Exploratory study of the characteristics of feedback in the reflective dialogue group given to medical students in a clinical clerkship

Purpose Structured narrative reflective writing combined with guided feedback is an efficient teaching method for enhancing medical students’ reflective capacity. However, what kinds of feedback offered and reflection presented in a reflective group remain unclear. The aim of this study was to investigate the characteristics of feedback in a reflective dialogue group. Methods Fifth-year medical students on a monthly interval rotation at the pediatric department of a medical center in eastern Taiwan during the 2012 academic year completed their reflective writing regarding patient and family psychosocial issues, and were subsequently debriefed in a 2-h group discussion session to receive feedback from a clinical tutor and peers. Content analysis was conducted to explore the characteristics of feedback and reflection presented in the reflective dialogue. The evaluative questionnaire regarding the benefits of reflection with others was administrated following the group session. Results Forty students participated in five reflective groups and 108 psychosocial issues were discussed and identified. The tutor played an initiating role in the group discussion by providing six equal feedback types involving exploring new knowledge, initiating advanced discussion, highlighting the issues, and encouraging the students. The students provided eight types of feedback that involved a substantial deep discussion on psychosocial issues and action plans based on the complex interactive ecological network of clinical encounters. Each student attained 1.25 times the depth or breadth of reflection after receiving feedback and experienced the benefits of reflection with others. Conclusion Through structured narrative reflective writing combined with pluralistic group discussion with a tutor and peers, the medical students had time to think deeply and broadly about psychosocial issues among patients and their family members. Facilitative feedback providing new knowledge, deeper discussion, and exploring new ways of action planning for psychosocial issues was recommended to promote students’ reflective capacity.

broad and deep reflections. Second, most reflection involves reflection-on-action after the event has occurred in medical practice; thus reflective writing, such as an unstructured diary or a guided reflective journal, could recapture patientÁphysician encounters in elaborative, analytic approaches for transforming the experience further. Finally, guided reflection by a faculty mentor or peer group members encourages people to share their thoughts, feelings, or reactions to their experiences reciprocally, and promotes multiple perspectives and alternative actions.
Medical curricula that apply these teaching methods to facilitate medical student reflection have been widely developed and effectively examined. DasGupta and Charon (11) designed a 6-week seminar of reflective writing for 16 second-year medical students to describe their own illness story, or that of a family member or friend. The written essays were shared and discussed each week, and qualitative evaluations at the final session showed increasing amounts of student empathy and self-awareness. The other qualitative study combined structured reflective writing of field notes and individualized faculty feedback for preclinical students in a 'doctoring' course for firstand second-year students. The thematic qualitative evaluations of field notes and feedback showed the following advantages: deep reflection, valued feedback, promoting later group process, and personal and professional development (12). A randomized comparison study examined the teaching effectiveness of guided critical reflection and faculty feedback for 149 third-year students. One question about reflective ability was evaluated and indicated that the teaching methods of either critical reflection guidelines or feedback on the reflective process have more benefits for improving student reflective ability than teaching without any guideline or feedback (13). Another study compared two teaching designs, the traditional written format with a small-group discussion, and the blog format following an online commentary and feedback for clerkships, and found that the level of reflection did not differ between the two teaching designs, whereas students' preferences for blogging or essay writing influenced their learning experience (14).
Both qualitative and quantitative studies have indicated that combined structured reflective writing and guided feedback from a mentor or a group is the most efficient teaching method for enhancing student reflective capacity (11Á14). Theories have also suggested that group work or one-to-one discussions facilitate the later stage of the reflective learning cycle; however, few studies have specifically explored the characteristics of effective feedback (15). Sandars indicated that the role of the facilitator during feedback should include counseling and mentoring skills for building a supportive environment to enable students to process the reflection (5). Branch and Paranjape addressed the general rule that facilitating reflection involves elaborating on a topic and transferring it to a deeper or more complex level of reflection (16). The initial study found that elaborate questions and a positive tone were effective writing-feedback comments to stimulate student reflection (17). However, what kinds of feedback offered and reflection presented in a reflective group remains unclear.
In this study, we conducted content analysis to explore the characteristics of feedback and reflection presented in the reflective dialogue group composed of a clinical tutor and medical students during their clinical clerkship in a pediatric department. The following research questions were addressed: What were the characteristics of feedback provided from clinical mentor and students in the reflective dialogue group?; Did medical students think broader and deeper about their practice after receiving feedback from tutor and peer?; and, How did medical students perceive the benefits after the reflective group session.

