Figure 1. Flow Diagram for Assembly of the Analytic Sample for all Three Outcomes.
The sample included ICU hospitalizations of ≥ 1 day from participants enrolled in the NHATS 2011 cohort between the years of 2011–2017 to allow for follow-up assessment through 2018. Only the 1st ICU admission in the interval between two annual NHATS interviews for a participant (annual interval) was eligible for inclusion. After excluding participants who were admitted from a nursing home or spent ≥ 100 days in a nursing home between the pre-ICU interview and the ICU hospitalization, and those with missing data on race, 989 ICU admissions remained for consideration. Of these, 332 ICU hospitalizations were excluded because of in-hospital death (n=106), discharge to hospice (n=56), the follow-up interview occurring >365 days after discharge (n=73) or missing the entire follow-up interview (n=97). After excluding those with maximal impairment in each domain at pre-ICU interview and those with missing data because of participant death after hospital discharge, 641 admissions were eligible for the analysis of function, 458 admissions were eligible for the analysis of cognitive function, and 519 admissions for mental health. ICU = Intensive Care Unit, NH = Nursing Home.