There is substantial variability in intensive care unit (ICU) utilisation and quality of care. However, the factors that drive this variation are poorly understood. This study uses a novel adaptation of positive deviance approach-a methodology used in public health that assumes solutions to challenges already exist within the system to detect innovations that are likely to improve intensive care.We used the Philips eICU Research Institute database, containing 3.3?million patient records from over 50 health systems across the USA. Acute Physiology and Chronic Health Evaluation IVa scores were used to identify the study cohort, which included ICU patients whose outcomes were felt to be most sensitive to organisational innovations. The primary outcomes included mortality and length of stay. Outcome measurements were directly standardised, and bootstrapped CIs were calculated with adjustment for false discovery rate. Using purposive sampling, we then generated a blinded list of five positive outliers and five negative comparators.Using rapid qualitative inquiry (RQI), blinded interdisciplinary site visit teams will conduct interviews and observations using a team ethnography approach. After data collection is completed, the data will be unblinded and analysed using a cross-case method to identify themes, patterns and innovations using a constant comparative grounded theory approach. This process detects the innovations in intensive care and supports an evaluation of how positive deviance and RQI methods can be adapted to healthcare.The study protocol was approved by the Stanford University Institutional Review Board (reference: 39509). We plan on publishing study findings and methodological guidance in peer-reviewed academic journals, white papers and presentations at conferences.
View details for PubMedID 28615274