IN-VITRO PHASE Sample Clauses

IN-VITRO PHASE. Since this phase was conducted using data from the local EHR and without contacting clinicians or patients, it was termed “in-vitro”. This phase had two manifold aims: (i) to address implementation barriers according to the Consolidated Framework for Implementation Research (CFIR)38 (Figure 3-2), and (ii) to integrate the transdiagnostic risk calculator into the local EHR. Figure 3-1 Overview of in-vitro and in-vivo phases Figure 3-2 CFIR Implementation barriers addressed during the in-vitro phase. The CFIR frameworks identifies five core implementation domains: characteristics of the intervention (the ‘core components’—that is, the essential elements of the intervention— and the ‘adaptable periphery’—that is, the adaptable elements in which the intervention occurs); outer setting (the ‘economic, political, and social context within which an organisation resides’); inner setting (the ‘structural, political, and cultural context through which the intervention proceeds’ and the relationship between these elements); individuals (the individuals responsible for carrying out the intervention or otherwise related to the intervention, their agency, and their relationships to each other and the intervention) and process (the active process through which the desired changes are achieved).38 Crucially, several aspects were carefully planned at the stage of model development to facilitate its subsequent implementation. The “core components” of the CFIR characteristics of the transdiagnostic risk calculator (i.e. lifespan-inclusive, individualised, clinically-based, transdiagnostic, automatic data acquisition and risk estimation, cheap, use of predictors widely available in clinical routine, usability in a sequential assessment framework) 39 were selected a priori to match the CFIR “inner” and “outer” settings (Figure 3-2). The “adaptable periphery” of the CFIR characteristics of the intervention (Figure 3-2), required to adapt the transdiagnostic risk calculator (developed using retrospective data) to prospective use, after checking for data quality and missingness (which may differ in real-world prospective vs retrospective EHR datasets). In the retrospective version of the risk calculator, individuals who developed psychosis within three months following their index diagnosis were excluded to mitigate for ICD-10 diagnostic instability. However, prospective implementation of this diagnostic lag was inefficient. Subsequent analyses confirmed that a refined version of...