Thorax, Trachea, and Lung Ultrasonography in Emergency and Critical Care Medicine: Assessment of an Objective Structured Training Concept

Background and Study objective. Focused lung ultrasound (LUS) examinations are important tools in critical care medicine. There is evidence that LUS can be used for the detection of acute thoracic lesions. However, no validated training method is available. The goal of this study was to develop and assess an objective structured clinical examination (OSCE) curriculum for focused thorax, trachea, and lung ultrasound in emergency and critical care medicine (THOLUUSE). Methods. 39 trainees underwent a one-day training course in a prospective educational study, including lectures in sonoanatomy and -pathology of the thorax, case presentations, and hands-on training. Trainees' pre- and posttest performances were assessed by multiple choice questionnaires, visual perception tests by interpretation video clips, practical performance of LUS, and identification of specific ultrasound findings. Results. Trainees postcourse scores of correct MCQ answers increased from 56 ± 4% to 82 ± 2% (mean± SD; P < 0.001); visual perception skills increased from 54 ± 5% to 78 ± 3% (P < 0.001); practical ultrasound skills improved, and correct LUS was performed in 94%. Subgroup analysis revealed that learning success was independent from the trainees' previous ultrasound experience. Conclusions. THOLUUSE significantly improves theoretical and practical skills for the diagnosis of acute thoracic lesions. We propose to implement THOLUUSE in emergency medicine training.


Introduction
Focused lung ultrasound (LUS) examinations are increasingly important diagnostic tools in emergency and critical care medicine [1][2][3][4]. A broad area of medicine, including internal medicine [5,6] and traumatology [1,7], can benefit from its time and cost-effectiveness and absence of radiation exposure, as well as the reduction of need for transportations 2 Emergency Medicine International of ventilated patients. Using portable ultrasound devices, LUS is now available in virtually all in-and out-of-hospital scenarios [8].
However, unlike focused assessment with sonography in blunt trauma (FAST) that has become a standard procedure, ultrasound examination of thorax and lungs has only recently been established for workup of critically ill patients [9]. Still, when a lesion is suspected, chest X-rays (CXR) or CT scans remain the diagnostic procedures of choice. Apart from echocardiography, noninvasive examination of the chest by ultrasound is not routinely performed in most centers although many investigations support its use [4,10,11].
While the detection of fluid such as in hematothorax or pleural effusion (PLE) has been described as relatively simple [12] and ultrasound-guided thoracocentesis can be safely performed in ventilated patients [5], sonographic diagnosis of pneumothorax (PTX) represents a challenge to most examiners. Using a panel of previously validated ultrasound findings and artifacts, for example, absence of lung sliding and comet tails [13,14] or presence of lung point phenomenon in the M-mode scan, it was demonstrated that diagnosis of PTX could be achieved in real time and with high sensitivity and specificity [4,15,15,16]. There is evidence that LUS allows reliable detection of PTX under conditions (e.g., prehospital) where CXR or CT is not available [1].
To our knowledge, no focused training method for chest sonography, particularly for the use in emergency situations, has been validated until today.
Therefore, we designed an objective structured clinical examination (OSCE) [17] training concept for focused thorax, trachea, and lung ultrasound in emergency and critical care medicine (THOLUUSE). In the present study, we aimed to test if our training concept can improve (1) factual knowledge of theoretical background, (2) visual perceptive skills, and (3) practical imaging performance, that is, the ability to steer the ultrasound probe [18], obtain reproducible image scans, and interpret structures or artifacts during ultrasound examination of the chest.

Study Design.
We performed a prospective educational study with a standardized OSCE curriculum. A total of 54 trainees were enrolled into the program in 4 independent training days. Group A included fourteen medical students (age 25 ± 2 years (mean ± standard deviation)) and group B thirty-two anesthesiologists (32 ± 5). Both groups had none or very limited ultrasound knowledge. Group C consisted of eight trauma surgeons (37 ± 6), who all previously underwent a FAST training [19]. Instructors were ultrasound-trained pulmonologists ( = 3), internists ( = 3), cardiologists ( = 4), anesthesiologists ( = 2), and trauma surgeons ( = 2). Medical interns ( = 3) were specifically trained to give instructions for use of ultrasound phantoms. A formal institutional review board approval was obtained at the University of Frankfurt, Medical faculty. All trainees gave informed consent for anonymous analysis of their test results. Patients or relatives of ventilated patients gave their informed consent to serve for teaching purposes according to local ethical standards at our institution.

