Lecture based versus peer assisted learning : quasi-experimental study to compare knowledge gain of fourth year medical students in community health and nutrition course

1Department of Community Medicine, Lahore Medical & Dental College, University of Health Sciences, Lahore, Pakistan 2Department of Medical Education and Educational Research, Lahore Medical & Dental College, University of Health Sciences, Lahore, Pakistan 3Department of Surgery, Lahore Medical & Dental College, University of Health Sciences, Lahore, Pakistan 4Faculty of Health Sciences, Aga Khan University, Karachi, Pakistan


Introduction
2][3][4] Medical curricula are now being modified relevant to the competencies required of medical graduates.This entails a paradigm shift from the established teacher focused instructional approaches to more student-centered learning activities 5 and making a judicious selection of learning tools. 6The strategic choice of teaching learning activities plays a major role both in the perceived satisfaction and also in the knowledge gained by medical students. 7,8mong the instructional tools identified for effective information transfer to learners, lectures and peer assisted learning (PAL) are two diverse educational approaches, each supporting different learning needs. 91][12] Specifically, lecture is easy to arrange, 10 can effectively convey information to a large audience, 9,[13][14][15] have the potential to simplify complex concepts and produce the desired results in terms of academic achievements. 16he popular critique of lectures include disengaged and passive audience 14,15,17 and inability to develop active learning skills in students. 13In contrast to expert-led coaching, PAL is a dynamic learning stratagem, where students learn in collaboration with each other rather than taking the direct assistance of an instructor. 181][22] Peer teaching is also a valuable resource to meet the increasing number of medical students. 21,23The main shortcoming of PAL is the uncertainty about the quality of the educational process. 18,20,24,256][27][28][29][30] Nevertheless, the objective learning benefits of this approach have yet to be fully established in different learning environments.There is insufficient empirical evidence which can assess the knowledge gained by medical students through lectures or PAL.In studies, where peer assisted and expert assisted learning were compared in terms of test scores, a consensus on the superiority of one method on or the other was never reached.Some studies found PAL to be better than lectures, 8,27 others concluded that both strategies produced similar academic outcomes [31][32][33] while there is also evidence that PAL groups score less than those taught in a didactic fashion. 34,35In Pakistan, the gap in knowledge about the effectiveness of PAL as a learning tool is very wide.This void results because there are very few medical schools in the Pakistan which have incorporated PAL in their teaching learning program or have researched its efficacy in enhancing knowledge gained by students . 9,22,33ahore Medical & Dental College (LMDC) was founded in 1999 as a private sector institution, which follows a traditional discipline-based curriculum.Following the trend of adopting new innovative learning strategies, Department of Community Medicine at LMDC introduced PAL program in the fourth year MBBS class.The selected PAL model was the same level or same class, with equal status of learners.In this model, all participants were learners and teachers at the same time.PAL was incorporated as an adjunct to the traditional teaching methods, including lectures and tutorials.The present study was conducted to assess the effectiveness of PAL in comparison with the lecture, via the end of course test scores.The purpose of the comparison was to generate evidence for making the decision about the adoption of PAL as a complimentary learning aid to lectures within the traditional medical syllabus.

Materials and Methods
A quasi-experimental study was carried out at the Department of Community Medicine, LMDC, Lahore in 2014, spread over a span of 21 weeks.

Study design
The study design used was single centre, parallel group, randomized control trial (RCT).

Inclusion and Exclusion criteria
The inclusion criterion for participation in the study was medical students with exposure to both lectures and PAL.The exclusion criterion was students who have never participated in PAL.At LMDC, the inclusion criterion was only fulfilled by the fourth year MBBS class in the subject of Community Medicine.A convenience sampling technique was used and the entire fourth year MBBS class of 2014 (125 students) shaped the sampling frame.The objectives and the methodology of the study were explained to the class.The students who agreed to take part in the study were considered the study sample (N = 99).

Randomization
A simple random allocation of 99 students was conducted for the lecture group (control group; n = 50) or PAL group (intervention group; n = 49).In order to minimize selection bias, the allocation concealment (Masking) technique was used in the allocation of participants to the two groups.The PAL students were further divided into five smaller clusters, with ten learners in four subgroups and nine learners in the fifth subgroup.The 26 students who did not consent to participate in the study were allowed to attend the lectures and these students were not included in the analysis of the study results (Figure 1).

