A Cross-sectional Study Investigating Learning Approaches in Undergraduate Medical Education

Objective The primary objective of this proof-of-concept cross-sectional study was to identify a framework for appraising the learning-approaches of undergraduate medical students in a competency based medical curriculum and correlating the results with teaching-approaches, as well as academic performance. The study was pursued at MBRU, which is a medical school in the Middle East with an undergraduate entry medical program. Results Our framework was blueprinted using the Approaches and Study Skills Inventory for Students (ASSIST) questionnaire, to which we made some modifications such that the overall cogency of the questionnaire wasn’t affected. Initial results with modified ASSIST at MBRU showed that most of our students adopted Deep or Strategic-learning approaches. This observation is in line with other studies in the literature, which shows that modified ASSIST is a suitable tool for mapping generic learning approaches with teaching approaches. Further, based on the insights from our initial results following the implementation of modified ASSIST, we have considered specific pedagogical strategies, in practice at MBRU, which cater to the generic learning approaches of majority of our undergraduate medical students. These pedagogical approaches, A. Feynman’s Technique; and B. Blended learning strategies, if implemented suitably in a curriculum will transform “Surface-learners” to “Deep/Strategic-learners”.

competency-based learner-centred approaches to teaching to spur active learning [3]. The effectiveness of such learner-centred pedagogical strategies will be augmented if they are fittingly mapped, with generic learning approaches adopted by medical students in a given cohort. This implementation is pivotal for medical schools in the Middle East, which admit students from diverse high-school curricula, based on matriculation scores, and experience high dropouts [4]. This proof-ofconcept study is aimed at identifying a framework for appraising the learning approaches of undergraduate medical students in a competency based medical curriculum and correlating the results with the availed teaching-approaches as well as academic performance.

STUDY LANDSCAPE
The study was conducted at Mohammed Bin Rashid University of Medicine and Health Sciences (MBRU). MBRU is a new medical school with an undergraduate entry medical program, in which the curriculum is founded on a competency-based model and spans six-years. The curriculum is divided into 3 phases (Figure: 1) [1].
Each phase of the curriculum includes integrated courses and builds on the preceding one, such that the curriculum is "spiral" (Figure: 1) [5]. The school has a diverse student population, drawing students from more than 19 countries across the globe.

Ethical Considerations
The study was approved MBRU-Institutional Review Board (MBRU-IRB) (Approval No. MBRU-IRB-SRP2018-048). Participation was voluntary, with a declaration of informed consent. The study spanned between January and September of 2018.

Participants
The study population comprised 84 undergraduate medical students (belonging to Year-2 or 3) enrolled in the MBBS-program at MBRU. A purposive sampling method was applied [6], whereby candidates were eligible to participate based on their status as an undergraduate medical student in MBRU. Year-1 students were excluded from the study, as the short-period of time in medical school was insufficient to assess their learning approach.

Evaluation of Learning and Teaching Approaches
Our study employed the Approaches and Study Skills Inventory for Students (ASSIST) questionnaire to evaluate the predominant learning and teaching approaches, among the participants (Table: 1). This questionnaire was developed by Entwistle [7], for university students' conceptualizations of learning, approaches to studying, and preferences for different types of instructional methods.
Available research on ASSIST has identified three primary approaches to studying: Deep, Strategic, and Surface ( Figure: 2). Construct validity has been supported by several studies which have linked approach to learning to academic performance [8]. Deep and Strategic approaches usually are related to greater success, while Surface approaches may put students at risk for poor academic performance. As a result, curriculum design and pedagogy that improve Deep and Strategic approaches may be beneficial to improving performance outcomes.
In the present study we effected a modified version of the ASSIST questionnaire.
The initially published ASSIST questionnaire consists of 3 different sections designated A, B, and C. All three sections consist of items with a Likert scale rating for response. As shown in (Table: 2), in the proposed research the 66 items have been reduced to 41 items (Sections B and C). We modified the questionnaire to include a smaller number of questions, while maintaining an equal number of questions across the learning approaches. In the modified ASSIST, an additional 4 items were added. These items recorded the demographics of the participants, i.e. age, sex, year of study in the MBBS program, and high-school education of the participant.
The last 8 items of the modified ASSIST enquired about preferred teaching approach. They measured whether students preferred a deep or a surface teaching approach.

Study Variables
The key aim of this study was to identify the predominant learning approaches among undergraduate medical students. We also investigated the correlations between different teaching and learning approaches, as well as, correlations between learning approaches and different demographic factors such as age, gender, year of study and high-school curriculum. Additionally, the correlations between learning approaches and perceived academic performance were also investigated. All of these variables were collected and tabulated as questionnaire results in an encrypted spreadsheet. All variables were numerical continuous variables.

Statistical Analyses
The questionnaire response file was converted to a spreadsheet and the questions for each approach were reorganized into adjacent columns, with the value of each response in the respective row. The average-score for each approach was calculated for each participant by taking the average of all responses recorded for a certain approach. The predominant learning approach for each participant was chosen as the highest score per learning approach. Descriptive statistics were utilized to describe the demographics of the study population. In order to identify whether a significant difference existed between the mean scores for predominant learning approaches, paired t-tests [9], were used.
The Spearman correlation coefficient [10], was calculated for Likert-scale variables to compare if students prefer the teaching approach that matches their learning approach. This was done by analysing correlations between learning approach and preferred teaching approach, as well as a correlation between learning approach and perceived academic performance. The Spearman correlation analysis was conducted utilizing the average scores for each learning and teaching approach and scale scores for perceived academic performance as data values. Ordered logistic regression was used to analyse the effect of predominant learning approach on perception of academic performance, controlling for age, gender, and cohort. The level of significance (p-value) considered was 0.05. Data was analysed using STATA, version 15.1. [11].

