Stage one: identifying the research question

Stage two: identifying relevant studies

A literature search was carried out to identify studies and reports between 1966 and February 2017 using the databases Medline (Ovid), ERIC, EBSCO and Google Scholar. The main search term “self-regulated learning” was combined with key terms and variants of “medical student”, “medical school” and “medical education”. The search identified 911 potential papers. The following criteria were used for the review and selection of the studies: available in English, focused on SRL in the medical clinical learning environment and relevant to the topic after review of abstracts. Articles investigating similar but different theories such as self-directed learning were not included due to conceptual differences [16, 17]. The first author (KC) screened the titles and abstracts for relevance to the research questions, and 890 articles were excluded as they did not meet the inclusion criteria.

Stage three: study selection

The full texts of the remaining 21 articles were read by KC and 7 articles were excluded as the studies focused solely on non-clinical medical students. The reference lists of all selected publications were then hand searched for any additional relevant studies. Authors of key papers were also contacted for relevant reports or publications. Fourteen articles were selected for inclusion in the review (see Fig. 1).

Stage four: data charting and collation

Stage five: summarising and reporting the results

The following Results section summarises and discursively reports the review findings.

What have studies found regarding self-regulated learning in the medical clinical years?

There were 3 major findings regarding SRL in the clinical years: 1) changes in SRL occur in the clinical learning environment, 2) higher levels of SRL are associated with higher academic achievement, more success in clinical skills and better mental health outcomes and 3) certain factors can influence SRL levels. These are further explained below.

Changes in self-regulated learning occur in the clinical learning environment

Three studies qualitatively explored how students adapt to different clinical environments. White found that students from a PBL curriculum adapted more effectively and used more SRL skills when transitioning to the clinical environment than students from a traditional curriculum [18]. In Woods et al.’s study, students showed high levels of SRL as they adapted to a surgical rotation but lacked critical self-reflection [10]. Berkhout et al.’s observed in their study that experienced students required less support and were more likely to create their own learning goals compared to novice students [19].

Self–regulated learning and academic achievement

Five studies reported correlations between levels of SRL and academic achievement during clerkship [20–24]. Sobral’s and Song et al.’s studies found associations between aspects of SRL and achievement in written exams during the clinical years [21, 22], whilst Turan and Konan found no such relationship between SRL scores and written examination results [20]. In regard to success in clinical skills three studies found associations [20, 22, 23]. Turan and Konan found weak correlations between SRL scores and Objective Structured Clinical Examination (OSCE) results [20], Sobral found associations between SRL scores and diagnostic skill [22] and Cleary et al. found a link between SRL and success in cannulation [23]. One study found SRL was negatively correlated with procrastination and positively correlated with mastery goal structures, two variables linked with academic success [24].

Self-regulated learning and mental health

Nguyen et al. [25] observed that most SRL scores were negatively associated with depression.

Factors influencing self-regulated learning levels

This review found the SRL of medical students can be affected by a wide range of variables [5, 18, 22, 26–28]. Two studies found that experience was linked with higher levels of SRL [19, 26]. Alegria et al.’s study suggests that technological resources such as tablet computers can further cultivate SRL levels by allowing students to access clinical and educational information [5]. Another study reports that a PBL curriculum can support the SRL of students especially during the transition to the clinical environment [18, 19]. Sobral’s study suggests reflection, one aspect of SRL, can be systematically improved through specific interventions [22]. Berkhout et al.’s 2015 study reports the SRL of clinical medical students is influenced by a range of variables including personal, contextual, social factors as well as experienced autonomy [27]. Lyons-Warren et al., suggest SRL is hindered when students feel there is a lack of flexibility when pursuing their learning needs [28]. Berkhout et al.’s 2017 study proposes five SRL patterns exist for clinical students, with each pattern requiring a unique approach to support learning [29].

How has self-regulated learning in medical students in the clinical learning environment been measured?

Methods used to study SRL were varied and included qualitative, quantitative and mixed-method approaches.

Qualitative processes such as semi structured interviews and focus groups were used in five studies [5, 10, 18, 19, 27]. One other study used microanalytic assessment [23]. Microanalysis was originally used to explore the differences in SRL between novice and professional sportsmen [30], and is a structured process which involves verbal responses to open–ended questions, targeting forethought, performance and reflection processes during a specific task. The responses are recorded verbatim then coded [31].

Quantitative approaches included the Self-Regulated Learning Measure for Computer-based learning (SRMC) [21], the Self–Regulated Learning Perception Scale [26] and the Motivated Strategies for Learning Questionnaire (MSLQ) or a modified version of this instrument [20, 25].

The SRMC tests the use and frequency of 10 SRL subcategories (self-evaluation, organizing and transforming, goal setting and planning, seeking information, keeping records and monitoring, environmental structuring, self-consequences, rehearsing and memorizing, seeking social assistance, and reviewing records) and two classes of non-self regulation behavior (will power and non-applicable statements). The tool has high internal consistency and high inter-rater reliability [21].

The Self–Regulated Learning Perception Scale uses 41 items to measure 4 domains; 1) motivation and action to learning 2) planning and goalsetting 3) strategies for learning and assessment and 4) lack of self–directedness [26]. The scale is highly reliable.

Three articles used the MSLQ [20, 24, 25], which was developed by Pintrich in 1991 [32]. The MSLQ contains 81, 7-point Likert type scale questions measuring 15 subscales of SRL. The MSLQ has variable internal consistency depending on the subscale and high validity [2].

One study used Q-methodology, a method that uses features of both quantitative and qualitative measures [29]. In the Q-methodology participants sort a set of statements along a continuum of a fixed grid (from not at all applicable to me, to very applicable to me) and are asked to explain their rationale for their sorting choices. Similar patterns are identified in the population, and the resulting patterns are interpreted and described as shared perspectives.

