If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Prehospital medicine has struggled to manage critical patients without the resources available to hospital-based teams. Point-of-care ultrasound could bridge this resource gap by providing critical insight into the pathology of trauma patients. This study aimed to determine if early positive extended focused assessment with sonography in trauma (eFAST) identification would lead to improved patient outcomes.
Methods
This is a prospective observational trial that took place from February 1, 2019, to August 13, 2021. Paramedics, with no prior ultrasound experience, at a single ground ambulance agency were trained in obtaining and interpretating eFAST examinations.
Results
Thirty-seven paramedics were trained and performed a total of 502 eFAST examinations with a total correct interpretation rate of 97.35%. There was a sensitivity of 30.0%/75.0%, specificity of 98.75%/94.05%, a positive predictive value of 33.33%/37.5%, a negative predictive value of 98.55%/98.75%, a positive likelihood ratio of 24.05/12.6, and a negative likelihood ratio of 0.71/0.27 for all exam/patient-only scans. The time spent on scene for eFAST and non-eFAST calls was not significantly different (F3, 2,512 = 2.59, P = .051, = .003).
Conclusion
Although we were able to show successful training and interpretation of eFAST with paramedics, given the low prevalence of disease, our study did not show eFAST use improving patient outcome. However, the large likelihood ratio suggests its benefit may lie with appropriate trauma resource utilization.
The implementation of emergency ultrasound (US) performed by nonradiologists has long been considered the standard of care in emergency departments (EDs) across the United States.
US allows emergency physicians to answer focused clinical questions at the bedside, which translates into faster, more accurate diagnoses of time-sensitive emergencies, leading to improved patient outcomes.
The American College of Emergency Physicians has long endorsed the use of US by physicians for multiple uses, including thoracic and abdominal trauma, abdominal aortic aneurysm, and determining cardiac wall activity.
Furthermore, the American College of Surgeons, in the 2008 Advanced Trauma Life Support guidelines, recognize the extended focused assessment with sonography for trauma (eFAST) as the standard of care for trauma patients.
The use of US in the prehospital environment has continued to evolve with its implementation by physician and nonphysician providers both in the United States and abroad, including the United Kingdom, Europe, and Australia. Using point-of-care ultrasound (POCUS), emergency medical service (EMS) providers gain added insight into medical and trauma pathology not reliably detected by physical examination. This insight allows providers the ability to diagnose, initiate treatment, and stabilize life-threatening conditions within the first critical minutes of a patient decompensating.
Furthermore, multiple studies have demonstrated that paramedics can easily and accurately acquire US skill with training lasting from 6 hours to 2 days depending on the diagnostic applications learned.
Prehospital medicine has long been challenged to manage critical patients without the resources and tools available to hospital-based emergency teams treating those same patients. By implementing US, a noninvasive intervention, paramedics will be provided a tool capable of furnishing valuable diagnostic information that could alter the resuscitative efforts in critically ill or injured patients with the potential to reduce morbidity and improve outcomes of patients with life-threatening emergency conditions.
In the critically injured patient, an accurate physical examination is critical to the management of that patient. However, it is generally recognized that the physical examination is relatively insensitive compared with advanced imaging modalities.
Although the identification of free fluid in the abdomen is not readily managed in the prehospital environment, the identification and management of life-threatening conditions such as hemo/pneumothorax and pericardial effusion require immediate intervention. Similarly, although the identification of free fluid in the abdomen may not necessitate prehospital intervention, that knowledge could allow for the transfer of the patient to an appropriate trauma-ready facility and/or allow such a facility to mobilize the resources needed, which has morbidity and mortality benefit.
Despite research demonstrating the benefits of US and that paramedics can be trained to successfully obtain and interpret US images, US has been slow to be adopted by EMS agencies. Despite a decrease in cost and improvements in size and usability, a paucity of evidence demonstrating patient benefit is likely the primary reason for slow EMS adoption. This study aimed to determine if early positive prehospital eFAST identification would lead to improved patient outcomes (ie, shortening the time to the surgeon at bedside, the time to blood product administration, a decreased ED dwell time, and an improved Mainz Emergency Evaluation System with capnometry (MEESc) score in adult patients suffering blunt and/or penetrating trauma).
