POCUS was zero. However, when examining individual management changes in subjects in which a particular management option was initially considered (Figure 2), frequent management changes were observed. The intervention most frequently added to management was IV fluids in 13 subjects. IV fluids were considered in only 17 patients initially. The intervention most frequently removed from management was use of inhaled bronchodilators, in 24 subjects, and it was initially considered in 67 subjects. Nitroglycerin was part of initial management in 18 subjects and was removed 50% of the time post-POCUS. TPA, intubation, and chest tube/thoracentesis were overall infrequently considered as initial management strategies pre- or post-POCUS. TPA was never considered in reported management. Intubation was considered in initial management in 5 subjects and removed in one of the 5 cases; chest tube/thoracentesis was considered part of management only once, in which it was added to management post-POCUS.
Figure 2. Change in Differential by Diagnosis.
Post-POCUS change in each of 9 individual diagnoses in A) overall cohort, B) US team performed studies, C) primary team performed studies. Percentage reported is percent of total studies for the overall cohort (N = 165), US team performed studies (N = 91) and primary team performed studies (N = 65). Frequency of each change if it was not in the original differential, remains in the differential, removed from the differential or added to the differential after POCUS are reported. ARDS = acute respiratory distress syndrome; CHF = congestive heart failure; COPD = chronic obstructive pulmonary disease.
2.5. Image quality
All images reviewed for quality were from the university site. Of these 24 studies that underwent blinded review by four fellowship trained experts, the mean overall quality of POCUS studies, on a 4-point scale (Supplemental Figure S2), was rated 3.0 and there was a small non-statistically significant difference in mean quality for POCUS performed by the US team over the primary team (3.2 vs 2.8, p = 0.05). There was moderate interrater reliability for image quality between the four blinded reviewers (ICC = 0.60, 95% CI 0.40 – 0.77) using the same 4-point scale. There was also moderate interrater reliability between raters for diagnosis (Fleiss’ kappa = 0.56, 95% CI 0.48 – 0.64).
3. Discussion
This multi-center, prospective, non-inferiority cohort study found that use of POCUS changes the most likely diagnosis in cases of undifferentiated dyspnea in the emergency department and can narrow the differential diagnosis when performed by the primary treating team. This effect is non-inferior to POCUS performed by a separate US team. These findings suggest that POCUS can aid providers in their initial management of acutely dyspneic patients and can decrease the cognitive burden of the initial patient evaluation. This is extremely important in the emergency setting as rapid and accurate diagnosis improves efficiency of care.
Our findings are similar to those of previous observational trials of POCUS in undifferentiated dyspnea, which have shown that the use of ultrasound can narrow a provider’s differential diagnosis and change the most likely diagnosis.9,27,28 These previous trials had relatively narrow applicability, as a separate team led by an ultrasound faculty member performed all ultrasounds. A study by Umuhire et al., demonstrated a 66% change in primary diagnosis after POCUS29. However, in this study a separate physician sonographer team performed all POCUS studies and the greater percent change in primary diagnosis can likely be attributed to different patient populations and greater availability of other diagnostic testing at our centers. Most practice environments lack a dedicated emergency ultrasound team even during daytime hours. Our study demonstrates that POCUS performed by the primary provider is non-inferior in modifying the primary diagnosis being considered and contributes to a narrowing of the differential diagnosis.
This is an important distinction because criticisms of prior studies utilizing a dedicated US team have implied that primary treating teams are more likely to utilize POCUS in their decision-making if able to offload the performance of the study to another team or if they have expert interpretation of studies. Our study demonstrates that the effect of POCUS on most likely diagnosis and narrowing of the differential was similar when either team performed the study. Importantly, the increase in confidence of the primary team in their diagnosis was equivalent when they performed the POCUS study themselves or when the US team performed it, suggesting that reliance on a separate US team to acquire and interpret images does not impart a greater effect on medical decision-making.
Furthermore, the rapid diagnosis of rare, critical pathology such as pericardial effusion with concern for tamponade is often touted as a demonstrable benefit of POCUS in the evaluation of the dyspneic patient. In our study, post-POCUS pericardial tamponade remained in the differential in 10 patients and notably in 3 (2 performed by the primary team) it was not considered in the initial differential diagnosis. This diagnosis can easily be missed by other modes of evaluation such as chest X-ray and requires a drastically different management than other more common diagnoses that we investigated. Though POCUS rarely led to a change in management in our study, this may have been due to a limited number of subjects in which initial management options were declared by attending physicians prior to POCUS. Within the subset of subjects where management options were selected initially, 44% of the time there was either an addition or removal of a management option.
