Performance of visual methods for estimating dietary intake in the hospital setting - A systematic review and meta-analysis.

Abstract

Inadequate dietary intake is one of the main factors contributing to hospital malnutrition, leading to unfavorable clinical outcomes. Visual estimation methods (VEMs) offer a practical alternative for monitoring dietary intake, but their accuracy compared to reference methods remains under debate. This systematic review with meta-analysis aimed to evaluate the performance and accuracy of VEMs in estimating dietary intake in hospitalized patients. This study followed the PRISMA guidelines. A comprehensive search was conducted on October 1 and 2, 2025, and subsequently updated on August 5, 2026, using the same search strategy in the PubMed, Scopus, Embase, and Web of Science databases. Two researchers independently selected, extracted, and assessed data quality using the JBI tools (PROSPERO ID: CRD420251151609). Random-effects meta-analyses calculated pooled mean differences with 95% CIs. Intra-subject correlation was accounted for using paired standard errors derived from Bland-Altman limits, directly reported difference, or an imputed mean correlation. Publication bias and sensitivity were evaluated using funnel plots, Egger's test, and leave-one-out analyses. This systematic review included 27 studies on visual tools for estimating dietary intake in hospitalized patients. Of these, 6 were included in the meta-analysis. Publication bias analysis, using Egger's test, did not indicate statistical significance (p = 0.347), providing no evidence of publication bias. Agreement between dietary estimation tools and the reference method demonstrated an overall mean difference (MD) of -14.91 kcal (95% CI: -45.83 to 16.00). The negative MD indicated a trend towards underestimation; however, the pooled effect was not statistically significant. High heterogeneity was observed among studies in the overall analysis (I² = 88.4%; p < 0.001). The use of visual estimation methods for dietary intake in hospitalized patients, despite a mild tendency toward underestimation, suggests potential utility as a complementary approach for monitoring intake. These tools may offer a practical alternative in daily hospital routines, particularly in clinical settings where routine meal weighing is less feasible to implement.