Saudi Arabian students in postgraduate dental programs: Investigating factors associated with burnout

Research questions

This research will address the following research questions: 1) What level of burnout, on average, do Saudi Arabian students enrolled in postgraduate programs report as measured by the emotional exhaustion (EE), depersonalization (DP), and (diminished) personal accomplishment (PA) subscales of the Maslach Burnout Inventory (MBI)?; 2) What proportion of variation in levels of burnout among Saudi Arabian students enrolled in postgraduate programs is accounted for by the set of predictors which includes: country where studying, hours of work, being sponsored or not, marital status, and prior work experience?; and, 3) Which of the following predictors accounts for unique variation in levels of burnout among these students? The predictors are (1) country (US vs UK), (2) hours of work, (3) sponsorship status, (4) marital status (5) gender, and (6) prior work experience. These questions will be assessed with a quantitative comparative research design assessing Saudi students enrolled in postgraduate dental programs in the UK and the US using the Maslach Burnout Inventory - Human Services Survey for Medical Personnel (MBI-HSS MP),¹⁵ which is a variation of the MBI tool adapted for medical personnel to measure burnout by addressing three scales: 1) feelings of emotional exhaustion due to work duties, 2) depersonalization measuring lack of sympathy and impersonal attitudes toward patients, and 3) personal accomplishment measuring feelings about success and achievement related to the profession.

Research significance

Program designs have a significant impact on stress and burnout in postgraduate dental students, and studies indicate that more individualized support for students is needed to reduce stress and burnout.¹ Different approaches have been applied in an effort to reduce risks of stress and burnout. Ahmed et al.,¹ compared a case-based curriculum and a subject-based curriculum for fifth and sixth-year dental and medical students at Kuwait University and found the case-based format correlated with higher stress levels. Some of the leading sources of stress were related to inconsistent feedback from varying instructors, challenges in effective communication with instructors, and workload amount.¹ Factors relating to inconsistent feedback, communication challenges are likely to be heightened for international students, in addition to gaps in knowledge and standards in US programs compared to the home country.⁶ The American Dental Education Association specifies guidelines for robust learning environments for diverse students, as well as measures to address linguistic and cultural backgrounds.¹⁶ However, it is unclear what measures are generally applied to address stress in postgraduate programs. To the researchers’ knowledge, there has yet to be a study comparing the stress and burnout among a specific international population in postgraduate dental programs in the United States and the United Kingdom. The study was designed to inform Saudi students about possible predictors of burnout when planning on studying abroad and can also guide US and UK. postgraduate programs in regard to addressing stress and burnout considerations for international students.

Definition of terms

The terms used in this article pertain to the psychological experiences of burnout and stress. The author also clarifies in this section the elaborate educational requirements for a dental student in the locations of interest of this study. This clarification highlights the educational journey of postgraduate dental residents leading to their enrollment in residency programs in the US or UK. The educational timeline for the US and Saudi Arabia were acquired based on the author’s experience of studying in both countries. The UK dental educational timeline was acquired from the General Dental Council (GDC), the licensing body for dentists practicing in the UK.¹⁷

Burnout: According to Maslach and Leiter,⁸ burnout is a psychological syndrome emerging as a prolonged response to chronic interpersonal stressors on the job. The three key dimensions of this response are an overwhelming exhaustion, feelings of cynicism and detachment from the job, and a sense of ineffectiveness and lack of accomplishment.

Stress: According to Baum,¹⁸ stress is an uncomfortable emotional experience accompanied by predictable biochemical, physiological, and behavioral changes.

Undergraduate dental student: Undergraduate students enrolled in a dental school.

Postgraduate dental student: Students who graduated from a dental school awarding a dental degree, who successfully passed licensure exams and are enrolled in a higher education specialty program. These students are referred to as ‘residents’ when the type of postgraduate program they are enrolled in requires a clinical residency.

Clinical postgraduate dental specialty programs: Those programs with a heavy focus on the clinical procedures of a certain branch in dentistry. The postgraduate students enrolled in this type of program are referred to as ‘residents’. These programs typically require residents to diagnose and treat a vast number of patients throughout the duration of the program. They award graduates degrees in the clinical specialties of dentistry, e.g., orthodontics, endodontics, periodontics, etc.

