Testing the Transportability of Sepsis Subtypes to Patients With ARDS for Postdischarge Outcomes

Area of Science:

• Critical Care Medicine
• Pulmonology
• Data Science

Background:

• Survivors of acute respiratory distress syndrome (ARDS) face significant long-term health challenges.
• Previous research identified 5 sepsis survivor subtypes but lacked a method for applying them to ARDS patients.

Purpose of the Study:

• To determine if sepsis survivor subtypes can be operationalized in ARDS patients.
• To assess if these operationalized subtypes predict post-discharge mortality and readmission risk.

Main Methods:

• Secondary analysis of ARDS patients from the ROSE trial.
• Compared two methods for subtype operationalization: transported algorithm vs. de novo latent class analysis (LCA).
• Used random-effect logit models to link subtypes with 12-month mortality.

Main Results:

• 580 participants were assigned to 5 subtypes using the transported method.
• Both transported and de novo subtypes were significantly associated with mortality and readmission.
• Agreement between methods was moderate (kappa=0.361), strongest in extreme risk groups.

Conclusions:

• Sepsis survivor subtypes are applicable and can be operationalized in ARDS patients.
• Both external transport and internal rederivation methods identify patient groups with varying mortality and readmission risks.
• The choice of operationalization strategy may depend on the specific clinical application.