Blinding
The participants were not blinded to their allocation but blinding of analyst was ensured until the whole data was analyzed.

Intervention
Community Health and Nutrition was the course chosen for the study.Both lecture and the PAL sessions were held simultaneously and the length and content covered in each session were the same for both groups.The Community Health and Nutrition course was divided into eight topics and the whole course was completed in eight, two hour sessions.The learners in both the groups were handed their personal Community Health and Nutrition workbook 30 minutes before the session ended, which was specially designed as an aid to re-examine and summarize the learning that takes place in the course.The workbook was also divided into eight sections, with each section corresponding to the course covered in one session.This exercise was an open book activity and the workbooks were gathered at the end of each session, to be distributed again at the beginning of the next session.The lecture group followed the below mentioned teacher-led structured program.In one session, each topic was covered by two lectures, conducted in a lecture theatre.Structure of each two hourly session followed by the Lecture group: Lecture 1, 40 minutes; Break, 15 minutes; Lecture 2, 40 minutes; Workbook exercise, 25 minutes.In contrast, each PAL subgroup planned its own learning strategy, which was also spread over a 2-hour period, with workbook activity in the last 25 minutes.Each subgroup was allocated a separate room and monitored by a faculty member.The monitor did not participate in the PAL session but made sure students were working within their allocated groups.The concurrent sessions, along with the presence of a monitor and collection of workbooks before adjourning the sessions, prevented the Lecture and PAL groups from mixing and sharing information for that particular session.

Pre-test
Learners' baseline knowledge was assessed at the beginning of the course via a class test (assessment tool) of 100 marks, consisting of Multiple Choice Questions (MCQs), Short Answer Questions (SAQs) and Objectively Structured Practical Examination (OSPE).The marks allocated to the various components of the assessment were: MCQs = 20, SAQs = 40 and OSPE = 40.

Post-test
Learner knowledge gained by the end of the course was appraised by the same assessment tool that was administered as pre-test at the beginning of the course.

Data analysis
All analyses were performed according to the original assigned groups (intention-to-treat).The assessment scores of the pre-test and the post-test comprised the data.Data was analyzed using SPSS 20.Variables studied were gender and high school qualifications (independent variables) and test scores (dependent variable).Descriptive statistics was used to present: 1. Test scores as mean and standard deviation (scores of each assessment type i.e.MCQs, SAQs, OSPE and the total test scores).2. Demographic variables (gender and entry qualifica-tions) as number and percentage.Prior to data analysis, data on test scores were checked for assumptions of homogeneity of variance and normality using Levene's and Shapiro-Wilk tests.Independent t test was used to analyze the mean difference of pre-test scores between the Lecture and PAL groups and then again to assess the difference in the post-test scores in the two groups.Paired t test was used to check the difference between pre-test and post-test scores in the Lecture and PAL groups.A one-way analysis of covariance (ANCOVA) was also conducted for this study.The independent variable was educational strategies (Lecture and PAL), the dependent variable was post-test scores and the covariate was the pre-test scores.Chi-square test was applied to assess the association between gender and high school qualifications with test scores in the Lecture and PAL groups.Cronbach α was applied to assess the reliability of the assessment tool.A P value <0.05 was considered the cut off point for statistical significance.

Results
The response rate in the present study was 79%.Among 99 study participants, 36 (36%) were male and 63 (64%) were female.The high school qualification of 77 (78%) participants was FSc., 19 (19%) completed GCSE A levels and only 3(3%) had completed American Board.The lecture group had 47% males, 52% females, 47% students who completed FSc., 63% who had A levels as high school qualifications.In the PAL group there were 53% males, 48% females, 53% with FSc.qualifications and only 37% of those who had done A level (Table 1).The reliability coefficient of the end of course test scores (assessment tool) was calculated using Cronbach α, which was 0.81.

