Patient and clinician factors influencing the choice of breast cancer surgery : a qualitative and quantitative study.

Background and aims: In women with breast cancers up to 5 cm diameter, breast conservation
therapy (BCT) and mastectomy are equivalent for survival and morbidity, although recurrence
and body image vary. This mixed-method study in a UK region (population 5 million) aims to
identify reasons for mastectomy rate (MR) variation between units.
Methods and findings The study comprised five components; two in 14 units:
(1) An audit of the NHS Breast Screening Programme (n=5060 cases from 11 screening units)
established MR variation was not due to case-mix or caseload (P=O.OOl).
(2) A Discrete Choice Experiment (n=68/93) employing multinomial logistic regression
confirmed clinicians surgical preferences (25 scenarios, n=1695) vary based on cancer (size, site
and centricitv, P<O.OOl){ patient (age and breast size, P<O.OOl) and clinician variables (gender and
clinician role, P=0.015 and P<O.OOl respectively).
Three within units representing hiqh, medium and low case-mix adjusted MRs:
(3) A validated questionnaire established that patients (n=356) preferred and achieved more
autonomy in treatment selection than before; particularly among those choosing mastectomy
(P<O.OOl).
(4) Interviews with clinicians (n=26; 13 nurses and 13 doctors) highlighted variation in local
ethos. Clinicians' focus in the low MR unit was the promotion of BCT and in others, autonomous
decision-making. Communication strategies and processes optimised this.
(5) Interviews with patients (n=65) demonstrated varied experiences between breast units.
While patient factors influenced decisions, breast team factors predominated. Patients from the
high and medium MR units described more informed autonomous decision-making processes and
support.
Conclusions In this study low MRs were associated with clinicians preferring BCT and higher
MRs with clinicians supporting patients' decision-making. Clinician factors related to treatment
preferences associated with high MRs were not identified. This does not preclude their existence
in other regions. Understanding surgical variation factors could facilitate treatment decision-
making equity, but is unlikely to reduce MRs.
-