This cross-sectional study invited postgraduate clinical trainees from all SCFHS affiliated residency programs within the KSA to participate during February 2024. Undergraduate medical students were excluded.

The institutional review board of Imam Abdulrahman Bin University had approved the study with reference number (IRB-2024-01-087). The participation in this study was voluntary. Consent to participate in the study was obtained prior to answering the questions. No personally identifiable information was asked. The study complied with the principles of the Declarations of Helsinki.

The minimum required sample size was 1,013. This was based on a total population of postgraduate trainees in the KSA that was 63,434 according to the statistical yearbook of the Ministry of Health (12). An expected frequency of 40.59% was used in the calculation which was from a previous study of medical students that had belonged to the deep approach, at 95% CI with a margin of error of 3% (1). The sample size was calculated via the Epi Info software V.5.5.2.

Postgraduate clinical trainees of both sexes were invited to participate. A non-probability sampling technique was used where several data collectors were assigned to overview the data collection process. This was achieved by sending a link which contained the survey to trainees registered phone numbers. To avoid duplication of responses, the link did not allow for any multiple responses.

The data was collected by means of an online-based, self-administered questionnaire. The questionnaire covered both sociodemographic characteristics such as age (mean ± SD) and sex (males, females). Also, academic and training specific characteristics were collected, these included whether the participant had previously failed during the undergraduate medical school (yes, no), and the type of the current postgraduate training program (medical, e.g., pediatrics, emergency medicine, dermatology, surgical, e.g., plastic surgery, obstetrics and gynecology, urology or others such as a master’s degree), the year of study (years 1–2, years 2–3, year 5 and above), whether the participant had failed during a promotion exam (yes, no).

The ASSIST short version tool was used which when scored was used as the main outcome of the study. The ASSIST tool consists of 18 questions measured on a Likert scale ranging from 1 to 5 (1 = disagree, 2 = somewhat disagree, 3 = unsure, 4 = somewhat agree, 5 = agree). The responses allow the investigators to identify the three learning approaches of participants where the deep learning approach includes items 2, 6, 10, 12, 15 and 17. The strategic approach includes items 3, 5, 7, 9, 11 and 13. The surface approach includes 1, 4, 8, 14, 16 and 18. The tool is scored by summing the responses to those specific items to a new variable, and the highest score to one of those approaches meant that it was the predominant approach to that participant. Participants who had scored equally to two approaches are removed from the analysis (1, 11). The reliability score for the ASSIST questionnaire was 0.87 indicating good internal consistency.

The data was analyzed in STATA software version 15 (13). Variables were described as either means ± SD after checking for normality or by frequencies and proportions. Bivariate associations were performed through a series of independent samples t-tests and ꭓ² tests where appropriate. Unadjusted and adjusted multinominal logistic regression analyses were used to compute the relative risk ratios (RRRs) along with their 95% confidence intervals (CIs) while using the strategic approach as the base outcome. The level of significance was set to less than 0.05. Model diagnostics were performed to ensure good model fit.

A total of 2,010 had accessed the online questionnaire, of those 1,187 had responded, which gave a response rate of 59.05%. However, 279 were excluded from the analysis due to having equal scores to two learning approaches. Of the remaining 908 trainees, the average age was 27 years ± 2.79 SD. Also, 55.29% were females compared to 44.71% males. Over 75% had not previously failed medical school. With regards to the type of postgraduate program, 60.13% were medical, 24.78% surgical and only 15.09% were enrolled in other types of postgraduate training such as a master’s program. Over half the participants were in their first year of training (54.74%), and only 13.95% had failed a promoting exam (Table 1).

Table 2 describes the sociodemographic and educational characteristics of postgraduate trainees according to the three learning approaches. Most trainees were found to adopt the strategic approach followed by the deep approach (39.98 and 37.22% respectively). No statistically significant difference was observed for age or sex. However, a higher proportion of males adopted the strategic approach, whereas females preferred the deep approach (41.63 and 40.04% respectively). A statistically significant association was found between previously failing during medical school and the learning approach, where trainees with no history of failure during medical school mostly adopted the strategic approach (41.20%). With regards to the type of postgraduate program, medical trainees equally adopted the deep and strategic approach (40.11%), whereas for surgical trainees, a clear preference for the strategic approach was observed (37.78%) (p = 0.03). As for previously failing a postgraduate promotion exam, of the total 595 trainees, those who had failed mostly adopted the surface approach whereas those who had not failed mostly adopted the strategic approach (43.37 and 42.19% respectively) (p < 0.001).

The multinominal logistic regression analyses found that age was not a significant predictor for the predominant learning approaches (Table 3). With regards to the deep approach in comparison to the strategic approach, females were observed to be more likely to adopt it (Adjusted RRR = 1.51, 95% CI = 1.02–2.24). According to the unadjusted model, trainees who had previously failed during medical school were observed to be more likely to adopt the surface approach in learning in comparison to the strategic approach (Unadjusted RRR = 1.73, 95% CI = 1.08–2.77), however this was not significant after adjustment. Previously failing a promotion exam was a significant predictor for adopting the surface approach more than the strategic approach, both before and after adjustment (Adjusted RRR = 2.85, 95% CI = 1.44–5.61).

