Abstract
Background: The academic field of personalized integrated care is still growing. The number of publications indexed in PubMed has tripled over the last 10 years, to nearly 8.000 publications in 2021. Still, in literature there is no agreement on the concept of personalized integrated care (personalised care institute, 2023; Goodwin, 2016; Valentijn et al., 2013; Zonneveld et al., 2018).
In its strategy for 2022 – 2026, IFIC notes that personalized (or person-centered) integrated care is usually defined by the problems it aims to solve, which are fragmentation of care and an under-resourced, medically oriented primary care (IFIC, 2021). IFIC recognizes a continuum of definitions, and suggests care systems adopt a definition that meets the needs of the community (IFIC, 2021). However, from research as well as clinical practical perspective, we believe a general, or universal definition of integrated care is useful, or even necessary. Therefore, with our research we aim to present ingredients for a definition of personalized integrated care, as a step towards a general definition.
Approach: We performed a Systematic Concept Analysis using Rodgers’ seven-stage evolutionary methodology (Rodgers, 1989), combined with the practical implications proposed by Foley & Davies (2017). In our search strategy, we used the narrow search rule (IFIC, 2022) next to the Mesh-terms on Integrated Care and Person-Centered Care (Pubmed, 2022) and search terms based on an existing concept analysis on person-centered care (Morgan & Yoder, 2012). Included papers were analyzed using content analysis in Altas.ti focusing on the four categories of Rodgers.
We used Artificial Intelligence (AI), in Miro (OpenAI, 2023), to perform primary thematic analysis on the extracted data. Next, thematic analysis was performed manually, through identifying themes in the data per category, followed by identification of overarching themes. In synthesis, the AI-generated definition was evaluated using the manual extracted themes.
We engaged with practitioners, researchers and policymakers during conferences and projects, to set the scope. Our target-audience consists of service-users, professionals, researchers, policymakers in personalized integrated care.
Results: After application of in- and exclusion criteria, our search in the databases CINAHL, PsychINFO, Pubmed, Medline, Academic Search Premier, Business Source Premier, Psychology and Behavioral Sciences Collection and Cochrane Library, resulted in a realm of 65 papers.
The AI-analysis resulted in a definition of personalized integrated care. Manual thematic analysis resulted in a wide array of themes of personalized integrated care in the four categories of Rodgers. We expected to see an evolution over time of the concept, but this seems less clear-cut than anticipated. The majority of included studies are from the last decade.
Implications: The application of AI was helpful to start data-extraction, but did not result in a suitable definition. Further thematic analysis by the researchers was needed to ensure the quality of data-extraction and results.
Next steps: We will present and discuss our findings with the participants of ICIC25 who form our target audience. This will help to finetune our findings, and add in the development of a clear, balanced definition, which in turn can inform practice, research and policy.