Development of a method for estimation of food and fluid intakes by nursing assistants in long-term care facilities: a pilot study.

The objective of this study, conducted at a 180-bed nursing home in Miami, FL, was to develop and evaluate a food and fluid intake estimation method designed specifically for practical, day-to-day use in nursing homes. Nursing assistants assisted in the development of the Food and Fluid Estimation Diagram (FFED) by providing feedback on existing estimation methods. The FFED employs depictions of generic foods/food groups commonly served in nursing homes that are partially to fully shaded to represent 5 consumption levels (0%, 25%, 50%, 75%, and 100%). To evaluate the FFED, nursing assistant (n=12) estimates of 32 food and fluid items across 4 simulated resident trays (n=384 total estimates) were compared to weighed values. Percent agreement and Cohen's kappa were used to evaluate validity, and Cronbach's alpha and the kappa statistic were used to evaluate reliability. Results revealed that 85% of the 384 intake estimates for the simulated food items were correct (Cohen's kappa=0.80, P<.001). Estimates of simulated intake were most accurate for drinks (90% correct) and least accurate for semisolid food items (69% correct). Cronbach's alpha indicated an acceptable level of reliability (0.70, P<.001). All of the nursing assistants were able to effectively use the FFED for estimating intake for the 34 simulated food items (kappa ranged from 0.65 to 0.92, P<.001). These results suggest that it is possible for nursing assistants to achieve accurate estimations of individual food and fluid items. A practical approach to convert individual item estimates into meaningful estimates of overall meal intake remains to be determined.