STRENGTHS AND LIMITATIONS OF THIS STUDY
-
Although there are global studies available with similar sample sizes, to the best of our knowledge, this is the only study that has recruited a large sample size to measure the learning environment in medical colleges in Pakistan.
-
The study explored the learning environment of medical colleges, using the Johns Hopkins Learning Environment Scale (a relatively new tool compared with the conventionally used Dundy Ready Educational Environment Measure (DREEM)) and its relationship with students’ well-being. Data were adjusted for important confounders to see true associations.
-
We could not investigate other background factors that may have affected the quality of life of students even though they gave high scores to the learning environment.
-
Because of the COVID-19 pandemic, face-to-face data collection was not possible and access to the students became difficult, therefore, sampling was conducted online.
-
Since the study was conducted during the pandemic, this might have impacted the usual (prepandemic) learning environment of the institutions.
Introduction
Although a complex array of personal and professional factors influences student well-being, their satisfaction with specific features of the learning environment appears to be a critical factor.1 In the past few decades, researchers have laid the foundation for strengthening evidence for measuring the learning environment; thus, highlighting the importance of regular measurements and evaluation for informing the required changes and improvements.2
The impact of a positive and supportive learning environment on student well-being cannot be ignored. Understanding medical students’ perceptions can provide useful information for making changes to the existing learning environment, comparing different institutions, and effectively planning and implementing the curricula.3 Student well-being can also be measured in terms of the quality of life (QoL) and reduced burn-out. QoL is ‘an individual’s perception of his/her position in life in the context of the culture and value systems in which they live and concerning their goals, expectations, standards, and concerns’.4 The operational definition is the self-reported perception of students about their overall QoL.
Burn-out is a measure of physical and psychological exhaustion as well as mental distress, triggered and increased by professional or occupational demands. Burn-out has two basic characteristics: emotional exhaustion and depersonalisation.5 Maslach et al defined emotional exhaustion, operationally, as the level of burn-out students feel each day at the college, and depersonalisation as how much they feel callous towards other students.
Studies have shown that the experiences of medical students in a supportive and positive learning environment greatly influence their academic performance, well-being and satisfaction.6 Although a great concern prevails in medical education that a negative or suboptimal learning environment can lead to greater distress among students, investigations into how the learning environment is related to QoL and burn-out among students are grossly lacking.1 7
The learning environment in a medical school is dynamic. Changes in the learning environment may occur over a period due to competitiveness, the curriculum, the expectations of stakeholders and the students. Encountering students who are stressed and have poor QoL is not uncommon. Students’ experiences are influenced by different factors and this makes them interpret their personal experiences as good or bad and consequently make a judgement of the overall learning environment of their institution. Their perception of the learning environment in a medical college closely relates to two logical and important sequences, students’ QoL and burn-out.
Until recently, the association between the learning environment and student well-being was not studied in Pakistan. With the increase in the number of medical colleges in the country and the rising cost of education, there may be a change in the expectations of stakeholders, potentially leading to more distress among them. A positive learning environment may indicate good QoL and lesser burn-out for students, which is a fundamental requirement for their development, and the opposite may happen if the environment is negative. An assessment of the learning environment may become more meaningful to stakeholders when all its components are optimally captured and its association with reported QoL and burn-out among students is determined.
This research was based on this background. There is limited literature available globally, and in Pakistan, on the measurement of learning environment using the relatively newer tool developed by Shochet et al, the Johns Hopkins Learning Environment Scale (JHLES);8 and its association with student well-being, that is, reported QoL and burn-out.9 This study, therefore, aims to measure the association of the learning environment with the reported QoL and burn-out (measured as emotional exhaustion and depersonalisation) among students from selected medical colleges in Lahore, Pakistan. The research helps in identifying the factors and corrective measures to improve the learning environment in an institution and consequent enhancement of student well-being. A single-item Linear Analogue Self-Assessment (LASA) Scale has been used to measure QoL,10 11 and a validated single-item linear question to measure burn-out among students.11 12 This provides meaningful insight into the learning environment and student well-being in our environment and adds new information to the literature. The objectives of this study were to:
-
Measure the learning environment in selected public and private sector medical colleges in Lahore, Pakistan, using the JHLES.
-
Determine the association of learning environment with QoL and burn-out among students in selected public and private medical colleges in Lahore, Pakistan.
The hypothesis was that with higher learning environment scores, students should have higher QoL and less burn-out.