Course Curriculum and Training System.
Theoretical training was scheduled for 2.5 hours and included six brief lectures of anatomy, physiology, and pathology of the thorax, as well as four case presentations on clinical scenarios related to PLE and PTX (Table 1). Practical training was scheduled for 2.5 hours and included two units of hands-on training (HOT), following modifications of previously described OSCE protocols [20,21]. In HOT-1, each trainee had to pass six and in HOT-2 seven different thematic stations (Table 2). There was a strictly organized circuit system between the HOT stations. The ratio of instructor per station and trainee was 1 : 1 and 1 : 2 offering each trainee at least 5 min of training per station. Each instructor gave a short introduction to the specific objective of the station, followed by a oneminute demonstration of the standardized sonographic views or procedures. In total, twelve major sonographic views or artifacts were studied in healthy models. The same healthy models were used in HOT-1 and -2, but views and structures alternated between individuals. Two stations in HOT-2 included patients with selected pathologies ( Table 2). Patients with chronic or malignant lung disease were included to demonstrate PLE; patients that recently underwent sterno-or thoracotomy due to cardiac or pulmonary disease were used to demonstrate PTX. Both HOT-1 and HOT-2 included one virtual station, where typical pathologic findings and artifacts were demonstrated in the form of electronic pictures and video clips on a laptop computer (Table 3).

Phantoms and Ultrasound Equipment.
In order to support the trainees active learning process, we included custom made gel phantoms into the HOTs. Gel was made by commercially available pork skin leafs (Dr. Oetker Nahrungsmittel, Bielefeld, Germany). Phantoms were prepared with 20 grams (equivalent to 12 pieces) of gelling leafs per half a litre of distilled water, carefully heated to 60 ∘ C, followed by stirring for 1 min. Next, gels were casted into plastic containers (95 × 15 × 15 cm) and allowed to cool down for two hours at room temperature. While still viscous, water and content filled rubber balloons were carefully submerged in the gel body by avoiding insertion of air bubbles. Next, gels were incubated in a refrigerator overnight until solid. Stiff gel phantoms were used without antimicrobiological additives for 10 days. In HOT-1, one station contained a series of rubber balloons. Each balloon was filled with specific contents empirically chosen to mimic typical phenomena found in chest ultrasound: (a) pure water to visualize transmission and reflexion of ultrasound on interfaces with different acoustic impedance, for example, thorax wall, pleural line, and PLE; (b) air to mimic reverberation artifacts found in PTX; (c) parboiled rice grains and starch to mimic partially consolidated PLE and fibrinous structures; (d) olives as substitute for soft but solid tissues; (e) a stone as substitute for carbonrich structures such as bones with complete ultrasound absorption and dorsal extinction. Finally, balloons with small jelly babies sized less than 1 cm and a tiger duck made of  wood of a size of about 2 cm were included as target structures and to practice recognition of specific objects (Figures 1(a)-1(g)). In HOT-2, one station contained a thoracocentesis set (Pleurofix no. 1, B. Braun, Melsungen, Germany), a 10 mL syringe, and 20 G cannulae (size 0.9 × 0.9 cm). A puncture model was custom made of gel cuboids with immersed rubber balloons (approx. 7 cm in diameter) and prefilled with a clear yellowish liquid, resembling PLE. Each balloon could be used at least two times when reinjected with liquid. Trainees received instruction on how to use thoracocentesis equipment and on techniques of ultrasound-guided puncture, for example, in-line and out-of-plane visualization of the needle (Figure 1(b)). Additionally, trainees were taught a standardized sonographic examination sequence to screen for PLE or PTX (Figures 2(a) and 2(b)). Ultrasound equipment used in this study was SonoSite Titan and MicroMaxx, equipped with a linear array (L38e, frequency range: 10-5 MHz), micro convex array (C15 7.5, 4-2 MHz), or convex array (C60e, 5-2 MHz) probe, kindly provided by SonoSite GmbH, Frankfurt, Germany.

