An empirical investigation of the reliability and validity of the U.S.D.A. model to determine least-cost hospital food service systems

The U.S.D.A. Model which compares food service systems was tested to determine its reliability and validity. The Model lists the costs and needs of each type of food service system for different hospital bed sizes and demonstrates how two systems can be compared. The Model concludes that the ready foods system is the least-costly food service system.

The U.S.D.A. Model was tested using data from eleven hospitals which use the ready foods system. Labor, material and overhead costs were compared. A description of the cost analyses is included.

A significant difference was found between the actual cost data from the elven ready foods hospitals and the costs the U.S.D.A. Model predicts these hospitals should have anticipated. The major area of difference was in the area of labor costs. The Model predicted that the man-hours required for each operation are much less than what actually occurred in the eleven hospitals tested.

The U.S.D.A. Model was based on data from six conventional and six ready foods hospitals. The U.S.D.A did not look at these twelve hospitals in terms of complete systems, but rather looked at the dishroom from one and the tray delivery system from another. This research looked at eleven ready foods systems in detail as complete systems. This may help explain the significant difference between the actual costs from eleven hospitals with the same system and the U.S.D.A. Model which was based on calculated costs from twelve hospitals only six of which used the ready foods system.

It is concluded that the U.S.D.A. Model is neither a reliable nor a valid tool for a food service manager to use in determining the type of food service system to integrate into his operation. It is further recommended that this Model not be used in determining the costs of a food service system because it may give an inaccurate measurement and cause a hospital to spend more money than planned.