Figures 1a,b shows females’ and males’ adoption of learning approaches according to the type of postgraduate program. For females, medical trainees were seen to prefer the deep approach and least prefer the surface approach, whereas for surgical trainees, they were seen to prefer the strategic approach and least prefer the surface approach. Trainees enrolled in other types of programs such as a master’s program were also seen to prefer the strategic approach (Figure 1a).

For males, and unlike females, medical trainees were observed to prefer the strategic approach in learning. Surgical trainees were seen to mostly adopt the strategic approach, followed by the surface approach, whereas for other types of programs, the strategic approach was also preferred (Figure 1b).

In this study, the strategic approach was the most common approach, followed by the deep approach. This contrasts with the only other Saudi research which utilized the ASSIST tool among undergraduates in which the deep approach was the most common (1). However, internationally, two previous studies, one in Singapore and the other in Sri Lanka, had found similar results to the current study in which the predominant learning approach was the strategic closely followed by the deep approach (10, 11). These were also among postgraduates. Postgraduate trainees may feel pressured to prioritize job completion over in-depth understanding due to heavy workloads, prolonged shifts, and patient care duties (15). This pressure could cause trainees to adopt a strategic learning approach where they prioritize completing tasks right away over fully interacting with the course material.

The current findings also show that trainees with no history of failure during medical school mostly adopted the strategic approach. On the other hand, trainees who had previously failed during medical school were observed to be more likely to adopt the surface approach to learning in comparison to the strategic approach. However, adjustment analyses revealed that this difference was not statistically significant. Similarly, a previous failure in a postgraduate promotion exam was a significant predictor of adopting the surface approach more than the strategic approach, whereas those who had not failed mostly adopted the strategic approach. In agreement with results from the current study, a study among pre-clinical undergraduate students in Dominica showed a significant association between learning approaches and academic performance of students. When compared to low-achieving students, high-achieving students expressed a preference for strategic and deep learning approaches. Additionally, low achievers reported using the surface approach to learning most of the time (16). One more study using a sample of occupational therapy students in Norway found a correlation between greater exam scores and lower surface approach scores and higher strategic approach scores (17). A comparable and very recent study with 1,010 Egyptian nursing students by Dogham et al. revealed a significant association between surface learning approaches and academic stress. Elevated levels of academic stress can have a negative impact on students’ motivation and involvement in the learning process. This can result in feelings of overwhelm, lack of interest, or burnout, which may cause students to adopt a surface learning strategy (15).

In contrast to the studies that identified sex as a significant factor influencing students’ learning approaches, no statistically significant difference was observed in the present study (15, 16). However, a higher proportion of males adopted the strategic approach, whereas females preferred the deep approach. A similar observation was reported in a Saudi study among undergraduate medical students (1). This observation was also seen in other parts of the world. For example, in Egypt and in India females were also more likely to adopt the deep approach when compared to males (18). Understanding the disparities in approaches to learning between males and females may facilitate postgraduate programs in providing sex-specific resources and support. For instance, offering focused seminars or counseling services that address female and male trainees’ unique learning needs may help to alleviate such differences (15). Upon further analysis of the sex-based differences between the postgraduate clinical trainees’ programs and the different learning approaches, our results highlight the complex interactions between sex and specialty-related variables that influence the learning approaches of postgraduate trainees. By tailoring educational strategies to align with trainees’ specialty-specific needs and learning approaches, medical educators can foster deeper knowledge retention, improve clinical performance, and support the development of well-rounded, competent physicians who are better prepared to meet the complex challenges of modern healthcare systems.

The findings of this study offer critical insights for medical educators in the KSA, particularly within the framework of Vision 2030. The predominance of deep and strategic learning approaches among postgraduate trainees underscores the necessity for educational strategies that foster critical thinking, self-directed learning, and reflective practice. Vision 2030 emphasizes the transformation of the healthcare sector through workforce development and the enhancement of medical education programs to align with international standards. This includes a shift toward competency-based education, which necessitates curricula that support diverse learning approaches and cater to the individual needs of trainees (5).

The observed variations in learning approaches based on sex, specialty, and academic experience highlight the importance of personalized educational interventions. For instance, the higher prevalence of deep learning approaches among female trainees and those in surgical specialties suggests that tailored pedagogical methods could enhance learning outcomes across different demographics. Furthermore, targeted educational strategies, including mentorship, reflective practice modules, and leadership development initiatives tailored to female trainees, may not only enhance individual academic outcomes but also contribute to broader systemic goals of health sector advancement under Vision 2030. This aligns with the health sector transformation program of improving the quality of healthcare education and ensuring that all healthcare professionals are equipped with the necessary competencies to meet the evolving needs of the population (19).

Moreover, the strategic learning tendencies among trainees who are involved in research-based training programs suggest that integrating further research opportunities and teaching responsibilities into postgraduate programs could promote more effective learning strategies. This approach supports a culture of research and innovation within the healthcare sector (19).

The study does have some limitations. Firstly, the findings’ generalisability may be impacted by the use of a non-probability sampling technique. Second, participant subjectivity may have an influence on study findings. Moreover, the academic performance of trainees may be impacted by other factors such as individual variances, mitigating circumstances and learning motives that were not taken into account during this study. Other limitations include the cross-sectional design, which cannot establish the causality of the associations, and the reliance on self-reported data, which may have introduced biases.