This study was conducted in July 2019 (endocrine session – pituitary case scenarios) and October 2019 (diabetes session – diabetes case scenarios), as part of a continuous professional and educational development for clinicians-in-training in Diabetes and Endocrinology in Health Education West Midlands (HEWM) deanery. All specialist trainee registrars specializing in Endocrinology and Diabetes or Metabolic medicine participated in the study.
SIMBA was based on interactive SBL through WhatsApp®. We initially identified five real-life case scenarios for endocrine and diabetes sessions. Following approval from specialists, anonymised transcripts were created on these case scenarios. These transcripts included medical history, clinical examinations, investigation results, imaging and other relevant information that would enable trainees to diagnose the case, propose management and follow-up plans. No patient identifiable data was included in the transcript. These transcripts were validated and approved by a consultant endocrinologist with specialist expertise to ensure that they portray real-life scenarios of respective cases. While the images used in endocrine session were approved by a consultant neuro-radiologist with a special interest in pituitary pathology, similar approval was obtained for continuous glucose monitoring and Libre readings for diabetes session (Figure 1).
For endocrine session, standardized transcripts of five anonymized pituitary cases — Non-Functioning Pituitary Adenoma (NFPA), craniopharyngioma, macroprolactinoma, acromegaly and Cushing’s disease — were prepared. For diabetes session, standardized transcripts of four anonymized diabetes cases — interpreting Libre readings, interpreting continuous glucose monitor (CGM) readings, hypoglycaemic unawareness, and monogenic diabetes — were created. Along with the simulated case scenarios, non-simulated case scenarios were identified by the consultant specialist supervising the specific session based on their prevalence and relevance to daily clinical practice. In both sessions, every effort was made to achieve a healthy balance between simulated and non-simulated cases.
In endocrine session, five moderators were chosen to participate in the study, and the number of moderators was increased to ten for diabetes session based on the feedback from endocrine session. The moderators were recruited considering their interest in the field of endocrinology and their motivation to participate in the innovative method of learning. To ensure their proficiency, the selected moderators were provided with the finalized transcripts three weeks prior to the session. The moderators familiarized themselves with the transcripts followed by at least five mock simulation sessions amongst each other. In order to ensure there was no heterogeneity in the responses, these moderators were then tested by the senior authors of the study. The role of the moderators was to simulate a patient, a senior clinician and a multi-disciplinary team (MDT) liaison at different points of the simulation. At the start of simulation, the moderator took up the role of a patient from whom trainees requested history of presenting complaints and relevant associated medical history. Where the trainees requested physical, biochemical, radiological or any other relevant examination results, moderators simulated a senior clinician to provide this information. Lastly, the moderators played the role of MDT liaison when the trainee combined all relevant information to arrive at the diagnosis, management and follow-up plans. Moderators were instructed to give the relevant information that are provided in the transcripts. If they were asked for information not provided in the transcript, they would reply saying “this information is not available”.
On the day of simulation, each moderator was assigned to a small group of trainees (two to four) with whom they would be replying in parallel. All trainees had been asked in advance to bring their own computers/ laptops/ notebooks through which they connected with the moderators via WhatsApp® Web application. The session started with the information in figure 2 via WhatsApp®. In summary, the trainees were instructed to approach the cases as they would in their daily clinical practice. Once the trainees were ready, the simulation was initiated by providing them with the presenting complaint of the patient. At this point, the moderator played the role of a patient answering relevant questions from the trainees to provide medical history. Once the history taking was complete, the moderators took up the role of a senior clinician providing necessary and relevant examination findings when asked by the trainee. The moderator prompted the trainee when they completed history and examination advising them to move to the next step of evaluation. If the trainee then proceeded to request blood tests, they were sent an electronic blood tests form to request the necessary investigations. Once the completed form was returned via WhatsApp, the moderators replied with the results of requested blood test. If the trainee proceeded for a dynamic function test and/or radiologic investigation, the process for request and provision of results as for blood investigations was repeated. Once these steps were complete, the moderator prompted the trainee to summarise the findings and propose the diagnosis and management plan to MDT. As these cases had been discussed in MDT in real-life, the moderator compared the trainee’s reply to the factual outcome. Should they matched, the moderator informed the same to the trainee. If they did not match, the moderator advised the trainee with the correct diagnosis. In both scenarios, the moderator asked the trainee for a follow-up plan. Once the trainees provided follow-up plans, the simulation ended. If, at any point during simulation, a trainee requested information that was unavailable on the transcript (e.g. ordering an inappropriate investigation or skipping a crucial step in diagnosis/management), they were prompted by the moderators that the information was not available or were appropriately guided back to the relevant step.
