Pre-clinical medical student cardiac point-of-care ultrasound curriculum based on the American Society of Echocardiography recommendations: a pilot and feasibility study

Background Cardiac point-of-care ultrasound (POCUS) training has been integrated into medical school curricula. However, there is no standardized cardiac POCUS training method for medical students. To address this issue, the American Society of Echocardiography (ASE) proposed a framework for medical student cardiac POCUS training. The objective of this pilot study was to develop a medical student cardiac POCUS curriculum with test scoring systems and test the curriculum feasibility for a future definitive study. Methods Based on the ASE-recommended framework, we developed a cardiac POCUS curriculum consisting of a pre-training online module and hands-on training with a hand-held ultrasound (Butterfly iQ, Butterfly Network Inc., Guilford, CT, USA). The curriculum learning effects were assessed with a 10-point maximum skill test and a 40-point maximum knowledge test at pre-, immediate post-, and 8-week post-training. To determine the curriculum feasibility, we planned to recruit 6 pre-clinical medical students. We semi-quantitatively evaluated the curriculum feasibility in terms of recruitment rate, follow-up rate 8 weeks after training, instructional design of the curriculum, the effect size (ES) of the test score improvements, and participant satisfaction. To gather validity evidence of the skill test, interrater and test-retest reliability of 3 blinded raters were assessed. Results Six pre-clinical medical students participated in the curriculum. The recruitment rate was 100% (6/6 students) and the follow-up rate 8 weeks after training was 100% (6/6). ESs of skill and knowledge test score differences between pre- and immediate post-, and between pre- and 8-week post-training were large. The students reported high satisfaction with the curriculum. Both interrater and test-retest reliability of the skill test were excellent. Conclusions This pilot study confirmed the curriculum design as feasible with instructional design modifications including the hands-on training group size, content of the cardiac POCUS lecture, hands-on teaching instructions, and hand-held ultrasound usage. Based on the pilot study findings, we plan to conduct the definitive study with the primary outcome of long-term skill retention 8 weeks after initial training. The definitive study has been registered in ClinicalTrials.gov (Identifier: NCT04083924). Supplementary Information The online version contains supplementary material available at 10.1186/s40814-021-00910-3.


Introduction
As technology advances, cardiac point-of-care ultrasound (POCUS) with hand-held ultrasound devices (HHU) is being used increasingly in clinical practice by multiple specialties including cardiology, emergency medicine, critical care medicine, anesthesiology, and family medicine [1][2][3][4][5]. In 2019, HHU that operate with smartphones or tablets became commercially available (https://www.butterflynetwork.com/). The image quality is high enough for cardiac assessment and the costs are touted as affordable to even medical students [6]. There is a foreseeable future where medical student stethoscopes could be replaced with HHU or "ultrasound stethoscopes" for learning cardiovascular clinical examinations [7].
With growing student interest in cardiac ultrasound examination [8], multiple studies have demonstrated feasibility and benefits of integrating cardiac POCUS into the undergraduate medical education (UME) curriculum [9][10][11][12][13][14]. A comprehensive review article on cardiac POCUS in medical school education described that 12 medical schools in the United States, Canada, Norway, United Kingdom, Israel, Poland, and China have incorporated cardiac POCUS curricula [15]. However, instructional design for teaching cardiac POCUS to medical students is highly variable, without validated or standardized methodology regarding training duration, teaching methods, and competency evaluation [15].
In an effort to address this issue, the American Society of Echocardiography (ASE) proposed a framework for medical student cardiac POCUS teaching for UME program directors wishing to institute POCUS curricula at their schools (https://aselearninghub.org/) [15]. The ASE framework includes a pre-training didactic education with e-learning, hands-on training, and a competency evaluation. The goals are to enhance cardiac physical examination skills and augment learning of normal anatomy, rather than learning advanced pathology. At this time, the effectiveness of the ASE-recommended framework has not been evaluated.
Therefore, the primary objective of this pilot study was to develop a medical student cardiac POCUS curriculum based on the ASE recommendations and evaluate the curriculum feasibility for a definitive study. A secondary objective was to assess the curriculum learning effects in order to identify a single primary outcome for a definitive study.