Methods
Study design
This was a cross-sectional study that assessed the associations between learning environment, students’ QoL and burn-out.
Study setting and duration
The study was conducted in six medical colleges (three from the private and public sectors each, selected randomly from 21 medical colleges in Lahore, Pakistan): Shalamar Medical and Dental College, Combined Military Hospital (CMH) Lahore Medical College, University College of Medicine and Dentistry from the private sector; and Allama Iqbal Medical College, Fatima Jinnah Medical College and King Edward Medical College from the public sector. Infrastructure in the private sector colleges was up to date with approximately 150 enrolments per class, while the public-sector colleges had relatively congested infrastructure with up to 300 students per class.
Ethical approval
Ethical approval for the study was taken from the University College of Medicine and Dentistry, Lahore, Pakistan. See details at the end of the article.
Study participants
The target population included all undergraduate medical students enrolled in all levels of the Bachelor of Medicine and Bachelor of Surgery (MBBS) programme in public-sector and private-sector medical colleges throughout Pakistan. The accessible population was students currently enrolled in the medical colleges in Lahore. Students in allied health sciences, dentistry and nursing were excluded.
Sample size and sampling technique
From the sampling frame of 21 medical colleges located in Lahore, three public-sector and three private-sector medical colleges were selected randomly. Each selected medical college was taken as a cluster, and five MBBS programme levels were taken as mini clusters. All students in the mini cluster were included in the sample.
Our sampling frame consisted of a total of approximately 6750 students (4500 from the public-sector colleges and 2250 from the private-sector colleges). It was not possible to collect data from all students face to face due to COVID-19 restrictions. It was, therefore, decided to adopt a single-stage cluster sampling method for data collection using online Google Forms. Although an overall sample of only 692 students was required (expected frequency of outcome factor 50%, confidence limits 5% and design effect 1.8 for cluster surveys), we decided to take a minimum of 50% response from each class for better representation of students at the class and college levels. We sent invitations to all students from the six medical colleges and shared the link to access the questionnaire. Data collection was closed once 50% response was achieved from each class. Using this strategy, data were collected from 3400 students, with adequate representation from each class level of the six medical colleges (table 1).
Sociodemographic characteristics of the study participants (n=3400)
Study variables
The baseline sociodemographic variables were included as independent variables (gender, age, class, college type and residential status of the student). The learning environment was the predictor variable, while QoL and burn-out were the outcome variables for this study.
Data collection instruments
The questionnaire consisted of three sections. The first section included the demographic characteristics of the respondents. The second section consisted of 28 items from the JHLES Scale divided into seven domains and one item for global overall perceptions of students, on a Likert Scale of 1–5 and five categories.8 The seven domains of the JHLES were labelled as the community of peers, faculty relationships, learning environment, meaningful engagement, mentoring, inclusion and safety, and physical space.8 The 28 JHLES items for learning environment measurement (score range 28–140); and one item for global overall perceptions of students, on a Likert Scale of 1–5 and five categories. The description against each category of overall global perception was: (1) Terrible—not learner-centred, no opportunities for reflection, authoritarian, not trustworthy, disrespectful of diversity and alternate perspectives, predominantly negative aspects, positive aspects few and not mediated by negative ones; (2) Poor—overall mostly negative environment with some positive aspects; (3) Fair—an equal mix of positive and negative features; (4) Good—overall mostly positive with some negative aspects; and (5) Exceptional—environment marked by safety, trust, respect, welcoming of diversity, flexible, provides opportunities for the learner to challenge themselves with appropriate supervision and feedback, opportunities to reflect, and predominantly positive aspects which mediate the negative ones.8 After two surveys and exploratory factor analyses, the final instrument had factor structure as seven subscales including: (1) The community of peers, (2) Faculty relationships, (3) Academic climate, (4) Meaningful engagement, (5) Mentoring, (6) Inclusion and safety, and (7) Physical space (factor coefficient of ≥0.40 was taken as the threshold for item inclusion and a significant Bartlett’s test); with 28 items in them. The possible scoring of the instrument was 28–140; higher scores indicated a positive learning environment. Cronbach’s α for this study was 0.88.
