Food Satisfaction Survey as a predictor of significant unplanned weight loss
Significant unplanned weight loss in aged care residents is associated with increased frailty, hospitalisation, and mortality, and in Australia it is tracked nationally as a quality indicator. Residents’ experience of food and dining, including quality, choice, and staff support at mealtimes, may plausibly influence nutritional intake and therefore weight outcomes. In this analysis I explore whether there is an association between an aged care home’s Experience Rating in MOA’s Food Satisfaction Survey and its prevalence of significant unplanned weight loss reported in the National Aged Care Mandatory Quality Indicator Program (NACMQIP).
Methods
Weight-loss data submitted for NACMQIP via MOA and results data for the Food Satisfaction Survey, both for the Oct–Dec 2025 quarter, were extracted from the MOA database. Homes were retained if they appeared in both data sources and had a Food Satisfaction Survey completion rate of at least 90%.
In this analysis prevalence of significant unplanned weight loss was the dependent variable and Experience Rating (0–100%) was the predictor. See the definitions in Table 1.
| Definition | ||
|---|---|---|
 |
Experience Rating | The aggregate of agree/strongly agree (or their equivalent) percentages |
 |
Significant unplanned weight loss | A resident who experienced a drop in weight of 5% or more compared to the previous quarter |
A locally quadratic LOESS (locally estimated scatterplot smoothing) curve with Experience Rating on the x-axis and prevalence of significant unplanned weight loss on the y-axis was plotted. Each home’s contribution to the fit was weighted by its number of residents assessed for significant unplanned weight loss. A span of 0.75 was chosen—a common default—which means that 75% of the data was used in each local fit. The whole range was used for fitting, but only Experience Ratings between 70% and 100% were visualised as data is relatively sparse outside this region, so the curve cannot be reliably estimated.
A logistic regression was fitted, where the outcome was the number of residents that experienced significant unplanned weight loss, the binomial denominator was the number of residents assessed for significant unplanned weight loss, and the predictor was the Experience Rating from the home’s Food Satisfaction Survey. A model that used Jan–Mar 2026 weight-loss data so the Food Satisfaction Survey was lagged by one quarter was also considered as a sensitivity analysis to reduce the risk of reverse causality.
Statistical significance was assessed at a conventional 5% level. The statistical computing software R version 4.5.3 (2026-03-11) was used to fit models for the analysis.
Results
There were 394 homes that used both the NACMQIP tool and Food Satisfaction Survey in Oct–Dec 2025, with at least a 90% completion rate for the latter. This corresponds to 27,239 residents assessed for significant unplanned weight loss.
Figure 1 shows that prevalence of significant unplanned weight loss declined steadily as Experience Rating increased, with some flattening out near an Experience Rating of 100%.
The odds of significant unplanned weight loss were reduced by an estimated 6.5% (95% CI: 3.1% to 9.8%; p<0.001) for every 10-percentage-point increase in a home’s Food Satisfaction Survey Experience Rating. The equivalent figure is 5.2% (95% CI: 1.4% to 8.8%; p=0.007) for the sensitivity analysis using a one-quarter-lagged Food Satisfaction Survey Experience Rating as the predictor.
Discussion
There was a relationship between how a home scores in the Food Satisfaction Survey and how many residents experience significant unplanned weight loss. This might indicate that homes whose residents have a positive experience of food quality, staff and service, choice and autonomy, meal enjoyment, have their needs and preferences met, and show positive sentiment about food, tend to experience significant unplanned weight loss less frequently. The consistency of the effect when using a later quarter’s weight-loss data (Jan–Mar 2026) as the outcome makes simple reverse causality an unlikely explanation.
This analysis has several limitations. As an observational, facility-level study, it does not rule out confounding by factors such as staffing levels or resident case-mix, nor does it establish that the relationship holds at the individual resident level. The Experience Rating also aggregates several distinct domains, so this analysis does not identify which component drives the association.
Although this is an associational finding and not necessarily surprising, the analysis is consistent with food and dining experience being one modifiable factor among several contributing to unplanned weight loss. Because the Experience Rating aggregates multiple domains, further work identifying which specific domains drive the association would be needed before recommending a targeted intervention.