3.1. Limitations
This study is limited by potential selection bias. Since subject enrollment was based on the decision of the primary attending EP, it may be influenced by the attending physician’s perception of the utility of ultrasound. This may have caused more subjects to be enrolled by attendings with more ultrasound experience, including ultrasound fellowship experience, or who have undergone residency training more recently. The identity of attending was not linked to each survey form per our study protocol due to concerns for confidentiality of the medical decision-making process and patients’ protected health information. It is therefore possible that some attendings may have been enrolled proportionately more often. This may be a reflection of certain attendings having higher likelihood of incorporating ultrasound into their medical decision-making. However, it is not certain that this is correlated with prior ultrasound training.
In addition, patient encounters were included in this study based on the attending physician’s diagnostic uncertainty and the perceived potential for bedside ultrasound to impact the patient’s differential or management. This may result in increased likelihood of enrollment in cases where certain diagnoses were higher on the initial differential. The practical need for providers to fill out a survey in real time likely limited enrollment of the highest acuity patients, where limited time and cognitive load are of even greater importance. This seems to have been an issue given the limited number of intubations considered in our data set. Also, the number of patients eligible for this study is difficult to determine as there was no practical method to measure diagnostic uncertainty in all patient encounters for dyspnea, and patient encounters were included largely based on convenience to the primary treating attending to complete the surveys.
While the presenting age, acuity, vital signs and disposition of patients in each group were similar, only 119 out of 156 total patient encounters studied were linked to the patient record via images saved with patient information. The remaining 37 patient encounters did not have this information available because performing providers did not enter patient identifiers at the time of the POCUS study, and thus selection bias may have been introduced via differential patient characteristics. In our study, 81% (74/91) of US team and 69% (45/65) of primary team performed studies were linked to patient data and any uncaptured differences in presenting characteristics are likely to be small.
It is also difficult to distinguish whether changes in differential and management observed in the study were due to bedside ultrasound alone. Interestingly, ultrasound frequently appears to have been used to remove diagnoses from the differential, for example in the consideration of pulmonary embolism. However, bedside ultrasound alone is not often sufficient to rule out a PE. COPD was added to the differential in 2 cases, despite the inability of ultrasound to rule in this diagnosis. These changes may reflect the temporal aspect of this prospective study and underlying maturation bias while other non-POCUS diagnostic data, such as laboratory studies, are accumulated. Also, depending on the length of this time interval, the patient’s response to initial management and their clinical course can provide the physician with additional information that may change differential or management. This allows the provider to add or remove various diagnoses from the differential and make changes in their management that may not be a result of bedside ultrasound alone. This limitation is a pragmatic reflection of how POCUS fits into actual practice patterns, where multiple sources of information simultaneously enable the provider to make informed decisions and no diagnostic test result exists in isolation. Furthermore, in primary team performed POCUS, the reporting of results was more likely to be immediate and thus the effect on medical decision-making is less influenced by other diagnostic information. In our study, the percent of encounters where the post-POCUS primary diagnosis was changed was actually higher when POCUS was performed by the primary team (40% of encounters vs. 32%, Table 3a) and this may reflect the influence of non-POCUS diagnostic information. Information regarding the ultimate primary diagnosis at discharge was not collected in this study given common disagreement in this outcome and thus could not be compared to the primary diagnosis from our surveys. However, our study’s goal was to evaluate the changes in differential diagnosis due to POCUS rather than the diagnostic accuracy of POCUS. A final limitation to our study regarding primary diagnosis change is the selection of non-inferiority margin. While 20% is relatively wide, we believe that this margin accurately reflected the limits of a potential change in primary diagnosis due to POCUS given the multitude of other diagnostic modalities available at both centers. Given our results, future studies may define narrower limits of non-inferiority.