Coverage of the SRL theoretical framework by the included studies

Changes in self-regulated learning occur in the clinical learning environment

Two studies that investigated clinical transition highlighted the importance of self-regulating learning to maximise experiential learning [10, 18]. The importance of self–regulated learning during the transition to clerkship is not surprising as supervisors are often pre-occupied with patient care and sometimes not interested to teach [35, 36]. However, this scoping review did not find any quantitative studies that explored the preclinical to clinical transition using a validated tool. Thus future studies could explore this transition using a validated quantitative approach to provide further insights into the topic.

Beyond the transition period, Berkhout et al.’s study suggests SRL continues to develop in the clinical environment, with novice and experienced learners having different needs to support their SRL. The study notes that novice students require more support from others, specifically residents and peers, to help them formulate learning goals and navigate the new learning environment. In the broader medical education literature, the significance of resident teaching is mixed, with some studies highlighting the correlation between resident teaching and medical student learning [37] and academic performance [38], whilst other studies not finding such an effect [39, 40].

Self–regulated learning, academic achievement and mental health

This review indicates that higher levels of SRL may be beneficial for medical students due to its positive correlation with academic achievement and clinical skills [20–24] as well as its negative association with depression [25].

The positive correlation between SRL and diagnostic and clinical skills is important for students not just when they are in medical school but also as practising physicians. Furthermore, as doctors must maintain their competencies and skills throughout their careers in order to consistently meet the high standards of patient care [41–43], they must develop an awareness of their learning needs and use appropriate learning strategies to achieve their goals. As lifelong learners, doctors should adopt SRL and be motivationally, behaviourally and meta-cognitively proactive in their own learning process [44–46]. Developing SRL in medical students in the clinical environment is important not only for the short term, so that medical students may achieve better grades and clinical proficiency, but also for the long-term so that graduates can participate in lifelong learning and provide patients with effective care.

The negative association between SRL and depression is significant because medical students as a whole face higher levels of psychological stress than the general population [47, 48], with studies suggesting that rates of depression increase during the clinical years [49, 50]. Moreover some evidence suggests that mental distress during medical school is associated with problems as physicians [51, 52], which might negatively affect patient care [53]. Although the link between SRL and decreased depression is correlational and not causative, the association between aspects of SRL and mental health has been found in other studies involving preclinical medical students [54] as well as in studies outside medical education [55, 56]. Thus it is may be relevant to promote SRL strategies to help students optimise their mental well-being.

What factors influence the self-regulated learning levels of medical students in the clinical learning environment?

In their study, Turan et al. found that medical students in the clinical years appeared to intrinsically have higher levels of SRL than preclinical medical students [26]. Although it is reassuring to believe self-regulating learning naturally develops and increases as students progress through medical school and postgraduate training, caution must applied when interpreting the results as the study explored the changes in SRL in separate cohorts rather than following its development in the same student cohort. Indeed numerous authors in the broader literature have suggested that self-regulated learners are not always successful when left to develop their own strategies [57–60].

Finally it appears that at least some aspects of SRL can be targeted and improved, with one study suggesting that levels of reflection in students can be increased after an intervention. This finding that SRL can be improved by certain interventions is supported by research outside the transitions literature [42, 61–63], with one author suggesting that explicitly teaching metacognition is efficacious [61].

How has self-regulated learning in medical students in the clinical learning environment been measured?

Several approaches have been used to measure SRL in clinical medical students and there appears to be no single best approach with each approach having its advantages and disadvantages. For example, studies using the qualitative approach were effective in identifying and exploring contextual factors related to SRL in the clinical environment whereas the quantitative approach could identify the strengths of associations to test hypotheses. Two of the three studies using quantitative methods developed their own instruments specifically for use in the clinical context [21, 26]. In terms of microanalytic assessment, the tool has been reported in the literature to be effective in examining motor and cognitive skills that have a clear beginning, middle and end section [64]. Thus microanalytic assessment appears to be a suitable method to analyse the SRL of students during specific tasks such as taking a history and examining a patient or skills such as cannulation, but not to measure global SRL that may occur with informal learning in the clinical environment. Q methodology measures subjective experiences through features of qualitative and quantitative approaches, and has been reported to be a more robust technique than Likert-type scales to study attitudes in health education [65].

Studies which investigated SRL across multiple cohorts reported the cross-sectional design as a limitation. These authors suggest longitudinal studies may allow the transition to be better measured. Studies using questionnaires as their collection tool reported reliability and validity limitations, especially as they measured self-reported learning behaviours, which are subject to social desirability and recall bias. However each survey instrument had reasonable psychometric properties. Some studies also reported the possibility of selection bias either due to small sample sizes [24] or due to inherent differences in the selected students [18]. Future studies could thus consider using longitudinal study design using a range of methods to better understand how SRL develops and is maintained in the clinical environment.

Not identified in our search were novel qualitative methods used in the broader SRL literature, such as think-aloud protocols [66, 67], video-taped events [68], structured study-diaries [69] and computer trace analysis [70, 71]. Future studies could use these novel methods to measure the SRL of clinical medical students.

Limitations

This review was restricted to articles published in the English language and those concerning the medical profession. We also did not include studies with similar terms to SRL such as self-directed learning in our review due to conceptual differences in perspectives and constructs. We did include medical and educational databases as well as Google Scholar to widen the search to where SRL research might be catalogued. Future research should consider including nursing and other allied health professionals as learning in the clinical environment is also a core part of health professional education and findings about effective learning could be transferrable across professions. As this article is a scoping review and not a systematic review, we did not critically appraise the selected studies. We did however locate findings which addressed our aims of examining the nature and extent of the literature, and to identify potential future research directions.