Design
The study was approved by the University of Hawaii Human Studies Program and was conducted between July 2019 and September 2019 (Protocol number: 2019-00265).
This pilot study was comprised of 4 elements: (i) Development of a cardiac POCUS curriculum for pre-clinical medical students, (ii) Evaluation of curriculum feasibility and further modifications needed for a definitive study, (iii)Development of an assessment tool to evaluate curriculum learning effects and gathering validity evidence of the tool, and (iv) Assessment of the curriculum's multiple learning effect outcomes in order to determine a single primary outcome for the definitive study. We conducted this pilot study based on guidelines for reporting pilot and feasibility studies [16][17][18] with the Consolidated Standards of Reporting Trials (CONSORT) extension to pilot and feasibility trials checklist (Additional File 1).

Participants and setting
We planned to recruit 6 pre-clinical medical students to determine the curriculum feasibility. Eligible participants were first-or second-year medical students who had completed a 12-week pre-clinical cardiovascular and pulmonary core curriculum at the University of Hawaii, John A. Burns School of Medicine (JABSOM) in the United States. We used e-mail and public postings at JABSOM for recruiting participants. One of the investigators (BS) was a medical student at JABSOM and approached his colleagues by word of mouth to invite them to participate.

Cardiac POCUS curriculum development
We developed a basic cardiac POCUS curriculum for pre-clinical medical students based on the ASErecommended framework that encourages using a blended learning system with a flipped classroom model [15,19]. The notions of curriculum design, teaching methods, and competency assessments were underpinned by educational principles for effective learning and skill retention [20], which include concurrent feedback, deliberate practice, mastery learning, and range of difficulty. Student goals for this curriculum were to independently obtain basic cardiac POCUS views in a healthy volunteer and identify the normal anatomic structures seen in the POCUS views. Curriculum developers were educational methods and echocardiography subject matter experts; a cardiovascular anesthesiologist specializing in perioperative transesophageal echocardiography and critical care echocardiography (SJ), a pediatric emergency physician and simulation-based education specialist (JL), a cardiac surgeon (AN), 3 cardiologists specializing in echocardiography and echocardiography education in their university hospitals (AK, MI, KK), a Fellow of the American Society of Echocardiography (FASE) certified echocardiographer (KK), and an adult intensivist and simulation-based education specialist (BB).The curriculum timeline is shown in Fig. 1 [15], while the ASE online module provides 10 cardiac ultrasound views. However, teaching all 10 views in the first 4 ASE modules was inappropriate for initial medical student training [10]. Selection of the 5 views was based on review of 2 guidelines and 3 studies: the World Interactive Network Focused on Critical UltraSound (WINFOCUS) recommendations [4], the European Association of Cardiovascular Imaging (EACVI) consensus for focused cardiac ultrasound (FoCUS) [23], a Focused Echocardiography Entry Level (FEEL) training study [24], a medical student FoCUS training study [14], and an initial transthoracic echocardiography (TTE) training study for anesthesiologists [25]. The WINFOCUS recommendations and the EACVI consensus selected the same 5 views as an ideal standard or basic for cardiac POCUS examination, and the 3 studies utilized the same 5 views for hands-on training. The ASE recently published cardiac POCUS and critical care echocardiography (CCE) training recommendations that also supported our 5 views selection for initial cardiac POCUS learning tasks [3].

