The Efficacy of Medical Students to Identify Pathological Lung Sound Over a Period of Time

Purpose: The primary objective is to evaluate the accuracy, sensitivity and specificity of 2nd and 3rd-year medical students' accuracy in identifying pathological sounds after 3 days of training.

Secondary Objective: To evaluate the loss of the ability of the student to correctly identify pathological lung sounds over a period of time when she/he does not undergo regular training.

Background: Auscultation is one of the four pillars of patient examination [1, 2, 3], and is particularly during a pulmonary objective investigation. Today, the assessment of auscultation sounds remains largely subjective and depends on the qualifications and expertise of a specialist [4, 5]. Arts et al. [6] recommend leaving the practice of traditional auscultation to experienced specialists since sensitivity is as low as 37% when identifying between normal cases and several lung pathologies. Furthermore, auscultation skills require continuous practice and can even degrade over time [7], but can be improved through training [2]. Hence in our research were are interested to investigate the effect of training on medical students' accuracy to identify lung sounds correct. The research is created to investigate skill deterioration over a period of time. Investigator's expectations are that the research will provide us with a better understanding of how often healthcare staff should undergo training to maintain a good standard.

Goals of the study:

To evaluate the specificity, sensitivity and accuracy of 2nd and 3rd-year medical students whilst identifying lung sounds.

To evaluate the deterioration of auscultation skills over time, to better understand how often training should be undertaken by a medical student.

Duration of the Study: The study is estimated to complete enrolment within 3 days from study initiation; however, enrolment will remain open until the study goal is met. The duration of this study for each subject will be a maximum of 7 months.

Product Description: Commercially available various earphones, mobile phones and PC will be used by the students. These products have been acquired by students before the research for personal use.

Product Intended Use: The students' mobile phones/laptops/PCs will be used to access SAITI proprietary Website (https://saititests.github.io/).

Product Acquisition: The product was acquired by students before the study with their funds and SAITI Website was developed by our research team.

Methods. Study design: A double-blind study involving 140 students and 540 lung sound recordings. The lung sounds will be used for training students and then assessing their accuracy to identify pathological sounds correctly. All students studying in LSMU in the 2nd- and 3rd-years will be invited to participate in the study.

Creation of a lung sound database for teaching and assessment purposes: a total of 84 patients \& 504 lung sounds were recorded by an Internal Medicine physician. Permission was obtained for the study from Kaunas' Regional Bioethics committee (P1-BE-2-57/2021). All participating parties signed the bioethics agreement forms. The lung sounds were recorded in the wards of the LSMU Kaunas Hospital using a 3M Littmann model 3200 electronic stethoscope. Each of the 504 recordings was 15 seconds long. Once all lung sounds were recorded, a double-blind review was initiated. During this process, 101 lung sounds were selected by a family physician. The family physician had to listen to the lung sounds and judge if she heard crackles, wheezes, and no adventitious. The additional criterion in judging was if the lung sounds were "good enough" quality for teaching purposes or not. The cases were used for training purposes if both the Internal Medicine physician and the Family physician agreed.

The students will be randomised into control (CNT) and exposed group (EXP) groups. Both of the groups will do a test after 3 days. There will be an introductory presentation for both groups about the research project. Though, only the EXP group have sections in the presentation on lung sounds. This presentation will run for about 30 - 35 min. The main topics will be to introduce the students to how the study is conducted, and what sort of pathological and none pathological sounds they can expect during the learning and assessment process. The EXP group receive access to the SAITI website with 21 questions to practice and home in their skills. Each of the 21 questions will contain one patient that was auscultated at the back at six different locations. The student will have to listen through all six channels and then she/he will be able to see the right answer. The recommendation for practice will be 90 min (per day with a 15 to 30 min break or as necessary for a student to feel comfortable. They will be allowed to spend more or less time on the SAITI website, but times will be monitored. These questions will be used to learn pathological and none pathological lung sounds.

The test will have 9 questions (6 sounds associated with pathologies and 3 normal lung sounds). A follow-up test will be conducted after 1 week and 1 month, 6 months. The exam format will be kept the same during all assessments. Students will not perform training during this period on SAITI. Though students will not discourage from participating in auscultations, they will be asked to inform investigators if they did so. This will allow the assessment of pathological sound recognition skills degradation over a period of time.

Data will be collected on an Excel spreadsheet system as the primary source of patient response collection. A dedicated hard-copy paper copy of the excel will be produced and kept as a backup.

Results will be analysed with the SPSS (version 27) and Jamovi (version 2.3.21). A p-value < 0.05 will be considered statistically significant. The groups (CNT vs EXP) will be compared with the independent Student's t-test to see if there is a significant difference between the mean of the two groups. Though, if the data does not adhere to the parametric test's criteria a Mann-Whitney test will be applied. Whilst for the measurement made over 6 months of students' sensitivity, specificity, and accuracy (at intervals of 4, 10, 34, 184 -days) a one-way analysis of variance (ANOVA) statistical test will be applied for normally distributed data. Whilst, if data is not normally distributed a none parametric test will be applied such as Kruskal - Wallis method. McNemar's test will be applied to compare the performances of the same EXP students between their second- and third-year peers.

