Article, Radiology

BRIPPED scan for evaluation of ED patients with shortness of breath

a b s t r a c t

Objective: The BRIPPED scan is an ultrasound evaluation of pulmonary B-lines, right ventricle size, inferior vena cava collapsibility, pleural and pericardial effusion, pneumothorax, left ventricle ejection fraction, and lower extremity deep venous thrombosis. The primary goal was to evaluate the effect of the BRIPPED scan on the physician’s list of differential diagnoses for patients presenting with shortness of breath.

Methods: This prospective randomized control trial was performed on patients presenting to the emergency department with shortness of breath. Primary data analysis was performed using an ordinal quasi-symmetry model to compare the magnitude of change in the differential diagnoses between 2 groups. Secondary outcome measures included changes in physician orders or interventions, time to disposition, time to perform the BRIPPED scan, and the interrater reliability of the interpretation of the scan.

Results: A total of 104 patients and 24 physicians were enrolled in this study. Fifty-two patients were randomly assigned to each cohort. Among the BRIPPED cohort, there was significant movement of likelihood for several eti- ologies of shortness of breath on the physician differential. There was no significance in the change of differential diagnosis between the BRIPPED and control cohorts. The average (SD) time to perform the scan was 5.7 (1.3) mi- nutes (95% confidence interval, 5.4-6 minutes).

Conclusion: The BRIPPED scan is a rapid ultrasound evaluation of shortness of breath in the emergency depart- ment. BRIPPED influenced physician differential diagnoses to the same degree as laboratory and radiographic testing. BRIPPED did not alter the final diagnosis in this patient population.

Introduction

Background

Patients presenting to the emergency department (ED) with short- ness of breath (SOB) can be among the most challenging due to the di- verse underlying pathology and urgency of appropriate treatment.

? Author contributions: H.B. and V.S. conceived and designed the trial. V.S. obtained research funding and institutional review board approval. V.S., M.C., and D.B. supervised the conduct of the trial, recruitment of participants, and the data collection. Along with support from the biostatistics department who provided statistical advice, V.S. and M.C. analyzed the data. V.S. drafted the manuscript, with all authors contributing substantially to revisions. V.S. takes responsibility for the manuscript as a whole.

?? Meetings: Poster presented at ACEP; Denver, CO; October 2012.

? Grant: EMF Grant Awarded 2011-12 ($5000).

?? This study was registered at ClinicalTrials.gov (NCT01662843).

??? Conflicts of interest: None.

* Corresponding author at: Department of Emergency Medicine, 500 J Clyde Morris Blvd, Newport News, VA 23601. Tel.: +1 757 510 8197.

E-mail address: vms0419@gmail.com (V.M. Stewart).

Emergency ultrasound (EUS) has been shown to be a valuable adjunct to the standard clinical examination. Research is still limited on how to integrate EUS in the clinical setting and how it impacts diagnostic thinking [1].

Among other purposes, EUS is used for the evaluation of symptom or sign-based complaints [1]. Multiple point EUS examinations have been an integral part of the evaluation of undifferentiated hypotension [2-5]. The role of pulmonary ultrasound has been considered in acute circulatory failure [6]. Several EUS algorithms have previously been sug- gested for the evaluation of undifferentiated SOB [7-11].

The BRIPPED scan (pronounced bee-ript) is a standardized ultra- sound evaluation of pulmonary B-lines, right ventricle (RV) size and strain, Inferior vena cava collapsibility, pleural and pericardial ef- fusion, pneumothorax, ejection fraction of the Left ventricle , and lower extremity deep venous thrombosis. These EUS applications were chosen by the investigators based on a desire to build upon existing sonographic algorithms, personal experience with ultrasound, and because they provide diagnostic clues about the cause of a patient’s SOB. A summary of the considerations for the components of BRIPPED is

http://dx.doi.org/10.1016/j.ajem.2015.11.006

Table 1

The BRIPPED examination

B-lines: Sonographic pulmonary B-lines have been shown to correlate with congestive heart failure [11]. A 2-zone scanning protocol has been shown to perform similarly to an 8-zone protocol [12].