Course Assessment.
For scientific evaluation of the THOLUUSE program, trainees had to pass three types of tests: (a) a theoretical test with multiple-choice questions (MCQs) before (precourse) and after (postcourse) completion of the program. Each MCQ test set contained fifteen textural questions and five questions with scanned ultrasound pictures where specific sonoanatomical structures had to be identified. The maximum time to answer each question was 60 seconds. As one of the posttests, (b) a recognition quiz to test the visual perception skill with fifteen video clips (see Appendix Table 1 in Supplementary Material available online at http://dx.doi.org/10.1155/2013/312758), each loop had a duration of 10 seconds followed by a 5-second break to note the results onto a standardized answering form. A self-running DVD was produced with MAGIX 5.5 deluxe (MAGIX AG, Berlin, Germany). The DVD was started once and ran continuously without replay or break until all video clips were shown. The order of the clips, questions, and answers for postcourse testing was changed in a randomized fashion to prevent memory of order or answers. Finally,    at the end of HOT-2, there was (c) a practical postcourse examination in which trainees had to demonstrate their skills in correct positioning of the ultrasound probe and to visualize and identify sixteen predefined sonoanatomic items considered particularly relevant to chest ultrasound (see Appendix Table 2). Over 15 min, trainees performance was observed by an instructor and rated on a standardized score sheet. Instructors were not allowed to help or modify the positioning of the probe, and noted results were blinded to the participants. A cut-off level of 60% correctly performed tasks/identified structures was arbitrarily defined to passing before the onset of the study.

Statistical Analysis.
Our null hypothesis was that training would not significantly increase trainees skills. A finding was considered significant when error probability was less than 5% and null hypothesis could be rejected. Results are given as box plots, scatter plots, or mean and confidence intervals or standard deviation if not indicated otherwise. For comparison of the training effects of the different test groups, nonparametric Wilcoxon matched-pairs signed rank test (within groups) or Mann Whitney -test (between groups) was applied. Originally, we had scheduled a precourse test for practical skills as well. However, we had to omit precourse testing because of insufficient knowledge and practical skills of most participants before entering the study. Therefore results of the postcourse test were computed against zero.
Test results are shown in two ways: (a) mean result of all trainees passing a test was regarded as general learning effect by examinees and (b) number of trainees who scored positive for a specific question or observation. The latter was interpreted as learning success on a specific item. Statistical analysis and figures were produced with GraphPad Prism software (San Diego, CA, USA).

Assessment of Immediate Learning Effects.
Medical students (group A), anaesthesiologist residents (group B), and trauma surgeons (group C) mean precourse scores of correct answers were 11.2 (10.5-13.1, 95% CI), 11.5 (9.8-13.1), and 11.8  (Figure 3(a)). When MCQs were sorted according to the learning target of the course curriculum, the gain of theoretical knowledge in sonoanatomy, physiology, and pathology of the chest, particularly regarding PLE and PTX, was homogeneous in all groups (Figure 3(b)).

Recognition and Interpretation Skills.
The ability of the trainees to evaluate a specific physiological or pathological ultrasound finding under pressure of time was tested by recognition and interpretation of video clips. The number of correct answers markedly increased from pre-to postcourse tests in all groups: group A from 7.9 (6.8-9.1) to 11.6 (10.8-12.5, < 0.001 in comparison with precourse testing), in B from 8.5 (6.7-9.2) to 12.3 (10.3-13.2, < 0.001), and in C from 9.1 (7.1-11.2) to 12.6 (11.4-13.9, < 0.01), respectively (Figure 3(c)). When the scores were sorted by pathology, best results were obtained for detection of fluids. Following completion of the course program, most trainees were familiar with basic sonographic signs for detection of PTX as compared to normally inflated lungs by using both Band M-mode, (Figure 3(d)).

