Estimation of food and fluid intakes by nursing assistants in long-term care facilities: a pilot study - Evaluating Dietary Intake & Diet Composition in Seniors

Nutrition Research Newsletter, August, 2003

Nutrition and nursing professionals rely on food intake reports to identify residents who are at risk and to plan and evaluate nutrition interventions. An assessment of food intake (that is, whether a resident is eating less than 75% of most meals) is required for completion of the Minimum Data Set. Researchers and clinicians agree that it is critical to develop a reliable method for tracking food intake in nursing homes, because none of the commonly used assessment methods provide conclusively accurate estimates of food intake. For example, one commonly used intake estimation method, which requires nursing assistants to assess the food items on a meal tray as a whole and assign a value of 0%, 25%, 50%, 75%, or 100% consumed, was only 44% accurate in estimating intake (under both simulated and routine conditions) and failed to identify 38% of actual resident meals that were less than 75% consumed. Although an effective method for estimating food intake in nursing homes has not been identified, accurate estimates of intakes of individual foods have been achieved under some conditions. In numerous school plate waste studies, good estimations of individual foods have been achieved with six-point scales using trained observers. Food intake estimates have also been shown to be accurate in nursing home settings when well-trained observers used either a seven-point scale or an integer system to report intake for each individual food item. These results indicate that accurate estimations of food intake are achievable when a method based on individual food estimates is used and when estimators are well trained for the assigned task.

Given that accurate estimates of individual food items are achievable, the goal of this study was to develop and evaluate a method that nursing assistants in long-term care facilities could use to accurately estimate the amounts of individual foods consumed by residents. This study was conducted in a 180-bed nursing home in Miami, FL. The Food and Fluid Estimation Diagram (FFED) was designed so that information on both food and fluid intake could be obtained. Two slightly different versions of the FFED, adapted from the children's plate waste estimation method developed by Comstock, were developed to be printed on the back of residents' meal tickets: one for estimating breakfast intake and the other for lunch and dinner.

The FFED employs depictions of generic foods/food groups commonly served in nursing homes. There are five pictures of each food or drink; each is partially-to-fully shaded to represent one of five consumption levels (0%, 25%, 50%, 75%, or 100%). The estimator simply selects the pictures/hat best represent the amount of food and drink consumed. All instructions pertinent to the FFED were printed in English, Spanish, and Creole. The FFED was designed to be used in a two-step process, wherein a nursing assistant marks the pictures best representing intake for specific food items and another staff member computes the estimates of overall meal intake. Estimates for the overall fluid intake can be calculated by adding the cubic centimeter amounts indicated for all marked energy-containing and non-energy-containing drink items.

Seven female nursing assistants--three Hispanic (primary language Spanish) and four African American--participated in two discussion sessions in which they were asked what they liked and disliked about a number of food intake estimation methods currently in use by nursing facilities and hospitals across the country. Methods evaluated included tray diagrams, point scales, and a one-step assessment method where food items on a meal tray are assessed as a whole and assigned a value of 0%, 25%, 50%, 75%, or 100% consumed. In general, nursing assistants found these methods difficult to use because they either required many foods on the tray to be considered at one time or because they required the addition of points, fractions, or percentages from individual foods. They also found the instructions to be wordy and inflexible about various possibilities of residents' meal intake. For example, if a resident ate all of the meat on the main plate or all of a supplement but did not consume anything else, nursing assistants found it difficult to estimate the overall meal intake.

In response to comments of the nursing assistants, the FFED was developed to allow foods to be estimated individually, eliminating the need for nursing assistants to use numbers, fractions, or mathematical computations while being flexible and sensitive to various possibilities of residents' meal intake. A second staff member would then use the individual food estimates to calculate overall meal intake.

A delay between mealtime and when estimates can be recorded in the medical record has also been shown to be a barrier to accurate intake estimation. Nursing assistants' estimates are less likely to be susceptible to preconceived notions of what a resident usually eats, subjective judgments as to what constitutes adequate intake, and memory lapses about what the resident ate earlier in the day. To avoid error due to memory lapse, the FFED was designed to be placed on the back of meal tickets and marked while a meal tray is still in view.