Observational cohort study to determine the degree and causes of variation in the rate of surgery or primary endocrine therapy in older women with operable breast cancer

Background: In the UK there is variation in the treatment of older women with breast cancer, with up to 40% receiving primary endocrine therapy (PET), which is associated with inferior survival. Case mix and patient choice may explain some variation in practice but clinician preference may also be important. Methods: A multicentre prospective cohort study of women aged >70 with operable breast cancer. Patient characteristics (health status, age, tumour characteristics, treatment allocation and decision-making preference) were analysed to identify whether treatment variation persisted following case-mix adjustment. Expected case-mix adjusted surgery rates were derived by logistic regression using the variables age, co-morbidity, tumour stage and grade. Concordance between patients’ preferred and actual decision-making style was assessed and associations between age, treatment and decision-making style calculated. Results: Women (median age 77, range 70e102) were recruited from 56 UK breast units between 2013 and 2018. Of 2854/3369 eligible women with oestrogen receptor positive breast cancer, 2354 were treated with surgery and 500 with PET. Unadjusted surgery rates varied between hospitals, with 23/56 units falling outside the 95% confidence intervals on funnel plots. Adjusting for case mix reduced, but did not eliminate, this variation between hospitals (10/56 units had practice outside the 95% confidence intervals). Patients treated with PET had more patient-centred decisions compared to surgical patients (42.2% vs 28.4%, p < 0.001). Conclusions: This study demonstrates variation in treatment selection thresholds for older women with breast cancer. Health stratified guidelines on thresholds for PET would help reduce variation, although patient preference should still be respected.