Validation of the 4AT for assessing recovery from delirium in older hospital patients

Abstract

Background A crucial part of delirium care is assessing for recovery, yet there are no validated methods for this. The 4AT is a widely used delirium assessment tool, but its performance in assessing recovery remains unstudied. This study evaluated the 4AT’s performance in assessing recovery from delirium. Materials and methods In this prospective diagnostic accuracy study, older hospitalised patients (≥70 years) with reference standard delirium on enrolment were assessed 2–4 times over ≤9 days. Paired researchers independently conducted blinded assessments of (i) a reference standard (Diagnostic and Statistical Manual for Mental Disorders, 5th edition), including the Delirium Rating Scale-Revised-98 and neuropsychological tests and (ii) the 4AT (index test, score ≥ 4 positive) plus brief measures of distress and psychotic symptoms. Results A total of 120 people with delirium participated [median age 86.3, range 70–99, 67 (55.8%) female and 55 (45.8%) with dementia]. All of them completed the first two assessments, 103 (85.8%) completed three and 69 (57.5%) four. Reference standard delirium was present in 102/120 (85%), 72/103 (69.9%) and 53/69 (76.8%) cases at assessments two to four, respectively. In Receiver Operating Characteristic analyses, the 4AT’s sensitivity for detecting delirium was 0.95 (confidence interval 0.91–0.99), 0.96 (0.91–1) and 0.94 (0.88–1), and specificity was 0.67 (0.13–1), 0.88 (0.71–1) and 1 (1–1) at assessments two to four. In total 18 (15%) participants recovered from delirium. Distress was common in delirium and decreased with recovery. Conclusion The 4AT maintains diagnostic accuracy on repeated admissions and may effectively assess delirium recovery in acute hospital settings. Fewer patients than expected recovered within 9 days, suggesting more studies on the natural history of delirium in different settings would be informative.