Evaluation of mammography screening programs: treatment-related outcomes and importance of selected confounders

Abstract

Evidence from randomized controlled trials (RCTs) of mammography screening indi-cates a reduction in breast cancer mortality of about 20%, but real-world evidence for some organized screening programs is still being evaluated, as is the case in Germa-ny. In most cases, this has to be done with observational data comparing participants and non-participants. Addressing self-selection is therefore of utmost importance in these studies. Correction factors based on data from outdated RCTs became a simple way to deal with this issue, but often face stark compatibility issues regarding the un-derlying study populations as well the inability to take into account time-varying con-founding. Beyond mortality reduction, less intense treatment of breast cancer detected through screening compared to clinically detected breast cancer may be another poten-tial benefit of mammography screening, but there is currently a lack of comprehensive data on this topic. The present dissertation uses menopausal hormone therapy (MHT), an important breast cancer risk factor, as an example to illustrate the particular chal-lenges in the evaluation of mammography screening programs with observational data. Further, it aims to highlight the importance of assessing treatment (-related) outcomes of mammography screening. Three studies were conducted. The systematic review showed a near unanimous association of MHT use with participation in mammography screening. The study on MHT use in Germany showed that from 2004 to 2016, MHT use changed substantially for most types and routes of administration. Finally, the third study showed considerable differences in initial and long-term breast cancer treatment by mode of detection (screen-detected vs. not screen-detected). This dissertation highlights that great care should be taken in examining and discussing how self-selection can be adequately dealt with in any given study. Further, it may be beneficial to make use of data available outside of screening programs in order to bet-ter monitor the benefits and harms, especially with regard to the advances in breast cancer treatment in the future.