Analyzing Patterns of Participant Behavior in Integrated and Distributed Simulation Using Semi-Structured Learning Design (SSLD) Statements

Introduction: Although simulation has normalized around single mannequin team scenarios, there are many other forms. The authors have in recent years been exploring the possibilities of new web technologies connecting previously isolated simulators to provide innovative structured learning environments. Two key research questions have been tackled by the HSVO Project: (1) how do one or more groups of teachers and learners behave when using multiple simulators in a single educational activity and (2) how do they behave when participating in multi-site synchronous simulation activities? Methods: A series of ten different sessions were run involving learners from four geographically distant medical schools. Scenarios were developed that used multiple simulators (virtual patients, mannequins and part task trainers) in different combinations (single group, site-specific groups, cross-site groups, etc) to be run concurrently across two sites at a time. Each session involved a different scenario or a different configuration of a scenario. A web-based platform was developed to support the execution of multiple device and multiple location scenarios1. This study was passed by the ethics boards (or equivalent) of all participating institutions. Groups of learners at each site were recruited through the local research lead. Each session was defined using a structured script called a semi-structured learning design statement (SSLD). Video recordings were made of all sessions. Participants were also given a pre-session evaluation to gauge their prior experience along with a post-session evaluation of their experiences. Results: the video recordings were selected as the primary source of analysis. The mean session duration was 90 minutes and the median 76 minutes. The mean size of a learner group (at a single site) was 4.9 and the median 4. Prior experience and attitude was relatively normalized within a site but variant between sites. Analysis: A set of codes representing different events and behaviors was developed from an initial review by two coders and refined through a first round analysis of all 10 videos. Coding involved watching each video and recording the observed behaviors and events in two-minute segments. Codes were then grouped under three categories: engagement, communication and environment. The coder pool was expanded to 5 and a second pass of coding was conducted. This yielded a reasonably congruent model of key events and behaviors in each session. At the same time, a time-based sequence of each key event in the session was recorded and compared to the planned sequence and timing in the corresponding SSLD. Discussion: technical issues did not prove to be significant inhibitors; indeed several tutors turned such events into opportunities for further teaching and reflection. Scenarios and sessions that involved more active learning led to greater reflection and engagement amongst the learners. Using more simulators also led to higher level thinking about the nature of practice and their preparation for it. All scenarios where adapted and extemporized with more time given over to briefing and orientation and less to feedback than had originally been planned. The role of the tutor was identified as critical with those that took a more facilitative role proving more effective than those that took a didactic role. Analyses of this rich data set are ongoing. Conclusions: this is a pilot study investigating a new and complex area of simulation-based education and, as such, all findings should be considered provisional. Nevertheless the study has developed new simulation techniques as well as tools and methodologies to explore them. While some findings are not particularly surprising (active learning is better than passive), others show greater promise for further exploration, in particular reflective opportunities and the ways that designs are improvised on in practice.

引言：尽管模拟教学已在单模拟人（mannequin）团队场景中趋于标准化，但仍存在诸多其他形式。近年来，本研究团队一直在探索借助新型网络技术，将此前相互独立的模拟器（simulator）互联互通，以打造创新性结构化学习环境的可能性。HSVO项目已解决两项核心研究问题：（1）在单次教育活动中使用多台模拟器开展教学时，教师与学习者组成的单个或多个团队会呈现何种行为模式；（2）参与多站点同步模拟教学活动时，参与者会呈现何种行为模式。
研究方法：本研究共开展10场不同的教学活动，参与者来自4所地理相隔较远的医学院校。研究团队设计了多种组合方案，将多台模拟器（虚拟患者、模拟人（mannequin）以及部件任务训练器）搭配使用，涵盖单团队、站点专属团队、跨站点团队等形式，可同时在两个站点同步开展。每场教学活动对应不同的教学场景，或同一场景的不同配置方案。本研究开发了一款基于网页的平台，以支持多设备、多地点场景的教学活动1。本研究已获得所有参与机构的伦理委员会（或同等职能机构）批准。各站点的学习者团队均通过当地研究负责人招募。每场教学活动均通过一份名为半结构化学习设计说明（semi-structured learning design statement, SSLD）的结构化脚本进行定义。所有教学活动均进行了视频录制。参与者需完成课前测评以评估其既往学习经历与相关背景，并在活动结束后完成针对本次体验的课后测评。
研究结果：本研究以视频录制内容作为主要分析数据源。每场教学活动的平均时长为90分钟，中位数为76分钟。单个站点的学习者团队平均规模为4.9人，中位数为4人。各站点内部参与者的既往经验与学习态度相对统一，但不同站点之间存在差异。
数据分析：本研究首先由2名编码员通过初步审阅制定了一套用于表征不同事件与行为的编码体系，并通过对全部10段视频的首轮分析对该体系进行优化。编码工作需逐段观看视频，以2分钟为单位记录观测到的行为与事件。随后，编码被划分为三大类别：参与度、沟通交流与教学环境。随后将编码员队伍扩充至5人，开展第二轮编码工作。最终得到了能够较为一致地反映每场教学活动核心事件与行为的分析模型。同时，本研究记录了每场教学活动中各核心事件的时间序列，并将其与对应半结构化学习设计说明中预设的序列与时间安排进行比对。
讨论：技术问题并未成为显著的阻碍因素；事实上，多名指导教师将此类技术故障转化为开展深化教学与反思的契机。包含更多主动学习环节的教学场景与活动，能够让学习者产生更深入的反思并提升参与度。使用更多台模拟器也能促使学习者对临床实践的本质及其准备工作展开更高层次的思考。所有教学场景均进行了调整与即兴发挥，相较于原计划，本次研究投入了更多时间用于情况介绍与岗前培训，而用于反馈的时间则有所减少。本研究明确了指导教师的核心作用：采用引导式教学的指导教师比采取灌输式教学的教师取得了更优异的教学效果。目前针对该丰富数据集的分析工作仍在进行中。
结论：本研究属于探索模拟教学领域全新复杂议题的先导性研究，因此所有研究发现均应视为初步结论。不过，本研究已开发出新型模拟教学技术，以及用于探索此类技术的工具与方法论。尽管部分发现并不出人意料（如主动学习优于被动学习），但其他发现仍为后续研究展现出了可观的前景，尤其是关于反思契机以及实际教学中对教学设计进行即兴调整的相关议题。