Development of a Consensus-Based Definition of Focused Assessment With Sonography for Trauma in Children

This qualitative study describes the use of a modified Delphi process for soliciting the input of a panel of experts in defining a standardized examination protocol for pediatric patients with a blunt torso trauma.


Introduction
Intra-abdominal injury after blunt abdominal trauma is a leading cause of preventable deaths in children in the US. 1 However, early and accurate diagnosis of intra-abdominal injury in children is challenging.Current diagnostic strategies are suboptimal because of the trade-off between missed injury and resource overutilization, including children's exposure to ionizing radiation from computed tomography (CT) scans.
Focused Assessment With Sonography for Trauma (FAST) is a point-of-care ultrasonography (POCUS) study that uses no radiation.The FAST method was introduced in the US in the 1990s to describe a set of ultrasonographic views for the rapid evaluation of free fluid (hemorrhage) in patients with injury. 2 In adult patients, use of FAST decreases time to surgical intervention, patient length of stay, surgical complications, and CT scan and diagnostic peritoneal lavage rates. 35][6][7] For this reason, FAST has not been ubiquitously incorporated into diagnostic strategies for children with injury. 6,8evious studies have suggested that clinicians with more experience in performing FAST in children have higher diagnostic yields. 9,10In addition, experienced clinicians have been found to be more likely to capture complete, high-quality images and feel confident about integrating their results into their clinical strategy. 11,12Currently, there is no agreed-on standard for a complete protocol, adequate image quality, and accurate interpretation for FAST in children with injury. 13This lack of a standardized pediatric FAST definition is a critical factor in the variability in its use, image quality, and diagnostic accuracy.Therefore, we conducted this qualitative study to define a complete, high-quality, and accurate interpretation for FAST and extended FAST (E-FAST) in children with injury using an expert, consensus-based modified Delphi technique.

Methods
For this qualitative study, the 2-round, mixed-methods, modified consensus Delphi technique 14 was conducted between May 1 to June 30, 2021, and consisted of 2 web-based surveys per guidelines 15 and 1 live webinar consensus meeting between rounds.The institutional review board at the University of California, San Francisco approved this study.Panelists provided verbal consent to participate and were allowed to withdraw at any time.The reporting of this study follows the Consolidated Criteria for Reporting Qualitative Research (COREQ) reporting guideline. 16e modified Delphi technique is a consensus-based approach that systematically assembles statements from a group of experts. 15The method is iterative and encourages participants to share opinions but uses anonymity to reduce participants' dominant impact.The RAND/UCLA Appropriateness Method is a modified Delphi technique that, unlike the original Delphi, allows panelists the opportunity to discuss their judgments between rounds. 17

Panel Selection and Survey Instrument
Those of us in the Steering and Executive Writing Committee (A.E.K., N.A., M.P., A. Shaahinfar, M.L.-M., N.S., and D.G.) outlined the objectives of the FAST consensus panel and invited an international group of pediatric emergency POCUS experts from the P2Network Research Committee writing group. 18The P2Network is a platform for sharing information and collaborating on pediatric emergency medicine POCUS initiatives.The 26-member panel was chosen to represent a geographically diverse sampling of experts.Age and race and ethnicity data were not collected.
Experts were defined as those who completed 1500 or more POCUS studies or served in an institutional POCUS leadership role. 19,20The intended participants received an email outlining the objectives, methods, and requirements of the study.
The Steering and Executive Writing committee conducted an initial scoping review to assess contemporary and historical literature that could be used to define a complete FAST protocol, study

Live Webinar Discussion
After the initial survey, the panelists were provided with a summary of responses that described the means, medians, IQRs, and accompanying histograms for appropriateness ratings.The free-text responses were also summarized, and each participant received their individual responses.To prevent bias, the committee maintained the anonymity of replies from other participants.
During the live webinar discussion on May 17, 2021, panelists discussed the first-round results and anonymously provided verbal or written feedback on survey items.In addition, all panelists were given 7 days to access an online document (Google Docs; Google) to anonymously and asynchronously comment on all results, respond to other panelists, suggest new items, and edit existing survey items.The committee iteratively reviewed, refined, discussed, and summarized into statements the input from the webinar discussion and online document.

Second-Round Survey
The committee adjusted the initial survey items for content and phrasing according to participant feedback.No survey items were removed.Panelists were sent a hyperlink to the second-round survey and asked to rate the FAST and E-FAST definitions, landmarks, quality, and accuracy statements for appropriateness (using a 1 to 9 scale, with 1 indicating extremely inappropriate and 9 indicating extremely appropriate); completeness according to evaluations, views, and landmarks; and importance (using a 1 to 9 scale, with 1 indicating not all important and 9 indicating extremely important).
The panelists had 10 days to complete the second survey.The committee then reviewed all of the survey results, final statements, and comments.Final statements within similar domains were merged into hybrid statements and reviewed for accuracy by all panelists.