Materials and Methods
Study Design and Setting
This is a prospective observational trial that rigidly maintained the current standard of care in critically injured trauma patients. This study was conducted by a single hospital-based ground EMS agency in northern Arizona. The agency is the sole ambulance provider, covering almost 6,400 square miles, and is responsible for all emergent and nonemergent ambulance transport within that area, totaling approximately 20,000 calls for service annually. The response area is geographically diverse including the city of Flagstaff (with a population of 80,000) as well as rural and wilderness areas. The population within the response area is culturally diverse with extremes of age and socioeconomic status. The EMS agency transports close to 450 trauma patients (activated as trauma guarded or critical) annually. All patients were transported to a single level 1 trauma center by ground ambulance or flown to the same facility if air transport was deemed necessary because of the patient's condition. The study has received institutional review board evaluation and approval (project number 2018-020) by the Northern Arizona Healthcare Institutional Review Board (IRB). The Northern Arizona Healthcare IRB also approved a waiver of informed consent because of the potential critical nature; patients would be unable to provide verbal or written consent.
Training and data collection took place between February 14, 2019, and August 13, 2021. All paramedics at the agency with 2 years or greater experience as a paramedic were trained to obtain and interpret eFAST examinations. Requiring paramedics to have a minimum of 2 years of experience was thought pertinent to ensure that that there would be a baseline level of skill, knowledge, and comfort in prehospital care. Including paramedics who had just graduated from a paramedic academy or those with less than 2 years of experience was thought to be detrimental because those providers might not have the clinical skill or gestalt that comes with time of interacting and treating patients in the prehospital environment. All paramedics were new to US and had no previous experience obtaining or interpreting examinations. An emergency physician with a fellowship in POCUS designed and provided paramedics with an initial 4-hour training including didactic and hands-on training with assistance by several emergency physicians also fluent in POCUS. Training included US physics, image acquisition, machine functionality, and image interpretation. Scans were conducted on both male and female colleagues who were paid for their time to demonstrate the following views: subcostal cardiac, parasternal long axis, right upper quadrant, left upper quadrant, transverse pelvis, longitudinal pelvis, and bilateral pleura. A 20 question pre- and posttraining test was also performed to measure competency. All US examinations were performed using the Philips (Amsterdam, Netherlands) Lumify (Lumify S4-1 with 1-4 MHz sector/phased array transducers were used), a Food and Drug Administration–approved medical device currently in use in health care settings across the world. In January 2020, a 4-hour refresher training was completed that mimicked the initial training with instruction from the 4 previous physician instructors. Eight total machines were used during the study period. An outline of the training schedule can be found marked as appendix 1. Paramedics were encouraged to scan each other at the start of each shift to build competency and comfort with the procedure. Paramedics were trained to not give priority to obtaining an eFAST examination over other immediate lifesaving interventions or prolong scene time simply to obtain the examination. Paramedics were encouraged to use their judgment as to where to fit the eFAST examination into their assessment and treatment of patients. Primary data points for eFAST to be obtained are 1) the presence or absence of organized cardiac activity, 2) the presence or absence of a pericardial effusion with gross estimation of size (large vs. small), 3) the presence or absence of free intra-abdominal fluid, and 4) the presence or absence of hemo/pneumothorax with gross estimation of size (large vs. small). A copy of the data collection form is attached as appendix 2.
During phase 1, any suspected positive eFAST findings were communicated to the ED during the prehospital notification using agreed-upon verbiage to identify the finding as part of a research project and not a definitive diagnosis. Trauma guarded or critical activation was based solely on predetermined anatomic/physiologic findings, not the result of the eFAST examination. Upon arrival to the receiving hospital, patient handoff to the ED staff would proceed according to standard practice. Phase 2 went live on October 14, 2020. During phase 2, any positive eFAST finding would be communicated to the ED during the prehospital notification, which would immediately trigger a trauma critical activation and the additional resources that such an activation included.
Patients 18 years of age and older suffering blunt or penetrating trauma were eligible for enrollment in the study. Those under 18 not suffering blunt or penetrating trauma or those who the paramedic felt that the time taken to complete an eFAST examination would negatively affect the patient through a delay in providing lifesaving assessments or interventions were excluded.
Data Collection and Handling
All examinations were uploaded onto a secure Health Insurance Portability and Accountability Act–compliant server for later evaluation by 1 of 3 quality improvement (QI)/quality assurance (QA) physicians. The QI/QA physicians were made up of 3 of the 4 emergency physicians who provided the initial and continued training, including the organization's prehospital medical director as well as an ED physician with a fellowship in POCUS. The uploaded files were renamed with a study identifier to maintain patient privacy, and the images were not attached to the patients’ medical record. A data collection form was then filled out by the paramedic and placed in a collection tray for the QI/QA physicians to use when evaluating and grading paramedic interpretation of eFAST examinations. Data were entered into an Excel spreadsheet (Microsoft Corporation, Redmond, WA), and statistical analyses was conducted using SPSS Version 28 (IBM Corp, Armonk, NY).