The providers performing POCUS in this study had variable levels of ultrasound training from ultrasound novices to ultrasound fellowship trained faculty. There were also multiple sites studied, including a large university hospital and a large community hospital. Both sites showed similar outcomes. Per our study protocol, we did not link individual patient encounters to attending physicians and thus the ability to discern an effect on differential diagnoses or management based on experience of the attending provider is limited. Furthermore, while each site is staffed by separate attending physician groups with different practice patterns, the presence of the same resident physicians at both sites limits generalizability to sites without trainees or different residency programs. Prior studies have shown that the presence of an ultrasound fellowship enhances resident POCUS education30 and thus our results may be skewed towards higher utilization of POCUS in medical decision making due to the presence of an ultrasound fellowship at our university based site and presence of the same residents at both sites. Additionally, 120 out of 156 (97 out of 119 with linked patient data) patient encounters occurred at the university site which may further skew our results and limits generalizability.
At both sites, there was no study or clinical protocol dictating the specific POCUS evaluation of acutely dyspneic patients. This was thought to be more representative of real-world practice scenarios and differs from prior studies with protocolized POCUS evaluation. Our intent was not to dictate a specific type of practice but rather to evaluate the effect of POCUS on medical decision-making when performed by the treating team who have other clinical responsibilities. It is also worth noting that while US team performed POCUS studies were generally rated higher image quality, and more often contained all of echocardiography, lung and IVC ultrasound (Table 2), the post-POCUS change in medical decision-making was similar to primary team performed studies. The means of image quality scoring (2.8 in primary team studies and 3.2 in US team studies) suggests that on average more of the US team performed studies were of diagnostic quality where a score of 3 is graded as a diagnostic (Figure S2). The results of image quality analysis may be limited by consecutive selection of first 12 studies in each group rather than random selection or full coverage. However, of these 24 total studies, only 2 occurred on the same day and the team structure of both the primary and US teams at both sites make it unlikely that the same sonographer acquired the images.
4. Conclusion
In typical ED practice settings, physicians must balance the diagnostic and management needs of an entire department of patients, many of whom require rapid diagnosis and management. POCUS can not only provide EPs with immediate diagnostic information but can dramatically reduce their cognitive load by narrowing the differential diagnosis. Our study is the first that demonstrates this effect in both a practice setting where a dedicated US team is present and where POCUS is performed by the primary team. While our study suggests that POCUS performed in settings without access to a dedicated ED US team can be equally efficacious in terms of reducing cognitive load, our findings should not obviate the need for these teams as their presence likely enhances the overall use of POCUS within any practice setting. Future studies at centers without trainees are needed to further generalize the effect of POCUS on diagnostic decision making.
Supplementary Material
Figure 3. Change in Management.
Post-POCUS change in each of 7 management options in A) overall cohort, B) US team performed studies, C) primary team performed studies. Percentage reported is percent of total studies for the overall cohort (N = 165), US team performed studies (N = 91) and primary team performed studies (N = 65). Frequency of each change if it was not in the original management options being considered, remains in management removed from management or added to management after POCUS are reported. IV Fluids = intravenous fluid bolus. NIPPV = non-invasive positive pressure ventilation; TPA = tissue plasminogen activator.
Article Summary.
Why is this topic important?
Point-of-care ultrasound (POCUS) has revolutionized the evaluation of dyspneic patients in the Emergency Department. However, in prior studies demonstrating its utility, POCUS is often performed by ultrasound (US) teams or experts without patient care responsibilities.
What does this study attempt to show?
This study evaluates the effective of POCUS performing team – the primary team with other patient care duties versus a dedicated POCUS team – on medical decision making in patients presenting with dyspnea.
What are the key findings?
Using a before and after POCUS survey of differential diagnosis and planned management, this study evaluated the change in most likely diagnosis and other medical decision making elements as a result of POCUS. We found that when the POCUS was performed by the primary treating team, the change in primary diagnosis was non-inferior to when it was performed by the dedicated POCUS team. Size of differential diagnosis, confidence in diagnosis, and image quality were equivalent as well.
How is patient care impacted?
Continued expansion of the use of ultrasound to evaluate dyspneic patients, especially for more typical practice settings without a dedicated POCUS team, may be of benefit to patients and providers. Demonstration that POCUS can reduce cognitive load and impact medical decision making in a variety of practice settings will likely drive further adoption.
Acknowledgements:
The authors acknowledge and thank the faculty of the Departments of Emergency Medicine at the University of Michigan and St. Joseph Mercy Hospital Ann Arbor for their participation in this study.
Funding Sources/Disclosures: VL received funding for this work from the University of Michigan Emergency Department Resident Research Development Grant for this work. BF received short term research funding from the National Heart, Lung, and Blood Institute (5T35HL007690). CF has received research support unrelated to this work from the National Heart, Lung, and Blood Institute (1K12HL133304). No authors have reported relevant disclosures.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest Disclosure: All authors report no conflicts of interest.