Statistical Analysis
The committee used the RAND/UCLA Appropriateness Method to analyze the survey responses from the first round and the comments from the webinar and online document in the second round. 17In round 1, survey responses were grouped according to numeric ratings and recurring themes.Then, the round 2 comments were iteratively reviewed, refined, discussed, and summarized into statements for final rankings.
Panel consensus was ascertained after round 2 using the interpercentile range adjusted for symmetry (IPRAS) method, as described in the RAND/UCLA Appropriateness Method User's Manual. 17All quantitative responses were evaluated for agreement using the IPRAS method and presented with their median ratings and IQRs.Median ratings were classified as appropriate or important (7-9), uncertain (3.5-6.5), and inappropriate or not important (1-3) for each survey item in which there was no disagreement (eTable 1 in the Supplement).The IPRAS method used the distribution of results instead of a threshold to establish agreement.Disagreement on survey items was defined where the distribution of responses was bimodal or where the calculated interpercentile ranges (difference between 30th and 70th percentiles) were higher than the IPRAS. 17,24Moreover, the survey results reported the proportion of respondents who considered items as appropriate or important.

Results
Of the 29 invited pediatric FAST experts, 26 (90%) agreed to participate in the panel and 3 responded as unavailable.All 26 panelists (11 women [42%] and 15 men [58%]) completed the surveys in both rounds, and 24 of 26 panelists (92%) participated in the live and asynchronous online discussions.The FAST consensus panel consisted of physicians with board certification in 4 specialties: emergency medicine, pediatrics, pediatric emergency medicine, and internal medicine (Table 1).More than half of the panelists had greater than 5 years of postgraduate POCUS experience, and most panelists worked at a Level 1 trauma center.
A summary of the panel responses and changes from the first round to the second round of surveys is presented in eTable 2 in the Supplement.Briefly, round 1 included 125 survey items, of which 84 were deemed appropriate and reached consensus.In addition to suggested modifications, 26 new survey items were included in the revised survey for a total of 151 items.An agreement was achieved on FAST and E-FAST study definitions, protocol evaluations, views, and landmarks; 14 statements on image quality; and 20 statements on accurate interpretation.

Proposed Study Definitions and Evaluations
The consensus definitions for FAST and E-FAST (Box) were rated as appropriate, with both definitions receiving a median (IQR) rating of 9 (8-9).There were several notable comments regarding the FAST and E-FAST definitions.Some panelists commented that the word "limited" might suggest a lack of quality; therefore, replacing the word with "focused" was proposed.Ultimately, the panel decided to keep "limited" to avoid repeating "focused," which was used to define the acronym FAST.Similarly, the term "peritoneal" was suggested instead of "intraperitoneal."However, the FAST or E-FAST does not evaluate the retroperitoneal cavity, and thus "intraperitoneal" was deemed as the more precise choice.

JAMA Network Open | Emergency Medicine
Development of a Consensus-Based Definition of FAST in Children Intraperitoneal free fluid, pericardial fluid, and pleural fluid were considered essential evaluations for FAST.These same 3 evaluations, with the addition of pneumothoraces, were deemed essential for E-FAST.Inclusion of the evaluation for cardiac activity or cardiac standstill was unclear after 2 rounds of surveys for FAST (median [IQR] rating, 6 [4.25-7.75])and E-FAST (median [IQR] rating, 6 [4.25-9]).Evaluations for pneumopericardium and bladder injury were not considered important for FAST or E-FAST.

Completeness by Views
The panelists found that a complete FAST is dependent on whether it has a negative or positive result (Box).A negative FAST result must include an adequate evaluation of all FAST views.In contrast, a positive FAST result must evaluate each anatomic region, with at least 1 view demonstrating abdominal, thoracic, or pericardial free fluid (pathology).To examine the anatomic regions, the panelists found the following views to be appropriate and important for FAST: right upper-quadrant abdominal view, left upper-quadrant abdominal view, suprapubic views (both transverse and sagittal views), and subxiphoid cardiac view (Figure 2A).The same views were appropriate and important for E-FAST with the addition of the lung or pneumothorax view (Figure 2B).The parasternal cardiac long view was rated as appropriate for both FAST and E-FAST, but its importance did not reach an agreement.The panel commented that the parasternal long cardiac view could be considered an acceptable substitute if the subxiphoid cardiac view was technically limited or challenging to perform.
Both suprapubic views (transverse and sagittal) were found to be appropriate and important.In contrast to the cardiac views, 19 panelists (73%) responded that 2 suprapubic views (transverse and sagittal views) were required, whereas 7 panelists (27%) responded that 1 suprapubic view was sufficient.