Quality Assurance
All paramedic performed eFAST examinations were reviewed by 1 of 3 emergency physicians providing QI/QA for the study in a timely manner to determine both quality and accuracy of the scans. Scans were graded on a study conceived scale from 1 to 5. Scan grades 1 and 2 were of poor quality with anatomy difficult to identify. A grade of 3 indicated organs were identifiable, but images were of moderate quality with basic anatomic landmarks identifiable. A grade 4 scan indicated all organs were identifiable with good quality, and grade 5 indicated excellent quality, allowing for the identification of pathology beyond what is typically identified in an extended FAST examination. Physicians would also provide timely feedback to paramedics about the quality of scans. Any equipment issues were tracked and reported to the IRB administrator.
Outcome Measures
This pilot project was designed with 2 outcome goals. The first (phase 1) was to determine the efficacy of EMS providers obtaining eFAST USs with a target of paramedics correctly interpreting examinations at 95%. The second (phase 2) was to determine if early positive eFAST identification and the bundle of care that would follow (ie, shortening the time to the surgeon, a decreased time to blood administration, a decreased ED dwell time, and an improved MEESc score) lead to improved patient outcomes.
Results
Training: Participants
Thirty-seven paramedics participated in the training conducted by the 3 physicians. Of the 37 paramedics, 11 (29.7%) were female, and 26 (70.3%) were male. The mean age of the women (mean = 33.9 years, standard deviation [SD] = 5.52 years) and years of experience as paramedics (mean = 7.6 years, SD = 5.73 years) were not significantly different than the mean age of the men (mean = 34.1 years, SD = 8.83) and years of experience (mean = 7.27 years, SD = 8.83 years; age: t35 = 0.069, P = .945; 95% confidence interval [CI], –5.832 to 6.25; years of experience: t35 = −1.27, P = .900; 95% CI, −6.27 to 5.52).
Training: Test Results
The paramedics who participated in the training took a pretest and a posttest. They also took a test after completing the refresher course. The maximal score on all 3 tests was 20 points. Thirty-seven paramedics took both the pretest (mean = 8.27, SD = 2.365) and the posttest (mean = 17.35, SD = 2.017), and the scores ranged from 3 to 14 on the pretest and 12 to 20 on the posttest. The mean difference (−9.081) for the 2 tests was significant (t36 = −17.596, P <.001, d = 3.14; 95% CI, −10.128 to −8.024). The refresher course occurred 1 year after the initial training. Thirty of the paramedics completed the refresher course and took the refresher 20-point test. The scores ranged from 11 to 18 and averaged 14.07 (SD = 1.760). The refresher mean score was higher than the pre-test, but not as high as the post-test from the initial testing. The refresher tests were taken anonymously so they could not be matched to the pre- and post-tests from the initial training and statistical testing could not be performed to determine if these scores are significantly different or not. The difference in participants in the initial and refresher training was due to providers leaving the organization.
Accuracy of Scans
Ninety examinations were performed on patients in the field between January 2019 and February 2021. Two of the scans had incomplete data and were not included in the analyses (Fig. 1). Because the study had no positive examinations in phase 2 and thus was unable to meet its stated aim, patient demographics like sex, age, and the subsequent diagnoses of the patients were not collected because they were no longer pertinent. The 88 scans were performed by 12 paramedics. The 12 paramedics performed between 2 and 21 scans, with a median of 5 scans performed. Three physicians rated the quality of the scans by the paramedics. The 3 physicians rated 8, 27, and 55 scans. For most patients, 7 different areas were scanned. The descriptive statistics of the quality scores for each area are presented in Table 1.
Of the patient-only scans (88), paramedics rated 8 scans to be positive for an anatomic problem, and the physicians rated 3 of those to be positive. The sensitivity for the scans was 75%. Paramedics indicated that 79 of the scans were negative; the physicians agreed with the assessments, except for 1 patient they rated as positive. The specificity for the scans was 94.05%. The positive predictive value (PPV) was 37.5%, and the negative predictive value (NPV) was 98.75%. The prevalence of an anatomic problem in the sample was 4.55%.
Scene Time
Data for time spent on scene were used to determine if more time was spent at the scene when performing the eFAST compared with when the eFAST was not performed. Data were available for the patients for whom the eFAST protocol was followed and in 2018, 2019, and 2020 when the eFAST was not used. Outliers greater than 3 z scores and times equal to 0 were removed from the data. A 1-way analysis of variance was used to compare the on-scene times for the eFAST and non-eFAST calls. The Levene test was used to determine that the variances were homogeneous (P =.464). The amount of time spent on scene for the eFAST and non-eFAST calls was not significantly different (F3, 2,512 = 2.59, P = .051, = .003).
