Medical students as teachers: impact of peer-led teaching sessions on NBME exam performance

Abstract

Introduction/Background

Peer-led teaching sessions (PLTS) supplement conventional medical education, allowing students to share knowledge effectively. PLTS have shown significant academic benefits for student instructors. However, the impact on attendees, especially concerning NBME-style examinations, remains underexplored.

Methods

This study analyzed de-identified data from 67 first-year medical students at Kirk Kerkorian School of Medicine, divided into PLTS attendees (study group) and non-attendees (control group) based on session participation. Data were collected before two NBME examinations, with attendees defined as those attending at least five Peer-led teaching sessions. Academic performance was measured using NBME Customized Assessment Services (CAS) exam scores. Statistical analyses, including p-tests, chi-square, and t-tests, assessed differences in performance and potential biases based on MCAT scores, first-generation status, and socioeconomic background.

Results

The study group initially had 26 students, and the control group had 31 students. Post-first examination, the study group adjusted to 22 students, and the control group to 35 students. MCAT scores and demographics showed no significant differences between groups. The first examination revealed a significant higher score in PLTS attendees’ scores (4.28% increase, p = 0.036), indicating a positive impact of PLTS on basic science knowledge. The second examination on organ systems physiology showed no significant difference (p = 0.722).

Discussion

PLTS effectively enhance foundational science understanding but may be less impactful for advanced topics. The diminishing effect may result from increased use of diverse study resources and adaptation to medical school. Future research should explore long-term benefits, optimal implementation strategies, and the impact of instructor quality and engagement. Therefore, the overall conclusion is that PLTS positively influence NBME-style exam performance in early medical education, but benefits decrease over time.

Introduction/Background

Peer-led teaching sessions (PLTS) are used in medical teaching institutions to supplement conventional teaching methods. In these sessions, students that accomplished certain milestones in their education teach other students in the same program. Though not a groundbreaking innovation in graduate education, PLTS leverages the principle of allowing students to share knowledge with their peers, knowledge they have garnered from the same curriculum or program. Typically, these student instructors are upperclassmen or individuals who have attained certain curriculum milestones, empowering them to teach others. This pedagogical approach traces its roots back to the 1800s [1]. Reflecting on the Latin roots of the word “Doctor,” derived from “docere,” meaning to teach, it’s inherently fitting that medical educators have long embraced this etymological principle through implementing PLTS. In fact, the method has been utilized in healthcare for decades. The most recent meta-analysis dates some studies being done on this topic as early as in the 1990’s [2]. Since its introduction in medical education, PLTS has spurred various scientific inquiries, particularly regarding its benefits for both the student instructors and their peers.

The findings of several studies showed a definite positive effect for students who act as mentors for their peers. One study in particular, Wong et al., looked at performance between the students acting as teachers and the control group. The study showed that the educator group had significantly higher USMLE Step 1 and Step 2 scores and medical school Grade Point Average (GPA) [3]. The concept behind such a benefit likely stems from the concept of tutors re-learning the material and understanding the information they relay to their peers. Meta-analysis consisting of 23 studies done by Ribosa et al. concluded that students can learn and academically improve simply by creating learning material for others [4]. It is not surprising then how teaching became an important skill for medical students. In fact, the study by Soriano et al. highlighted institutional support for peer teaching, revealing that 99% of surveyed US medical schools incorporate student teaching responsibilities into their programs [5]. These findings suggest that the act of teaching not only reinforces the educators’ knowledge and skills but also garners broad institutional endorsement. However, there is also evidence behind the benefit for the students in this method of education.

The complexity and breadth of the medical school curriculum pose significant challenges for students. Peer-led sessions offer a unique avenue for students to gain insights into managing academic demands and engaging deeply with the content. A recent systematic review and meta-analysis in 2021, involving 21 randomized controlled trials, showed significant improvement in participants compared to traditional study methods. However, the majority of the studies involved clinical skills knowledge rather than basic science concepts [2]. Examples of the studies included are Ong et al. and Tolsgraad et al. both of which found that students performed better in clinical skills such as suturing and catheterization after tutoring sessions by peers (6–7). Additional study by Trevino et al. in 1978, concluded that there appeared to be a benefit for students engaging in peer tutorial sessions. However, the study used a crossover model by comparing the performance of students receiving the tutoring to their prior scores on the same subjects [8]. Additionally, due to the age of the study, 1978, their examination was likely professor created, which carries a significant difference and associated confoundings compared to NBME style questions.

Our literature review identified a gap in knowledge specifically evaluating the impact of PLTS on NBME-style examination scores of students attending these sessions compared to students who did not. While previous studies mainly focused on the impact of PTLS on presenters, we focused our study on the impact of these sessions on attendees. This study aims to fill that gap by quantitatively evaluating the correlation between participation in peer-led sessions and improvement in NBME-style multiple-choice exam scores among first-year MD students of different content areas.

