Search

Objective: To examine the variation in surgical and adjuvant treatment of breast cancer of known histology and detected on screening in a large cohort of patients treated by the surgeons of a health region., Design: Part prospective, part retrospective observational study using the databases of a region's breast screening programme and of the cancer registry., Setting: The former South East Thames region., Subjects: 600 women aged 49-79 who presented during 1991-2 with invasive breast cancer up to 20 mm in diameter that had been detected on screening. These patients were treated by 35 surgeons., Main Outcome Measures: Mastectomy rate by surgeon and the use of adjuvant treatment (radiotherapy, tamoxifen, and chemotherapy) were compared with risk factors, tumour grade, resection margins, and axillary node status., Results: The mastectomy rate varied between nil and 80%, although the numbers at these extremes were small (0/13 v 8/10). Surgeons operating on more than 20 such cases had a lower mastectomy rate (15%) than surgeons treating fewer cases (23%), but this difference was confounded by variation in casemix. There were also wide variations in mastectomy rates and in axillary sampling rates that were independent of casemix or caseload. There was broad agreement on the use of adjuvant tamoxifen (94%), but few patients received chemotherapy (2.5%). 78 patients (19%) did not receive radiotherapy, including 51 out of 317 patients with unfavourable tumours, and 26 patients did not receive tamoxifen. Whether the patient received adjuvant treatment was more dependent on referral by the surgeon than the risk factors for local recurrence and was independent of caseload., Conclusion: Mastectomy rates for similar tumours vary widely by surgeon independently of casemix or caseload, but surgeons with a higher caseload tend to have a lower mastectomy rate. Omission of postoperative radiotherapy or tamoxifen after conservative treatment is not related to risk factors for local recurrence or caseload. Confidential feedback of treatment profiles to individual surgeons has been used, but when benefit has been established treatment should be guided by evidence based protocol.