As a relatively new phenomenon in 2009, Swedish nonprofit social service providers proposed quality improvement as a way to reduce mistakes, use resources more effectively and meet the needs and expectations of clients in a better way. Although similar experiences have been studied in health care, the transfer of quality improvement to nonprofit social services gives a possibility for more knowledge on what enables, and constrains, systematic quality improvement in this specific context.

This thesis is based on five years of supporting quality improvement in the Swedish nonprofit welfare sector. Specifically, it builds knowledge on which active mechanisms and enabling or constraining structures exist for nonprofit social service quality improvement. By studying quality improvement projects that have been conducted in the development program Forum for Values, critical cases and broad overviews are found valuable. These cases have resulted in four papers on quality improvement in nonprofit social services. The papers include: critical cases from a nursing home for elderly and a daycare for disabled children (Paper I); a critical case from a sheltered housing (Paper II); an overview of performance measurements in 127 quality improvement projects (Paper III); and an analytical model of how improvement policy and practice are bridged by intermediaries (Paper IV). In this thesis, enabled or constrained events and activities related to Deming's system of profound knowledge are identified from the papers and elaborated upon. As a basis for transforming practice into continuous improvement, profound knowledge includes the four knowledge domains: appreciation of a system, theory of knowledge, understanding of variation and psychology of change. From a realist perspective, the identified events are seen as enabled or constrained by mechanisms and underlying regularities, or structures, in the context of nonprofit social services.

The emerging mechanisms found in this thesis are: describing and reflecting upon project relations; forming and testing a theory of action; collecting and displaying measurable results over time; and engaging and participating in a development program. The structures that enable these mechanisms are: connecting projects to shared values such as client needs; local ownership of what should be measured; and translating quality improvement into a single practice. Constraining structures identified are: a lack of generalizable scientific knowledge and inappropriate or missing infrastructure for measurements.

Reflecting upon the emergent structures of nonprofit social services, the role of political macro structures, reflective practice, competence in statistical methods and areas of expertise becomes important. From this discussion and the findings some hypotheses for future work can be formulated. First, the identified mechanisms and structures form a framework that helps explain why intended actions of quality improvement occur or not. This frameworkcan be part of formulating a program theory of quality improvement in nonprofit social services. With this theory, quality improvement can be evaluated, reflected upon and further developed in future interventions. Second,new quality improvement interventions can be reproduced more regularly by active work with known enablers and constraints from this program theory. This means that long-lasting interventions can be performed and studied in a second generation of improvement efforts. Third, if organizations integrate quality improvement as a part of their everyday practice they also develop context-specific knowledge about their services. This context-specific knowledge can be adopted and further developed through dedicated management and understanding of variation.

Thus, if enabling structures are invoked and constraining structures handled, systematic quality improvement could be one way to integrate generalizable scientific knowledge as part of an evidence-creating practice.