A systematic review and meta-analysis of the criterion validity of nutrition assessment tools for diagnosing protein-energy malnutrition in the older community setting (the MACRo study)

Background & aims: Malnutrition is a significant barrier to healthy and independent ageing in older adults who live in their own homes, and accurate diagnosis is a key step in managing the condition. However, there has not been sufficient systematic review or pooling of existing data regarding malnutrition diagnosis in the geriatric community setting. The current paper was conducted as part of the MACRo (Malnutrition in the Ageing Community Review) Study and seeks to determine the criterion (concurrent and predictive) validity and reliability of nutrition assessment tools in making a diagnosis of protein-energy malnutrition in the general older adult community. Methods: A systematic literature review was undertaken using six electronic databases in September 2016. Studies in any language were included which measured malnutrition via a nutrition assessment tool in adults ≥65 years living in their own homes. Data relating to the predictive validity of tools were analysed via meta-analyses. GRADE was used to evaluate the body of evidence. Results: There were 6412 records identified, of which 104 potentially eligible records were screened via full text. Eight papers were included; two which evaluated the concurrent validity of the Mini Nutritional Assessment (MNA) and Subjective Global Assessment (SGA) and six which evaluated the predictive validity of the MNA. The quality of the body of evidence for the concurrent validity of both the MNA and SGA was very low. The quality of the body of evidence for the predictive validity of the MNA in detecting risk of death was moderate (RR: 1.92 [95% CI: 1.55–2.39]; P < 0.00001; n = 2013 participants; n = 4 studies; I : 0%). The quality of the body of evidence for the predictive validity of the MNA in detecting risk of poor physical function was very low (SMD: 1.02 [95%CI: 0.24–1.80]; P = 0.01; n = 4046 participants; n = 3 studies; I :89%). Conclusions: Due to the small number of studies identified and no evaluation of the predictive validity of tools other than the MNA, there is insufficient evidence to recommend a particular nutrition assessment tool for diagnosing PEM in older adults in the community. High quality diagnostic accuracy studies are needed for all nutrition assessment tools used in older community samples, including measuring of health outcomes subsequent to nutrition assessment by the SGA and PG-SGA