Study background and procedures
This study was implemented in a required clinical course that involved writing about patient and family psychosocial issues, designed for fifth-year medical students on a monthly interval rotation (7Á9 students on each rotation) at the pediatric department of a medical center in eastern Taiwan during the 2012 academic year. All 50 students planned to rotate at the pediatric department, completed the reflective writing, and participated in one of the six group sessions separately in 2012.
During the process of primary care and history taking interviews in each rotation, each student separately chose and observed the psychosocial issues of one patient and his family members, completed their reflective writing at the third week of rotation, and subsequently discussed in a 2-h group session to receive feedback from the course tutor and peers. The reflective writing was structured by guidelines developed from Gibbs' concept of reflective cycle (10), which comprises six components: description, feelings, evaluation, analysis, conclusion, and action plan.
During the reflective group section, 7Á9 medical students sat in a circle and received a short orientation regarding the group activity process. Each debriefed student demonstrated psychological issues in a reflective approach, and received one peer feedback that particularly focused on alternative action plans. Tutor feedback was given using reinforcement teaching skills, the sandwich feedback method, and possible alternative approaches for solving identified psychosocial issues. Following group session, the students were asked to complete an evaluative questionnaire about the benefits of reflection with others.
This study was conducted after obtaining approval through informed consent from the institutional review board at Buddhist Tzu Chi General Hospital (IRB102-20).
The group was designed and led by the corresponding author. Group sessions were recorded by a tape recorder after receiving consent from the students. In addition to the instructor, at least one member from the research team joined the group dialogue session and observed the whole process.

Data analysis
Group process analysis All group sessions were anonymously transcribed by two research assistants. Each session was subsequently divided into three analyzed parts: characteristics of feedback, dialogues involving considerable depth or breadth of reflection, and psychosocial issues discussed during the session.
Content analysis is a method used to deal primarily with verbal materials to extract desired information or identify specific characteristics through a multistep process. The advantage of the content analysis lies in its ability to transform the qualitative information into quantitative information such as category frequencies or ratings for further comparison (18). In the present study, we applied content analysis to identify and code the characteristics and frequencies of feedback, and dialogues involving considerable depth or breadth of reflection in the reflective dialogue group.
First, the research team undertook a trial analysis on one group's transcript and found certain types of feedback emerged consistently as to develop a preliminary coding list of feedback. We then used this preliminary coding list to code the second group transcription until the taxonomy of feedback had been discussed elaborately and fully. As a result, we defined eight types of feedback between the students, and six types of feedback from the tutor to the students (Appendices 1 and 2). Second, we developed the coding manual regarding the dialogues involving depth or breadth of reflection based on the Mezirow theory (19). We defined 'the depth of reflection' as the presence of analyzing, exploring, and reflecting all assumptions or problem-solving strategies regarding psychosocial issues or action plans. 'The breadth of reflection' was defined as the presence of discussing derived topics related to psychosocial issues or action plans. Third, two researchers as a team coded the last group materials separately by using the coding manual and then discussed the codes consensually. If there was a disagreement, the third researcher was consulted to reach an agreement.
All psychosocial issues discussed during the group sessions were openly reviewed and identified, and named consensually by researchers. The frequencies of psychosocial issue, feedback, and the depth or breadth of reflection were calculated as category frequencies.
The evaluative questionnaire Jasper stated reflective practice has many advantages for professional development especially by using reflective strategies with others (20). We further conducted a 10-item evaluative questionnaire, with items such as 'They provide me with immediate feedback', with a five-point scale to measure the degree of agreement on the reflective benefits with peers evaluated by students, where 1 stood for strongly disagree and 5 stood for strongly agree.
The chi-square test was used to analyze the frequency and differences of feedback types and the depth or breadth of reflection. The mean and standard deviation of the 10 items of the evaluative questionnaire were also statistically analyzed. All data were computed using the Statistical Package for Social Science, version 15.0 (SPSS Inc., Chicago, IL, USA).