Cardiac POCUS hands-on training session
The cardiac POCUS hands-on training session was comprised of a pre-training skill test (10 min), a cardiac POCUS lecture and hands-on training (60 min), and an immediate post-training skill test (10 min). All hands-on training sessions were conducted by a single instructor (SJ). Immediately prior to the start of the hands-on training session, we assessed student baseline 5-view cardiac POCUS image acquisition skill (Pre-training skill test). After the baseline skill assessment, the instructor guided a 30-min interactive lecture using PowerPoint slides and a model heart (Cardiac POCUS lecture). The lecture included didactic content regarding basic 5 cardiac POCUS views and acoustic windows using the ASE online module slides and pre-recorded video instructions. The video instructions, which we created, demonstrated the 5-view image acquisition on the healthy volunteer. The lecture content is summarized in Additional File 2. Image acquisition instructional design in the video instructions was created with reference to an imaging protocol in the ASE comprehensive TTE guidelines [26] and a point-of-care ultrasound textbook [27]. The main instruction points for the 5-view image acquisition are summarized in Additional File 3. Following the cardiac POCUS lecture, students engaged in a 30-min supervised hands-on training of the 5-view image acquisition on a thin, healthy volunteer (Cardiac POCUS hands-on training). The instructor (SJ) assumed the role of the healthy volunteer during the hands-on training and, at the same time, provided concurrent, verbal, and tactile feedback. In the hands-on training, students practiced the image acquisition repeatedly until they were able to obtain each image clearly. In order to determine a feasible group size for the hands-on training session, we trialed 3 different group sizes: one 3-student group, one 2-student group, and one 1-student group. Immediately after and 8 weeks after the hands-on training session, we reassessed student cardiac POCUS skill with the same skill test (Immediate post-training skill test and 8-week post-training skill test, respectively).

Skill test scoring system Skill test
We assessed student cardiac POCUS skill at pretraining, immediate post-training, and 8-week posttraining using a skill test scoring system that we developed. During the skill test, students independently demonstrated the 5 cardiac POCUS views in a fixed sequence (PLAX, PSAX, A4C, S4C, and SIVC) on the same single healthy volunteer without guidance. Students were limited to 2 min to obtain each view, for a total of 10 min for 5 views. During the 2-min image acquisition interval, students self-identified their "best" view and pressed a recording button to acquire a 5- second video clip. Students were allowed to record a maximum of 2 video clips for each view. If they recorded 2 video clips, they chose the single best video clip without review as the recording to be evaluated for this study. We utilized the Butterfly iQ application predefined cardiac ultrasound preset for gain and other ultrasound imaging parameters [28]. The healthy volunteer took the left decubitus position for PLAX, PSAX, and A4C, and the supine position with bent knees for S4C and SIVC, and controlled his respiratory rate at 6 per minute and held his breath for 5 seconds when the video clip recording started. Students, therefore, focused only on probe manipulation to obtain each view and did not adjust ultrasound imaging parameters, the healthy volunteer's position, or breathing cycle during the skill test.

10-point maximum skill test scoring system
We developed a 10-point maximum scoring system by modifying an existing assessment tool for TTE views [25]. The scoring system was designed to assess 5-view image quality for the purpose of rapid bedside cardiac assessment, not for a formal diagnostic comprehensive echocardiography examination. Scoring system development details are shown in Additional File 4. The 10point maximum skill test scoring system rated the 5 views; each view was assessed as excellent (2 points), acceptable (1 point), or poor (0 point) for cardiac POCUS use (Table 1). Excellent quality reference images of the 5 views obtained by the cardiologist (MI) on the healthy volunteer (SJ) are shown in Fig. 2A. Examples of acceptable and poor quality images are shown in Fig. 2 B and C, respectively. After de-identifying the skill test video clips including removal of pre-, immediate post-, and 8week post-training status, we arranged the clips in a randomized order using the Microsoft Excel random number table for blinded assessment. Three independent blinded raters scored image quality using the 10-point maximum skill test scoring system, and the average score from the 3 raters was utilized as a representative score. Interrater reliability of the skill test scoring system was assessed with intraclass correlation coefficient (ICC) using 10-point maximum skill test scores (pre-, immediate post-, and 8-week post-training) and 2-point maximum scores from each of 5 views (PLAX, PSAX, A4C, S4C, and SIVC) from all students. Test-retest reliability of the 3 raters was assessed with ICC by re-assessing the skill test video clips after de-identification. The 3 raters were echocardiography experts who developed the skill test scoring system and were blinded to student identity and timing of the assessment. If interrater reliability and/or test-retest reliability were moderate or poor (ICC < 0.75), we planned to re-evaluate the scoring system or perform rater training.

