Problem-based learning in medical degree teaching: a scoping review

Problem-based learning (PBL) is a pedagogical approach that shifts the role of the teacher to the student (student-centred) and is based on self-directed learning. Although PBL has been adopted in undergraduate and postgraduate medical education, the effectiveness of the method is still under discussion. We employed a scoping review to appraise available international evidence concerning to the effectiveness and usefulness of PBL methodology in undergraduate medical teaching programs. We applied the Arksey and O’Malley framework to undertake a scoping review. A search of literature published in English and Spanish identied one hundred and twenty four publications eligible for this review. Despite the fact that this review includes many studies, their design is heterogeneous and only a few provide a high scientic evidence methodology (randomized design and/or systematic review with meta-analysis). Furthermore, most are single-center experiences with small sample size and there are no large multi-center studies. PBL methodology obtains a high level of satisfaction, especially among students. It is more effective than other more traditional (or lecture-based methods) at improving social and communication skills, problem-solving and self-learning skills. Knowledge retention and academic performance are no worse (and in many studies are better) than with traditional methods. PBL is not universally widespread, probably because it requires greater human resources and continuous training for its implementation. More comparative and randomized studies and/or other systematic reviews and meta-analysis are required to determine which educational strategies are the most suitable for training the doctors of the future. the

With the goals of the study established, the four members of the research team established the research questions. The primary research question was "Which is the effectiveness of PBL methodology for teaching in undergraduate medicine?" The secondary questions include "Which is the perception and satisfaction of medical students and tutors in relation to PBL methodology?" Stage 2: Identifying relevant studies After the research questions and a search strategy were de ned, the searches were conducted in PubMed and Web of Science using the MeSH terms "problembased learning" and "Medicine" (the Boolean operator "AND" was applied to the search terms). No limits were set on language, publication date, study design or country of origin. The search was carried out on 14th February 2021. Citations were uploaded to the reference manager software Mendeley Desktop (version 1.19.8) for title and abstract screening, and data characterization. Stage 3: Study selection The searching strategy in our scoping study generated a total of 2399 references. The literature search and screening of title, abstract and full-text for suitability was performed independently by one author (JCT) based on predetermined inclusion criteria. The inclusion criteria were: 1) PBL methodology was the major research topic; 2) participants were undergraduate medical students or tutors; 3) the main outcomes were one of the following: knowledge retention, social and communication skills and/or student/tutor satisfaction; 4) all types of studies were included including descriptive papers, qualitative, quantitative and mixed studies methods, perspectives, opinion, commentary pieces and editorials. Exclusion criteria were studies including other types of participants such as postgraduate medical students, residents and other health non-medical specialties such as pharmacy, veterinary, dentistry or nursing. Studies published in languages other than Spanish and English were also excluded. Situations in which uncertainty arose, all authors (CB, BS, RP) discussed the publication together to reach a nal consensus. The outcomes of the search results and screening are presented in Fig. 1. One-hundred and twenty-four articles met the inclusion criteria and were included in the nal analysis.
Stage 4: Charting the data A data extraction table was developed by the research team. Data extracted from each of the 124 publications included general publication details (year, author, and country), sample size, study population, design/methodology, main and secondary outcomes and relevant results and/or conclusions. We compiled all data into a single spreadsheet in Microsoft Excel for coding and analysis. The characteristics and the study subject of the 124 articles included in this review are summarized in Tables 1, 2 and 3. The detailed results of the Microsoft Excel le are also shown in Table 4.