Non-clinical postgraduate dental specialty programs: Those programs with less focus on direct patient-doctor interactions. They vary in nature and requirements according to the field of study. These programs award degrees in the nonclinical specialties of dentistry, e.g., public health and community dentistry, oral biology and biomaterials etc. This term also potentially includes non-clinical degrees in clinical specialties, e.g., Doctor of Philosophy degrees (PhD) in endodontics.

Timeline of dental studies in Saudi Arabia: Undergraduate dental student, intern, licensure exam, dentist, postgraduate dental student.

Timeline of dental studies in the United States: Pre-dental student, dental student, licensure exams, dentist, postgraduate dental student.

Timeline of dental studies in the United Kingdom: Undergraduate dental student, foundation dental training, registration for licensure procedure, dentist, postgraduate dental student.

Note: For the sake of simplifying, the term ‘undergraduate dental student’ is used in this study to include all dental students pre-graduation from dental school regardless of location.

Ethical approval

This study has been performed in accordance with the Declaration of Helsinki regulations and guidelines. Participation in the study was voluntary and confidential through anonymous, non-identifying survey design. Written informed consent was obtained from all study participants. Approval from the institutional review board (IRB) was obtained on 2/27/2019 with protocol #19-53.

Participants

Direct communication with the Saudi Arabian Cultural Mission was conducted to identify programs with Saudi Arabian dental residents enrollment. Communication with Saudi Arabian dental residents enrolled in these programs was then initiated by the study representatives in the US and the UK directly whether face-to-face, e-mail, or social media accounts. The target population is Saudi Arabian dental residents enrolled in UK-based and US-based post graduate programs in the following clinical specialties: endodontics, fixed and removable prosthodontics, oral and oromaxillofacial surgery, orthodontics, oral medicine, operative and restorative dentistry, advanced education in general dentistry, oral radiology, and pediatric dentistry and periodontics. These specialties were selected due to their clinical nature with repeated resident-patient interactions. The programs award graduates either a certificate of advanced graduate study (CAGS) or a Master of Science degree (MS, MSc) upon completion. All programs require residents to be proficient in the English language as indicated by their Test of English as a Foreign Language (TOEFL) or International English Language Testing System (IELTS) scores. Non-clinical programs with lowered or no direct patient contact were excluded due to the nature of the study requiring repetitive direct contact with patients to report the burnout experience.

Sampling procedure

The sampling method used is the non-probability quota sampling technique. Although ideal, a stratified random sample from the population is not feasible. Based on a power analysis using the software G*Power, it was determined that to detect a small-medium effect (f² = .09), a sample size of 158 is needed (α = 0.05, power = .80, # of predictors = 6). Given the main focus of this study being the comparison between residents in two countries (i.e., UK and US), the minimum number of participants desired was 79 participants for each country. The G*Power software can be freely downloaded from http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/. As it turned out, the sample size obtained was just 87 cases with complete data (from both countries combined) and given the R² values obtained (which were translated into f² values of .23, .02, and .08), the power was .91, .12, and .43 for the regression models predicting EE, DEP, and PA, respectively. This limitation to statistical conclusion validity will be noted in the discussion.

Instrumentation

To address the research questions, the criterion variables are emotional exhaustion (EE), depersonalization (DP), and professional accomplishment (PA); the control variables are hours of work per week, marital status, gender, prior work experience, and sponsorship status; and the predictor variable is country.

The survey starts with an informed consent page followed by the first part of the survey. This first part was created by the author and includes demographic questions regarding program location, age, gender, marital status and family living arrangements. Other questions address specialty program type and length, resident position/year, hours of work per week, sponsorship status and work experience prior to joining the program.

