The use of thoracic ultrasound has expanded widely within the las couple of years, and several studies have proved a high diagnostic accuracy for many of the most common causes of respiratory failure and dyspnoea.

The ultrasound scan is a bed-side, and dynamic examination, which demands sufficient theoretical and practical knowledge and competence by the operator, but so far, no studies have explored the effect of simulation-based training for gaining adequate competence compared to traditional hands-on training on healthy figurants.

The aim of this study is to examine whether TUS training on a simulator is superior to training on healthy figurants. Secondly, to examine whether the choice of hands-on training has an effect on the number of examinations performed by the trainees from baseline to 4 months follow-up.

Thoracic ultrasound (TUS) differs from ultrasound examinations in other organ systems, because it is not possible in the healthy, ventilated lung to visualize structures or anatomical parts of the lung, like it is when examining abdomen or performing an echocardiography. It is therefore not possible to transfer results directly from educational studies using simulation-based training in other areas of medicine, to thoracic ultrasound.

TUS examinations are considered safe, and without pain, exposure to radiation or delay of patients' course, but ultrasound in general is highly operator dependent and lack of theoretical knowledge or practical skills could potentially lead to incorrect diagnosis and thus treatment. Therefore simulation-based training, theoretical and practical tests could be the key to a "pre-trained novice" with a level of competency higher than a complete novice when performing the first TUS examination on a patient in a clinical setting.

The objective of this trial is to examine whether TUS training on a simulator is superior to training on healthy figurants, which today is a commonly used method for gaining skills and competencies in TUS. Secondly, to examine whether the choice of hands-on training has an effect on the number of examinations performed by the trainees from baseline to 3 months follow-up.

Methods The design is a three-armed, multicentre, blinded randomized controlled trial. The trial takes place at three simulation centres at university hospitals in Denmark; Odense University Hospital, Rigshospitalet, and Aarhus University Hospital. The intervention period is scheduled to run from August 2018 to May 2019.

Participants All physicians employed at public hospitals in Region of Southern Denmark, Capital Region of Denmark, Region Zealand, and Central Denmark Region, are eligible for inclusion in the trial. Because physicians from a wide range of specialities can benefit from TUS examinations, no exclusion criteria are established based on the specialities from the physicians.

Trainees sign up for participation, and inclusion by mail, and will receive a reply including information about the trial. Exclusion criteria are; lack of informed consent, physicians with connection to the trial, or involvement in the design or conduction.

Prior to intervention Prior to the randomization all participants will complete an online survery and educational programme in TUS in order to reach sufficient theoretical knowledge. Included participants must pass a theoretical test.

All included participants will receive a study identification number that makes it possible to pair the results from the questionnaire to the intervention and performances.

All materials needed prior to the theoretical test, including online educational material, log-in to questionnaire and test will be send to the participants by mail. The online educational material comprises access to Munksgaards' online e-portal in basic ultrasound. The theoretical part is estimated to 2-3 hours. The theoretical test is administered in Research Electronic Data Capture (REDCap) provided by Odense Patient data Explorative Network (OPEN). Participants will be excluded if they do not complete the test.

Part two of the trial will take place in one of the simulation centres. All participants will receive an introduction to the ultrasound machine. A medical student working on the project will do the introduction. Subsequently the randomization is done in REDCap, which allows an online, computer-generated allocation sequence concealed to the project leaders. The ratio is 1:1:1 with no stratification for site (location).

Trial intervention The trial intervention includes a new experimental educational approach; in vitro simulation-based TUS training. The TUS module for the US Mentor Simulator is made in collaboration with 3D Systems (3D Systems Healthcare, Littleton, USA, formerly known as Simbionix).

This simulation model will serve as first intervention arm. The group randomized for simulation-based training is allowed to practice for 2.5 hours prior to assessment. The second intervention arm in the trial is the conventional hands-on training method; examination of healthy volunteers (figurants), in this case medical students who signed up for the job, and who is a part of the research group. Trainees are as well allowed to train for up to 2.5 hours, and the medical students are not allowed to help or guide the trainee during the examinations. Last group will not receive any hands-on training other than the general information, and will serve as controls. Figure 1 present a flowchart of the trial.

When trainees have finished the hands-on training, they will continue to assessment of competencies, done by an instructor blinded to the intervention. The assessment will take place in the emergency department examining real patients suspected to have thoracic pathology/pathologies. As assessment tool the LUS-OSAUS (Lung Ultrasound - Objective Structured Assessment of Ultrasound Skills) score sheet is used. The assessment will be repeated twice. The instructor assessing the participants is blinded to the intervention. The data-managers providing the statistical analysis are going to be blinded when performing the analyses and when drawing conclusions of the results.

