3-item DSD — Three-item Delirium Superimposed on Dementia screener

What this tool is and how to interpret its evidence

The 3-item DSD is a candidate ultra-brief screen for delirium superimposed on dementia, not a fully validated clinical instrument. It was derived in a secondary analysis of first-day data from 391 older medical and surgical inpatients with dementia. The proposed combination uses two patient tasks—days of the week backwards and identifying the type of place—plus an observation of sleepiness. In the derivation analysis, any error or failure to answer a patient task, or observed sleepiness, made the screen positive.

The authors compared candidate combinations with a study reference assessment based on CAM criteria, structured interview, observation, record review and nurse information. The selected three-component combination had reported sensitivity of 94% and specificity of 42%. Its main potential value is therefore as a sensitive first step that sends positive cases to a more specific delirium assessment; it is not a stand-alone diagnosis.

Important limitations are that multiple candidate combinations were examined, selection and performance estimation used the same dataset, and the candidate components came from the broader assessment dataset. The paper itself presents this as a promising screener requiring future clinical validation. This page therefore classifies it as an emerging record rather than a mature independent family. No informant question is part of the selected three-component screen, formal administration time was not reported, and no separate authorised clinical form has been verified; the open primary article is linked for the complete methods and results.