To assess perceived stress and burnout, the second part of the survey includes the English version of the Maslach Burnout Inventory Human Services Survey for Medical Personnel MBI-HSS (MP). It was distributed to participants via a web link. According to Maslach et al.,¹⁹ sufficient levels of stability estimates and internal-consistency reliability estimates (Cronbach’s alpha) are reported for the three MBI-HS scores from a broad sample of workers in professions related to human-services. Cronbach’s alphas were .90, .79, and .71 for emotional exhaustion, depersonalization, and personal accomplishment, respectively. Multiple studies were cited in which test-retest coefficients for the three scale scores were reported for a variety of samples; for example, over a few weeks (.82, .60; and .80, respectively); three months (.75, .64, and .62, respectively); and up to one year (.60, .54, and .57, respectively).

The MBI section is the second part of the survey,¹⁵ which contains 22 items to assess burnout on the three identified scales: emotional exhaustion, depersonalization and professional accomplishment. Participants will be asked to rank item responses on a seven-point Likert scale where 0 means ‘never’, 1: ‘a few times a year or less’, 2: ‘once a month or less’, 3: ‘a few times a month’, 4: ‘once a week’, 5: ‘a few times a week’, and 6: ‘every day’. The survey was not piloted prior to this study.

Procedure

Contact was made via e-mail with informal representatives of the Saudi Arabian dental residents in the US and the UK, given their spread out locations from the author, located in Western US and the Northeastern US while the third was in the UK. Moreover, their network connections with Saudi Arabian dental residents who are enrolled in a variety of dental programs in both countries.

The two representatives handed out paper recruitment flyers to residents within both of their schools and sent out emails with the flyers attached to acquaintance residents in other schools. These flyers invited potential participants to provide the representatives with their email addresses for the online survey link to be sent out to. The flyers included assurance statements about the anonymous and voluntary nature of the survey. Once residents responded to the flyer in person or by email by providing their preferred email address, the email addresses were added to a list of recipients on the survey website, SurveyMonkey, and the survey link was forwarded to the list through the website itself. The website used to administer the survey is an online survey tool (SurveyMonkey.com, Palo Alto, CA, US) allowing for anonymous responses and voluntary participation. Once they received the link via email, clicking on it takes the respondent to the consent form page first. Reading and signing the consent by clicking on the appropriate button was mandatory to view the survey questions. Otherwise, the link automatically redirected responders to exit the survey. The anticipated response was 80%. However, due to low participation rate, contact with prominent social media figures in the Saudi dental field on Twitter was initiated. These influencers have many followers who are mostly dental field professionals and students. A web link was distributed in a Twitter post by two social media figures in an attempt to reach out to as many potential respondents as possible. This helped improve response rate significantly. The surveys were available online for a period of three months. Data collection started towards the end of February 2019 and continued until early May 2019. To improve response rate, an incentive of $5 Starbucks™ e-gift cards were to be offered to 20 participants who were going to be chosen randomly. Participants were given the choice to sign up for these gift cards after filling out the survey where they were asked to provide any e-mail address to send the e-gift card randomly to it. A physical randomization technique was going to be used where participants who provide e-mail addresses were going to be written on pieces of papers, mixed in a bowl and chosen 20 randomly. However, only three participants signed up for the gift cards, which were sent to the provided e-mail addresses without the need for randomization since the number of gift card interested participants was below 20. Subsequently, data was downloaded to a Microsoft Excel (2019) spreadsheet and the statistical package for the social sciences (SPSS) software (IBM SPSS Statistics for Windows, version 23 (IBM Corp., Armonk, N.Y., USA) which can be found at (https://www.ibm.com/sa-en/analytics/spss-statistics-software) was used to run the appropriate statistical analyses.

Data analysis

The data for this research is available in the Harvard Dataverse repository.²⁰ In order to evaluate the research questions, an α level of 0.05 was employed. Descriptive statistics along with t-tests and multiple regression analysis was used. A table of descriptive statistics and zero order correlations for all variables is provided as shown in Table 2.

The first research question (RQ) involves descriptive statistics including means, medians, and standard deviations which are reported for the three scales of the MBI and compared to existing norms for medical professionals²¹ using a series of three one-sample t-tests for the three subscales. To address the second RQ, a simultaneous multiple regression analysis was performed for each of the three MBI subscales utilizing the five control variables and the key predictor (country) in the model. The R² was used to gauge the proportion of variation in levels of burnout that the set of variables explains with similar analyses for the other subscales. To address the third RQ, three sequential multiple regression analyses were performed predicting the scores of each subscale separately. In Block 1, the five control variables were entered. In Block 2, the key predictor (country) was entered. The change in R² (ΔR²) and its level of significance was used to answer the research question. The results for both RQ₂ and RQ₃ are summarized as shown in Tables 4, 5, and 6.

