Bridging the Silos: A Comparative Analysis of Implementation Science and Improvement Science.

Background: Implementation science and improvement science have similar goals of improving health care services for better patient and population outcomes, yet historically there has been limited exchange between the two fields. Implementation science was born out of the recognition that research findings and effective practices should be more systematically disseminated and applied in various settings to achieve improved health and welfare of populations. Improvement science has grown out of the wider quality improvement movement, but a fundamental difference between quality improvement and improvement science is that the former generates knowledge for local improvement, whereas the latter is aimed at producing generalizable scientific knowledge. Objectives: The first objective of this paper is to characterise and contrast implementation science and improvement science. The second objective, building on the first, is to highlight aspects of improvement science that potentially could inform implementation science and vice versa. Methods: We used a critical literature review approach. Search methods included systematic literature searches in PubMed, CINAHL, and PsycINFO until October 2021; reviewing references in identified articles and books; and the authors' own cross-disciplinary knowledge of key literature. Findings: The comparative analysis of the fields of implementation science and improvement science centred on six categories: (1) influences; (2) ontology, epistemology and methodology; (3) identified problem; (4) potential solutions; (5) analytical tools; and (6) knowledge production and use. The two fields have different origins and draw mostly on different sources of knowledge, but they have a shared goal of using scientific methods to understand and explain how health care services can be improved for their users. Both describe problems in terms of a gap or chasm between current and optimal care delivery and consider similar strategies to address the problems. Both apply a range of analytical tools to analyse problems and facilitate appropriate solutions. Conclusions: Implementation science and improvement science have similar endpoints but different starting points and academic perspectives. To bridge the silos between the fields, increased collaboration between implementation and improvement scholars will help to clarify the differences and connections between the science and practice of improvement, to expand scientific application of quality improvement tools, to further address contextual influences on implementation and improvement efforts, and to share and use theory to support strategy development, delivery and evaluation.