Knowledge test scoring system
We assessed student cardiac POCUS knowledge at pretraining, immediate post-training, and 8-week posttraining using an identical knowledge test on Google Forms. The knowledge test consisted of 40 multiplechoice questions regarding the identification of normal anatomic structures seen in the 5 cardiac POCUS views. The 40-point maximum knowledge test scoring system is shown in Additional File 5. Students independently completed the pre-training knowledge test prior to reviewing the ASE online module. Students independently completed the immediate post-training knowledge test after the hands-on training session and the 8-week post-training knowledge test after the 8-week posttraining skill test (Fig. 1).

Butterfly iQ hand-held ultrasound
The Butterfly iQ HHU measures 185 × 56 × 35 mm and weighs 313 g, which is connected by wire to a smartphone or tablet with the Butterfly iQ application. The application features 19 selectable presets of optimal imaging parameter values including abdomen, cardiac, lung, and vascular options. The application automatically operates in accordance with a selected preset. The application has touch-screen controls including those for image recording and freezing, and depth and gain adjustment. The battery is designed for run time of over 2 h in B-mode. Ultrasound image recordings can be saved in the application, uploaded to an secure commercial cloud, and downloaded to a computer after data encryption (https://www.butterflynetwork.com/) [28].

Curriculum learning effect outcomes
We measured 7 curriculum learning effect outcomes:

Curriculum feasibility criteria
We semi-quantitatively evaluated the feasibility of the curriculum in terms of recruitment rate, follow-up rate 8 weeks after training, instructional design of the hands-on training session, the effect size (ES) of the test score improvements, and participant satisfaction. Feasibility criteria for recruitment rate and follow-up rate 8 weeks after training were 80% or more. We subjectively assessed the feasibility of the instructional design as not feasible, feasible with modifications, or feasible as is [18]. Identified instructional design flaws were modified where possible. We utilized student curriculum feedback to guide modifications. With respect to the ES of the test score improvements, the criteria were that the ES of the 4 skill and knowledge test score improvements [i-iv] were moderate or large (ES greater than 0.5). With regard to participant satisfaction, the criteria were that the students report satisfaction with the overall curriculum [vii] at post-and 8-week post-training, with mean 5-point Likert scale ratings greater than 3. Based on these feasibility outcomes, we evaluated the overall curriculum as not feasible, feasible with modifications, or feasible as is to implement for a definitive study.

Sample size
The main focus of this study was to develop the cardiac POCUS curriculum with test scoring systems, and then  pilot test it on a small sample size of 6 participants to confirm the curriculum feasibility and gain feedback before progressing to a larger scale study. Numerical data of this study could be utilized for sample size estimation for a future definitive study, but the accuracy of the estimation is limited because of the small samples [29].

Statistical analysis
All statistical analyses were performed using BellCurve for Excel (Social Survey Research Information Co., Ltd.).
All numeric variables are presented as mean ± SD. Mean difference between pre-and immediate post-or 8-week post-training data were calculated with paired t-test and presented as mean and SD with 95% confidence interval (CI) and the ES. We interpreted the clinical significance of ES according to Cohen's effect size guidelines (0.2 to 0.5 = small ES, 0.5 to 0.8 = moderate ES, and more than 0.8 = large ES) [30,31]. ICC estimates and their 95% confidence intervals were calculated based on a mean rating (k = 3), absolute-agreement, 2-way random-effects models for interrater reliability, and a mean rating (k = 2), absolute-agreement, 2-way mixed-effects models for test-retest reliability. We interpreted the ICC estimates according to ICC reporting guidelines (ICC of less than 0.5 = poor, 0.5 to 0.75 = moderate, 0.75 to 0.9 = good, and greater than 0.9 = excellent reliability) [32]. Due to the small sample size of 6 participants, all numerical data should be treated with caution and treated as preliminary.

Characteristics of participants
The first 6 second-year medical students we approached agreed to participate in this study. Student demographic data and previous ultrasound training experience are shown in Table 2. Three students attended an unstructured 1-h introductory focused assessment with sonography in trauma (FAST) hands-on training using student volunteers during a first-year medical school elective before participation in this study. No students had ultrasound experience on patients or simulators. Between the immediate post-training test and the 8-week posttraining test, 3 students reviewed relevant textbooks or websites and 1 student participated in a FAST examination hands-on training on a student volunteer with the medical school ultrasound interest group.