RV strain: RV enlargement can be caused by a pulmonary embolus (PE), acute RV infarct, congestive heart failure, pulmonary valve stenosis, or pulmonary hypertension, and is a risk factor for early mortality in PE [13].

IVC size and collapsibility: Using an IVC size cutoff of 2.0 cm has been shown to have a sensitivity of 73% and specificity of 85% for a right atrial pressure (RAP) above or below 10 mm Hg. The collapsibility during forced inspiration b 40% has even greater accuracy for elevated RAP (sensitivity 91%, specificity 94%, negative predictive value 97%) [14].

Pneumothorax: Bedside ultrasound is more accurate than supine chest x-ray with Diagnostic ability approaching that of computed tomography [15,16].

Pleural effusion: EUS has been shown to have an accuracy similar to a chest x-ray for evaluation of pleural effusion [16].

Pericardial effusion: EUS has a sensitivity of 96% and specificity of 98% compared with formal echocardiography [17].

Ejection fraction: The qualitative assessment of left ventricular ejection fraction by emergency physicians has been shown to correlate well with an assessment by a cardiologist [18-20].

DVT in lower extremities: Ultrasound was performed by emergency physicians using a 2-point compression venous ultrasound on patients with suspected lower extremity DVT. This approach had a 100% sensitivity and 99% specificity in diagnosing DVT, compared with a reference venous ultrasound in radiology [21].

condensed in Table 1. Similar to the BLUE, FALLS, FATE, and other proto- cols, the BRIPPED scan evaluates varying etiologies of SOB [6-11]. BRIPPED differs from previous similar algorithms in that it uses a more concise evaluation of sonographic lung windows, does not require a pa- tient to be supine, does not involve Tissue Doppler imaging, and considers IVC collapsibility. These differences between BRIPPED and other protocols allow similar information to be gained while requiring fewer lung windows to be imaged, less technical experience of a physi- cian sonographer, and considering a more comprehensive differential diagnosis for SOB.

The primary goal of this investigation was to evaluate the effect of BRIPPED on the physicians’ lists of differential diagnoses for patients presenting with SOB. Secondary outcome measures included changes in physician orders or interventions, time to disposition, time to per- form the BRIPPED scan, and the interrater reliability of the interpreta- tion of the BRIPPED scan. We hypothesize that there will be no difference in physician differential diagnosis ranking between the BRIPPED scan performed in the first minutes of the patient encounter as compared with the diagnosis ranking after the standard of care work- up results are available several minutes to hours after patient arrivals.

Materials and methods

Study design

This prospective randomized control trial was performed on a convenience sample of patients presenting to the ED with a chief concern of SOB or difficulty breathing. This study received a full review and approval by the institutional review board (1107FB0204) and was registered at ClinicalTrials.gov (NCT01662843).

Study setting and population

This study was performed at an academic ED (65 000 patients/y) from October 2011 to May 2012. This study involved both a patient sub- ject group and a physician subject group. The physician subject refers to the primary treating physician of the patient subject. For clarity, “physicians” refers to physician subjects and “patients” refers to patient subjects. Physicians included senior year (PGY-3) emergency medicine (EM) residents and attending EM physicians (all American Board of Emergency Medicine board certified) who were blinded to the study hypothesis. Patients aged 18 to 89 years presenting with a chief concern

of SOB were included after identification of eligibility by 1 of 2 study in- vestigators. Enrolled patients did not present with isolated SOB, and many had other secondary complaints such as chest pain, cough, edema, or other symptoms. Patients were excluded if they had a singu- lar history of asthma, were 20 or more weeks pregnant, had a respirato- ry rate less than 18 breaths/min, or had thoracoabdominal trauma in the previous 72 hours.

