PhD defence M. (Margrietha) van der Linde

Below is a brief summary of the dissertation:

Hospitals are increasingly being compared based on quality. These comparisons serve as guidelines for improvement projects, inform patients, and can be used to hold hospitals accountable for their performance. It is therefore important that quality indicators measure what they are supposed to measure. This thesis investigates what registry data can and cannot say about the quality of hospital care, and demonstrates that the use of this data stands or falls with a good methodology.
 
First, the dissertation shows that only a limited part of the variation in quality indicators can be attributed to the hospitals themselves; the majority is due to differences between patients or chance. Consequently, improvement initiatives aimed at hospitals may focus on the wrong level. Indicators were assessed for validity and reliability using national data on breast, esophageal, and gastric cancer. Many indicators proved suitable for internal learning but unreliable for transparent reporting.
 
The last chapter of this dissertation is about the rol of providers in health disparities. We propose a new methodology that can decompose health disparities between groups in healthcare into three categories: differences that already exist before treatment, differences resulting from unequal access to better-performing hospitals, and differences in treatment within hospitals. This method was subsequently applied to knee prostheses, where it was found that the differences were largely already present before treatment.