Objectives: To develop and validate an abbreviated version of the Cognitive Failure Questionnaire that can be used by patients as part of self-assessment to measure functional cognitive outcome in ICU survivors.
Design: A retrospective multicenter observational study.
Setting: The ICUs of two Dutch university hospitals.
Patients: Adult ICU survivors.
Interventions: None.
Measurements and main results: Cognitive functioning was evaluated between 12 and 24 months after ICU discharge using the full 25-item Cognitive Failure Questionnaire (CFQ-25). Incomplete CFQ-25 questionnaires were excluded from analysis. Forward selection in a linear regression model was used in hospital A to assess which of the CFQ-25 items should be included to prevent a significant loss of correlation between an abbreviated and the full CFQ-25. Subsequently, the performance of an abbreviated Cognitive Failure Questionnaire was determined in hospital B using Pearson's correlation. A Bland-Altman plot was used to examine whether the reduced-item outcome scores of an abbreviated Cognitive Failure Questionnaire were a replacement for the full CFQ-25 outcome scores. Among 1,934 ICU survivors, 1,737 were included, 819 in hospital A, 918 in hospital B. The Pearson's correlation between the abbreviated 14-item Cognitive Failure Questionnaire (CFQ-14) and the CFQ-25 was 0.99. The mean of the difference scores was -0.26, and 95% of the difference scores fell within +5 and -5.5 on a 100-point maximum score.
Conclusions: It is feasible to use the abbreviated CFQ-14 to measure self-reported cognitive failure in ICU survivors as this questionnaire has a similar performance as the full CFQ-25.