Results
Except for one piloting group, this study analyzed the reflection dialogue process and evaluative questionnaires of five group sessions including 40 medical students. Male to female ratio was 24:16, and the mean age was 24.7 years old. There were 108 psychosocial issues identified from five group coding transcripts which categorized into 7 major themes, including medical communication, the intricate medical ecological system, role and function of a family, medical students' competence toward a medical profession, ethical dilemma in clinical setting, and extreme diverse patient perspectives from different sociocultural background underlying the complexity of clinical encounter. One sample case demonstrated how student A debriefed controversial psychological issues involving young adolescents, dropping out of school and family, misconduct, and economic problems along with group dialogues.
One 17-year-old girl dropped out from high school working as a waiter at a night shop for more than one year, admitted due to dysuria and lower abdominal pain. She had a history of artificial abortion recently; urinary tract infection and pelvic inflammatory disease were diagnosed this time. None of her family members visit her during the whole admission course except her boyfriend and the boss of the night shop. I think she must come from low social economic status, facing difficulties in schooling, poor family relationship, struggling for living and with behavioral problems. She may encounter bad friends, be exploited, engaged in sexual trades, controlled by drugs, and unable to appeal her hell . . . My difficulties were: I have so different growing environment as compare to patient. I can't seem to understand this girl's living environment and her thoughts. I don't know how to communicate with her in the first setting. I am afraid that I may hurt her unintentionally. And most importantly, I don't know how to stand for her and change this complex situation . . .. Group dialogue Phrases in parentheses are types of feedback identified. Feedback from students (Deeply discussing psychosocial issues) Feedback from the tutor and students Some of the psychosocial issues were subject to cultural bounds, such as daughters or daughters-in-law in the families identified with traditional Taiwan or Chinese culture being expected to take care of ill parents, which usually makes them feel angry and frustrated. One student shared a story about a long-term care case of an old woman. While the son of the patient was given most of the inheritance of his father (the woman's deceased husband) and never showed up to visit the sick, her daughter took the duty of her care and complained about the unfairness. Both instructor and students in this study went onto discussing the male-preference phenomenon, and showed their sympathy for those daughters-in-law or daughters who took the most responsibility of patients' care. Explanations of such psychosocial issues should take into account the sociocultural perspectives. Some other examples were shown as Appendices 1 and 2.

Group feedback
The texts of five reflective dialogue groups were analyzed, with 107 types of feedback 'from the tutor to the students' and 73 'between the students' ( Table 1). Six types of feedback 'from the tutor to the students' were frequently categorized as follows: exploring new knowledge about psychosocial issues (25.2%), deeply discussing psychosocial issues (17.8%), exploring new knowledge regarding action plans (16.8%), deeply discussing action plans (15.9%), highlighting psychosocial issues (14%), and encouraging or approving students (10.3%). No significant differences were found between these six types of feedback (X 2 07.90, p00. 16), indicating that tutor feedback was equally distributed among various categories.
Eight types of feedback 'between students' were frequently categorized as follows: deeply discussing psychosocial issues (50.7%), deeply discussing action plans (21.9%), confronting or debating action plans (8.2%), engaging in self-disclosure (6.8%), encouraging or approving (5.5%), confirming psychosocial issues (4.1%), confirming action plans (1.4%), and confronting or debating psychosocial issues (1.4%). Significant differences were found between these eight types of feedback (X 2 0114.73, p00.00), indicating that the feedback between the students in the reflective dialogue group was more centered on discussing psychosocial issues and action plans than the other categories.
Depth and breadth of group reflection Table 1 shows 50 dialogues coded as having the depth or breadth of reflection on psychosocial issues or action plans, meaning that each of the 40 students reported the psychosocial issues would have an average of 1.25 reflection statements that showed their increase in the depth or breadth of reflection after receiving feedback from the tutor and peers. These enriched or enlarged perspectives were shared in the group. No significant differences were found between these four levels of reflection (X 2 07.44, p00.06); depth of reflection on psychosocial issues (n 020, 40.0%); depth of reflection on action plan (n 013, 26.0%); breadth of reflection on psychosocial issues (n010, 20.0%); and breadth of reflection on action plan (n07, 14.0%).