Study protocol

All eligible patients were evaluated by the treating physician on ar- rival to the ED, and initial orders were placed for workup and treatment. Informed consent of both the patients and physicians was then obtain- ed. Fig. 1 summarizes this study protocol. Physicians were asked to com- plete differential diagnosis ranking forms after their initial evaluation of all patients. The initial “pre” differential diagnosis ranking list form was identical for both the BRIPPED and control groups. This ensured that the physicians were blinded to the treatment group assigned until after the predifferential ranking form was completed and the sealed intervention group packet was opened that contained the data collection sheets for the next phases of the study. The sealed intervention group packets were prepared prior to data collection by computer-generated, block randomization to assign the intervention group. The packets of both groups were opaque and equal in overall appearance, size, weight, and thickness.

After the predifferential ranking list was given to the physician subject, the BRIPPED ultrasound protocol was performed by a study

BRIPPED Protocol

104 Patient Consented, zero patients excluded

“Pre” differential ranking form completed

BRIPPED group (52 patients)

Control Group (52 patients)

Imaging studies and labs ordered per standard of care

BRIPPED

ultrasound performed

Imaging studies and labs ordered per standard of care

BRIPPED results Results known to given to treating treating physician physician as available

“Post” differential “Post” differential ranking form ranking form

completed completed

Fig. 1. BRIPPED protocol.

investigator on patients randomly assigned to the BRIPPED cohort. The concurrent completion of the predifferential ranking list and the ultra- sound scan ensured that the BRIPPED ultrasound scan was performed within minutes after the physician’s initial history and physical exami- nation. In addition, the study investigator was blinded to the results of the initial differential diagnosis ranking list completed by the physician subject. The ultrasound examination is summarized in Table 1. Immedi- ately after the BRIPPED scan, the study investigator informed the treating physician subject of the results. The physician was asked to repeat their differential diagnosis ranking form in the light of this infor- mation. Study investigators noted orders; medications, radiography added and cancelled after BRIPPED examination results were known; and overall time to disposition.

For patients not randomly assigned to the BRIPPED ultrasound group, the physicians were asked to repeat the differential diagnosis ranking forms once all results of the ordered workup were available.

Study investigators that performed the BRIPPED scan completed a minimum of 16 hours of didactic training for general EUS application prior to enrollment. In addition, they completed 2 hours of didactic training specific to BRIPPED (evaluation for pulmonary B-lines, RV size and strain, IVC collapsibility, pleural and pericardial effusion, pneumo- thorax, ejection fraction of the LV, and lower extremity deep venous thrombosis) and demonstrated 5 supervised practice BRIPPED scans. Study investigators had some prior EUS experience and were senior year (PGY-3) residents or current EUS fellows.

Methods of measurement of the BRIPPED scan

The BRIPPED ultrasound scan was conducted with patients semirecumbent as able. Supine views were performed to optimize images obtained only if tolerated by the patient. Sonographic images were obtained using a Sonosite M-Turbo or Edge (SonoSite, Inc, Bothell, WA) machine. Two separate transducers were used to obtain images, the phased array P-21 (4-2 MHz) and linear L38xi (10-5 MHz) transducers. The order in which images were obtained was left to the sonographer’s discretion. Scanning time was calculated as the difference between the initial and final image recorded by the machine’s time stamp. The required image criteria for each component of the BRIPPED scan are described below.

Pulmonary B-lines were measured by the presence of 3 or more B-line artifacts per rib space in a minimum of 2 zones scanned. B-lines, if present, were noted along anterior MIdclavicular lines while simultaneously scanning for pneumothorax. In addition, B-lines were noted if seen during scanning the lateral inferior quadrant on each side while evaluating pleural effusion. Still images with the maximum number of B-lines present or a video clip were recorded. All saved clips were 4 sec- onds in duration in an anterograde direction. Saving anterograde video clips increases scanning time as compared with retrograde saved clips. From the experience of the investigators with the pilot study, 4 seconds allowed scanning time to be minimized and was sufficient time to grossly evaluate ventricular diastolic collapse for tamponade physiology, Lung sliding for an average to increased respiratory rate, and for a patient to sniff on request for IVC collapsibility.

Right ventricular size was determined from Apical 4-chamber views by direct comparison with the LV. A diastolic RV/LV ratio equal to or greater than 1 was considered abnormal.