Assessment (test) scores in lecture and PAL groups
Test scores met the underlying assumption of homogeneity of variance as evidenced by F (1, 97) = 2.437, P = 0.122 and were normally distributed at 5% level of significance (W= 0.987, P = 0.477 > 0.05 = α).Table 2 shows no statistically significant difference between the pre-test scores of lecture and PAL groups.As depicted in Table 3, there was a highly statistically significant difference in the pre-test and post-test scores obtained by study participants in both the lecture and PAL cohorts (P ≤ 0.001).
After applying ANCOVA, a preliminary analysis evaluating the homogeneity of regression (slopes) demonstrated no significantly different relationship between the covariate and the dependent variable as a function of independent variable, F (1, 95) = 0.584, P = 0.447.The ANCOVA was also not significant, F (1, 96) = 0.055, P = 0.814.There- fore it was concluded that there was no statistically significant difference in the post-test scores between the Lecture and PAL groups after controlling the pre-test (Table 4).

Effect of gender and high school qualifications on the test scores of the study participants
As seen in Table 5, gender and high school qualifications had no bearing on post-test scores of learners either in the Lecture or PAL groups.

Discussion
Healthcare education requires a blend of traditional and contemporary teaching and learning practices that can enhance student competencies to align with the expected requirements and anticipations.One of the major criteria for choosing learning tools is their ability to produce positive academic results.The current study compared two different approaches, PAL and Lecture, but could not establish any significant superiority of one method on the other based on post-test scores (P = 0.81).However, the study results demonstrated that in terms of academic achievements, the collaborative strategy was equally effective to any established traditional method in knowledge assimilation and its application.All 99 study participants performed at the same level in the cognitive (MCQs and SAQs) and practical (OSPE) aspects of the test.
The comparable post-test analogy in medical education, with no difference in educational achievements in peerled and faculty-led tutoring, has also been described in previous studies.Bentley and Hill 31 compared test scores of students tutored through teacher based versus peer supported learning in an Anatomy course.Their study  27 assessed student learning in a Gastroenterology/Hematology course, where some of their study participants were in the PAL sessions and the rest were receiving didactic lessons.Learning outcomes were evaluated through MCQs.In this study, the PAL group's marks were notably better than the students taught through didactic method (P ≤ 0.01).
Poor performance in PAL cohorts was also reported by studies conducted by Knobe et al 4 and Walsh et al, 35 who noticed that students taught by peers obtained significantly lower marks than those taught by expert teachers in complex skills.
In the current study, gender and educational background had no bearing on the test scores.No research was available for the comparison of this finding.
Limitations of the study included small sample size, sampling bias, restricted generalizability and limited external validity.In the present study, no method was used to balance randomization.

Conclusion
The present study concludes that Lecture and PAL are both effective strategies for academic achievement.Students who were tutored by other students performed equally well compared with their peers in an expert led program.Therefore, it can be concluded that PAL can easily be incorporated in LMDC as a supplement to lectures in the evolving medical school curricula.It is suggested that more multi-centre, multi-subject research should be conducted, with larger sample sizes to obtain more reliable evidence on the suitability of PAL as an academic tool in medical education.

Table 1 .
Demographic characteristics of study participants in lecture and PAL groups (N = 99)

Table 2 .
Mean comparison of assessment scores and t test results in lecture and PAL groups (N = 99)

Table 3 .
Paired t test results for comparison of pre-test and post-test scores in lecture and PAL groups (N = 99)

Table 4 .
ANCOVA for assessment scores and educational strategies in lecture and PAL groups (N = 99)

Table 5 .
83i-square test results for effect of gender and high school qualifications on post-test scores of students in Lecture and PAL groups (N = 99) Manzoor 32 checked the effectiveness of lectures and PAL in the Community Medicine course on Prevention of Non-Communicable Diseases, in fourth MBBS class.Her results showed no significant difference between the two study groups on the basis of academic scores obtained in MCQ tests (P = 0.47).Ten Cate33conducted a long-term analysis of academic outcomes of test scores between peer instructed and the teacher taught groups in multi-disciplinary courses like Metabolism or Circulation.His results also reinforced that peer learning could not produce better results than teacher led courses, but at least it did not negatively affect academic achievement.In contrast to our findings, some studies established that students undergoing PAL produced better academic results than those who were in faculty-led teaching programs.Abedini et al8tested students in a Pharmacology course using MCQs.Their results indicated post-test marks obtained by PAL cohort was significantly higher than the lecture group (P ≤ 0.02).Similarly, Peets et al