Completeness by Landmarks
The panelists rated 32 anatomic landmarks for the FAST and E-FAST ultrasonographic views.The panelists agreed that essential anatomic landmarks within each view were necessary to ensure a  2). 25,26In contrast, the panelists were uncertain about the importance of  2).
The suprapubic view was divided into the transverse and sagittal suprapubic anatomic views.

Image Quality
Consensus was reached on 14 quality statements, with no disagreements in the second survey.These 14 quality statements were converted into 3 hybrid statements (Box).After round 1, the panelists included image quality related to optimizing depth and gain, transducer selection, and ultrasonography system settings.Other than improving the factors associated with the ultrasonography machine, the panel wanted clinicians to optimize image quality through proper probe manipulation and skin contact through gel use.Similarly, the panelists wanted the definition of free fluid to capture complex fluid collections with a heterogeneous hyperechoic or hypoechoic appearance (eg, blood clot).Abbreviations: E-FAST, Extended Focused Assessment With Sonography for Trauma; FAST, Focused Assessment With Sonography for Trauma.
a Landmarks with a score of 7 to 9 were rated as important; 3.5 to 6.5, uncertain; and 1 to 3, not important for a complete Focused Assessment With Sonography for Trauma examination.
b Landmark ranked as important.
c Disagreement in importance ranking per interpercentile range was adjusted for symmetry.

JAMA Network Open | Emergency Medicine
Development of a Consensus-Based Definition of FAST in Children

Interpretation Accuracy
The panel reached an agreement on 20 interpretation of accuracy statements in the second survey, which were converted into 4 hybrid statements (Box).Two statements on interpretation accuracy did not reach an agreement but created a divide between the 26 panelists.
One statement, "A FAST study can be considered a qualified negative if the operator does not adequately visualize one or more landmarks," was rated by 11 panelists as appropriate, whereas 12 panelists rated it as inappropriate.Panelists who agreed with the statement wanted to recognize the limitations of FAST and that patient-level factors may not allow a complete visualization of all landmarks.In contrast, panelists who disagreed with the statement suggested that ranked landmarks could lead to suboptimal FAST studies.The other statement, "Trace free fluid in the pelvis may be considered a negative study," was rated by 9 panelists as appropriate, whereas 10 panelists rated it as inappropriate.Panelists who agreed with the statement emphasized that trace free fluid could be considered physiological in specific pediatric populations.However, those panelists who disagreed with the statement were unclear on how clinicians could differentiate between physiological and pathological free fluid while acknowledging that the FAST result, whether positive or negative, should not uniquely dictate clinical next steps. 32