The IVC was measured from a longitudinal subcostal view to the right of the patient’s midline. Lack of collapsibility on forced inspiration indicated intravascular volume overload, whereas complete collapsibility of the IVC indicated Volume depletion.

Pleural effusions were evaluated by scanning the costophrenic angles with the transducer placed over the lateral lower thorax anterior to the midaxillary line. Long-axis views were obtained bilaterally and any fluid seen was noted.

Parasternal long-axis views were primarily used for detection of

pericardial effusion. Also, pericardial effusion was noted in the

longitudinal subcostal views and the apical 4-chamber views used for IVC and RV assessments, respectively. Circumferential and localized pericardial effusions were quantitatively characterized as less than 1 cm, or equal to or greater than 1 cm. Diastolic RV collapse, indicating tamponade, was also noted.

The presence of pneumothorax was simultaneously assessed while scanning for pulmonary B-lines along the anterior midclavicular lines, with the patient sitting upright. Still images or video clips were recorded to demonstrate lung sliding, comet tails, “waves on a beach” sign, or power Doppler slide.

Qualitative ejection fraction estimates were recorded as video clips from a parasternal long-axis view. Ejection fraction was characterized normal, reduced, or markedly reduced.

Deep venous thrombosis of the lower extremities was measured following a 2-region compression protocol. The patient sat or lay with both legs slightly externally rotated and the knee slightly flexed. Bilateral short-axis views of the common femoral and popliteal veins were obtained. Still images were recorded in a side-by-side comparison for compressibility. Deep vein thrombosis was considered present if either thrombus was directly visualized or the vein was not completely collapsible as the accompanying artery was seen starting to compress.

Data collection and processing

Sample size calculations of 52 per cohort were determined with Stuart-Maxwell testing using our case-series pilot data primary outcomes to evaluate marginal homogeneity of qualitative variables simultaneously. Consecutive numbered packets from 1 to 104 were prepared using computer-generated randomized blocks of 20. Within each packet was the presence or absence of the BRIPPED ultrasound data collection form to ensure that physicians and investigators were blinded to the treatment group (BRIPPED or control) until after the initial or predifferential diagnosis form was completed.

For both the “pre” and “post” differential diagnosis forms, the treating EM physician ranked 15 potential diagnoses from most to least likely per a 5-point Likert scale. For both the BRIPPED and control groups, laboratories and radiology were ordered as per the usual workup as considered appropriate by the treating physician. The physicians reranked their differential diagnosis on the “post” form after reviewing either the BRIPPED results in the BRIPPED group, or all laboratories and radiology results among the control group. Physicians based their ranking after verbal communication of results by the investigator and after reviewing the ultrasound images recorded. Disposition times were measured from provider assignment to final disposition selection. Medications, laboratory testing, radiography, interventions, and consultations ordered and canceled after the “pre” and prior to the completion of the “post” differential diagnosis rank form were recorded by the study investi- gators for both cohorts. All data were entered into a Microsoft Excel spreadsheet (Microsoft Excel 2010; Microsoft, Inc, Redmond, WA) before analysis.

Outcome measures

Our primary outcome measure was the magnitude of change in the differential diagnosis after the BRIPPED scan compared with the magni- tude of change of the differential diagnosis after routine evaluation with all accompanying laboratory tests and radiographic imaging for undif- ferentiated SOB. Secondary outcome measures included changes in physician orders or interventions, time to disposition assignment, time to perform the BRIPPED scan, and the interrater reliability of the inter- pretation of the BRIPPED scan.

Primary data analysis

Primary data analysis to compare the magnitudes of change within the BRIPPED cohort and between each cohort was performed from a 5-point Likert scale using an ordinal quasi-symmetry model analysis

Table 3 Magnitude of change in likelihood after the BRIPPED ultrasound protocol for a given differ- ential diagnosis

