Heemskerk‐Gerritsen 2015.

Methods	Prospective cohort
Participants	Used a combination of an ongoing nationwide Dutch study on risk assessment and gene‐environment interactions, Clinical Genetics/Family Cancer Clinics, Netherlands Cancer Institute, Foundation for the detection of Hereditary Tumors databases, and linkage to the Netherlands Cancer Registry and the Netherlands Pathology Database to identify eligible women were those with proven BRCA1 or BRCA2 female mutation carriers with BC diagnosed during the period 1980–2011 with no history of bilateral BC or ovarian cancer, no evidence of distant disease activity, and at least 1 unaffected breast in situ. CRRM = 242 Control (surveillance) = 341
Interventions	CRRM
Outcomes	Incidence of contralateral BC Mortality Overall survival
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Free of selection bias?	Unclear risk	Table 1 shows there are some differences in the proportion of population age groupings, surgery for primary BC, radiotherapy, chemotherapy, and RRSO.
Free of performance bias?	Low risk	Yes participants were included via a pathology and hospital record review. All participants were treated within a relatively moderate timeframe (i.e. 21‐year period). Because of the multiple sources of data, it is unclear how data collection was standardised across different sources. However because of the objective and fairly standard data being collected it is assumed that these data items are valid across all oncological databases in the Netherlands.
Free of detection bias?	Low risk	Data were collected from medical records from participating clinics and hospitals, and through data linkage to the Netherlands Cancer Registry and Netherlands Pathology Database. Survival, mortality, and incidence are object measures.
Free of attrition bias?	Low risk	The study authors reported that they excluded 85 participants based on missing baseline or outcomes data