Curriculum learning effect outcomes
All 6 students completed the pre-training ASE online module self-study, the hands-on training session, all the skill tests, and the post-training 5-point Likert scale questionnaires. Five students completed all the knowledge tests; one student did not complete the pretraining knowledge test. Individual skill and knowledge test scores are illustrated in Fig. 3, and mean skill test scores and the scores for each of the 5 views are summarized in Table 3.  (Fig. 3B).

Curriculum feasibility
The recruitment rate was 100% (6/6 students) and follow-up rate 8 weeks after training was 100% (6/6 students). We identified several instructional design flaws requiring modification of the hands-on training group size, content of the cardiac POCUS lecture, hands-on teaching instructions, and HHU usage. The instructional design modifications are detailed in the discussion section. The 4 skill and knowledge pre-post learning effects ([i] skill test score difference between pre-training and immediate post-training, [ii] the difference between pretraining and 8-week post-training, [iii] knowledge test score difference between pre-training and immediate post-training and [iv] the difference between pretraining and 8-week post-training) each demonstrated a large ES of 3.05, 0.86, 4.93, and 3.82, respectively. The students reported [vii] overall curriculum satisfaction with mean 5-point Likert scale ratings were greater than 3 points; 4.5 ± 0.5 at post-training and 4.5 ± 0.5 at 8week post-training (mean ± SD). Based on these feasibility outcomes, we evaluated the overall curriculum's feasibility as feasible with modifications to implement the future definitive study.

Discussion
This pilot study confirmed the feasibility of the preclinical medical student cardiac POCUS curriculum, without recruitment and attrition concerns, with large ES of learning effect outcomes, and favorable student satisfaction with the curriculum. We identified several curriculum design flaws that required modification including the hands-on training group size, content of the cardiac POCUS lecture, the hands-on teaching instructions, and HHU usage.

Curriculum design modifications
Hands-on training group size This pilot study trialed hands-on training group sizes of 3 students, 2 students, and 1 student to determine a feasible group size for a definitive study. In the 3-student group size, hands-on training required an excess of 20 min beyond the programmed training time to provide adequate hands-on practice for each student. Moreover, each of the students had 40 min additional waiting time while other students took the pre-and immediate post-training skill tests. We deemed this long waiting time was inappropriate and impractical for volunteer participants. Both the 2student group and the 1-student group did not raise concerns regarding time management and efficiency. However, during the 2-student group session, we observed the potential for positive or negative peer interactions affecting learning experiences, which could lead to intervention heterogeneity within or between groups, and interfere with consistency and continuity of individualized instructor feedback [33]. A study that compared constant instructor feedback to intermittent feedback with peer feedback in chiropractic skill training demonstrated that constant feedback was advantageous for accurate skill acquisition and initial hands-on practice, while the intermittent feedback was advantageous for long-term skill retention [34]. A study of cultural differences in simulation-based learning showed an individual learner's interaction pattern with instructors and colleagues during debriefing could affect learning effectiveness [35]. These studies suggest a multiple student group size may affect student learning and lead to intervention heterogeneity within a group or between groups. In addition, we speculated that recruitment of more than 1 volunteer for a single session could be difficult considering medical student schedules. Based on the pilot study observations, we selected the 1-student group size with no waiting time, less intervention heterogeneity, and ease of scheduling with student volunteers for the definitive study.  This study did not intend to determine an ideal or practical group size. The choice of 1-student group size was for feasibility to conduct a future study investigating accurate and unbiased learning effects, not to reveal the optimal group size for learning effectiveness or efficiency. This study also did not seek to identify a realistic or practical group size for curriculum integration into the class size of about 70 students per year in our medical school. We assume it is impractical to implement 1on-1 sessions for all. At present, there is no evidence regarding an ideal or practical group size for cardiac POCUS training [15]. The Teaching the Teacher (Module E) of the ASE online module recommends a ratio of 4 students to 1 healthy volunteer to 1 instructor as the ideal learning environment (https://aselearninghub.org/). The American Heart Association (AHA) basic life support (BLS) course may provide guidance regarding a practical group size. The BLS course recommends a ratio of 6 students to 2 manikins to 1 instructor which allows instructors to observe the students, provide feedback and guide the students' acquisition of skills [36]. These recommendations may serve as important references for medical school ultrasound training curriculum integration.
Content of cardiac POCUS lecture Based on student curriculum feedback, we modified the content of the cardiac POCUS lecture for the definitive study. In this pilot study, we included didactic teaching on the cardiac ultrasound beam shape, probe orientation marker, and probe manipulation during the hands-on training, since the pre-training ASE online modules (A, B, and C) did not provide content on these topics. We also included a question-and-answer review of anatomical structures seen in the 5 views during the hands-on training to confirm and reinforce student knowledge. Two students gave feedback that these didactic contents would be better provided during the lecture rather than during the hands-on training in order to confirm prerequisite knowledge before hands-on practice and focus on manual skill acquisition during the hands-on training time. In the definitive study, we plan to incorporate the items not involving direct hands-on practice into the cardiac POCUS lecture for effective use of hands-on training time. The modified content of the cardiac POCUS lecture is in Additional File 6. Although mentioned in the ASE online module, all 6 students did not understand that the papillary muscle level of parasternal short-axis view was helpful for left ventricle (LV) wall motion analysis. We plan to reinforce this clinically important concept in the definitive study.