BRIPPED group Difference between

pre- and post-Likert

[22-24]. Often quantitative values are ascribed to qualitative data and

scale likelihood

Likert data are treated as interval data. However, there are circumstances that require Likert data to be interpreted as ordinal data. In the case of considering different possible etiologies of SOB, the treating physician must commit to a diagnostic and treatment plan quickly. Often the treating physician initiates a treatment plan from the most likely etiologies before confirming the diagnosis to resuscitate or stabilize the patient. The distances between the Likert points in this study are not necessarily equal (as is assumed with interval data). The diagnosis order matters over the quantifiable difference between the di- agnoses, and therefore, the data should be treated as ordinal. The ordi- nal quasi-symmetry model was chosen because it enables the analysis of a multiway table with the same ordinal categories. Two EM attending physicians who met the requirements as American Registered Diagnostic Medical Sonographers independently and blindly reviewed ultrasound images for interrater reliability with Cohen ? coefficient [25].

Results

Characteristics of study subjects

A total of 104 patients presenting with SOB and 24 physicians were enrolled in this study. Fifty-two patients were randomly assigned to each cohort. Patient characteristics are presented in Table 2 with no statistically significant difference in vital signs, age, or sex distribution between the 2 groups. In general, patients were afebrile, had a respira- tory rate around 20 or higher, and were not hypotensive. Among physi- cians, 7 were senior EM residents and 17 were attending EM physicians. Physicians were enrolled for a median number of 3 patients, ranging from 1 to 10 patients each. No patients or physicians were excluded. There were no deviations from the protocol.

Main results

BRIPPED cohort

Among the BRIPPED cohort, there was significant movement of diag- nosis likelihood on the Likert scale after the ultrasound. Primary out- comes are reported in Table 3 and Fig. 2. Table 3 summarizes the magnitude of change in likelihood for each diagnosis independently after the BRIPPED ultrasound (Table 3). Fig. 2 summarizes the magni- tude of change for a diagnosis over other potential diagnoses. The odds of shift quotients in Fig. 2 and P values obtained in Table 3 were calculated using the quasi-ordinal symmetry model.

Image interrater reliability

All components of the BRIPPED examination were performed on all patients enrolled in the BRIPPED group. Each of the 832 ultrasound images obtained was compared and was determined to have a Cohen ? of 1. No technical limitations or difficulties were encountered either

Table 2

Summary of the patient study subjects

Differential diagnosis Change P

Acute coronary syndrome 1.5 .002

Anxiety 2.2 .028

Aortic aneurysm	0.29	.71
Asthma/chronic obstructions pulmonary disease	0.78	.081

Congestive heart failure 2.4 .001

Hypervolemia -0.3 .389

Intoxication -1.61 .142

Pericarditis 3 .003

Pleural effusion 2.2 b.001

Pneumonia 1.8 b.001

Pneumothorax 3 .003

Pulmonary embolism 1.7 .003

with scanning or on review of images. No treating physicians’ interpreta- tions of images immediately after the scan differed from the recorded data sheets completed by the investigators.

BRIPPED vs control

There was no significant change of differential diagnosis ranking between the BRIPPED group and the control group. As in the BRIPPED group, diabetic ketoacidosis, anemia, and pericardial effusion were not considered in significant numbers to be of statistical significance within the control group. Laboratory testing, radiography, interventions, and consultations ordered and canceled were compiled. No orders or con- sultations were canceled in the control group, and one consultation and no orders were canceled in the BRIPPED group after the treating physician knew the scan results.

Time measures

The average (SD) time to perform the BRIPPED scan was 5.7 (1.3) minutes (95% confidence interval [CI], 5.4-6 minutes). Fig. 3 summa- rizes the time to disposition among enrolled patients. The disposition time averaged (SD) 166 (93.8) minutes in the BRIPPED group (95% CI, 140.5-186.5 minutes) compared with 190 (108) minutes (95% CI, 161- 219 minutes) in the control group. Seven of 52 patients (13.5%) of pa- tients in the BRIPPED cohort were identified as assigned a disposition within 60 minutes or less, compared with less than 2% of the control

BRIPPED: Odds of Shift Quotient

Pulmonary Embolism

Pneumothorax Pneumonia Pleural Effusion Pericarditis

Congestive Heart Failure

Asthma/COPD

Anxiety

Characteristic Median BRIPPED (interquartile range)

