Sampling methods and considerations in quality improvement: a scoping review

Quality improvement (QI) in health care involves systematic, data-driven approaches to enhance service quality, safety, and efficiency. Sampling is critical to ensure that data collection is feasible, contextually appropriate, and aligned with improvement goals. However, sampling methods in QI-often pragmatic and non-probability based-are inconsistently reported and poorly justified.

This scoping review, the first to address this topic, aimed to identify and synthesize sampling strategies, frameworks, and sample size considerations for QI initiatives, situating them within the broader evidence implementation and implementation science context.

This review followed the JBI methodology for scoping reviews and was registered in the Open Science Framework (osf.io/rs83a). Peer-reviewed and gray literature from 2000 to 2024 was searched for in PubMed, Web of Science Core Collection, and CINAHL Ultimate (EBSCOhost), as well as organizational websites (e.g., Institute for Healthcare Improvement, Agency for Healthcare Research and Quality, National Institute for Health and Care Excellence, and the World Health Organization). Sources offering conceptual, methodological, or theoretical insights into sampling in QI were included, while empirical QI studies were excluded. Two reviewers independently screened and extracted data, with findings synthesized narratively and in tables.

Ten sources were included. Sampling in QI was primarily intended to support timely, relevant, and credible decision-making rather than statistical inference. Non-probability methods-particularly judgment, purposive, and convenience sampling-were dominant, valued for contextual fit and feasibility. Decisions were shaped by local constraints, perceived risks, and implementation stage. While limitations such as bias, generalizability, and unclear sample size guidance were acknowledged, few sources provided actionable frameworks.

The results indicate that QI sampling reflects a balance between pragmatism and statistical rigor. This highlights the need for clearer, fit-for-purpose guidance to support transparent, context-sensitive, and methodologically sound sampling decisions.