Comparing the 4AT and CAM for Delirium Screening in Older Adults: A Rapid Review of Diagnostic Accuracy

Introduction

Delirium is common and frequently under-recognised in older hospital inpatients. The 4AT and Confusion Assessment Method (CAM) are widely used bedside assessments, but their comparative accuracy against an independent reference standard is not clearly summarised. We compared their accuracy in older inpatients.

Methods

Rapid review. PubMed was searched in English for the previous 10 years using ("4AT" OR "4 A's test") AND ("confusion assessment method" OR CAM) AND delirium AND (accuracy OR sensitivity OR specificity), yielding 32 records. Eligibility required accuracy data for both the 4AT and the standard CAM against an independent reference standard; studies validating either tool alone were excluded by design. Two studies met criteria. The directly comparative primary study was appraised using QUADAS-2.

Results

Evidence types differed. A UK multi-centre prospective comparative diagnostic accuracy study (Shenkin SD, Fox C, Godfrey M. BMC Med 2019;17:138; n=785) found 4AT sensitivity 76% versus CAM 40%, and specificity 94% versus 100%; only 17% of eligible patients were recruited (mainly consent barriers), raising a selection concern. A systematic review and meta-analysis (Lin CJ, Su IC, Huang SW. Ageing Res Rev 2023;90:102025; 38 studies of delirium assessment tools; n=7,378) provided indirect pooled estimates: sensitivity was similar for 4AT and CAM (76% vs 80%), whereas CAM specificity was higher (89% vs 98%; P=.01). Estimates varied by setting, assessors, and sample characteristics.

Conclusions

Neither tool is uniformly superior. The 4AT may support initial case-finding because it is brief, does not require formal tool-specific training, and was more sensitive in the direct comparison. CAM may produce fewer false-positive screens where staff are trained. Neither replaces expert clinical diagnosis. Head-to-head evidence against an independent reference standard remains scarce; a single-database rapid search may also have missed comparative studies, so these findings are indicative rather than definitive.