To allow participants to familiarize themselves to the simulation model, the first case in each session was run as a trial. After the first case, the case and its approach was discussed in detail in line with current guidelines by an expert, which further helped the trainees to better understand the course of the simulation session. We include an example WhatsApp transcript as a supplement to the paper for better understanding of the model.
During the endocrine session, the case scenario of non-functioning pituitary adenoma (NFPA) was run as a trial. Following this, the trainees underwent simulation case scenarios for macroprolactinoma, craniopharyngioma, acromegaly and Cushing’s disease, followed by respective case discussions with consultant endocrinologist.
During the diabetes session, the case scenario of interpreting Libre readings was chosen as a trial similar to NFPA in endocrine session. This was followed by case scenarios for interpreting CGM readings, hypoglycaemic unawareness, and monogenic diabetes, followed by case discussions with relevant approaches in detail.
During post-simulation discussions, the consultant focused on the appropriate approach to the cases, in relation to the evidence-based international, national and local hospital guidelines as appropriate, in that order of hierarchy, for each specific condition (20–26). Structured feedback/debriefing occurred at the end of the session when the chair discussed the cases and highlighted the lessons learned by getting the trainees to reflect on their performance and discuss strategies for using these lessons to improve their daily practice. The discussions were interactive, and the trainees had ample opportunity to ask any further questions regarding the diagnosis or management of the simulated cases. The trainees were not ranked or scored on their performance. However, they were informed about the accuracy of their diagnosis during the simulations by the moderators as described above.
The confidence of the trainees (measured using a Likert scale ranging from strongly disagree to strongly agree) in approaching different pituitary and diabetes cases was assessed pre- and post-simulation (27,28). These data were then categorised into three groups: (i) confident: for those who responded with strongly agree and agree; (ii) not confident: for those who responded with disagree and strongly disagree; (iii) unsure: for those who responded with agree somewhat, disagree somewhat and undecided. The confidence levels of managing cases pre- and post-simulation are reported using frequencies, percentages, and are displayed in bar charts. Due to the nature of the data, Wilcoxon rank sum tests (significance set at p<0.05) were deemed appropriate (using STATA MP/4 (Statacorp 2017)) to statistically compare confidence levels pre- and post-simulation. Significant tests are highlighted using an asterix.
Improvements in trainees’ confidence levels pre- and post-simulation of simulated scenarios (endocrine session – NFPA, craniopharyngioma, macroprolactinoma, acromegaly, Cushing’s disease; diabetes session - interpreting Libre reading, interpreting CGM reading, hypoglycaemic unawareness, and monogenic diabetes) vs. non-simulated scenarios (endocrine session – microprolactinoma, pituitary apoplexy, thyrotropinoma, gonadotropinoma, pituitary carcinoma; diabetes session - neuropathy, gestational diabetes, blood glucose meters, and ketone meters) were also displayed using frequencies, percentages, bar charts, and were also statistically tested using Wilcoxon rank sum.
In addition to views on the management of the cases, trainees were also asked to comment on their overall impression of the session, the consultant’s contribution during discussion and their interaction with the moderators.