Long-term care facilities across Switzerland are required to report six national quality indicators (QIs) covering pain, polypharmacy, malnutrition, and physical restraints. These indicators are used to monitor care quality, support benchmarking, and inform public reporting. However, QIs are only meaningful if the underlying data are trustworthy: if care staff collect and record data differently, the resulting figures cannot reliably be compared across facilities or used for quality improvement. Prior to this dissertation, little was known about how consistently the Swiss national QIs were being measured in practice.

The dissertation first examined polypharmacy QI measurement in German-speaking LTCFs, before NIP-Q-UPGRADE began. Polypharmacy was defined as receiving nine or more active substances within seven days — a more demanding calculation than simply counting prescribed medications. Staff counts of active substances differed from those of the research team in nearly two thirds of assessed residents, due to ambiguous counting rules, incomplete medication master data, and inconsistent electronic algorithms. A second study assessed how consistently all six national QIs were measured across the German-, French-, and Italian-speaking regions of Switzerland. Interrater reliability varied considerably, ranging from poor for malnutrition to near-perfect for physical restraints. High variability in QI rates between facilities was also observed. Together, these studies defined the scope of the data quality problem.

Drawing on evidence gathered across NIP-Q-UPGRADE and existing literature, the dissertation describes the development of a multi-level data quality improvement intervention. Built around a train-the-trainer approach, it prepares facility-based champions to lead internal data quality monitoring, deliver staff training, and provide feedback. It comprises materials such as presentations, posters, factsheets, and checklists, a dedicated helpdesk. The intervention at facility level was complemented by system-level changes including adapted measurement rules and unified algorithms. Pilot implementation was evaluated across LTCFs in all three linguistic regions. All intervention components received high scores on acceptability and feasibility. Qualitative findings identified management support, protected time, and sufficient human resources as the main determinants of successful implementation in LTCFs, while trainer competence and participants' prior expertise shaped how the training was received. These findings informed targeted refinements to the training structure and materials.

This dissertation contributes to the field of long-term care quality monitoring by providing evidence on the reliability of nationally mandated QIs in Swiss LTCFs and by describing a replicable pathway from problem identification through intervention development to pilot implementation. It demonstrates that LTCFs can be meaningfully supported in monitoring and improving QI data quality, and lays the groundwork for broader scale-up.