Prior presentations: Components of this work have previously been presented at the SAEM 2019 Annual Meeting and were accepted for presentation at AIUM 2020 prior to its cancellation.
References
- 1.Laribi S, Keijzers G, van Meer O, et al. Epidemiology of patients presenting with dyspnea to emergency departments in Europe and the Asia-Pacific region. Eur J Emerg Med. 2019;26(5):345–349. doi: 10.1097/mej.0000000000000571 [DOI] [PubMed] [Google Scholar]
- 2.Shellenberger RA, Balakrishnan B, Avula S, Ebel A, Shaik S. Diagnostic value of the physical examination in patients with dyspnea. Cleve Clin J Med. 2017;84(12):943–950. doi: 10.3949/ccjm.84a.16127 [DOI] [PubMed] [Google Scholar]
- 3.Inglis AJ, Nalos M, Sue KH, et al. Bedside lung ultrasound, mobile radiography and physical examination: a comparative analysis of diagnostic tools in the critically ill. Crit Care Resusc. 2016;18(2):124. [PubMed] [Google Scholar]
- 4.Chavez MA, Shams N, Ellington LE, et al. Lung ultrasound for the diagnosis of pneumonia in adults: A systematic review and meta-analysis. Respir Res. 2014;15(1). doi: 10.1186/1465-9921-15-50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Alrajab S, Youssef AM, Akkus NI, Caldito G. Pleural ultrasonography versus chest radiography for the diagnosis of pneumothorax: review of the literature and meta-analysis. Crit Care. 2013;17(5):R208. doi: 10.1186/cc13016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Maw AM, Hassanin A, Ho PM, et al. Diagnostic Accuracy of Point-of-Care Lung Ultrasonography and Chest Radiography in Adults With Symptoms Suggestive of Acute Decompensated Heart Failure. JAMA Netw Open. 2019;2(3):e190703. doi: 10.1001/jamanetworkopen.2019.0703 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Pandharipande PV, Reisner AT, Binder WD, et al. CT in the Emergency Department: A Real-Time Study of Changes in Physician Decision Making. Radiology. 2016;278(3):812–821. doi: 10.1148/radiol.2015150473 [DOI] [PubMed] [Google Scholar]
- 8.Ultrasound Guidelines: Emergency, Point-of-Care and Clinical Ultrasound Guidelines in Medicine. Ann Emerg Med. 2017;69(5):e27–e54. doi: 10.1016/j.annemergmed.2016.08.457 [DOI] [PubMed] [Google Scholar]
- 9.Pirozzi C, Numis FG, Pagano A, Melillo P, Copetti R, Schiraldi F. Immediate versus delayed integrated point-of-care-ultrasonography to manage acute dyspnea in the emergency department. Crit Ultrasound J. 2014;6(1):5. doi: 10.1186/2036-7902-6-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zanobetti M, Scorpiniti M, Gigli C, et al. Point-of-care ultrasonography for evaluation of acute dyspnea in the emergency department. Chest. 2017. doi: 10.1016/j.chest.2017.02.003 [DOI] [PubMed] [Google Scholar]
- 11.Gargani L, Volpicelli G. How I do it: Lung ultrasound. Cardiovasc Ultrasound. 2014;12(1):25. doi: 10.1186/1476-7120-12-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pagano A, Numis FG, Visone G, et al. Lung ultrasound for diagnosis of pneumonia in emergency department. Intern Emerg Med. 2015;10(7):851–854. doi: 10.1007/s11739-015-1297-2 [DOI] [PubMed] [Google Scholar]
- 13.Shokoohi H, Boniface KS, Pourmand A, et al. Bedside Ultrasound Reduces Diagnostic Uncertainty and Guides Resuscitation in Patients With Undifferentiated Hypotension. Crit Care Med. 2015;43(12):2562–2569. doi: 10.1097/ccm.0000000000001285 [DOI] [PubMed] [Google Scholar]
- 14.Gargani L, Volpicelli G. How I do it: lung ultrasound. Cardiovasc Ultrasound. 2014;12:25. doi: 10.1186/1476-7120-12-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Baston C, West TE. Lung ultrasound in acute respiratory distress syndrome and beyond. J Thorac Dis. 2016;8(12):E1763–e1766. doi: 10.21037/jtd.2016.12.74 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Labovitz AJ, Noble VE, Bierig M, et al. Focused cardiac ultrasound in the emergent setting: a consensus statement of the American Society of Echocardiography and American College of Emergency Physicians. J Am Soc Echocardiogr. 2010;23(12):1225–1230. doi: 10.1016/j.echo.2010.10.005 [DOI] [PubMed] [Google Scholar]