Methods

Academic performance & data collection

All data used in this study were extracted from deidentified program evaluation databases, which included exam scores and High-Yield Review session attendance records. At Kirk Kerkorian School of Medicine, High-Yield Review sessions, equivalent to PLTS, involve second-year students teaching high-yield content on specific topics. The databases were utilized in compliance with ethical standards for educational research. Prior to data collection, the study protocol was reviewed by the Institutional Review Board (IRB), which determined that the use of deidentified data from existing educational records did not constitute human subjects research requiring full IRB review.

To gauge academic performance, scores from regular curriculum-based examinations were utilized. At Kirk Kerkorian School of Medicine, these exams are all MCQ using NBME CAS (Customized Assessment Services).

In addition to assessing academic outcomes, the study examined potential self-selection biases between the attendees’ and non-attendees’ groups. This analysis included a comparison of Medical College Admission Test (MCAT) scores to gauge pre-existing academic disparities. Additionally, we compared the proportion of first-generation college students and those from low socioeconomic backgrounds within each group.

Study population & inclusion/exclusion criteria

A first-year cohort of 67 students was divided into a study group of PLTS attendees and a control group of PLTS non-attendees based on their voluntary attendance. Data were collected for sessions prior to two NBME examinations. The study group for each exam consisted of students who chose to attend a minimum of five PLTS before the specific examination, while the non-attendees-group included students who did not attend any sessions before the NBME examination.

In the sessions before the first examination, there were 26 students in the study group and 31 in the control group. In sessions before the second exam, the study group had 22 students, and the control group had 35. The total cohort consisted of 67 students; however, the reason the numbers for both groups do not sum to 67 is due to our inclusion criteria. Specifically, students in the study group had to attend at least five Peer-Led Teaching Sessions (PLTS), while the control group consisted of students who did not attend any sessions. There were 10 students who attended between 1 and 4 sessions before each of the examinations, and thus, they did not meet the criteria for either group, which resulted in their exclusion from both the study and control groups for both examinations.

This readjustment in the number of participants in the study/control group between the first and second exams is due to the distinct content covered in the two exams. The first exam focused on basic science concepts, which are generally reflective of college or prerequisite material. In this context, we believe that students attending PLTS sessions tailored for the first exam did not gain any advantage for performance on the second exam, which focused strictly on organ system physiology. The two sets of content are distinct, with the first being more foundational and the second requiring a deeper understanding of medical physiology.

In this study, an intention-to-treat analysis was not utilized because keeping students who only attended sessions for the first exam, but not for the second, would introduce bias into the study. These students may not have performed well on the second exam simply because they did not attend sessions specifically tailored for that content. Including them in the study group for the second exam would likely have skewed the results and reduced the accuracy of the analysis, as their lack of attendance for sessions related to the second exam could negatively affect their performance, thereby artificially lowering the average score for the study group.

Instructors

In this study, only two instructors were consistently responsible for all teaching sessions. They collaboratively created the presentations and teaching materials, ensuring uniformity in content delivery. Both instructors were second-year medical students who had recently completed the material they were teaching and were familiar with it from their first year of medical school. They were high-achieving students, scoring in the upper 25th percentile on relevant exams.

Data analysis

Statistical analysis employed a variety of tests to assess the significance of the differences observed between the attendees’ and non-attendees’ groups. All statistical tests reported 2-tailed p-values with alpha set at 0.05. Chi-square tests were used to compare proportions, while t-tests were employed to compare means. Both chi-square and t-tests were also used to analyze pre-existing differences between the groups.

Results

Demographics of the groups

The analysis of MCAT scores revealed a marginal difference of 1 to 2 points between the two groups, a disparity that did not reach statistical significance, with a p-value of 0.158 and 0.431, for the two examinations respectively. This suggests that MCAT scores, and by extension prior academic preparedness, were comparable between the groups and unlikely to account for differences in the impact of PLTS sessions on exam performance.

Similarly, when examining the proportion of first-generation students and those from low socioeconomic backgrounds, the analysis did not uncover statistically significant differences between the study and control groups. The first examination cohort had the Chi-Square test p-values of 0.990 and 0.520 for first-generation status and socioeconomic status, respectively. The second examination cohort had the Chi-Square test p-values of 0.579 and 0.486 for first-generation status and socioeconomic status, respectively. Overall, this data suggests a homogeneous distribution of these variables across both groups and across the two examinations.

These findings, summarized in Table 1, underscore that the composition of the study and control groups was well-balanced in terms of MCAT scores, first-generation college student status, and socioeconomic background.