5-view hands-on teaching instructions
Mean skill test scores for PSAX at 8-week post-training and A4C score at immediate post-training were less than 1.0 point, which means poor or unacceptable for clinical use (Table 3). We plan to modify the PSAX teaching instruction with reference to a POCUS textbook [27] in a more understandable way with step-by-step instruction of probe manipulation using the script: First, you start from a clear image of PLAX. Second, you rotate the probe 90 degrees clockwise, and then you usually find the fish mouth appearance of the mitral valve level in the short-axis view. Third, you tilt the probe downward carefully until you see 2 papillary muscles and you should get a clear image of the papillary muscle level of the parasternal short-axis view.
During A4C teaching in the study, in order to demonstrate a true longitudinal LV cavity, we instructed students to tilt up a probe adequately after placing the probe on an intercostal space without distorting the image. However, we observed that distorted images after adequate probe tilting was unavoidable for the healthy volunteer A4C image acquisition with the HHU. Therefore, we accepted a mild LV foreshortening and plan to prioritize clear visualization of all 4 chambers in order to provide straightforward instructions to novice students for the definitive study. Our modification of the A4C teaching instruction was supported by ASE POCUS and comprehensive TTE guidelines [3,26].
The SIVC teaching instruction resulted in excellent quality image acquisition by all students immediately after training (Table 3). However, all students queried the significance of SIVC findings in clinical practice. In our curriculum, we did not focus on image interpretation but rather on proper image acquisition for initial training in cardiac POCUS skills for novice students at an appropriate level for learning effectiveness [15,20]. In order to balance student curiosity with maintaining an appropriate difficulty level for the curriculum, we plan to provide a brief commentary about volume responsiveness assessment and central venous pressure estimation using inferior vena cava (IVC) diameter based on our clinical experience, ASE TTE guidelines [26,37], and a POCUS textbook [27].

HHU (Butterfly iQ) limitations
We experienced automatic ultrasound scan shut-off due to probe temperature increase after continuous scanning during the hands-on training. Once the probe heated up and shut down, we had to suspend scanning for several minutes until the probe cooled down [28]. To manage this technical issue, we plan to utilize 2 Butterfly iQ probes in the hands-on training session for the definitive study.