Participants
Sixty-four (76%) responses were received. Details of exclusion at various stages of data analysis are shown in Figure: 3.

Learning Approaches
The distribution of the predominant learning approaches among the 60 students is shown in Figure: 4A.

Teaching Approaches
The results for preferred teaching approaches are shown in Figure: 4B. Five students were excluded from the analysis, as they ranked equally for both Deepteaching and Surface-teaching approaches. Of the 55 students, majority preferred Surface-teaching Approach Figure: 4B.
Next, we correlated the students' preferred teaching approaches with learning approaches (Figure: 4C). Here, Question 38 in ASSIST (Table: 1) concerning the surface-teaching approach, assessing information delivery/dissemination, had the highest score (s = 4.5). On the other hand, Question 39 in ASSIST (Table: 1) pertaining to the deep-learning approach, appraising learner understanding, had the lowest score (s = 2.9).

Spearman Correlation
A Spearman correlation was conducted to determine if students' preferred teachingapproaches match their learning-approach (Table: 3). Sixty students were included in this analysis. Deep-learners (R= 0.46, p<0.001) and Surface-learners (R=0.49, p<0.001), were found to prefer their respective teaching approaches.
Next, the association between learning approach and academic performance was also studied (Table: 4). Only Strategic-learners had a significant positive correlation with perceived academic performance (R=0.54, p<0.001).

Discussion
Our study using a modified ASSIST questionnaire (Table: 1) presents a tool for mapping the synchronicity between learning approaches and teaching approaches in undergraduate medical education, especially in locales where the student population is diverse with dissimilar academic foundation/backgrounds. Through a modified ASSIST questionnaire we found that the Deep-learning approach was the predominant learning approach in our students, a trend that has been observed in other similar studies [12,13]. Competency-based medical education has most likely impelled the integration of problem-solving and analytical thinking in the "typical medical curriculum" of today. Additionally, correlation analysis between the three learning approaches indicated a positive correlation between Strategiclearning and Deep-learning items, but a negative correlation between Strategiclearning and Surface-learning items. This indicates that Strategic-learners can adapt to Deep-learning approaches, but Deep-learners and Surface-learners cannot adapt to each-others' learning approaches. This finding in part supports the earlier observations of Ramsden, in which it was seen that strategic-learners employed "cues and clues" about assessment to engage both surface-and deep-learning strategies to achieve positive outcomes [14].
The 3P (Presage, Process and Product) teaching and learning model of Biggs and Moore, suggests that learning approach of a student is tempered and mitigated by: Accordingly, one of the aims of delivering a curriculum is to promote the "transformation" of "Surface/Deep-learners to Strategic-learners". This can effectively be achieved through the concerted and prudent use of active learning techniques [16,17]. Additionally, to provide the right milieu for this "transformation," teaching strategies should promote and foster self-directed learning through mental models such as the Feynman technique (Figure: 6); encourage collaboration; informally assess students; integrate technology in the learning process and; disseminate lessons with flexible learning paths.
One of the challenges of this study is that although most of our students are Deep/Strategic-learners, they prefer the surface-teaching approaches. This may be attributed to the fact that at MBRU, diverse blended learning strategies are employed ( Figure: 7), insufficiently examined by the items in the modified ASSIST.

Conclusions
We present a comprehensive tool designed using the framework of ASSIST to correlate learning approaches with teaching approaches. Initial results are similar to other studies in the literature. However, ASSIST is a self-reporting apparatus, and therefore may not always reveal the true approach to learning of students, especially if they responded in a way that they believed would have been the approved answers. Also, how learning approaches are predisposed in a multidimensional milieu, when resilience and stress-coping strategies of students are also included, is currently unknown. Future research should investigate these facets. The study spanned between January and September of 2018.

Consent for Publication: Not Applicable
Competing interest declaration: YB is the recipient of funding from Pfizer, Amgen and the Paragon Group to conduct medical education activities in the form of continuing professional development (CPD) and continued medical education (CME) activities. However, these funds haven't been used in this study depicted in the manuscript. Other authors declare no competing interest.

Funding: No Funding was obtained for this study.
Author Contributions: AA obtained IRB approval for the study, implemented the ASSIST questionnaire, curated the data and prepared the initial manuscript draft; WW analyzed the curated data using different statistical tools; FAAR and CT proof-read the manuscript and provided YB with constructive and practical inputs; DD helped in study design and provided logistical support; YB put-forward the initial design of the study, drafted the final version of the manuscript and explored the statistical data in the light of different learning theories. Acknowledgment: YB will like to thank Dr. Aida Azar, the course director of Student Research project. Data depicted in this manuscript was collected by AA to fulfill the requirements of her thesis submitted as part of the course requirement.

Availability of Data and Materials
The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author (YB) on reasonable request.

Authors' information
AA is a fourth-year medical student at MBRU. She has basic science research experience as a student intern at the Queens University of Belfast and the Mayo Clinic. She has received awards in the region, for her work in medical education. AA's interest in medical education is particularly focused on the different approaches to learning, and the factors influencing them specifically resilience and stress-coping strategies; and on these lines, she is at present working on a research project with YB.  16 23 I put a lot of effort into studying because I'm determined to do well.    Details of exclusion at various stages of data analysis. The 3P Model of Teaching and Learning. The model describes the factors which influence whe Figure 6 The Feynman Technique. The Feynman Technique is a mental model that was coined by Nobe The different Blended Learning Strategies in practice at MBRU. These strategies couldn't be