Sample size and statistics Significance 5%, Power 90%, mean difference wanted between the two interventional groups is 8.5 point. Standard deviation 8.67. Sample size per group = 22 (total; 66 participants).

Data will be accessed using OPEN analyse, and analysed using STATA and SPSS. OPEN Analyse acts as terminal server solution from the researcher's private PC with logging if files, but data are stored and processed on a server at the Regional IT.

A two-sided significance level of 0.05 will be used. Statistical methods for the primary and secondary outcomes are presented in Table 1, and include Post Hoc ANOVA with Bonferroni correction, hands-on training facility as independent variable (SIM, FIG, controls), and LUS-OSAUS scores as dependent variable.

All data will be analysed as intention to treat, therefore missing data will be handled by multiple imputation technique, even though missing data are expected to be minimal because instructors are registering a great amount of the data to the database.

Ethical considerations The patient scanned at the assessment will be given oral information about the trial including aim, running and assessment of participants, and that ultrasound is a non-invasive and radiation-free radiological examination with no risk of complications or side effects. Subsequent, orally informed consent will be given from the patient. If the ultrasound examination provides further information to the patient inquiry, which is suspected as new information, the physician in charge of the patient will be informed orally.

The trial complies with the Declaration of Helsinki on biomedical research and with the act on processing personal data. The Regional Committees on Health Research Ethics for Southern Denmark has been given the project description and protocol, and found that in accordance to Danish regulations, ethical approval is not required for carrying out the trial (S-20172000-44).

The trial is notified to the Danish Data Protection Agency under the in Region of Southern Denmark, and, as prescribed, a data management contract is going to be signed with authors and supervisors outside the region.

Discussion The presented trial is set to investigate whether a TUS simulator as hands-on training facility, can provide a higher level of competencies after 2.5 hour training, than training on healthy figurants, which is a commonly used model today.

Simulation-based medical education (SBME) has several advantages, and is a complex educational intervention that enables both immersive and experimental learning, and makes it possible to acquire and maintain skills in a calm and safe environment without putting patients at risk if a wrong decision or interpretation is made. The use of this approach has increased within the last decade, and in various specialities using technical procedures.

Furthermore, all pathologies considered mandatory to the content or course are possible to explore, and a trainee is able to practice a particular case or high-risk cases, over and over again, if doubt arises or if the trainee does not provide a satisfying result. Seen in a research perspective, SBME makes it possible to compare results of different trainees for research purpose, because of a standardized set-up. On the other hand, disadvantages in simulation training appear; e.g. if the fidelity drops for a short moment, the trainee may use a lot of effort to move back into the simulation setting, the technical models require updates, maintenance, and an instructor is often necessary in the beginning for introduction, and for emphasizing trainee reflection and peer review elements. Last, but not least simulation training cannot replace traditional apprenticeship or stand alone, but must be seen as and add-on approach prior to supervised training in a clinical setting.

Today in TUS, it is often a fixed number of examinations on patients that determines whether a physician can do a sufficient ultrasound examination sufficiently. Otherwise a supervisor, subjectively, accepts and approves a trainee's skills, but none of those methods ensure the professional level of competencies, and can be affected by external factors and cause feasibility problems. There will be situations where a patient is seriously ill, and rapid start of treatment is basis for a good outcome. In these situations education is not first priority, and is easily put in line.

Secondly, due to the varying incidence of various pulmonary diseases and pathologies, hands-on training on patients in a clinical setting does not ensure examination of all important sonographic findings.

To our knowledge no previous studies have been published, comparing the effects of different hands-on training facilities in TUS. Several studies have shown a positive effect of one particular modality, e.g. laboratory animals, cadavers, phantoms, or like in this trial, healthy figurants or simulators, but did not compare the effects.

Previous studies that have investigated simulation training in ultrasound of other organ systems than lungs or thorax, or ultrasound guided procedures, have showed large and significant effect when compared to no training, but in order to provide realistic and transferable comparison the control group should not be no training, but training as is the case today; training on healthy figurants. In order to implement simulation training for gaining TUS competencies, and rethink the educational tasks for physicians, the results for the simulator group are expected to be better than the group of comparison and controls.

Limitations The study will suffer from a number of limitations. Given the nature of the trial it is not possible to blind the participants, but the instructor assessing the participant is blinded to the intervention. During the analyses and reporting, the allocation will be blinded to the data-managers.

Inclusion Criteria:

• Physicians with no previous experience in thoracic ultrasound

Exclusion Criteria:

• Physicians with relation to the trial or study
• Physicians with previous experience in thoracic ultrasound
• Physicians who do not provide written consent