4AT: 4 'A's Test

The original prospective validation compared the 4AT with a geriatrician's DSM-IV delirium assessment and reported 89.7% sensitivity and 84.1% specificity. The title reports 234 participants whereas the abstract reports 236, and a corrigendum changed the Table 1 non-delirium count to 209, so a single denominator or 2 by 2 table should not be reconstructed from the publication.

This emergency-department study enrolled 419 older adults and compared the 4AT with a geriatrician reference assessment, with 350 participants having the collateral history required by the standard analysis. Sensitivity was 0.93 and specificity 0.91 both in that group and in an operational analysis of 378 participants in which missing collateral history was handled by a prespecified rule.

This post-anaesthesia-care study assessed 543 adults, of whom 22 had delirium by the reference assessment. The 4AT had 95.5% sensitivity and 99.2% specificity, but the low delirium prevalence and study exclusions make the very high estimates setting-specific.

This blinded multicentre study included 785 analysable older acute medical inpatients, with 392 allocated to the 4AT and an independent DSM-IV reference assessment. For the 4AT, sensitivity was 76% and specificity 94%, illustrating lower sensitivity than in several single-centre validations.

This 4AT-specific review included 17 studies and 3,702 observations across acute medical, emergency, surgical and stroke settings. Pooled sensitivity and specificity were both 0.88, with similar estimates after excluding stroke cohorts, although study methods and thresholds still varied.

In the research-assessment phase, 137 cardiac-surgery patients yielded 85% sensitivity and 90% specificity for the 4AT. During a 179-patient nurse-delivery phase sensitivity fell to 58% while specificity remained 94%, showing that implementation can materially alter performance.

This review synthesised 38 studies and 7,378 older adults across several delirium instruments rather than reviewing the 4AT alone. For the 4AT, pooled sensitivity was 0.76 and specificity 0.89, supporting good overall discrimination but also variation by setting and study method.

This prospective specialist-palliative-care study used random test order and a blinded DSM-5 reference assessment; 134 participants were included in the main analysis and 44 had delirium. Sensitivity was 89% and specificity 94%, providing useful evidence in a population often under-represented in validation studies.

This prospective hip-fracture study compared the 4AT with an independent DSM-5 assessment within three hours in 103 older patients. Sensitivity was 87.5% and specificity 91.1%, extending direct bedside evidence into a contemporary orthogeriatric cohort.