Evaluation of the benefits of reflection with others
Forty students evaluated the degree of agreement on the benefits of reflection in groups. The questionnaire showed good internal consistency (a 00.87). Two factor were extracted by exploratory factor analysis which contributing to 60.56% of the variance. Table 2 shows the students indicated averagely 'agree' to 'strongly agree' on the 10 benefits of reflection in groups (M 04.33, SD 00. 16). There were no significant differences between these 10 items (X 2 01.2, p00.99). However, the students indicated the highest agreement on the benefit of 'They act as a sounding board for my ideas' (M 04.6, SD00.54).

Discussion
In this study, we designed a required course combining structured reflective writing and a reflective dialogue group to assist medical students in broadening and deepening their reflective thinking on psychological issues, further inducing a transformative action plan on humanistic caring.
The main results of this study showed how feedback operates and facilitates reflection on psychosocial issues and action plans from the tutor and peers. This study found that compared to student feedback, tutor feedback (107:73) tended to focus on varied feedback types, whereas the students learned to give limited types of feedback. Thus, the role difference derived from teaching and learning in reflective groups was salient. The tutor in this study was equipped with facilitative repertoires such as providing new knowledge or information, initiating discussion, highlighting topics, and encouraging the students. The tutor was particularly involved in sharing new knowledge, additional reading suggestions, and facilitating discussion from multiple perspectives about psychological issues, whereas the students were highly involved in acting as sounding boards and making meaning for other students. Positive feedback may prompt and facilitate building a topic focus and a friendly atmosphere to deepen the reflection (5,16,17).
As indicated, the role of the peer-assisted learning (PAL) phenomenon can be clearly observed in reflective dialogue-group dynamics (21). This study also found that most of the students were highly involved in deeply discussing psychosocial issues and action plans. Psychosocial issues reported by the students generally involved stories and controversial topics that caused them to realize during disclosure that they had similar reflections or faced similar clinical difficulties or complexities, which drew considerable attention and induced deep   (Appendices 1 and 2). The group also facilitated discussing an action plan and engaged in favorable exchanges of ideas regarding problem-solving and future implications. However, students engaged in few confrontations or debates with each other, which might be partially attributed to Chinese culture, which stresses humility and avoiding conflict. However, whether there was sufficient assertiveness in expressing opinions, particularly those that facilitate diversity, would be worth exploring. Thus, the long-term benefits and limitations of learning from unplanned PAL require further investigation.
Although there was no significant difference between the greater depth or breadth of reflection on psychosocial issues or action plans, this study found significant differences involving 66% of the reflection being related to 'depth' and 34% being related to 'breadth' (X 2 05.12, p00.02). In such a reflective group, greater depth either on psychosocial issues or action plans could be more easily addressed than the level of breadth on reflection. The reflective group enabled members to concentrate on the presented issues or actions and to respond with indepth discussion. By contrast, although the discussion with increased breadth was less than in-depth, it still required a large proportion of time and should make effective contributions. However, this could be inferred only from cases in which the tutor provided equal feedback types and students were focused on a specific feedback type. Further study is necessary to confirm whether depth of reflection would be salient in other reflective groups, and to explore how to improve the breadth of reflection for medical students.
Reflection behaviors can be facilitated during group discussions. Each student attained 1.25 times greater depth or breadth of reflection after receiving feedback from the tutor and peers. Receiving feedback from the peers and tutor was also highly appreciated; the students indicated 'agree' to 'very much agree' on the benefits of reflection with others. The reflective dialogue group could be a welcome and effective course design for enhancing the reflection of medical students. While maintaining its effectiveness in increasing a discussion on psychosocial issues and action plans, facilitating optimal confrontation and debate through modeling or reinforcement may be necessary and worth attempting. Medical teachers should also assist students in organizing the perspectives generated in the group, regardless of the depth or breadth of consolidating their learning from peers and reflection.
The first limitation of this study is that we could not differentiate whether tutor or peer feedback contributes the greatest benefits and advanced level of reflection. Second, we could not infer whether the students showed more humanistic caring for patients and their families, as expected in the reflective course.

Conclusion
Through structured narrative reflective writing combined with pluralistic group discussion with the tutor and peers, the medical students in this study had time to think deeply and broadly about psychosocial issues among patients and their family members. Facilitative feedback providing new knowledge, deeper discussion, and new ways of action planning for psychosocial issues was recommended to promote students' reflective capacity.