Alexis MacDonald, MD
Critical Care Attending
Bassett Medical Center
Cooperstown, NY
Disclosure information not submitted.
Andrea Berger, N/A
Biostatistical Analyst III, Biostatistics Core
Geisinger, United States
Disclosure information not submitted.
Karen Korzick, MD, MA,FCCM
Program Director CCM Fellowship
Geisinger Medical Center Danville
Danville, Pennsylvania, United States
Disclosure information not submitted.
Title: Evaluation of Frailty in Post ICU Survivor Clinic Data: Analysis of mortality and readmission
Introduction: Frailty is a term used to conceptualize the vulnerability to stressors due to a reduction in physiologic reserve. Frailty as a syndrome is complex, multifactorial, and dynamic. The purpose of our study was to describe the relationship between Frailty, as measured by Risk Analysis Index (RAI), and mortality and in the intensive care unit. Prompt recognition of frailty before and during critical illness may support interventions or timely goals of care conversations with ICU patients and families.
Methods: A retrospective chart review was conducted from November 1, 2016, through December 31, 2019, with approval for the exempt study by Geisinger’s Internal review board. The cohort comprised our post ICU survivor clinic patients admitted to the ICU with GHP insurance product, KACO, or Medicare insurance with a Geisinger primary care provider diagnosed with sepsis or respiratory failure. The RAI instrument was applied retrospectively to evaluate the relationship between frailty and our primary endpoint, mortality. Secondary endpoints include disposition, length of stay, and ICU/hospital readmission.
Results: There were 466 encounters for 394 unique patients. The mean patient age was 62.5, and 54.1% were male. The mean RAI score was 19.5, with a standard deviation of 15.3. Thirty-six patients (7.7%) died during their hospital stay, 69 (14.8%) died while hospitalized, or within 30 days of discharge, 90 (19.3%) died while hospitalized or within 60 days of discharge, and 104 (22.3%) died while hospitalized or within 90 days of discharge. As RAI scores increased, the risk of mortality outcomes significantly increased (all p < 0.05). RAI did moderately well at discriminating inpatient mortality (AUC: 0.647; 95% CI: 0.553, 0.740). RAI performance at predicting 30, 60, and 90-day mortality was better (all AUCs > 0.70). After adjusting for the effects of age and sex, RAI was still significantly associated with mortality, and AUC estimates were slightly higher than those from crude models.
Conclusion: To our knowledge, this is the first time the RAI was used to evaluate frailty in a critical care patient population. Our study also contributes to the growing body of literature that supports frailty as an independent indicator for adverse outcomes in the ICU.