Discussion
In this study over 30 months, 37 paramedics were trained to perform eFAST examinations. A total of 502 scans were performed, 90 of which were on patients (Fig. 1). Paramedics were able to successfully achieve the study goal of greater than 95% correct interpretation with a total accuracy of 97.35%. A major focus of this study was attempting to show patient benefit in obtaining prehospital eFAST examinations, something previous studies have not done. There was much discussion during the planning phase of the project in trying to identify a meaningful way to measure “patient benefit.” The study authors in conjunction with local trauma surgeons believed that shortening the time to the surgeon at bedside,
a decreased ED dwell time, and an improved MEESc score were all suitable surrogates for measuring benefit. However, with such a low prevalence and no positive eFAST examinations during phase 2, the study was unable to measure its stated goal of improving patient outcomes. A power calculation was not completed before study initiation because the partnership with Philips, who provided the US machines, was for a set period and would not be extended indefinitely to meet a population goal.
It rapidly became clear that the true benefit of eFAST in the field lies in the likelihood ratio (LR) observed (Table 2). In this study, the authors did not observe clinically significant positive and negative predictive values. This was anticipated given the low prevalence of disease in the sample size. Prevalence influences the positive and negative predictive values, and in this sample size, the prevalence of an anatomic problem (eg, free fluid, hemo/pneumothorax, etc) in the sample was 2.04%. The PPV will be increased, and the NPV will be lower with a higher prevalence value. Conversely, the PPV will be lower when the prevalence is low. Given the low prevalence of anatomic problems in this study, the PPV will likely be low, and the NPV will be high.
The positive LR is used to rule in diagnoses, and negative LRs are more effective for ruling out diagnoses. Diagnostic tests with a positive LR greater than 10 are considered good diagnostic tests, whereas tests with a negative LR less than 0.1 are the most effective at ruling out conditions.
Interestingly, the LRs were quite impressive, with a patient only positive LR of 12.06 and a negative LR of 0.27. The current criteria for trauma critical activations are often based on vital sign abnormalities or mechanism. These criteria have LRs all less than what is demonstrated here and in other similar US studies
and patient comorbidities = 0.81-3.1318). It is important to note that US is not to be used as a test to rule out pathology but as an extremely helpful one to raise concern for positive findings. With a positive LR of 12.06 in the field, this indicates that those with abnormal anatomy as visualized via eFAST are 12 times more likely to truly have free fluid, hemo/pneumothorax, or tamponade. This could very well be useful to inform incoming hospitals of possible needs and utilization of appropriate resources.
Table 2All Scans Versus Patient-Only Statistics
All Examinations
Patient-Only Examinations
Statistic
Value
95% CI
Value
95% CI
Sensitivity (%)
30.0
6.67-65.25
75.00
19.41-99.37
Specificity (%)
98.75
97.30-99.54
94.05
86.65-98.04
Positive LR
24.05
6.98-82.81
12.60
4.54-34.98
Negative LR
0.71
0.47-1.06
0.27
0.05-1.45
Disease prevalence (%)
2.04
0.98-3.71
4.55
1.25-11.23
PPV (%)
33.33
12.68-63.26
37.5
17.77-62.49
NPV (%)
98.55
97.84-99.03
98.75
93.53-99.77
Accuracy (%)
97.35
95.51-98.58
93.18
85.75-97.46
CI = confidence interval; LR = likelihood ratio; NPV = negative predictive value; PPV = positive predictive value.
Previous US studies that reported sensitivity and specificity as well as PPV and NPV were studies in which scans were performed by physician or mixed provider levels, making generalizability to prehospital results difficult. There is a significant lack of published research looking at results when the scanning probe is solely in the hands of the paramedic. A recent publication by Strony et al
reported a sensitivity of 0.53, specificity of 0.98, PPV of 89%, and NPV of 88%, which further supports the concept that US can be used proficiently in the hands of paramedics in the prehospital setting.
Despite the smaller size, improved ease of use, and decreased cost of US machines, 2 major concerns that remain for prehospital medical directors and trauma surgeons are that implementing EMS POCUS would result in increased unnecessary resource allocation due to false-positive activations and extend the scene time in order to perform the examination. Despite previous studies
demonstrating paramedics’ ability to be trained and accurately interpret POCUS examinations, many emergency and trauma physicians may still be uncomfortable changing trauma activation criteria without first demonstrating their local paramedics’ proficiency in obtaining and interpreting eFAST examinations.