The third section included the measurement of student QoL using a validated LASA question adopted from larger inventories for QoL. LASA measures the overall QoL as reported by the respondent on a 5-point Likert Scale; it has been validated in a broad range of populations and medical conditions, has a Cronbach α=0.775, and is widely used in QoL research.10 11 13 QoL is investigated with the question, ‘How would you describe your overall quality of life?’ on a scale of 1–5. Categorisation of the scale is given as 1 as bad as it can be, 2 as somewhat bad, 3 as neutral, 4 as somewhat good and 5 as good as it can be. A low QoL is ≤2 score, a 4–6 score is average QoL, and an 8–10 score is a high QoL. For this analysis, the research outcome was taken to be binary as low QoL and average to high QoL.
Student burn-out was measured using a validated single-item linear question (adopted from the Maslach Burnout Inventory or MBI) to measure the two parts (emotional exhaustion and depersonalisation) of burn-out, on a Likert Scale of 0–6. The question on emotional exhaustion was: ‘How often do you feel burn-out from work?’ on a scale of 0–6. Categorisation of the scale was given as 0 daily, 1 a few times a week, 2 once a week, 3 a few times a month, 4 once a month, 5 a few times a year or less, and 6 never. The score was multiplied by 9 (to equate it to an MBI Score of 0–54). Cut-offs were given as low burn-out 0–18; average burn-out 19–26; and high burn-out 27–54. The depersonalisation question asked: ‘How often do you feel callous towards other people?’, on a scale of 0–6. Categorisation of the scale was done for emotional exhaustion. The score was multiplied by 5 (to equate it to an MBI Score of 0–30). Cut-offs were given as low burn-out 0–5, average burn-out 6–9 and high burn-out 10–30.11 12 Single-item measures of emotional exhaustion and depersonalisation showed strong and consistent associations with key outcomes among medical students, internal medicine residents and practising surgeons.12 For this analysis, the research outcome was taken to be binary as low burn-out and average to high burn-out for both components. The questionnaire used in this study is attached as the online supplemental file 1.
Supplemental material
Data collection procedure
Data were collected between November and December 2020, using Google Forms. The researcher contacted each class representative, explained the research idea and its significance to them, and requested them to add her name to their class WhatsApp group. The researcher then shared a brief message and a Google Forms link in the group. Once the desired response was obtained in a class group, the researcher thanked the students and exited the group.
Statistical analysis
Descriptive statistics were calculated using mean and SD, frequency trends noted for all sociodemographic variables, QoL, and burn-out. Analysis of the JHLES Score and subscale scores was done using stratification at the level of public and private medical colleges and by gender. Calculation of the Pearson correlation of JHLES mean scores with QoL and burn-out scores was done by gender stratification. A t-test was used to measure the significance of the differences and regression analysis was applied to remove the effect of confounders (the predictor was learning environment, the outcome was QoL and burn-out). All sociodemographic variables were independent variables (age, gender, class level, residential status and type of college).
Pretesting of the questionnaire was done before data collection; with 30 randomly selected students (15 each from both sector colleges). The external and internal validity of the study was maintained by controlling the biases and confounders. Potential biases were controlled using different strategies. Selection bias was controlled using randomisation and cluster sampling, information bias was controlled using standardised and validated data collection instruments, and location and observer biases were controlled using Google Forms to collect data because the colleges were physically closed due to the COVID-19 lockdown.
Patient and public involvement
None.
Results
Data were collected from 3400 students, with proportionate representation from both genders and from the first-year to final-year students. The average age was 21.52±1.6 years; 2410 (70.9%) were ≤22 years old and 990 (29.1%) were >22 years old. JHLES showed strong internal consistency in our study (Cronbach’s α=0.88). Table 1 shows the sociodemographic characteristics of the respondents.
JHLES Scores
The mean JHLES Score for all students was 81.7±13.5 (male=82.0 and female=81.6). The independent sample t-test did not show any significant difference (p=0.48). However public-sector (82.1) and private-sector (81.1) colleges showed a significant difference in their means (p=0.03) with a small effect size (Cohen’s d-effect size=0.083). As many as 15.6% of the students rated the learning environment as terrible or poor, while 37.3% rated it as good or exceptional, and the majority (47.1%) labelled it as fair.
Students’ self-reported well-being
Student well-being (in terms of QoL and burn-out) was measured using two self-reported questions; one on QoL and the second on the level of burn-out. QoL was measured using a single-item linear question, categorised as low QoL, average QoL and high QoL. The distribution of students’ levels of burn-out was noted using a validated single-item question that measured it at low, average and high levels of burn-out. It was important to identify the level of burn-out among students because higher levels were anticipated, which could affect them personally and academically. Table 2 shows the responses for QoL and burn-out in three categories.