Results
Stage 5: Collating, summarizing and reporting the results As indicated in the search strategy ( Fig. 1) this review resulted in the inclusion of 124 publications. Publication years of the nal sample ranged from 1990 to 2020 and the distribution of publications by years is shown in Fig. 2. The majority of the publications (51, 41%) were identi ed for the years 2010-2020 and the years in which there were more publications were 2001, 2009 and 2015. Countries from the six continents were represented in this review. Most of the publications were from Asia (especially China and Saudi Arabia) and North America followed by Europe, and few studies were from Africa, Oceania and South America. The country with more publications was the United States of America (n = 27). The most frequent designs of the selected studies were surveys or questionnaires (n = 45) and comparative studies (n = 48, only 16 were randomized) with traditional or lecture-based learning methodologies (in two studies the comparison was with simulation) and the main outcome was performance followed by student satisfaction (45 studies measured more than one outcome). The studies with the highest level of scienti c evidence (systematic review and meta-analysis and randomized studies) were conducted mostly in Asian countries (Tables 1 and 3). The study subject was speci ed in 81 publications nding a high variability but at the same time great representability of almost all disciplines of the medical studies ( Table 2).
The sample size was available in 99 publications and the median [range] of the participants was 132 . According to study population, there were more participants in the students' focused studies (median 134 and range 16-2061) in comparison with the tutors' studies (median 53 and range 14-494) Finally, after reviewing in detail the measured outcomes (main and secondary) according to the study design (Tables 3 and 4) we present a narrative overview and a synthesis of the main ndings.

Learning and knowledge retention
Seventy-ve of the 124 publications had learning and/or knowledge retention as their main or secondary outcomes, most of them (45) were comparative studies with traditional or lecture-based learning and 16 were randomized. These studies were varied in their methodology, were performed in different geographic zones, and normally analysed the experience of just one education centre. Most studies (53) reported superiority of PBL in learning and knowledge acquisition (Sokas et al. 1990;Richards et al. 1996;Gresham & Philp 1996;Hill et al. 1998;Blake & Parkison 1998;Hmelo 1998;Finch 1999;Casassus et al. 1999;Purdy et al. 1999;Farrell et al. 1999;Finch 1999;Curtis et al. 2001;Trevena & Clarke 2002;Astin et al. 2002;Whit eld et al. 2002;Whit eld et al. 2002;McParland et al. 2004;Casey et al. 2005;Gurpinar et al. 2005;Tamblyn et al. 2005;Abu-Hijleh et al. 2005;Distlehorst et al. 2005;Distlehorst et al. 2005;Hoffman et al. 2006;Kong et al. 2009;Tsou et al. 2009;Tsou et al. 2009;Wang et al. 2010;Abou-Elhamd et al. 2011;Urrutia et al. 2101;Tian et al. 2012;Hoover et al. 2012;Li et al. 2013;Ding et al. 2014;Meo 2014;Khoshnevisasl et al. 2014 (Steadman et al. 2006). It is noteworthy that the four systematic reviews and meta-analysis included in this scoping review, all carried out in China, found that PBL was more effective than lecture-based learning in improving knowledge and other skills (clinical, problem-solving, self-learning and collaborative) (Ding et al. 2014;Zhang et al. 2018;Ma et al. 2019;Liu et al. 2020). Another relevant example of the superiority of the PBL method over the traditional method is the experience reported by (Hoffman et al. 2006) from the University of Missouri-Columbia. The authors analysed the impact of implementing the PBL methodology in its Faculty of Medicine and revealed an improvement in the academic results that lasted for over a decade.

Social and communication skills
We found ve studies in this scoping review that focussed on these outcomes and all of them described that a curriculum centred on PBL seems to instil more con dence in social and communication skills among students. Students perceived PBL positively for team work, communication skills and interpersonal relations (Seneviratne et al. 2001;Suleman et al. 2010;Hande et al. 2015;Al-Shaikh et al. 2015;Mughal & Shaikh 2018).