This requirement, to prove competency, proved to be a necessary time and resource burden in this study. Despite allowing for broad inclusion criteria to enroll patients in the study, the authors surmised that providers would be unlikely to perform eFAST examinations on patients if they felt uncomfortable with the skill or felt like performing it took too long. The decision was made to encourage providers to scan each other at the start of their shifts. These scans would also be graded by the QI physicians with the goal of providing timely feedback and thus improving provider confidence and competence more quickly than simply relying on patient-only examinations. The requirement to prove local competency will likely be a necessary hurdle for any future agency looking to perform prehospital POCUS research.
Although the adage of “load and go” has shifted in recent years to one of “stay and stabilize,” where the focus is put on completing critical interventions and recognizing that the quality of prehospital care plays a pivotal role in patient outcomes,
Association of intra-arrest transport vs continued on-scene resuscitation with survival to hospital discharge among patients with out-of-hospital cardiac arrest.
concern for unnecessarily prolonging scene times is a common adage from prehospital medical directors and trauma surgeons alike whenever a new treatment or procedure is introduced, and performing an eFAST was no different. For this reason, it was critical to show that scene times would not be extended due to performing an eFAST examination on patients. Data for time spent on scene was used to determine if more time was spent at the scene when performing the eFAST compared with when the eFAST was not performed. Data were available for the patients for which the eFAST protocol was followed and in 2018, 2019, and 2020 when the eFAST was not used. Although the values were not significantly different, the average time spent on scene was about 2 minutes longer for those receiving an eFAST. Although the sample sizes greatly differed across the samples, the effect size was small, indicating the differences most likely would not be significant if the sample of eFAST patients was greatly enlarged. Unfortunately, it was not possible to determine if the 2-minute increase demonstrated was caused by the performance of the eFAST examination or simply a correlation because these patients were potentially sicker requiring extrication, transport via helicopter, or other treatments that required prolonged scene times.
Limitations
Unfortunately, this study had several important limitations; 2020 saw the COVID-19 pandemic hit the world, and the energy and focus necessary to run a study like this were forced to shift to other areas of prehospital care. Although the study persisted, provider fatigue, burnout, and COVID-induced protocol changes likely led to fewer patients being scanned than might have otherwise pre-COVID. The pandemic saw local ED and prehospital trauma volumes drop substantially, at 1 point by as much as 27% locally. There was an almost 5-month delay in initiating phase 2 of the study. Phase 2 was meant to begin on June 1, 2020. This would have allowed for a year's worth of data to be collected, evaluated, and presented to the trauma surgeon group who needed to approve the automatic upgrade of all positive eFAST examinations to a trauma critical. Unfortunately, because of COVID, the in-house statistician who was working on the project was pulled away to other projects and was unable to evaluate our data within the predefined time frame. Because there was no budget to pay for an interim statistician, the authors were forced to wait. This led to the evaluation not being completed until the end of July 2020. Further delay was met by turnover within the trauma surgeon group leadership, leaving only interim directors who did not want to approve a project that they were not originally a part of; phase 2 was not ultimately approved until a new director was named and they had time to review the study data. Although this study is promising in terms of demonstrating a positive LR, it was ultimately underpowered and limited by the number of positive examinations. Future POCUS research should focus on larger, trauma-rich EMS systems, which would be more likely to generate the numbers needed to see benefit from the previously stated outcome measures.
Conclusion
This study further supports the current body of research that paramedics can be trained to obtain and interpret eFAST examinations with a success rate equal to that of emergency physicians. Unfortunately, because of a low prevalence of disease during the study period, the authors were unable to evaluate the study's aim to determine if early positive prehospital eFAST identification would lead to improved patient outcomes. Further research is required to clearly identify patient benefit if prehospital US is to make its way into the mainstream.
Association of intra-arrest transport vs continued on-scene resuscitation with survival to hospital discharge among patients with out-of-hospital cardiac arrest.
The authors would like to thank all emergency medical service personnel for their dedication to continued training, quality improvement, and improving patient care.
This project was funded by 2 grants from the Northern Arizona Healthcare Research Department to provide funding for paramedic training and salary for physician oversight. All ultrasound machines were provided in partnership with Phillips. Other than providing the ultrasound machines, Phillips did not take part in study design, data collection, data interpretation or any other part of the project.
00189-4&id=gr1.jpg){kind=link}