Distribution of responses for quality of life and burn-out in medical students (n=3400)
The majority of the students had average QoL and high QoL (96.4%) and very few reported having low QoL. The emotional exhaustion component showed very little difference between the number of students having high burn-out and low burn-out, while the majority of the students reported depersonalisation. None of the students reported having an average level of burn-out, and medical students in both sector colleges had high levels of burn-out.
Correlation of JHLES mean scores with QoL and burn-out scores
The relationship between the JHLES mean scores and the QoL and burn-out was calculated for all respondents, stratified by gender. Pearson correlation coefficient was applied to measure the correlation between the scores of two variables. Table 3 shows the results for the correlation between JHLES and outcome variables, stratified by gender.
Relationship between JHLES mean scores and the QoL and burn-out, stratified by gender (n=3400)
A moderately strong positive correlation was observed between JHLES mean scores and QoL for both genders with a statistically significant value of p. The result suggests that higher JHLES Scores correlated with reported QoL in both genders. A weak positive correlation was observed between JHLES mean scores and burn-out (emotional exhaustion and depersonalisation) for both genders, with a highly significant value of p. This study showed a weak correlation between the learning environment score and the two burn-out variables (emotional exhaustion and depersonalisation).
We applied logistic regression analysis to identify the significance of JHLES Scores and other predictors for QoL and burn-out (emotional exhaustion and depersonalisation) among the students. Conventional confounders (age, gender, education, socioeconomic status) and other factors were adjusted to determine the association between the predictor factor and the outcome. Hence, the hypothesis was: higher JHLES Scores result in high QoL and low burn-out among students.
In the adjusted model, Adjusted Odd Ratio (AOR) was <1, and with a statistically significant value of p for the two predictors, higher class level and overall JHLES Score. The odds of a high QoL Score were less with a higher class level and with a higher JHLES Score.
In the adjusted model, AOR was <1, and the value of p was statistically significant for the predictor of female gender. AOR was >1 and the value of p was statistically significant for every successive higher class level and boarding status of the students, with the outcome (table 4). A statistically significant value of p was found with overall JHLES Scores too with a minimal rise in the value of AOR (1.05). The odds of having a high emotional exhaustion score were less likely among women and minimally high with higher JHLES Scores. There were odds of a higher emotional exhaustion score with each successive higher class and residence status of the students. The adjusted model showed a statistically significant value of p for the predictor of overall JHLES Scores, with the outcome of depersonalisation. The odds of having high depersonalisation were minimally raised with higher JHLES Scores (table 5).
Association of the predictors with the reported outcome of QoL
Association of the predictors with the reported outcome of burn-out (emotional exhaustion and depersonalisation)
Discussion
Our study shows an overall mean JHLES Score for all students of 81.7±13.5 with statistically significant differences in public and private sector colleges (p=0.03). The difference was statistically significant in the community of peers (p=0.009), faculty relationship (p=0.007), academic climate (p=0.036), meaningful engagement (p=0.01), and inclusion and safety (p=0.01) subdomains. The difference was not significant for the mentoring and physical space subdomains. A study conducted in three medical schools in Malaysia found higher overall JHLES mean and subdomain scores than our study.14 Our findings are consistent with the study conducted by Sengupta et al in India that reported higher overall mean scores with significant values of p for three out of seven domains of JHLES.15 The findings are also supported by the original study of Shochet et al conducted for the validation of the JHLES tool.8
We investigated the hypothesis that a positive perception of the learning environment is associated with higher QoL and low burn-out among medical students included in our study. The Pearson correlation of the overall JHLES mean scores with QoL and burn-out (emotional exhaustion and depersonalisation) suggest a moderately strong positive correlation between JHLES Scores and QoL for both genders, with a significant correlation (r=0.59, p<0.01 for men and r=0.61, p<0.01 for women, respectively). A weak positive correlation was found with burn-out (emotional exhaustion r=0.40 and depersonalisation r=0.34 for men and emotional exhaustion r=0.34 and depersonalisation r=0.34 for women, p<0.01). Our results are consistent with the study conducted by Tackett et al in three medical schools in Israel, China and Malaysia, using similar instruments and parameters. Tackett et al found higher ratings of QoL in two out of three schools and the lowest rating for emotional exhaustion in all three, but no significant difference was seen for depersonalisation in all three medical schools.9
A narrative review of literature by Benbassat on the educational environment showed that medical training caused emotional distress that delayed students’ development and affected their clinical performance.16 Another study by Dyrbye et al on 1701 students in five medical schools in the USA concluded that the learning environment has a significant association with student burn-out.1 A multicentre study conducted by Enns et al on 1650 students from 22 Brazilian medical schools concluded that the educational environment appears to be an important moderator of medical student QoL.7
Our study showed a significant value of p (<0.05) between QoL and overall JHLES Score. It could be stated that the odds of high QoL were less with the overall JHLES Score (table 4). This finding contrasts with the study by Tackett et al that found higher odds of having high QoL with a better learning environment (AOR 3.2, p<0.001).9 Our study also showed a significant value of p (p<0.00) with higher class levels. It can be stated that the odds of high QoL were less with higher class levels (table 4). Our findings contradict our hypothesis possibly of other factors present within or outside the learning environment that could have influenced the QoL among students and could have affected the association between QoL and learning environment, for example, workload, the burden of studies, a balance between personal life and medical education, curriculum and assessment, teaching and learning. Such factors were not within the scope of our work.
