Systematic Review of Dementia Risk Screening Tools in Primary Care Settings.

Introduction: This preregistered (PROSPERO ID: 272529) systematic review analyzed the reliability, validity, acceptability, and feasibility of methods to screen and identify primary care patients at increased risk for dementia.

Methods: Inclusion criteria were studies of dementia risk screening methods applied, or designed for use, in primary care, or derived or validated with primary care data for use with community-dwelling adults (≥18 years) without dementia/cognitive impairment. Articles were limited to full text peer-reviewed studies published in English. Searches were conducted from database inception until 09.21.21/09.22.21 (updated 09.04.22, 09.16.24) in PubMed, MEDLINE, EMBASE, PsycINFO and COCHRANE with backwards snowballing. Risk Of Bias (ROB) was assessed using the Newcastle-Ottawa Scale for non-randomized studies, the Cochrane ROB tool for randomized trials-V2, and the Critical Appraisal Skills Program qualitative checklist.

Results: Eleven validated tools and scores were investigated in 34 studies. Participant numbers ranged from 100-94,000,000. Tools and scores were used or validated in one (six tools) to 13 studies (one tool). Most studies were from higher-income countries. ROB was low in most studies. C-statistics were most commonly used (range = .52-.86). Tools and scores generally had acceptable prediction but performed poorly in samples differing from the derivation sample. The CAIDE and LIBRA demonstrated acceptable validity in midlife samples, and the LIBRA in older samples <80 years of age.

Discussion: The CAIDE and LIBRA are recommended for midlife and older adults <80 years, but further research is needed for adults ≥80 and lower-middle income countries.