- 17.Anderson KL, Jenq KY, Fields JM, Panebianco NL, Dean AJ. Diagnosing heart failure among acutely dyspneic patients with cardiac, inferior vena cava, and lung ultrasonography. Am J Emerg Med. 2013;31(8):1208–1214. doi: 10.1016/j.ajem.2013.05.007 [DOI] [PubMed] [Google Scholar]
- 18.Borloz MP, Frohna WJ, Phillips CA, Antonis MS. Emergency department focused bedside echocardiography in massive pulmonary embolism. J Emerg Med. 2011;41(6):658–660. doi: 10.1016/j.jemermed.2011.05.044 [DOI] [PubMed] [Google Scholar]
- 19.Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and t. J Am Soc Echocardiogr. 2010;23(7):685–688. doi: 10.1016/j.echo.2010.05.010 [DOI] [PubMed] [Google Scholar]
- 20.Squizzato A, Galli L, Gerdes VE. Point-of-care ultrasound in the diagnosis of pulmonary embolism. Crit Ultrasound J. 2015;7:7. doi: 10.1186/s13089-015-0025-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Goodman A, Perera P, Mailhot T, Mandavia D. The role of bedside ultrasound in the diagnosis of pericardial effusion and cardiac tamponade. J Emerg Trauma Shock. 2012;5(1):72–75. doi: 10.4103/0974-2700.93118 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Blehar DJ, Dickman E, Gaspari R. Identification of congestive heart failure via respiratory variation of inferior vena cava diameter. Am J Emerg Med. 2009;27(1):71–75. doi: 10.1016/j.ajem.2008.01.002 [DOI] [PubMed] [Google Scholar]
- 23.Muller L, Bobbia X, Toumi M, et al. Respiratory variations of inferior vena cava diameter to predict fluid responsiveness in spontaneously breathing patients with acute circulatory failure: need for a cautious use. Crit Care. 2012;16(5):R188. doi: 10.1186/cc11672 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Farrington CP and Manning G ‘Test Statistics and Sample Size Formulae for Comparative Binomial Trials with Null Hypothesis of Non-Zero Risk Difference or Non-Unity Relative Risk. In: Statistics in Medicine Vol. 9.; 1990:1447–1454. [DOI] [PubMed] [Google Scholar]
- 25.Schuirmann DJ. A comparison of the Two One-Sided Tests Procedure and the Power Approach for assessing the equivalence of average bioavailability. J Pharmacokinet Biopharm. 1987;15(6):657–680. doi: 10.1007/BF01068419 [DOI] [PubMed] [Google Scholar]
- 26.Hallgren KA. Computing Inter-Rater Reliability for Observational Data: An Overview and Tutorial. Tutor Quant Methods Psychol. 2012;8(1):23–34. doi: 10.20982/tqmp.08.1.p023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mantuani D, Frazee BW, Fahimi J, Nagdev A. Point-of-Care Multi-Organ Ultrasound Improves Diagnostic Accuracy in Adults Presenting to the Emergency Department with Acute Dyspnea. West J Emerg Med. 2016;17(1):46–53. doi: 10.5811/westjem.2015.11.28525 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Buhumaid RE, St-Cyr Bourque J, Shokoohi H, Ma IWY, Longacre M, Liteplo AS. Integrating point-of-care ultrasound in the ED evaluation of patients presenting with chest pain and shortness of breath. Am J Emerg Med. 2019;37(2):298–303. doi: 10.1016/j.ajem.2018.10.059 [DOI] [PubMed] [Google Scholar]
- 29.Umuhire OF, Henry MB, Levine AC, Cattermole GN, Henwood P. Impact of ultrasound on management for dyspnea presentations in a Rwandan emergency department. Ultrasound J. 2019;11(1). doi: 10.1186/s13089-019-0133-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Adhikari S, Raio C, Morrison D, et al. Do emergency ultrasound fellowship programs impact emergency medicine residents’ ultrasound education? J Ultrasound Med. 2014;33(6):999–1004. doi: 10.7863/ultra.33.6.999 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.