Question special
Lead Moderator

The secondary analysis of electronic health records ('Big Data') is a promising avenue for investigation in critical illness. Randomized controlled trials, traditionally at the top of the research reliability hierarchy, can be difficult, costly, labor intensive and often return results that are not generalizable. Prior analyses have suggested that the results and effect sizes found in observational studies are similar to those found in randomized controlled trials (

The Sepsis III Task Force performed a secondary analysis of electronic health records to develop and validate the clinical criteria for the diagnosis of sepsis and septic shock. What are the strengths and potential limitations of this approach? What changes to existing electronic health records are needed to allow for implementation and prospective validation of the Sepsis III clinical criteria?