Analysis of meal census patterns for forecasting menu item demand

Area of Science:

• Healthcare Management
• Operations Research
• Statistical Modeling

Background:

• Accurate patient census data is crucial for hospital operations.
• Forecasting systems require understanding patient census patterns.
• Existing methods may not fully capture meal-specific census variations.

Purpose of the Study:

• To identify functional relationships between patient census and official midnight census.
• To inform the design of a statistical menu item forecasting system.
• To analyze patient census data for three meals (breakfast, lunch, supper) against the midnight census.

Main Methods:

• Graphical analysis of patient census data.
• Analysis of variance (ANOVA) to compare meal census with midnight census.
• Appraisal of three distinct forecasting system design options.

Main Results:

• Identified reliable, predictable patterns in patient census data.
• Established functional relationships between meal times and midnight census.
• All three forecasting design options yielded identical quantity predictions.

Conclusions:

• Patient census data exhibits patterns suitable for mathematical forecasting models.
• A statistical forecasting system can be effectively designed based on identified relationships.
• The complexity of the forecasting model did not significantly impact prediction accuracy in this study.