Our study showed that the odds of higher emotional exhaustion were less among female students (AOR 0.76, 95% CI 0.65 to 0.90, p<0.001); and with higher overall JHLES Scores (AOR 1.05, 95% CI 1.04 to 1.06, p<0.0001), while the odds of higher emotional exhaustion were found with successive higher class and residential status (AOR 1.18, 95% CI 1.02 to 1.37, p<0.05) of the students. The odds of high depersonalisation were minimal with the overall JHLES Scores (AOR 1.06, 95% CI 1.05 to 1.07, p<0.0001) (table 5). None of the other predictors was found to have an association with the two outcomes. These findings are consistent with the study by Tackett et al which also found lower odds of emotional exhaustion and depersonalisation with overall JHLES Scores.9
The study shows that there is a significant association between a better learning environment and less emotional exhaustion for the female gender and higher overall JHLES Scores. However, emotional exhaustion was higher with successive higher class and residential status of the students. With a better learning environment, there was less chance of higher depersonalisation. The study also found that there is a significant association between learning environment and lesser QoL with increasing level of class and overall JHLES Score. However, since this study was conducted during the pandemic, it is difficult to know if this was the situation pre-COVID-19 or whether it was due to the pandemic.
This study has two major methodological limitations. It was designed before the COVID-19 pandemic started. However, data collection was done during the pandemic, which impacted the data collection plan. Face-to-face data collection was not possible due to the closure of the colleges; therefore, data were collected online. Since the study was conducted during the pandemic, this might have impacted the usual (prepandemic) learning environment of the institutions. We could not investigate other background factors that may have affected the QoL of students even though they gave high scores to the learning environment.
The way forward
Further studies are suggested at the end of the pandemic, with a bigger scope of intrinsic and extrinsic influencing factors for QoL and burn-out associated with the learning environment. Policies can be developed based on the results of this study to improve students’ well-being (QoL and burn-out).
Conclusion
Although students rated JHLES Scores on the higher side, they are still less likely to have a high QoL as an outcome. Furthermore, they are more likely to have higher emotional exhaustion with successive higher classes and with their residential status as boarders. Also, students have minimally high depersonalisation even when they rate higher scores on the JHLES Scale. Other internal and external factors that were not within the scope of this study could affect the QoL, emotional exhaustion and depersonalisation of students, for example, the motivation level of the student, family support financially and otherwise, peer influence, individual responsibility for programme funding, student’s emotional and mental health, and so on.
Study guidelines
Strengthening The Reporting of Observational Studies in Epidemiology (STROBE) cross-sectional reporting guidelines have been used to report the study.17
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
Ethical approval was obtained at two levels. The initial ethical approval was obtained from the Research Ethics Committee of the University College of Medicine and Dentistry on 17 November 2020 vide letter Ref # ERC/ 07/ 20/11. For data collection purposes further IRB or administrative approvals were obtained from the IRB committees of Allama Iqbal Medical College, King Edward Medical College, CMH Lahore Medical and Dental College, Shalamar Medical and Dental College, University College of Medicine and Dentistry, and Fatima Jinnah Medical College. Written informed consent was obtained from the students at the start of the study through online sharing of the consent form in the class WhatsApp groups before data collection.
Acknowledgments
The authors thank the medical students for their cooperation in data collection. The authors also thank the class representatives who facilitated data collection during COVID-19 through the use of their class WhatsApp groups.
This post was originally published on https://bmjopen.bmj.com