Practical Imaging Performance.
Testing more complex practical capabilities (i.e., steering the probe to a standardized anatomic position, obtaining a clear view, and identifying the sonoanatomical findings/artifacts), a mean of 95% of the postcourse sonograms was conducted in a technically correct fashion (Figure 4(a)).
Subgroup analysis of practical performance skills revealed no statistical significant differences of visualization and interpretation of distinct anatomic sites/learning items (Figure 4(b)).

Discussion
In the present study, we developed and assessed an OSCEbased training program for THOLUUSE. For more than a decade, FAST has been established as a standard procedure in shock rooms or emergency departments for early diagnosis of cavity bleeding in blunt trauma. Kirkpatrick et al. [1] suggested an extension of FAST (the so-called EFAST) to include sonographic detection of posttraumatic pneumothoraces in the algorithm. However, in the absence of appropriate teaching models and concepts, it remains challenging for an individual examiner to gain sufficient expertise for correct sonographic detection of PTX. Techniques of chest ultrasound, including PTX, have previously been described; however, the effectiveness of those instructions was never assessed and/or validated [1,16,22,23].
It has been questioned if a one-day training course with a brief factual input and few essential exams can teach ultrasound skills that can be of practical relevance for the trainee [24]. It is well known that factual knowledge communicated by noninteractive lectures has only limited impact on gaining and keeping knowledge [25]. With respect to ultrasound examination, it seems rather unlikely that sufficient expertise can be gained from such lectures alone [23]. In the same token, Sisley et al. reported that simply assessing factual knowledge gives an incomplete picture of performance, and they observed low image interpretation skills in the ultrasound evaluation of trauma [23]. Therefore, in this study, we chose an OSCE-based approach that encourages active learning and practical performance. This method is also useful to assess different competencies and can identify/unmask areas with deficiencies that are of practical relevance [23].
Our results show that upon completion of a one-day LUS course participants were able to perform a standardized sonographic examination of the thorax and correctly identify and interpret the most relevant physiological and pathological findings, including PLE and PTX. Of note, we found no significant differences between medical students and postgraduates from different medical professions, and trainees postcourse performances were independent of previous levels of ultrasound experience. Our findings are supported by a project for experimental ultrasound learning, the Advanced Diagnostic Ultrasound in Microgravity (ADUM) [26], launched by the National Aviation and Space Agency (NASA). In this study, nonmedical astronauts were trained by On-board Proficiency Enhancer (OPE) software to send a sonogram to a remote ultrasound expert on earth [27]. The authors demonstrated that sufficient skills for technical conductance and image acquisition of ultrasound can be taught to ultrasound novices in a few hours [26]. a novice echocardiographer performed the exam [28]. For FAST, it has been shown that a 1-day training course allows trainees to perform a preclinical FAST with a high level of accuracy [19].
A limitation of this study was that postcourse performance of trainees could only be assessed shortly after completion of the program. Previous evaluations of FAST course programs demonstrated that trainees skills decreased as a function of time, particularly when ultrasound techniques were not practiced on a regular basis and further supervised teaching was not available [25]. Therefore, we and others currently are establishing a continuous educational program that combines modules for abdominal, heart, and thorax ultrasound in emergency and critical care medicine [29,30].
In this respect, the THOLUUSE concept proposed in this study was certified by the German Society of Ultrasound in Medicine (DEGUM) and incorporated substantially into a modular blended learning programme of the German Society of Anesthesiology and Intensive Care Medicine (DGAI) for official training in emergency and critical care medicine [31].

Conclusion
The results of our study demonstrate that a one-day training program like THOLUUSE significantly improves theoretical and practical skills for sonographic diagnosis of acute thoracic lesions, including PLE and PTX. Postcourse gain of competence was independent of previous ultrasound expertise of the trainees, and we propose to implement THOLU-USE in the training of medical emergencies. THOLUUSE long time impact on the management of patients in emergency and critical care medicine needs further investigation.