Reliability

Reliability was evaluated using Cronbach’s alpha for the three MBI-HSS (MP) scales yielding coefficients of 0.819 for EE, 0.642 for DEP, and 0.708 for PA. That the reliability for DEP was below .70 suggests a potential threat to statistical conclusion validity for analyses involving depersonalization, a limitation to be noted in the discussion, as well.

Demographics

A total of 93 residents responded to the survey. The number of complete responses was 87. Over two-thirds (68.8%) of respondents were studying in the United States while about one- third (31.2%) of the respondents were studying in the United Kingdom. Male respondents comprised 58.1% of the sample while 41.9% were female. More respondents were married (64.5%) than single (35.5%). Prior work experience, specialty and hours of work per week demographics are shown in Table 3. Most respondents had a few years of prior work experience before beginning their postgraduate programs abroad. Respondents from a wide range of specialties participated in the survey, where Fixed and Removable Prosthodontics residents were the most numerous at 21.5% while Oral Medicine residents were the fewest (2.2%). Most of the sample reported being dual sponsored by the Saudi Cultural Mission and an employer. Over a quarter of the sample reported a high number of work hours per week where they were required to work 51 hours per week in program related activities and requirements.

Results regarding RQ one

To address the first research question, as to the level of burnout, one-sample t-test analyses were performed, employing an alpha level of.05 to compare the means of Saudi dental residents to that of the medical profession for the three MBI scales. The means of the medical professionals (2.466, 1.424, and 4.566 for EE, DP, and PA, respectively) were obtained from the MBI manual.¹⁹ There is evidence to suggest a statistically significant difference between the means of Saudi Arabian dental residents (M = 2.73, SD = 1.20) and the medical professionals population means in EE (p = .043). Specifically, we are 95% confident that the dental residents are at least .009 and at most.52 points higher in emotional exhaustion. However, there was insufficient evidence to suggest a statistically significant difference in DP (p = .860) and PA (p = .510), based on the dental residents’ sample statistics (M = 1.45, SD = 1.16) and (M = 4.41, SD = .96) respectively as shown in Table 2.

Results regarding RQ two and three

Multiple regression analyses were conducted to address research questions two and three regarding the proportion of variation accounted for by the set of six predictors and which predictors account for unique variation in the levels of burnout. Tables 4, 5 and 6 summarize the results of analyses for the three subscales, respectively.

Emotional exhaustion

In Table 4, the multiple regression model for EE with all six predictors was statistically significant and explained 18.8% of the variation, F(6, 80) = 3.087, p = .009, R² = .188. Whereas the set of five variables in the first block accounted for 16.8% of the variance, knowing the country in which the student was a dental resident explained an additional 2.0% of the variance in EE. This increase, however, was not statistically significant, F(1, 80) = 1.954, p = .166. Three of the predictors accounted for unique variation in levels of EE. The unstandardized regression coefficient for hours worked per week, b = .229, t (80) = 2.993, p = .004, indicates that for each additional unit increase of the predictor (10 hours increments), EE scores increased by .229 points, controlling for other predictors. Of more interest was the coefficient associated with Sponsorship, b = .712, t (80) = 2.491, p = .015, where being sponsored is associated with an increase of .712 points on the EE scale while controlling for other predictors.

Once other variables are taken into account, gender was a statistically significant predictor for the EE scale as well. On average, the EE scores of females (coded 1) were .550 points higher than males (coded 0), b = .550, t (80) = 2.091, p = 040. Those studying in the US (coded 1) had lower EE scores, on average, than those studying in the UK (coded 0), but the difference was not statistically significant, b = -.383, p = .166.

