Abstract Background The mortality rate is low in endocrine surgery, making it a difficult outcome to use for quality improvement in individual units. Lessons from population data sets are of value in improving outcomes. Data from the Australian and New Zealand Audit of Surgical Mortality (ANZASM) were used here to understand and elucidate potential systems issues that may contribute to preventable deaths. Methods ANZASM data relating to 30-day mortality after thyroidectomy, parathyroidectomy, and adrenalectomy from 2009 to 2020 were reviewed. Mortality rates were calculated using billing data. Thematic analysis of independent assessor reports was conducted to produce a coding framework. Results A total of 67 deaths were reported, with an estimated mortality rate of 0.03–0.07 per cent (38 for thyroidectomy (0.03–0.06 per cent), 16 for parathyroidectomy (0.03–0.06 per cent), 13 for adrenalectomy (0.15–0.33 per cent)). Twenty-seven deaths (40 per cent) were precipitated by clinically significant adverse events, and 18 (27 per cent) were judged to be preventable by independent ANZASM assessors. Recurrent themes included inadequate preoperative assessment, lack of anticipation of intraoperative pitfalls, and failure to recognize and effectively address postoperative complications. Several novel themes were reiterated, such as occult ischaemic heart disease associated with death after parathyroid surgery, unexpected intraoperative difficulties from adrenal metastasis, and complications due to anticoagulation therapy after thyroid surgery. Conclusion This study represents a large-scale national report of deaths after endocrine surgery and provides insights into these rare events. Although the overall mortality rate is low, 27 per cent of deaths involved systems issues that were preventable following independent peer review.