Test scoring system validity evidence
We gathered validity evidence based on Messick's framework recommended by the American Educational Research Association (AERA) [38]. AHA also referenced Messick's framework to evaluate assessment tools for resuscitation in their scientific statement of resuscitation education science [39]. Messick's framework classifies sources of validity evidence into 5 categories: content, response process, internal structure, relations to other variables, and consequences. In our study, content evidence refers to the relationship between the test content and what is intended to measure; response process evidence assesses test quality control or the extent to which the test administration and scoring are controlled or standardized to reflects the observed performance; internal structure evidence evaluates the reliability and reproducibility of the test results; relations to other variables evidence evaluates discriminatory ability of the test between experts and novices or the correlation between the test results and known measures of competence; and consequences evidence refers to downstream effects of test results and the decisions that are made [40][41][42]. Our skill and knowledge test scoring systems were developed by echocardiography experts, through extensive review and evaluation of previously published studies, expert discussions, and an iterative design process, thus contributing to the content evidence. Our control measures for obtaining the ultrasound video clips involved asking the students to record when they perceived they had obtained their "best" view, allowed for a second chance to record another view, and had the student choose which of the 2 recordings to submit for assessment. The students controlled the recording, without instructor intervention, therefore adding to the response process evidence of the skill scoring system. Our skill test scoring system demonstrated excellent interrater reliability and test-retest reliability which contributes to the internal structure evidence of this tool. With regard to relations to other variables evidence, we did not assess the test scoring system discriminatory ability between experts and novices for echocardiography because of small novice samples. We planned to gather the evidence in the definitive study with sufficient novice samples. With respect to consequences evidence, we did not gather the evidence because our pilot study did not intend to establish a pass/fail cut-off score or identify the clinical impact of the curriculum. Given the evidence for validity demonstrated by our tools, we determine that these scoring systems can be utilized in the definitive study.

Post-8-week skill retention
Skill test score results demonstrated that 3 students (student C, E, and F) retained skill 8 weeks after training and that 3 (student A, B, and D) did not (Fig. 3). From the post-training questionnaires, we determined 2 possible reasons for the inconsistent skill retention finding. Of the 3 students who retained skill, 2 (student C and E) attended an unstructured 1-h introductory FAST handson training elective using student volunteers before the participation in this study and 2 (student E and F) reviewed relevant websites or textbooks after hands-on training ( Table 2). We cannot reach decisive conclusions about the divergent skill retention finding due to the small pilot study sample size. If the definitive study demonstrates similar findings, we plan to formally examine demographic factors that could explain individual skill retention variation. We also found that 8-week posttraining skill test scores in the 3 (student A, B, and D) who did not retain their skill were lower than their pretraining scores. One possible reason for the lower scores is that the pre-training skill test was performed after pre-training self-study of the ASE online module, so the test scores did not necessarily represent a true baseline but rather skill after a recent review of the online module content. The 3 students may not have retained the online module content in addition to the skills acquired during hands-on training.

Single primary outcome selection for a definitive study
In this pilot study, we measured 7 learning effect outcomes. We needed to select the most important single primary outcome measure or question in order to establish a primary hypothesis for the definitive study. Our pilot study demonstrated short-term knowledge and skill retention and long-term knowledge retention with a large ES and the 95% CI not including zero, with consistently favorable student confidence and the curriculum satisfaction ratings. However, the skill retention remains unknown due to the small sample size, even though the ES of the skill retention was large. Most previous studies investigating the learning effect of cardiac ultrasound curriculum showed psychomotor or cognitive skill outcomes were improved immediately after an initial training [10,13,14,[43][44][45][46][47][48][49][50]. Assessments in these studies focused on very short-term knowledge and/or skill retention. Several studies have demonstrated longterm knowledge or confidence retention after several weeks or months [22,25,51,52], however, few studies have investigated long-term skill retention [22]. The importance of investigating long-term skill retention is evident both from the results of our pilot and our review of the literature. Therefore, we selected the outcome of the skill test score difference between pre-training and 8week post-training as a single primary outcome for the definitive study that is registered in ClinicalTrials.gov (Identifier: NCT04083924).

Conclusion
This pilot study confirmed the feasibility of the ASErecommended medical student cardiac POCUS curriculum, without recruitment and attrition concerns, with large ESs of skill and knowledge learning effect outcomes, and favorable student curriculum satisfaction. The skill and knowledge test scoring systems gathered validity evidence. We modified several curriculum design areas including the hands-on training group size, cardiac POCUS lecture content, the hands-on teaching instructions, and HHU usage. Based on the pilot study findings, we plan to conduct the definitive study with the primary outcome of the long-term skill retention 8 weeks after initial training.