시뮬레이션 교육에서 방향이 제시된 자기조절학습 vs 교수자 지도의 학습(Med Educ, 2012)

시뮬레이션 교육에서 방향이 제시된 자기조절학습 vs 교수자 지도의 학습(Med Educ, 2012)

Directed self-regulated learning versus instructor-regulated learning in simulation training

Ryan Brydges,1 Parvathy Nair,2 Irene Ma,3 David Shanks2 & Rose Hatala2

자기조절학습은 다음과 같이 정의된다. 많은 의료전문직 관련 기관들은 SRL을 비전과 미션에 포함시키고 있으며 이 때 SDL이나 평생학습과 같은 용어를 사용한다. SRL과 SDL이 여러 공통점이 있지만, 의학교육연구에서 이 둘은 구분되어 사용된다.

Self-regulated learning has been defined as a process involving ‘self-generated thoughts, feelings and actions that are planned and cyclically adapted to the attainment of personal goals’.15 Many medical professional agencies have included SRL in their visions and mission statements, using terms like ‘self- directed learning’ (SDL) and ‘lifelong learning’.16 Although SRL and SDL have many similarities,17 medical education researchers have used the terms differently: 

• SDL is often invoked when a learning environment is designed to promote autonomous learning (e.g. problem-based learning), whereas (학습환경이 자율적 학습을 촉진하도록 설계된 것)
• SRL is mentioned when the focus is on understanding the mechanisms of autonomous learning in order to identify how best to support learners when they engage in SRL.18 (자율적 학습의 메커니즘 이해에 초점을 두고 있으며, 이를 통해서 어떻게 SRL을 하는 학습자를 가장 잘 지원할 수 있는가를 알기 위한 것)

IRL과 DSRL을 비교하였다.

We compared this traditional approach of instructor-regu- lated learning (IRL) with what we call ‘directed self- regulated learning’ (DSRL).

절차 Procedure

참가자들은 baseline 설문지를 작성하였다.

Participants completed a baseline questionnaire on their previous experience with LP training (with both simulation and real patients) and reported their baseline confidence in performing LP on an 11-point Likert scale.

DSRL의 프로토콜 Protocol for directed self-regulated learning

IRL의 프로토콜 Protocol for instructor-regulated learning

네 명의 평가자가 global rating scale (GRS)와 checklist (CL)을 사용하여 pre, post, retention test를 평가하였다. 각 비디오에 대해서 두 명이 평가하였고, 공평하게 분배했다. GRS의 validity, reliability는 psychometric study로 supported 된다. 우리는 기존에 validity와 reliability가 확보된 CL을 가지고 그것의 변형된 버전을 만들었다. 일부 상관없는 문항을 삭제하였다. 비록 이러한 일부 문항의 삭제가 전문가 합의과정을 통해서 결정되었지만 validity를 다시 평가하지는 않았다. 대신 ICC를 계산하였다.

Four trained, blinded expert raters (authors RH, IM, PN, DS) used a global rating scale (GRS) (Appen- dix S1, online) and a procedural checklist (CL) to independently evaluate participants’ videotaped per- formances on the pre-test, post-test and retention test. Two raters evaluated each video; the rating load was distributed equally amongst the four raters. The concurrent validity, construct validity and reliability of the GRS are supported by a psychometric study on procedural skills.24 We created a modified version of a CL with demonstrated content validity and reli- ability,11 which consisted of 26 major and 44 minor actions. We removed some items fromthe original CL that we deemed as non-applicable to our simulation scenario, which shortened the overall CL, but main- tained an emphasis on major actions; specifically, we included 21 major actions and 14 minor actions (Appendix S2). Although the removal of items was based on expert consensus, we did not re-evaluate the validity of our modified CL; however, we did assess inter-rater reliability using the intra-class correlation coefficient (ICC).

