Background: Accurate monitoring of patient food intake supports effective clinical nutrition care, foodservice performance monitoring, and food waste reduction in hospital settings. Mater Hospital South Brisbane uses CBORD MobileIntake, a visual estimation method completed by dining host staff at tray collection for every patient at every meal. However, its accuracy when embedded in routine practice had not been formally evaluated.
Aim: To assess the accuracy of MobileIntake visual estimations compared with the gold‑standard Weighed Food Record, and to examine variability in estimation accuracy by staff member and food item.
Methods: A five‑day audit was conducted during breakfast and lunch services in the main kitchen at Mater Hospital South Brisbane. Intake data from 161 meal trays, comprising 1029 individual food items and 186 unique food items, were collected. Visual intake estimates recorded in MobileIntake were compared with weighed plate waste data. Accuracy was assessed overall, across 20 dining host staff, and for commonly served food items. Energy and protein intake estimates were also compared between methods.
Results: Overall MobileIntake accuracy was 67.4%, with a tendency to overestimate both energy and protein intake. Mean estimated energy intake was substantially higher using MobileIntake compared with Weighed Food Record for both breakfast and lunch. Staff accuracy varied, with 3 out of 20 staff achieving less than 50% accuracy. Estimation accuracy differed by food item, with the highest accuracy appeared to be for items served in portion controlled or fully filled vessels (e.g. served in full ramekins) and poorest accuracy for items typically served as pieces or mixed items (e.g. roast vegetables and sandwiches). Overall, the energy and protein intake calculation from MobileIntake overestimated by 15% and 14% respectively.
Conclusion: MobileIntake visual estimation demonstrated reasonable accuracy in routine hospital foodservice practice, with variability between staff and food items and consistent overestimation of intake. Targeted staff training, regular audit processes, and exploration of digital or AI‑assisted plate waste assessment may improve data quality. These findings support ongoing quality improvement initiatives to strengthen nutrition intake monitoring, foodservice evaluation, and sustainability outcomes in acute healthcare settings.