Table 1.

Table summarizing the demographics, MCAT scores, and background (socioeconomic Status & First-Generation College student) of included population

Exam 1				Exam 2
	Study (n = 26)	Control (n = 31)	Statistical Significance	Study (n = 22)	Control (n = 35)	Statistical Significance
First Generation (n)	30.7%	29.0%	χ²=0.020 p = 0.990	27.3%	34.3%	χ²=0.307 p = 0.579
Low Socioeconomic (n)	34.6%	48.4%	χ²=1.308 p = 0.520	36.4%	45.7%	χ²=0.485 p = 0.486
MCAT Mean Score	507	509	t = 1.433 p = 0.158	508	509	t = 0.793 p = 0.431

Exam scores

The analysis uncovered a notable difference in performance on the first examination, which concentrated on basic sciences. The PLTS attendees demonstrated an average higher score of 4.28% compared to the control group, a result that was statistically significant with a p-value of 0.036. This indicates a meaningful impact of the PLTS on students’ grasp of basic science concepts.

Contrastingly, the second examination, which was centered on organ systems physiology, did not exhibit a statistically significant difference in scores between the two groups, as indicated by a p-value of 0.722. This suggests that the advantage observed in the study group during the first exam did not extend to topics covered in the second exam.

These findings are visually presented in Fig. 1, which compares the performance of two groups across the two examinations. Additionally, Fig. 1 subscript summarizes the statistical outcomes for both exams, providing a clear and concise overview of the data analysis.

Fig. 1.

Mean Performance between PLTS attendees (Study) and PLTS non-attendees (Control) on Exam #1 (Basic Sciences & Foundation) and Exam #2 (Physiology of Organ Systems)

Discussion

Our findings provide compelling evidence that PLTS, specifically High-Yield Review (HYR) in the presented case, are significantly correlated with enhanced academic performance in foundational science subjects of medical curriculum. This correlation underscores the efficacy of PLTS in reinforcing students’ understanding and triaging of fundamental scientific concepts, which are critical in the early stages of their medical training. However, as the curriculum progresses towards more advanced and standardized material, such as organ systems physiology, the effectiveness of PLTS in boosting academic performance appears to wane. This trend could be attributed to several factors, one of which is the increased reliance on a variety of external resources by students, including textbooks, online platforms, and study groups, which may dilute the relative impact of PLTS. This assumption is not unfounded. Study done by Burk-Rafel et al. showed that only ∼ 18% of students relied on school material in preparation for USMLE content [9]. Similarly, the 2022 Association of American Medical Colleges (AAMC) Year 2 survey reported that up to 70% use third-party online resources at least weekly [10]. In fact, recent qualitative study by Lawrence found that students believe that medical education should formally adopt some of the third-party resources as teaching resources [11].

Another contributing factor to the trends observed in this study may stem from the behavioral adjustments of students at the onset of their medical education. Typically, the initial months of medical school involve a period of adaptation to the rigorous pace and volume of exams and content review expected of students. As they acclimate to studying in this demanding environment, the reliance on and perceived benefit of peer tutoring may diminish. This suggests that the initial utility of peer tutoring is potentially most impactful during the early transition phase into medical school, with its effectiveness possibly waning as students develop and refine their own study strategies and methods.

Regardless of the reason, the observed reduction in the impact of PLTS on academic performance in the second exam, which focused on organ systems physiology, highlights an important consideration for medical educators: the strategic implementation of PLTS in the medical curriculum. Given their indicated benefit in foundational science subjects, incorporating PLTS predominantly in the early phases of medical education could optimize learning outcomes and foster a deeper comprehension of complex scientific principles. This targeted approach would not only capitalize on the strengths of PLTS but also allow for the efficient allocation of educational resources and time.

Moreover, the observed shift in the efficacy of PLTS from foundational to advanced courses prompts a broader reflection on the adaptability of teaching methodologies in medical education. As students advance and as the content becomes more clinically oriented, educators may need to explore alternative or supplementary methods to PLTS that are better suited to the nature of the material and the evolving learning needs of students.

The findings of our study align with previous research in the field of PLTS that has predominantly presented qualitative data, focusing on the performance of the teachers or mentors themselves on comprehensive exams [4, 12]. However, this study provides quantitative evidence that medical students achieved significantly higher scores on their first basic science exam when they attended peer-led tutoring sessions. Examining student performance on a short-term basis, from one exam to another, can also assist students in recognizing the positive impact that peer-led teaching can have on their education.

