Variation between hospitals in surgical margins after first breast-conserving surgery in the Netherlands Margriet van der Heiden-van der LooLinda de Munck • Otto VisserPieter J. WestenendThijs van Dalen • Marian B. MenkeEmiel J. RutgersPetra H. Peeters

Surgical margin status after first breast-con- serving surgery (BCS) is used as a quality indicator of breast cancer care in the Netherlands. The aim is to describe the variation in surgical margin status between hospitals. 7,345 patients with DCIS or invasive cancer (T1-2,N0-1,M0) diagnosed between July 1, 2008, and June 30, 2009, who underwent BCS as first surgery, were selected from the Netherlands Cancer Registry. Patients were treated in 96 hospitals. Maximum target values were 30% 'focally positive' or 'more than focally positive' for DCIS and 10% 'more than focally positive' for invasive carcinoma. Results per hospital are presented in funnel plots. For invasive carcinoma, multivariate logistic regression was used to adjust for case mix. Overall 28.5% (95% CI: 25.5-31.4%) of DCIS and 9.1% (95% CI: 8.4-9.8%) of invasive carcinoma had positive margins. Variation between hospitals was substantial. 6 and 10 hospitals, respectively, for DCIS and invasive cancer showed percentages above the upper limit of agreement. Case mix correction led to significant different conclusions for 5 hospitals. After case mix correction, 10 hospitals showed significant higher rates, while 7 hospitals showed significant lower rates. High rates were not related to breast cancer patient volume or type of hospital (teaching vs. non- teaching). Higher rates were related to hospitals where the policy is to aim for BCS instead of mastectomy. The overall percentage of positive margins in the Netherlands is within the predefined targets. The variation between hos- pitals is substantial but can be largely explained by coin- cidence. Case mix correction leads to relevant shifts.