Abstract Objectives The objective of this paper is to describe a novel post-thyroidectomy haematoma simulator and present data on its face and content validity and curriculum applicability. Methods A simple simulation model of a post-thyroidectomy haematoma was made using a C-spine collar, silicone sheets and thickened jelly. The model was worn by a simulated patient in a hybrid simulation. A prospective evaluation study was performed. Medical, nursing and theatre staff were recruited to complete a simulation using the model. All participants completed a 16-item questionnaire using a 5-point Likert scale to assess the following: face validity (FV), content validity (CV), task-specific validity (TV) and curriculum applicability. Non-parametric statistics were used. Results 74 participants completed the questionnaire. The model achieved median face validity of 4 (interquartile range (IQR): 4-5), median content validity of 4 (IQR: 4-5) and median task-specific validity of 5 (IQR: 4-5). In 14 out of 16 domains, participants gave a score of 4 or 5 (agreed or strongly agreed) more than 80% of the time. The median score for curriculum applicability was 5 (IQR: 4-5). Conclusions This home-made simulator for post-thyroidectomy haematoma evacuation achieved face and content validity. It is an accessible, cheap and potentially life-saving training tool. This model may be useful for multidisciplinary simulation-based training in the light of the recent DAS, BAETS and ENT UK consensus guidelines on managing post-thyroidectomy haematoma. Keywords: Simulation, post-thyroidectomy haematoma, hybrid simulation, face validity, content validity Key points: 1. Post-thyroidectomy haematoma is a rarely encountered and potentially fatal complication of thyroid surgery. 2. No simulation models for this condition have previously been validated. 3. We developed a simple, home-made, cost-effective hybrid simulation model. 4. The model was tested at three separate training events for ENT registrars and the multidisciplinary team. Data was collected using Likert-scale questionnaires. 5. The model achieved good face and content validity.