Limitations

This study encounters several limitations that are important to acknowledge. First is the small sample size of 67 students and data from only one institution, which calls for further research with larger cohorts. Larger sample size will enhance the reliability and applicability of the findings. Additionally, attendance at PLTS does not inherently equate to active engagement; future studies could benefit from assessing individual attentiveness to better quantify the sessions’ effectiveness.

A second limitation includes the heterogeneity of study materials. The variation in content and difficulty across different PLTS may impact their effectiveness, underscoring the need for more detailed analysis of session content in future studies. As the same two instructors taught both the foundational and advanced sessions, we expect minimal differences in the quality of instruction or content coverage between these topics. The instructors’ consistent collaboration in developing materials further ensured alignment in instructional methods and delivery. Their exam performance and familiarity with the curriculum suggest that their expertise was similar across foundational and advanced content. However, no formal evaluation of teaching quality was conducted, and future studies can explore this limitation.

A third and crucial factor is instructor variability; the quality and teaching styles of peer instructors, which were not evaluated in this study, could significantly influence learning outcomes. This suggests a need for research into the qualifications and pedagogical approaches of peer instructors.

Additionally, while this study accounts for self-selection bias via comparison of the two groups, randomized control trial is a gold-standard for limiting such bias. Specifically, the voluntary nature of PLTS participation may introduce self-selection bias, with more motivated students possibly being overrepresented in the study group. Moreover, this study assumes that performance on the first exam did not significantly influence students’ performance on the second exam, and that switching study/control group membership between exams did not affect the outcomes. However, the change in group composition between the two exams—where some students attended Peer-Led Teaching Sessions (PLTS) before the first exam but not the second—could introduce variability in the results. In an ideal scenario, performance on the two exams should be analyzed with consistent group membership across both exams.

Lastly, this study did not account for students’ use of external resources, such as online platforms or study aids, which could influence the measured effectiveness of PLTS. Future research should consider these variables to isolate the unique contributions of PLTS to student learning outcomes. Addressing these limitations in subsequent studies will refine our understanding of PLTS’s utility in medical education and inform the development of optimized educational strategies.

Future directions

Future studies could explore the long-term benefits of established peer-review teaching programs, determining if students who attend do indeed score higher on USMLE exams or match into highly competitive residency programs. Determining what types of residency programs peer tutors match into could provide insight into the weight of peer teaching in the long term for tutors. Additionally, future studies in the fields of peer-led teaching could explore the demographics of students who become peer-led teachers, examining whether there are any benefits to previous teaching experience, and academic standing. Finally, it is important for future studies to explore the broader implications of PLTS across various measures of academic performance beyond NBME exam scores such long-term retention of knowledge.

For other medical schools seeking to implement successful PLTS, several recommendations could be drawn from our study. Early implementation of PLTS in medical education could be helpful for students adjusting their study routines. It could also help students cope with the newly found challenges of medical school. In fact, a study done at University of California, San Francisco (UCSF) School of Medicine showed that first year students especially found this aspect of PLTS helpful. The ability to interact with students who went through the same challenges as them serves as a therapeutic tool for students, which further encourages early implementation of these sessions [13]. Additionally, a robust yet qualified group of peer tutors could enable more students to seek tutoring.

Medical schools can leverage data collected from their own peer teaching programs, as well as those of other institutions, to demonstrate the benefits it offers to students’ education. Given the dense nature of the medical school curriculum, peer-led sessions aid students in navigating this challenging transition into medical education and beyond. In addition to its educational benefits, the connections formed between mentors and peers are valuable for all parties involved.

Conclusion

This study displayed evidence that peer-led teaching can have a positive impact on students’ performance on NBME style examinations, especially in the context of basic science knowledge at the start of the program. However, the improvement appears to fade with progress in medical school. Therefore, due to the limited sample size and data, more studies should be conducted to revisit this topic.

Abbreviations

PLTS

Peer-Led Teaching Sessions

NBME

National Board of Medical Examiners

MCAT

Medical College Admission Test

USMLE

United States Medical Licensing Examination

MD

Medical Degree

Author contributions

KA and VW co-created the study idea. KA wrote the main manuscript text, performed the literature search and review, created the tables, figures, and abstracts, and carried out the revisions. VW assisted with the revisions. TS contributed to the manuscript revisions. ES conducted all the statistical analyses, provided the data, and offered feedback for the revisions. All authors reviewed the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets generated and analyzed from de-identified institutional academic records of students are available from the corresponding author on reasonable request.

Declarations

Informed consent

As the study involved de-identified data from existing educational records, informed consent was not required. The study protocol was reviewed and approved by the Institutional Review Board (IRB) of the Kirk Kerkorian School of Medicine, which determined that the use of de-identified data did not constitute human subjects research requiring full IRB review.

Competing interests

The authors declare no competing interests.