Median control (interquartile range)

Acute Coronary Syndrome

Age (y) 59 (21) 57 (24)

Temperature (?F) 97.9 (1.4) 97.5 (1.4)

0 5 10 15 20 25

Odds of Shift Quotient

Heart rate (beats/min)	90 (30)	88 (27)
Respiratory rate (breaths/min)	20 (4)	18 (2)
Systolic blood pressure (mm Hg)	141 (35)	149 (44)
Diastolic blood pressure (mm Hg)	81 (18)	83 (19)
Oxygen saturation (%)	97 (5)	98 (3)

Fig. 2. Odds of shift quotient for predifferential diagnosis vs postdifferential diagnosis among the BRIPPED group. The estimated odds of shift quotient express the magnitude of change of the BRIPPED scan for a diagnosis over the other potential diagnoses.

0-60 min 61-120

min

121-180

min

181-240

min

241-300

min

301-360

min

361+ min

to why orders were not canceled. Both groups equally used nursing- ordered standard protocols for electrocardiogram, laboratory, and radiology studies; therefore, orders were not placed by the treating phy- sicians. Treating physicians may have felt an obligated workup momen- tum to complete studies ordered for patients with a complaint, such as SOB, which fita defined trigger for ordering algorithms as per depart- mental policy. Failure of cancellation of traditional studies in the BRIPPED group may be due to the lack of education and experience with the BRIPPED ultrasound algorithm by the treating physicians. Treating physicians did not receive formal didactic training in the BRIPPED algorithm immediately prior to the study. In addition, physicians in the BRIPPED group may have been reluctant to rely on a single testing modality. The study was not sufficiently powered to detect Changes in management among less common causes of SOB such as diabetic ketoacidosis, intoxication, anemia, or pericardial effusion.

There are several limitations of this study. Patients were enrolled on

BRIPPED Control

Fig. 3. Time to disposition. Time in minutes from door to disposition is arranged on the x axis as grouped in 60-minute increments. The y axis represents the number of patients per grouped disposition time.

group for the same time to disposition. Among the patients, 11.5% in the control cohort were assigned a disposition after 6 hours, as compared with less than 2% for the BRIPPED group.

Discussion

Because of the urgency of appropriate treatment, clinicians must concurrently consider various etiologies for SOB. Prior studies have de- scribed various ultrasound protocols for SOB [6-11]. Protocols such as the BLUE and FALLS do not evaluate the presence of lower extremity DVT, nor do they assess intravascular status in patients with pulmonary edema [6,8]. On the basis of prior studies, we attempted to develop a more comprehensive ultrasound scan to include additional etiologies of SOB. To improve on existing algorithms, we used Ultrasound techniques familiar to most emergency physicians with some ultra- sound experience. Apical and subcostal lung windows, parasternal and 4-chamber cardiac views, and vascular compression techniques are familiar to EM physicians with ultrasound experience and were chosen for the BRIPPED protocol.

In addition, this study builds on existing knowledge of ultrasound’s role in SOB to evaluate the impact on the physician’s differential diagno- sis. To date, no study has fit-ranked physician differential diagnoses to an ordinal quasi-symmetry model for comparison. This model is well suited for comparing Likert-ranked items [22,23,26].

BRIPPED significantly changed clinical suspicion for acute coronary syndrome, anxiety, asthma and chronic obstructions pulmonary disease, congestive heart failure, pericarditis, pleural effusion, pneumonia, pneumothorax, and pulmonary embolism. Clinically, this information is useful to the treating physician to rapidly narrow the differential diagnoses. This study found no statistical significance in the magnitude of change in the differential between the ultrasound group and the control group. BRIPPED, a bedside scan performed in less than 6 minutes, therefore was as effective at influencing physician differential diagnosis as the ordered studies among our control group. It is not our position to advocate BRIPPED replacing all laboratory and radiographic imaging. However, BRIPPED does not expose patients to ionizing radiation and may prevent ordering unnecessary or lower yield tests. BRIPPED may expedite earlier pharmacologic management in patients if the differential

diagnosis can be narrowed or confirmed more rapidly.