Student satisfaction
Fifty-six publications analysed student satisfaction with PBL methodology. The most frequent methodology were surveys or questionnaires (29 studies) followed by comparative studies with traditional or lecture-based methodology (19 studies, 7 of them were randomized). Almost all the studies (47) have shown that PBL is generally well-received (Sokas et al. 1990;Blosser & Jones 1991;Usherwood 1991;Bernstein et al. 1995;Kaufman & Mann 1996;Kalaian & Mullan 1996;Gresham & Philp 1996;Vincelette 1997;Kaufman & Mann 1999;Antepohl & Herzig 1999;Casassus et al. 1999;Purdy et al. 1999;Farrell et al. 1999;Ghosh & Dawka 2000;Dyke et al. 2001;Walters 2001;Leung et al 2001;Khoo et al. 2001;Villamor 2001;Curtis et al. 2001;Trevena & Clarke 2002;Chang et al. 2004;McLean 2004;Casey et al. 2005;Abu-Hijleh et al. 2005;Lucas et al. 2006;Burgun et al. 2006;Gurpinar et al. 2009;Suleman et al. 2010;Wang et al. 2010;Tian et al. 2012;Elzubeir 2012;Sulaiman & Hamdy 2013;Albarrak et al. 2013;Li et al. 2013;Meo 2014;Khoshnevisasl et al. 2014;Nosair et al. 2015;González et al. 2015;Tshitenge et al. 2017;Eltony et al. 2017;Yadav et al. 2018;Asad et al. 2019;Mpalanyi et al. 2020;Korkmaz & Ozcelik 2020;Li et al. 2020;Liu et al. 2020) but in 9 studies the overall satisfaction scores for the PBL program were neutral (Macallan et al. 2009;Grisham et al. 2015;Khan & Al-Swailmi 2015;Alduraywish et al. 2017;Yoo et al. 2019;Aldayel et al. 2019;Hu et al. 2019) or negative (DeLowerntal 1996;Tufts 2009). Some factors that have been identi ed as key components for PBL to be successful include: a small group size, the use of scenarios of realistic cases and good management of group dynamics. Despite a mostly positive assessment of the PBL methodology by the students, there were some negative aspects that could be criticized or improved. These include unclear communication of the learning methodology, objectives and assessment method; bad management and organisation of the sessions; tutors having little experience of the method; and a lack of standardisation in the implementation of the method by the tutors. Tutor satisfaction There are only 12 publications that analyze the satisfaction of tutors, most of them surveys or questionnaires (Bernstein et al. 1995;Vincelette et al. 1997;Khoo et al. 2001;Gurpinar et al. 2009). In comparison with the satisfaction of the students, here the results are more neutral (Macallan et al. 2009;Subramaniam et al. 2014;Khan & Al-Swailmi 2015;Chang 2016;Yoo et al 2019) and even unfavorable to the PBL methodology in two publications (Gri th et al 1996;De Lowerntal 1996). PBL teaching was favored by tutors when the institutions train them in the subject, when there was administrative support and adequate infrastructure and coordination (Navarro & Zamora 2014). In some experiences, the PBL modules created an unacceptable toll of anxiety, unhappiness and strained relations.

Other outcomes and descriptive experiences
The effectiveness of the PBL methodology has also been explored in other outcomes such as the ability to solve problems and to self-directed learning. All studies have shown that PBL is more effective than lecture-based learning in problem-solving and self-learning skills (Casassus et al. 1999;Seneviratne et al. 2001;Whit eld et al. 2002;Ding et al. 2014;Balendran & John 2017;Mughal & Shaikh 2018;Yadav et al. 2018;Villamor 2001;Demiroren et al. 2016).
Finally, other publications have reported the experience of some faculties in the implementation of the PBL methodology. Different experiences have demonstrated that it is both possible and feasible to shift from a traditional curriculum to a PBL program, recognizing that PBL methodology is complex to plan and structure, needs a large number of human and material resources, requiring an immense teacher effort (Desmarchais 2003;Chang et al. 2004;Tamblyn et al. 2005;Hoffman et al. 2006). In addition, and in spite of its cost implication, a PBL curriculum can be successfully implemented in resourceconstrained settings (Amoako-Sakyi & Amonoo-Kuo 2015;Carrera et al. 2003).

