Breaking bad news: an active learning method for medical students | BMC Medical Education

The study population

Fifth-year students of Université Paris Cité medical school performing internships at Robert-Debré Universitary Hospital or Necker Children’s Universitary Hospital were divided into an interventional group based on a discussion-based workshop on breaking bad news and a control group without any additional teaching on breaking bad news other as the usual curricula. In their fifth year (out of six) of medical studies, they have already had all main disciplines, such as all subspecialties, oncology, palliative care, emergencies, and ethics. Students were selected from voluntary units in both hospitals, and units were randomized from both hospitals in both groups, without mixing students from a same unit in different groups, to avoid bias. After the provision of a study information sheet, the students gave their written, informed consent to participation. The study protocol was approved by an independent ethics committee (CERAPHP, Paris, France; reference 2023-07-04).

Workshop on breaking bad news: the discussion group (for the interventional group only)

This workshop was prepared ahead of time during several multidisciplinary meetings comprising a clinical psychologist, a social worker, three physicians, and an expert patient (defined by the French High Authority of Health, as one who has developed over time a detailed knowledge of his illness and thus has real expertise in the daily experience of a disease, or a physical limitation linked to his condition). In order to prepare the teaching and assessment methods as thoroughly as possible, we organized a meeting with the teachers in charge of the “physician-patient relationship” course at Paris Cité University medical school. This discussion group constituted the active, participatory learning method evaluated in the OSCE. The group comprised a clinical psychologist, a social worker, two physicians, and the students from the interventional group. The workshop comprised four phases: 1/ three videos about breaking bad news (two videos about telling a person that their partner is in a critical condition in the intensive care unit in two different ways, which are used in the Sorbonne University’s course on the patient-physician relationship, and a short film with three patient scenarios (available at 2/ discussion and dialogue after the videos had been viewed (see details on the free discussion in Additional Methods); 3/ development of a guide on good practice in breaking bad news through collective intelligence (Table 1), and 4/ distribution of copies of the S-P-w-ICE-S guidelines [11].

The formative and summative OSCE (for the interventional and control groups)

The discussion group and the OSCE took place 6 weeks apart. We developed a summative OSCE station on breaking bad news based on the announcement of a diagnosis of breast cancer (Additional file 1). The main purpose was to evaluate the active learning method of the interventional group. During simulation, we assessed the students particularly on their communication skills (Additional file 1, evaluation grid). The station has been validated by the university board in charge of these faculty OSCES. The same day of the summative OSCE had been scheduled, the students of the interventional group were told that what they had learned during the first workshop was going to be assessed in this OSCE. The procedure was as follows: all the students of the intervention and the control group were distributed across seven circuits playing the same station including an evaluator and a standardized female patient. All the evaluators were qualified physicians, and all the standardized female patients were qualified physicians (n = 3) or other healthcare professionals (a resident, a nurse, an occupational therapist, and a psychomotor therapist). The evaluator was blinded to the students’ group allocations (i.e. interventional or control). Each student was rated by one single evaluator. Immediately after the OSCE station, the students were seen by a social worker and a psychologist for informal feedback on their initial impressions and completion of a self-questionnaire (Additional file 2).

Once all the students had been evaluated, they attended multidisciplinary debriefing meeting with the evaluators, the standardized female patients, and the teachers from the discussion group workshop. During debriefing session, we conducted group exchanges as follows: 1/ clarification of objectives with the grading grid (explaining and justifying important answers), correction of learning expectations, setting the environment, important items for a good diagnostic announcement; 2/ dynamic analysis of actions, thought processes and emotional states (with the help of a psychologist); 3/ systematic response to students’ free questions and 4/ at the end of session, the guide developed in the discussion workshop (including the S-P-w-ICE-S protocol) was returned to the students [11]. All four steps were designed to improve the students’ relational skills.

Variables evaluated

The following variables were evaluated for the two groups after the summative OSCE: 1/ the score for the OSCE on a 15-item grid covering the assessment of communication skills and attitudes (Additional file 1); 2/ the blinded evaluator’s response to the question: “In your opinion, has the student received additional training in how to break bad news?”; 3/ a self-assessment of the student’s performance during the station through a visual analogue scale (VAS)) and an assessment of the university courses on breaking bad news (Additional file 2); 4/ the student’s Rosenberg Self-Esteem Scale score [14, 15]; and 5/ the student’s Jefferson Scale of Empathy score [16,17,18].

Statistics

All statistical analyses were performed using GraphPad Prism software (version 6.0, GraphPad Software, Inc., La Jolla, CA, USA). The groups were compared by applying Student’s t-test, or Fisher’s exact test, as appropriate. The data were quoted as the mean ± standard deviation [min max] or effective (percentage). The threshold of statistical significance was set to p < 0.05 (in the additional files’ figures: ^*p < 0.05; ^**p < 0.01; ^***p < 0.001, and ^****p < 0.0001).