Discussion
The FAST consensus panel of heterogeneous international experts developed comprehensive definitions for a complete, high-quality, and accurate interpretation of FAST and E-FAST for children with injury.The definitions have implications for clinical use and quality improvement review and provide standard definitions for research on injuries in children.The panelists defined FAST and E-FAST as congruent with the current working definitions for adults with injury. 13,33e protocol definition of FAST has undergone alterations to match advances in POCUS applications.For example, FAST was initially proposed as Focused Abdominal Sonography for Trauma 34 ; however, with the addition of the cardiac view for evaluating pericardial fluid, FAST transformed to Focused Assessment With Sonography for Trauma. 13Similarly, advances in thoracic POCUS and the addition of pneumothorax evaluation led to the expansion of E-FAST. 31During the panel discussion, several panelists suggested adding novel POCUS applications to the FAST and E-FAST definitions, such as musculoskeletal evaluation for fracture.However, without sufficient evidence for these novel applications in children, the panel agreed to keep the definitions congruent with the protocol definitions for adults with injury.
During the panel discussion, one area of interest was the definition of FAST compared with the definition of E-FAST.Most panelists agreed with the working definition, but there were substantial discussions regarding the evaluation for hemothorax. 35Some panelists considered pleural evaluation only in the E-FAST definition.This opinion stemmed from a conceptualization of the 3 distinct body cavities and consideration for procedure reimbursement in E-FAST, including the Current Procedural Terminology (CPT) coding schema.The CPT schema divides ultrasonographic studies into body cavities, including abdominal, cardiac, and thoracic.Therefore, the addition of a pleural evaluation may alter the reimbursement criteria.However, the historical use of the FAST included the evaluation for hemothorax; thus, the panel agreed to include hemothorax evaluation in both the FAST and E-FAST definitions. 13Conversely, the panel discussed replacing the E-FAST definition and incorporating pneumothorax into the FAST definition.However, given that most children with injury are treated in general emergency departments, 36 where many practitioners train using the adult context or definitions of FAST and E-FAST, it was decided that both definitions would remain unchanged to avoid confusion between pediatric and adult POCUS studies.children.In contrast to the adult protocols, the children's protocols highlight the importance of the suprapubic view, which is the most sensitive for abdominal free fluid, especially in prepubescent children. 37In the round 1 survey, the panelists included more views and landmarks for defining a negative rather than a positive result.For example, only 1 view with intraperitoneal free fluid is necessary to define a positive abdominal FAST or E-FAST result.In contrast, all essential landmarks are important to define a complete negative FAST or E-FAST result.In addition, a negative result must be without free fluid in all required views.However, the panel noted that each view evaluates different anatomic regions; therefore, even an abdominal view with free fluid may still require cardiac and bilateral thoracic views.During the discussion, the panelists suggested that reporting FAST and E-FAST results should include any missing view or landmarks in the interpretation as well as patient factors (eg, stability, sex, and age) that could limit the examination.An example of a qualifying interpretation may state, "FAST result is negative but unable to visualize splenorenal recess in this unstable patient."Future work should consider the development of a qualifying schema to standardize FAST and E-FAST reporting.Similarly, future investigation should focus on whether certain landmarks are more critical to view than others for a complete FAST.
Historically, FAST or E-FAST has been measured against improper reference standards.For example, the consensus panel found it inappropriate to use FAST or E-FAST to detect abdominal solid organ injury.Instead, FAST is used to detect intraperitoneal free fluid, which is assumed to be hemoperitoneum in the setting of trauma.This distinction is important because the FAST or E-FAST is not a replacement for CT scans but instead may be a tool within a diagnostic strategy for identifying the need for advanced imaging, resource utilization, or acute interventions. 38This distinction is important for future clinical and research purposes.
The panelists recognized that, most often in the suprapubic view, the FAST or E-FAST may show small volumes of physiological free fluid in some children. 32A few studies have suggested that clinicians who perform the FAST or E-FAST may accurately and reliably differentiate physiological from pathological free fluid. 39,40However, larger, more comprehensive studies would need to be completed before the panelists could agree that isolated, trace amounts of free fluid in the pelvis may be recognized as a qualified physiological finding by clinicians.Furthermore, the panelists emphasized the importance of clinical context and patient factors (eg, age and sex) when qualifying trace free fluid.Similarly, the panelists identified the dynamic opportunities of FAST and recognized that volume limits the detection of free fluid. 41,42In adult patients, 50 to 250 mL of free fluid must be present before it could be reliably detected on the FAST or E-FAST. 28To overcome the volume threshold barrier, the panelists suggested performing serial FAST studies to help improve overall accuracy by recognizing the dynamic changes in volume. 43Future studies should assess the accuracy and reliability of the FAST and E-FAST by defining trace free fluid and conducting serial studies.

Limitations
This study has several limitations.First, the training and practice settings of this panel of POCUS experts may limit the generalizability of the findings.For the present modified Delphi technique, we included experts with substantial experience.Most panelists represented leaders in pediatric emergency POCUS within North America and were English speakers.However, these experts were from diverse geographic areas and represented highly respected institutions.Second, the focus of the consensus panel was to optimize the test characteristics of the FAST by defining an expert-level FAST protocol.The panel did not investigate clinical integration or address the psychomotor skills required for image acquisition.

Conclusions
In this qualitative study, the expert panel achieved consensus on the definitions for complete FAST and E-FAST studies with high image quality and accurate image interpretation in children with injury.
These definitions are similar to the adult protocol definitions.The consensus statements may be

Figure 1 .
Figure 1.Study Design With a Modified Delphi Approach

Table 1 .
Characteristics of the Consensus Panel a Not including fellowship or residencies; 23 of 26 physicians (88%) completed more than 1500 examinations in their career.All 26 physicians have either fulfilled this requirement or held an ultrasonography-related leadership role in their department (eg, ultrasonography director, POCUS director, and POCUS fellowship director).

Box. Definitions and Hybrid Summary Statements for FAST and E-FAST for Children With Injury
and accurate interpretation of FAST.However, they did not agree that some anatomic landmarks were not as important as others for a complete FAST.For the right upper-quadrant abdominal view, the panelists agreed on the importance of the Morison pouch (median [IQR] rating, complete

Table 2 .
Consensus Panel Ratings of Final Landmark Items and Ultrasonographic Views in FAST and E-FAST (continued)