Discussion
We conducted this scoping review to explore the effectiveness and satisfaction of PBL methodology for teaching in undergraduate medicine and, to our knowledge, it is the rst study of its kind that has been carried out. PBL methodology is implemented in medical studies on the 6 continents but there is more experience (or at least more publications) from Asian countries and North America. In spite of its di culties on implementation, a PBL curriculum can be successfully implemented in resource-constrained settings (Amoako-Sakyi & Amonoo-Kuo 2015; Carrera et al. 2003). Although it is true that the studies with the highest level of scienti c evidence (randomized studies and meta-analysis) were carried out mainly in Asian countries (and some in North America and Europe), there were no signi cant differences in the main results according to geographical origin.
In this scoping review we have included a large number of publications that, despite their heterogeneity, tend to show favorable results for the usefulness of the PBL methodology in teaching and learning medicine. The results tend to be especially favorable to PBL methodology when it is compared with traditional or lecture-based teaching methods, but when compared with simulation it is not so clear. There are two studies that show neutral (Wenk et al. 2009) or superior (Steadman et al. 2006) results to simulation for the acquisition of speci c clinical skills. It seems important to highlight that the four meta-analysis included in this review, which included a high number of participants, show results that are clearly favorable to the PBL methodology in terms of knowledge, clinical skills, problem-solving, self-learning and satisfaction ( Regarding the level of satisfaction described in the surveys or questionnaires, the overall satisfaction rate was higher in the PBL students when compared with traditional learning students. Students work in small groups, allowing and promoting teamwork and facilitating social and communication skills. As sessions are more attractive and dynamic than traditional classes, this could lead to a greater degree of motivation for learning.
These satisfaction results are not so favorable when tutors are asked and this may be due to different reasons; rst, some studies are from the 90s, when the methodology was not yet fully implemented; second, the number of tutors included in these studies is low; and third, and perhaps most importantly, the complaints are not usually due to the methodology itself, but rather due to lack of administrative support, and/or work overload. PBL methodology implies more human and material resources. The lack of experience in guided self-learning by lecturers requires more training. Some teachers may not feel comfortable with the method and therefore do not apply it correctly.
Despite how effective and/or attractive the PBL methodology may seem, some (not many) authors are clearly detractors and have published opinion articles with erce criticism to this methodology. Some of the arguments against are as follows: clinical problem solving is the wrong task for preclinical medical students, self-directed learning interpreted as self-teaching is not appropriate in undergraduate medical education, relegation to the role of facilitators is a misuse of the faculty, small-group experience is inherently variable and sometimes dysfunctional, etc. (Shanley 2007).

Limitations
Scoping reviews are not without limitations. Our review includes 124 articles from the 2399 initially identi ed and despite our efforts to be as comprehensive as possible, we may have missed some articles. Despite the fact that this review includes many studies, their design is very heterogeneous, only a few include a large sample size and high scienti c evidence methodology. Furthermore, most are single-center experiences and there are no large multi-center studies.
However, the adoption of a scoping review methodology was effective in terms of summarizing the research ndings, identifying limitations in studies' methodologies and ndings and provided a more rigorous vision of the international stat of the art.

Conclusions
This systematic scoping review provides a broad overview of the e cacy of PBL methodology in undergraduate medicine teaching from different countries and institutions. PBL is not a new teaching method given that it has already been 50 years since it was implemented in medicine courses. It is a method that shifts the leading role from teachers to students and is based on guided self-learning. If it is applied properly, the degree of satisfaction is high, especially for students. PBL is more effective than traditional methods (based mainly on lectures) at improving social and communication skills, problem-solving and selflearning skills, and has no worse results (and in many studies better results) in relation to knowledge retention and academic performance. Despite that, its use is not universally widespread, probably because it requires greater human resources and continuous training for its implementation. In any case, more comparative and randomized studies and/or other systematic reviews and meta-analysis are required to determine which educational strategies could be most suitable for the training of future doctors.