Edinburgh Napier University

  Introduction
Survivors of critical illness experience multidimensional disabilities that reduce quality of life, and 25–30% require unplanned hospital readmission within 3 months following index hospitalisation. We aim to understand factors associated with unplanned readmission; develop a risk model to identify intensive care unit (ICU) survivors at highest readmission risk; understand the modifiable and non-modifiable readmission drivers; and develop a risk assessment tool for identifying patients and areas for early intervention.

Methods and analysis
We will use mixed methods with concurrent data collection. Quantitative data will comprise linked healthcare records for adult Scottish residents requiring ICU admission (1 January 2000–31 December 2013) who survived to hospital discharge. The outcome will be unplanned emergency readmission within 90 days of index hospital discharge. Exposures will include pre-ICU demographic data, comorbidities and health status, and critical illness variables representing illness severity. Regression analyses will be used to identify factors associated with increased readmission risk, and to develop and validate a risk prediction model. Qualitative data will comprise recorded/transcribed interviews with up to 60 patients and carers recently experiencing unplanned readmissions in three health board regions. A deductive and inductive thematic analysis will be used to identify factors contributing to readmissions and how they may interact. Through iterative triangulation of quantitative and qualitative data, we will develop a construct/taxonomy that captures reasons and drivers for unplanned readmission. We will validate and further refine this in focus groups with patients/carers who experienced readmissions in six Scottish health board regions, and in consultation with an independent expert group. A tool will be developed to screen for ICU survivors at risk of readmission and inform anticipatory interventions.