Risk prediction in the community: A systematic review of case-finding instruments that predict adverse healthcare outcomes in community-dwelling older adults

5 years ago

Risk prediction in the community: A systematic review of case-finding instruments that predict adverse healthcare outcomes in community-dwelling older adults

Few case-finding instruments are available to community healthcare professionals. This review aims to identify short, valid instruments that detect older community-dwellers risk of four adverse outcomes: hospitalisation, functional-decline, institutionalisation and death. Data sources included PubMed and the Cochrane library. Data on outcome measures, patient and instrument characteristics, and trial quality (using the Quality In Prognosis Studies [QUIPS] tool), were double-extracted for derivation-validation studies in community-dwelling older adults (>50 years). Forty-six publications, representing 23 unique instruments, were included. Only five were externally validated. Mean patient age range was 64.2–84.6 years. Most instruments n =18, (78%) were derived in North America from secondary analysis of survey data. The majority n =12, (52%), measured more than one outcome with hospitalisation and the Probability of Repeated Admission score the most studied outcome and instrument respectively. All instruments incorporated multiple predictors. Activities of daily living n =16, (70%), was included most often. Accuracy varied according to instruments and outcomes; area under the curve of 0.60–0.73 for hospitalisation, 0.63–0.78 for functional decline, 0.70–0.74 for institutionalisation and 0.56–0.82 for death. The QUIPS tool showed that 5/23 instruments had low potential for bias across all domains. This review highlights the present need to develop short, reliable, valid instruments to case-find older adults at risk in the community.

Publisher URL: www.sciencedirect.com/science

DOI: S0378512215005952