Study population and selection criteria. All aspects of the study and consent forms will be the Lithuanian city of Kaunas regional bioethics committee approved before implementation. All participants will require full informed consent, be willing and able to comply with all study requirements and will meet the following criteria; male or female aged 18-40 (inclusive), must be LSMU students studying in 2nd and 3rd year.

Subjects will be excluded from the study based on the following criteria:

• Not a student at LSMU;
• Not studying in 2nd- and 3rd-year of medicine;
• Age over 40-year-old;
• Deaf;
• The subject does not provide full consent;
• Recruitment methods;
• Subjects will be identified and contacted through the Student Scientific Society, and further contacted via emails.

Data collection and reporting. Data will be collected at the following points: immediately after 3 days of training, after 1st -4th day, 2nd - 10 days after training, 3rd - 34 days after training and 4th - 184 days after training assessment. The assessment will have 15 questions (9 sounds associated with pathologies and 6 normal lung sounds).

Data from the study will be maintained for two (2) years after the date the investigation is completed, terminated or until the records are no longer required to support the protocol, whichever date is later.

Expected outcomes. The study has the potential to improve students' auscultation skills. By showing how often the students are trained to maintain a high level of accuracy, specificity and sensitivity in identifying pathological from none pathological lung sounds.

Adverse reactions: The is no expectation of any adverse outcomes or reactions.

Reasons for withdrawal or termination: A subject may be discontinued from the study at any time if the subject or Investigator feels that it is not in the subject's best interest to continue. The following is a list of possible reasons for study treatment discontinuation:

• Subject withdrawal of consent;
• The subject is not compliant with study procedures;
• Protocol violation requiring discontinuation;

All subjects are free to withdraw from participation at any time, for any reason, specified or unspecified, and without prejudice. Reasonable attempts will be made by the Investigator to provide a reason for subject withdrawals. The reason for the subject's withdrawal from the study will be specified in the subject's source documents and the Case Report Form (CRF).

Methods and Study Schedule: Subjects eligible for the study will review and undergo informed consent. Once consented, subjects will be randomly assigned on a 1:1 basis into two groups:

• EXP Group- where subjects will be trained for 3 days to identify pathological lung sounds before assessment/assessments. CNT group will not receive a 3-day training.
• EXP and CNT groups will additionally have a 25 - 35 min introduction session. Where students will be given an introduction to the project, auscultation sounds examples and their descriptions will be provided. Any additional questions about the project will be answered that might affect students' ability to do it.

Contact initiation with the subject (Day 0): The student will be contacted with an offer to participate in the study. In terms of subjects accepting to participate in the study and fulfilling inclusion/exclusion criteria, they will be randomized into two groups:

Control Group (CNT): 70 Students, Exposed group (EXP): 70 students.

Follow-Up Assessment: After an initial assessment, an examination will be performed at the following time points: 10 days, 34 days, and 184 days.

SAMPLE SIZE JUSTIFICATION Sample Size Calculations: The following type I error rates and decision boundaries for the study are specified: Primary outcomes are accuracy, specificity, and sensitivity measurements of the subjects.

Null hypothesis: There is no difference in accuracy, specificity, or sensitivity of CNT and EXP groups.

Type I error rate: α=0.05. Power calculations based on the pilot study's data for accuracy

The following formula was used to calculate the minimum sample size required for the actual study to find a difference in accuracy between CNT and EXP groups.

n=(S1^2+S2^2 )×(Zq1^2+Zq2^2 )^2/∆^2 where:

• S1^2 and S1^2 are variances of pilot samples
• ∆ is the smallest clinically important difference of means chosen by the researcher.
• Zq1 and Zq2 are quantiles of the standard normal distribution

Hence, in our study variance was 0.8258-0.7250= 0.1008 between the groups. The delta is taken at 0.05. Whilst normal distribution quantities are Zq1= 1.96 for the two-sided alternative and Zq2=20.8416 when f α=0.05 and β=0.2.

The calculations give us an answer of n≈32. During the pilot investigation, of 31 students that were invited 17 finished the study. That is ~54 %. Therefore, an assumption is made for the SAITI-T study, that attrition rates will be around 50 %. Therefore, the study needs 64 to 70 students in the enrolment stage for each group (140 in total).

Randomization: Subjects who meet all inclusion and exclusion criteria will be randomised into two groups. The subjects will be split into 1:1 ratio to either the exposed group (EXP) or the control group CNT.

ASSESSMENT OF SAFETY. Purpose: The primary objective is to evaluate the accuracy, sensitivity and specificity of 2nd- and 3rd-year medical students' accuracy in identifying pathological sounds after 3 days of training.