평가자 훈련 Rater training

Training protocol

Prior to rating the videotaped performances, the four raters engaged in training. The training protocol involved the rating of three randomly selected videos, a meeting to discuss disagreements, the rating of another three randomly selected videos and a meet- ing to discuss these, followed by a final rating and discussion of four randomly chosen videos.

여기서 사용한 10개의 비디오는 ICC에서 계산시 포함하지 않았음.

We did not use the ratings of those 10 videos in the ICC reliability calculation, but did include them in the remaining analyses. During the training, the four raters developed a shared understanding of how to use the GRS and CL. One significant training outcome was the raters’ mutual definition of resident competence as ‘a resident being capable of performing LP in the clinical context with direct supervision’.

통계 분석

For these two performance variables, we assessed group differences between the DSRL and IRL groups on the pre-test and post-test, and in the pre-test, post- test and retention test scores using two separate repeated-measures analyses of variance (ANOVAs) with test as the within-subjects factor and group as the between-subjects factor. We also computed Pearson correlation coefficients to assess the relationship between participants’ self-reported confidence and the performance variables at the pre- and post-tests. Finally, we used separate independent samples t-tests to examine group differences in participants’ self- reported LP experience, total number of LPs per- formed or observed, total time with an instructor, and total practice time. We used Tukey’s honestly signif- icant differences (HSD) test and calculated the bias- corrected Hedges’ g effect size for the appropriate post hoc comparisons. All data are reported as mean ± standard error (SE) and the alpha level was set at p < 0.05.

본 연구에 대한 설명 중 하나는, 교수자가 연습 중에 제공한 피드백이 학습에 큰 도움이 된다고 여기는 것과 관련되어있다. 그러나 안타깝게도 motor learning에 관한 수십년의 연구에서 동시적 피드백(concurrent feedback)이 오히려 장기 학습에 해롭다는 결과가 나왔다. 전형적인 형태는, 연습 중에는 그것이 도움이 되는 것으로 보이지만, 피드백이 불가능한 상황인 retention 평가에서는 급격한 하락이다.

One explanation for our findings is associated with the provision by instructors of feedback during practice, which is commonplace in many domains and is considered a potent aid to learning. Unfortu- nately, decades of motor learning research have shown that concurrent feedback can be detrimental to long-term learning.26 Typically, findings show a benefit during practice, resulting in rapid skill acquisition, followed by poor performance on reten- tion tests when the feedback is unavailable.27

연습하는 중간에 제공되는 동시적 피드백의 촉진적 특성은 무시하기 어려우며 IRL이 의학에서 왜 그렇게 널리 쓰이는지를 설명해준다. 그러나 이러한 직관에도 불구하고 우리의 결과는 performance-learning paradox를 지지하는 근거를 하나 더 추가했을 뿐이다. (즉각적인 이득이 나타나나, 이것이 장기적인 학습을 반영하지는 못한다는 것) 이러한 결과가 직관에는 반하는 것일 지 모르나, 수많은 연구들이 장기적 학습에 있어서 '바람직한 어려움'의 활용을 지지한다.

The facilitating nature of concurrent feedback during practice is difficult to ignore and probably explains how pervasive IRL is in medicine. Despite these intuitions, our results add to evidence supporting the performance–learning paradox, which refers to the common finding that immediate performance bene- fits (as demonstrated in post-test scores) do not always reflect long-term learning (as demonstrated in retention test scores).28 Counterintuitive as this paradox may be, a wealth of evidence supports the use of ‘desirable difficulties’ to depress immediate performance and improve long-term learning.28,29

두 번째 개념틀은 자기모니터링이다. Eva와 Regehr는 자기모니터링을 '자신이 특정 상황에서 필요한 기술과 지식을 유지할 수 있는 가능성에 대한 순간순간의 판단'이라고 정의한다.

A second conceptual framework for interpreting the findings comes fromresearch on self-monitoring. Eva and Regehr31 defined self-monitoring as ‘a moment- by-moment awareness of the likelihood that one maintains the skill or knowledge to act in a particular situation’.