Depersonalization

In Table 5, the multiple regression model for depersonalization (DEP), the set of six predictors was statistically significant but explained just 2.2% of the variation, F(6, 80) = .297, p = .937, R² = .022. Whereas the set of five variables in the first block accounted for 2.1% of the variance, knowing the country in which the student was a dental resident explained less than 1% more of the variance in DEP. This increase was not statistically significant, F(1, 80) = 0.039, p = .844. None of the predictors accounted for unique variation in levels of DEP. Those studying in the US (coded 1) had higher DEP scores, on average, than those studying in the UK (coded 0), but the difference was not statistically significant, b = .057, p = .844.

Personal accomplishment

In Table 6, the multiple regression model for personal accomplishment (PA) with all six predictors was not statistically significant and explained 7.8% of the variation, F(6, 80) = 1.128, p = .354, R² = .078. Whereas the set of five variables in the first block accounted for 7.0% of the variance, knowing the country in which the student was a dental resident explained less than 1% of the variance in PA. This increase, then, was not statistically significant, F(1, 80) = .686, p = .410. Gender was the only predictor that accounted for unique variation in levels of PA. On average, the PA scores of females (coded 1) were.490 points lower than males (coded 0), b = -.490, t (80) = -2.181, p = .032. Those studying in the US (coded 1) had lower PA scores, on average, than those studying in the UK (coded 0), but the difference was not statistically significant, b = -.193, p = .410.

Assumptions and limitations

Statistical assumptions include those underlying the use of multiple regression analysis such as: normality, linearity, homoscedasticity, and independence of the error terms. A methodological limitation of self-reporting assumes complete transparency in responding to the survey. However, the effect of this assumption is assumed to have been minimized by the anonymous and voluntary nature of this study. Moreover, responses from residents in different dental specialty programs limit the equal distribution of subjects across the specialty type category. This study is also limited by the concurrent political circumstances of the program countries and Saudi Arabia. It’s also important to note that generalizability or external validity is limited to the characteristics of the participants. Furthermore, some dental specialties were not included in this study due to the nature of these programs being only indirectly interactive with patients, e.g., oral histology and oral pathology.

Reliability testing of the MBI scales revealed sufficient internal consistency for EE and PA but not for DEP. This can be related to the number of items in the scale. DEP has five items only while EE and PA have nine and eight items respectively. This limitation has likely affected the multiple regression results pertaining to DEP where none of the individual predictors was statistically significant. This could also be explained by another limitation this study faced where the number of complete responses were insufficient to detect subtle effects. Increasing the incentive promised to participants or contacting residency program directors requesting recruitment of their residents in a formal way where they set aside time for residents to complete the survey during working hours could have increased the response rate. However, due to time and financial constraints and the researcher’s limited social network in dental schools in the US and the UK, these methods were not feasible. There was a difficulty in reaching out to all residents due to variation in physical location between different countries, schools and specialties. This results in limited statistical power to the analysis and poses a threat to statistical conclusion validity.

Suggestions for further research

More investigation is needed to evaluate the differences in reported burnout in the excluded specialties and across various program types, e.g., fellowship and PhD programs, to broaden our understanding of this phenomenon and inform future residents who are opting to pursue postgraduate studies. Although this is a comparative study, it is imperative to adopt a longitudinal study design to allow for causal inferences in understanding how burnout progresses over time in health care professionals and the effect it could have on patient care and medical errors. Cultural and social norms affecting female professionals should be explored to ensure better awareness in this critically affected subgroup, in particular. Moreover, qualitative study designs utilizing open ended questions could be used to elicit, for example, participants’ experience as a resident, including their feelings related to burnout and the expectations they perceive their sponsors to have for them during their programs.

Underlying data

Harvard Dataverse: Asiri, Amal, 2022, “Saudi Arabian Students in Postgraduate Dental Programs: Investigating Factors Associated with Burnout”, https://doi.org/10.7910/DVN/NTBYPX.²⁰

This project contains the following underlying data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).

Extended data

Harvard Dataverse: Asiri, Amal, 2022, “Saudi Arabian Students in Postgraduate Dental Programs: Investigating Factors Associated with Burnout”, https://doi.org/10.7910/DVN/NTBYPX.²⁰

This project contains the following extended data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).