Abstract Background Recommendations for mastectomy by multidisciplinary teams (MDTs) may contribute to variation in mastectomy rates. The primary aim of this multicentre prospective observational study was to describe current practice in MDT decision-making for recommending mastectomy. A secondary aim was to determine factors contributing to variation in mastectomy rates. Methods Consecutive patients undergoing mastectomy between 1 June 2015 and 29 February 2016 at participating units across the UK were recruited. Details of neoadjuvant systemic treatment (NST), operative and oncological data, and rationale for recommending mastectomy by MDTs were collected. Results Overall, 1776 women with breast cancer underwent 1823 mastectomies at 68 units. Mastectomy was advised by MDTs for 1402 (76·9 per cent) of these lesions. The most common reasons for advising mastectomy were large tumour to breast size ratio (530 women, 29·1 per cent) and multicentric disease (372, 20·4 per cent). In total, 202 postmenopausal women with oestrogen receptor-positive (ER+) unifocal tumours were advised mastectomy and not offered NST, owing to large tumour to breast size ratio in 173 women (85·6 per cent). Seventy-five women aged less than 70 years with human epidermal growth factor receptor 2-positive (HER2+) tumours were advised mastectomy and not offered NST, owing to large tumour to breast size ratio in 45 women (60 per cent). Conclusion Most mastectomies are advised for large tumour to breast size ratio, but there is an inconsistency in the use of NST to downsize tumours in patients with large ER+ or HER2+ cancers. The application of standardized recommendations for NST could reduce the number of mastectomies advised by MDTs.