3. Risk of bias for disease‐free survival.
| Study | Control | Confounding | Selection | Classification of intervention | Deviations from intended intervention | Missing data | Measurement of outcomes | Selection of reported results | Overall |
| Acea‐Nebril 2017 | S‐BCS | Serious | Low | Low | Moderate | Low | Low | Moderate | Serious |
| Some clinicopathological variables significantly different (age, menopausal status, tumour size, tumour stage, axillary lymph node status, location of tumour, multifocality). Some co‐interventions balanced (neoadjuvant CT and axillary management), some missing | All participants eligible included | Classification of interventions clear and determined at the start of intervention. Some aspects maybe determined retrospectively | Deviation from intended co‐intervention (adjuvant therapy time), co‐interventions significantly different | All patients followed up | Objective outcome measure | No indication of selected reporting | |||
| Borm 2019 | S‐BCS | Serious | Low | Low | Low | Moderate | Low | Moderate | Serious |
| Most clinicopathological variables significantly different: age, tumour size, tumour grade, axillary node status, immunohistochemical receptors (ER status). Important co‐interventions (adjuvant CT, adjuvant ET) not balanced across intervention group and may affect the outcome | All participants eligible included | Classification of interventions clear and determined at the start of intervention. Operative details given clearly | All patients received the surgical intervention described in the methods | Analysis unlikely to have removed risk of bias from missing data | Objective outcome measure | No indication of selected reporting | |||
| DeLorenzi 2016 (1) | S‐BCS | Low | Low | Low | Low | Low | Low | Moderate | Moderate |
| Important clinicopathological factors demonstrated balance (menopausal, histological type, grade, axillary node status, immunohistochemical receptors, lymphovascular invasion) or matched (age (within 5 years), year of surgery (within 2 years), tumour size. Important co‐interventions balanced across intervention group (adjuvant CT, adjuvant RT, adjuvant ET) | All participants eligible included | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in the methods | Most patients followed up | Objective outcome measure ‐ due to margin status | No indication of selected reporting | |||
| DeLorenzi 2018 | S‐BCS | Moderate | Moderate | Low | Low | Low | Low | Moderate | Moderate |
| Important clinicopathological factors demonstrated balance (age, BMI, tumour size, immunohistochemical receptors, multifocality), some significantly different (menopausal, grade). Some co‐interventions balanced across intervention group (adjuvant RT, any adjuvant therapy) | Selection may be related to the outcome (mastectomy eventually excluded) | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in the methods | Most patients followed up | Objective outcome measure ‐ due to margin status | No indication of selected reporting | |||
| Gulcelik 2013 | S‐BCS | Moderate | Low | Low | Moderate | Low | Low | Moderate | Moderate |
| Important clinicopathological factors demonstrated balance (age, tumour size, immunohistochemical receptor), most missing. Important co‐interventions demonstrated balance (adjuvant CT, adjuvant ET, axillary management, adjuvant RT) | All participants eligible included | Classification of interventions clear and determined at the start of intervention, operative details given clearly | All patients received the surgical intervention described in the methods. Included from the beginning of uptake of intervention | All patients included but some did not have sufficient follow‐up so excluded. Details not given | Objective outcome measure | No indication of selected reporting | |||
| Mansell 2017 | S‐BCS | Serious | Low | Low | Low | Low | Low | Moderate | Serious |
| Important clinicopathological factors significantly different: age, histological type, tumour size, grade, axillary node status, immunohistochemical receptor (ER, PR). Important co‐interventions balanced, some significantly different: adjuvant CT, adjuvant ET | All participants eligible included | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in the methods | All patients included followed up until June 2016 | Objective outcome measure | No indication of selected reporting | |||
| Mazouni 2013 | S‐BCS | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
| Important clinicopathological factors balance: histological type, tumour size, grade, axillary node status, immunohistochemical receptor (PR). Important co‐interventions predefined and uniform across studies (axillary surgery, neoadjuvant CT, adjuvant RT) | All participants eligible included | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in the methods | All patients included followed up | Objective outcome measure | No indication of selected reporting | |||
| Rose 2019 | S‐BCS | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
| Important clinicopathological factors statistically adjusted for. Location of surgeries different in intervention and control. Some co‐interventions balanced, some missing | All participants eligible included | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in methods | Most patients included | Objective outcome measure | No indication of selected reporting | |||
| Vieira 2016 | S‐BCS | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
| Important clinicopathological factors demonstrated balance. Matched for demographic and oncological aspects. Important co‐interventions demonstrated balance, missing data on axillary management of cases (97.4% for control group) | All OPS participants eligible included, matched BCS 'cases were matched to decrease a possible bias selection' | Classification of interventions clear and determined at the start of intervention: standard surgical treatment was quadrantectomy combined with level III axillary node dissection with was performed in 97.4% of patients | All patients received the surgical intervention described in methods | All patients included followed up | Objective outcome measure | No indication of selected reporting | |||
| Nakagomi 2019 | Mx | Serious | Low | Low | Low | Low | Serious | Moderate | Serious |
| Some variables demonstrated balance (histological type, axillary node status, immunohistochemical receptor status), some significantly different (age, tumour size, stage), many missing. Important co‐interventions missing (RT, axillary management), neoadjuvant CT significantly different | All participants eligible included | Classification of interventions clear and determined at the start of intervention: lattisimus dorsi mini flap or mastectomy | All patients received the surgical intervention described in methods | All patients included followed up | Objective outcome measure but details of follow‐up time not given | No indication of selected reporting | |||
| DeLorenzi 2016 (2) | Mx + R | Low | Low | Low | Low | Low | Low | Moderate | Moderate |
| Important clinicopathological factors demonstrated balance (grade, immunohistochemical receptors) or matched (age (within 5 years), year of surgery (within 2 years), number of positive axillary lymph nodes, tumour subtype). Important co‐interventions balanced across intervention group (adjuvant CT, adjuvant ET), adjuvant RT different | All participants eligible included | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in the methods | Most patients followed up | Objective outcome measure ‐ due to margin status | No indication of selected reporting | |||
| Mansell 2017 | Mx + R | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
| Some clinicopathological significantly different: age, immunohistochemical receptor (ER, PR). Other important clinicopathological factors balance: histological type, tumour size, grade, axillary node status, immunohistochemical receptor (HER2). Important co‐interventions demonstrated balance, adjuvant RT significantly different | All participants eligible included | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in the methods | All patients included followed up until June 2019 | Objective outcome measure | No indication of selected reporting | |||
| Ozmen 2020 | Mx + R | Serious | Moderate | Low | Moderate | Low | Low | Moderate | Serious |
| Important clinicopathological factors balance, some different (age, menopausal status, BMI, tumour size, grade, axillary node status, immunohistochemical receptor status (ER), multifocality), some missing. Important co‐interventions significantly different (adjuvant RT and axillary management), some missing (neoadjuvant RT + CT, adjuvant CT + ET) | Women chose their operation after being told the potential risks and benefits. Bias in assignment: "Both two procedures were explained to patients, and their choices were recorded." | Classification of interventions clear and determined at the start of intervention | All patients received the surgical intervention described in methods. All operations done by a single surgeon with more than 30 years of experience in breast surgery. | Most patients included: "Median follow‐up time was 56 (14‐116) months." | Objective outcome measure | No indication of selected reporting |
BMI: body mass index CT: chemotherapy ER: oestrogen receptor ET: endocrine therapy HER2: human epidermal growth factor receptor 2 PR: progesterone receptor R: reconstruction RT: radiotherapy