Background and Aims: Primary endocrine therapy (PET) is a common alternative to surgery for frailer older women with operable, oestrogen sensitive breast cancer but may result in treatment failure, contributing to the poor outcomes seen in this age group. Criteria for patient selection for such treatment are lacking with no clear guidance and wide variance in clinician opinion about the appropriate use of the nonsurgical option. There is debate about whether cancer specific and overall survival outcomes vary between treatment types with little high quality research published in the field. This mixed-methods study aimed to identify whether outcomes vary by treatment type and whether clinician preference contributes to the variation in treatment of older breast cancer patients in the UK and to explore some of the factors influencing clinician decision making. Methods and Results: This thesis used a range of methods to explore this issue including literature review and metaanalysis to assess published evidence of variance and its clinical impact, registry data analysis to assess the extent of the variance in current UK practice and whether this was significant when adjusted for case mix and then to explore the underlying reasons behind the variance using a combination of qualitative and questionnaire study of UK HCP in the field of breast care to determine why the variance exists. Each of these components is summarised below: A meta-analysis of data from six randomised controlled trials and 31 non-randomised studies demonstrated superior disease control and a likely survival benefit for surgery over PET in patients with predicted life expectancies of five years or more. Analysis of cancer registry data on 17154 women over 70 with ER+ operable breast cancer between 2002 and 2010 demonstrated considerable variation in surgery rates at hospital level which persisted despite case mix adjustment. Semi-structured qualitative interviews with 34 specialist healthcare professionals (HCPs) from 14 UK sites demonstrated a variety of factors HCPs consider when determining treatment. Opinion was divided regarding the best way to treat dementia patients and whether PET should be offered as a treatment option. A questionnaire survey of Association of Breast Surgery members demonstrated that comorbidities were most important in determining treatment of older breast cancer patients. Opinion was divided over the treatment of dementia patients. Only a quarter felt PET should be offered to all patients over 70 years. A Discrete Choice Experiment contained within the questionnaire demonstrated five variables (age, co-morbidity, cognition, functional status and cancer size) were independently associated with treatment preference (p<0.05) for surgery or PET. Conclusions: Meta-analysis of the published literature and registry data analysis corrected for case mix suggests that PET may result in inferior cancer specific and overall survival in older women. PET is however a valuable option in those with a short predicted life expectancy and case selection is therefore of critical importance in outcome optimisation. Analysis of registry data suggests that case mix does not fully explain treatment variation in older women with operable breast cancer which indicates that thresholds for selection vary widely and evidence based guidelines would be of value in standardising best practice. Clinicians vary in the factors they consider important in the decision making process and whether patients should be offered a choice of treatment themselves. This practice variance, coupled with the inferior outcomes associated with PET in poorly selected women could be a significant contributing factor to globally inferior breast cancer outcomes in older women. Clinical guidelines are urgently needed to address this variability.