Case Numbers and Process Quality in Breast Surgery in Germany: A Retrospective Analysis of Over 150,000 Patients From 2013 to 2014

Abstract

Background: Numerous studies from around the world have shown a positive association between case numbers and the quality of medical care. The evidence to date suggests that conformity to guidelines for the treatment of patients with breast cancer is better in German hospitals that have higher case numbers.

Methods: We used data obtained by an external program for quality assurance in inpatient care (externe stationäre Qualitätssicherung, esQS) for the years 2013 and 2014 to investigate seven process indicators in the area of breast surgery, including histologic confirmation of the diagnosis before definitive treatment, axillary dissection as recommended by the guidelines, and an appropriate temporal interval between diagnosis and operation. Case numbers were categorized with the aid of various threshold values. Moreover, subgroup analyses were carried out for patients under age 65, patients in good general health, patients without lymph-node involvement, and patients with a tumor size pT0 or pT1 or an overall tumor size less than 5 cm.

Results: Data on 153,475 patients from 939 hospitals were analyzed. Six of seven indicators had values that were better overall, to a statistically significant extent, in hospitals with higher case numbers. Although this relationship was not consistently seen, the worst results were generally found in the category with the lowest case numbers. Similar though less striking results were obtained in the subgroup analyses. An exception to the general finding was that, in hospitals with higher case numbers, the interval between diagnosis and operation was more often longer than three weeks.

Conclusion: Guideline adherence is higher in hospitals that treat more cases. The present study does not address the question whether this, in turn, affects morbidity or mortality. To improve process quality in peripheral hospitals, the quality assurance program should be continued.

Figure 1

QI 5 “Indication for sentinel lymph node biopsy (SLNB)“ The quality indicator 5 was calculated on the basis of the patient population without restrictions. The result for the indicator is shown as a scatter plot. Each hospital is visualized as a blue dot. The red curve represents a non-parametric regression, which can show up non-linear associations and is based on locally weighted smoothing algorithms (Lowess regression; [30]). The x axis shows the number of inpatient cases and the y axis the indicator result. DCIS, ductal carcinoma in situ

Figure 2

QI 4 “Lymph node biopsy in ductal carcinoma in situ (DCIS) and breast conserving therapy” The indicator was calculated on the basis of subgroup 2, which includes patients younger than 65 years with ASA1 or ASA2 and a total tumor size <5 cm. The result for the indicator is shown as a scatter plot. Each hospital is visualized as a blue dot. The red curve represents a non-parametric regression, which can show up non-linear associations and is based on locally weighted smoothing algorithms (Lowess regression; [30]). The x axis shows the number of inpatient cases and the y axis shows the indicator result. DCIS, ductal carcinoma in situ