In addition, this study found no statistical differences in laboratory testing, imaging studies, or consults ordered between the groups. In both groups, physicians were reluctant to cancel low-yield tests or imaging studies after the BRIPPED examination was performed, or higher-yield results returned earlier. Several possible reasons exist as

a convenience basis due to the limited availability of physician sonographers, therefore introducing selection bias. All 102 patients screened were enrolled. A further selection effect is due to the patient needing to self-describe a chief concern of “SOB” to the triage nurse for investigators to identify the subject for screening and enrollment. This inadvertently caused exclusion of patients too sick or unstable to be screened or consented. The location of the study within a single setting potentially limits the generalizability of our study to other settings.

Because outcomes measured were directly related to the physician’s organization of the differential diagnosis and Ordering patterns, we rec- ognized that a physician’s own experiences and level of training could introduce proficiency bias. We attempted to correct for proficiency bias by including physician subjects who were senior EM residents and attending EM physicians. Nearly all patients in the study received equivalent laboratory, electrocardiogram, and radiology imaging orders. It is likely that proficiency bias is further minimized because there were no treating physicians who refused to participate or review the ultra- sound images as they would other imaging modalities. This indicates a similar degree of comfort, training, or experience among the treating physician group with interpretation of the ultrasound images.

Cointervention bias occurred as physicians may have simultaneous- ly processed electrocardiograms or returning point-of-care laboratories at the time the BRIPPED scan results became known. It was not the in- tention of these authors to advocate BRIPPED replacing other diagnostic tests, but to evaluate its efficacy as a bedside tool incorporated into the larger picture of patient care. Cointervention bias was minimized by having the investigator perform the BRIPPED scan as soon as possible and for the physician to complete the postdifferential data sheet imme- diately after BRIPPED results were given.

During study image review, it was noted that the order varied in which the BRIPPED images were performed. Some investigators followed a “head to toe” approach with scanning, whereas others scanned all elements with one probe before switching to the second probe. There was no statistical significance in the time to perform the examination with either approach. Having a standardized training pathway for physician sonographers minimized variability of Image quality and content among performance of BRIPPED. American Registered Diagnostic Medical Sonographers physicians reviewing all scans also minimized interrater variability.

The sample size was calculated based on results from our smaller case-series pilot study. The pilot study did not include patients with SOB due to diabetic ketoacidosis, intoxication, or anemia. Pericardial effusion was not considered as a diagnosis with enough regularity among physicians to draw statistical significance using the ordinal quasi symmetry model. Therefore, the sample size would need to be increased to address BRIPPED’s utility across these less common categories. Patients enrolled were relatively hemodynamically stable. Another limitation of this study is the lack of unstable patients in both groups. However, treatment algorithms and disposition are often more focused

in unstable patients. immediate intervention is initiated, and the patient is rapidly admitted. Levels 3 and 4 patients often have the longest ED lengths of stay [24]. Patients who received the BRIPPED scan had shorter lengths of stay.

Inferior vena cava measurement has inherent limitations, with diameter variation in adults [14,25]. Because this protocol’s initial development, further research has demonstrated greater acceptability of diameter variation [25].

This intervention was not designed for patients with recent blunt or penetrating trauma. It was not our intention to replace existing algorithms for ultrasound in trauma. Future research is recommended to evaluate BRIPPED’s utility in patients with SOB due to traumatic injury. Additional research is also recommended to evaluate BRIPPED’s influence on patient outcomes, complications, and Associated costs.

In conclusion, the BRIPPED scan is a rapid, accurate approach to using ultrasound in the evaluation of SOB in the ED. BRIPPED influenced physician differential diagnoses to the same degree as laboratory and radiographic testing. BRIPPED did not alter the final diagnosis or management in this patient population.

Acknowledgments

The authors thanks Rebecca Ryszkiewicz, MD RDMS RDCS, and Bhaskara Ravi, PhD.

References

AJEM

BRIPPED scan for evaluation of ED patients with shortness of breath

a b s t r a c t

Introduction

Background

Materials and methods