본 연구에서 DSRL그룹은 여러 시뮬레이터 모델 사이에서 점차 발전해나가며 자신의 능력과 준비도를 자기모니터링할 수 있다.

In the present study, the DSRL group’s opportunity to progress between simulator models may have led participants to self-monitor their ability and ‘readiness’ frequently.

교수들이 사용할 수 있는 시간은 제한적이고 이것이 흔히 simulation-based training 도입의 장애로 여겨진다. 연구자들은 교육효율성을 표준 성과 변인으로 고려해볼 필요가 있다.

Given that limitedfaculty staff time is a frequently cited barrier to the implementation of simulation-based training,33 researchers may benefit fromconsidering training efficiency (e.g. use of faculty staff time) as a standard outcome variable.

그러나 총 hands-on simulator time으로 계산하면 이 주장은 전혀 반대가 된다. (DSRL에서 hands-on time이 훨씬 많았음)

However, when we consider the total hands-on simu- lator time, the argument is reversed.

특히 DSRL을 위해서는 더 많은 자원이 필요하고, 어떻게 이러한 자원 필요량의 균형을 맞출 것인지가 중요한 문제일 것이다.

Further, more resources (e.g. LP trays and equipment) were needed for the DSRL than the IRL group. Determining how to balance these resource concerns represents a major challenge for simulation researchers.

---

Med Educ. 2012 Jul;46(7):648-56. doi: 10.1111/j.1365-2923.2012.04268.x.

Directed self-regulated learning versus instructor-regulated learning in simulation training.

Brydges R1, Nair P, Ma I, Shanks D, Hatala R.

Author information

• 1Department of Medicine, University of Toronto, Toronto and The Wilson Centre, University Health Network, Toronto, Ontario, Canada. ryan.brydges@utoronto.ca

Abstract

OBJECTIVES:

Simulation training offers opportunities for unsupervised, self-regulated learning, yet little evidence is available to indicate the efficacy of this approach in the learning of procedural skills. We evaluated the effectiveness of directed self-regulated learning (DSRL) and instructor-regulated learning (IRL), respectively, for teaching lumbar puncture (LP) using simulation.

METHODS:

We randomly assigned internal medicine residents in postgraduate year 1 to either DSRL ('directed' to progress from easy to difficult LP simulators during self-regulated learning) or IRL (in groups of four led by an instructor). All participants practised for up to 50 minutes and completed a pre-test, post-test and delayed (by 3 months) retention test on the simulator. Pairs of blinded trained experts independently rated all videotaped performances using a validated global rating scale and a modified version of a validated checklist. Participants provided measures of LP experience and self-reported confidence. We analysed the pre-post (n = 42) and pre-post-retention performance scores (n = 23) using two separate repeated-measures analyses of variance (anovas) and computed Pearson correlation coefficients between participants' confidence and performance scores.

RESULTS:

Inter-rater agreement was strong for both performance measures (intra-class correlation coefficient > 0.81). The groups achieved similar pre-test and post-test scores (p > 0.05) and scores in both groups improved significantly from the pre- to the post-test (p < 0.05). On retention, a significant interaction (F(2,42) = 3.92, p = 0.03) suggests the DSRL group maintained its post-test performance, whereas that in the IRL group dropped significantly (p < 0.05). Correlations between self-reported confidence and post-test performance were positive and significant for the DSRL group, and negative and non-significant for the IRL group.

CONCLUSIONS:

Both IRL and DSRL led to improved LP performance immediately after practice. Whereas the IRL group's skills declined after 3 months, the DSRL group's performance was maintained, suggesting a potential long-term benefit of this training. Participants in the DSRL group also developed a more accurate relationship between confidence and competence following practice. Further research is needed to clarify the mechanisms of self-regulated learning and its role in simulation contexts.

© Blackwell Publishing Ltd 2012.

PMID:

 

22691145

 

[PubMed - indexed for MEDLINE]