Intermittent Demand Forecasting in a Tertiary Pediatric Intensive Care Unit

Area of Science:

• Healthcare Operations Management
• Supply Chain Analytics
• Pediatric Intensive Care

Background:

• Demand forecasting for medical supplies significantly impacts healthcare operating costs and patient care quality.
• Inaccurate forecasts lead to inventory surplus or shortages, potentially compromising patient safety.
• Uncertainty in medical supply consumption creates intermittent demand patterns (lumpy, erratic, smooth, slow-moving).

Purpose of the Study:

• To improve the accuracy of medical supply demand forecasting in a tertiary pediatric intensive care unit.
• To compare the forecast accuracy of various demand forecasting methods for intermittent demand patterns.

Main Methods:

• Classification of medical supply demand patterns into lumpy, erratic, smooth, and slow-moving categories.
• Comparison of multiple demand forecasting methods, including Croston's method, single exponential smoothing, and Simple Moving Average (SMA).
• Evaluation of forecast accuracy based on demand consumption patterns.

Main Results:

• Croston's method combined with single exponential smoothing demonstrated the highest accuracy for lumpy, erratic, and slow-moving demand.
• The Simple Moving Average (SMA) method was most effective for forecasting smooth demand.
• Integrating demand pattern classification with appropriate forecasting models significantly minimized forecasting errors.

Conclusions:

• Classifying demand patterns is essential for selecting the optimal forecasting model.
• The combined approach enhances forecasting accuracy, leading to improved patient safety.
• This framework reduces inventory management costs in healthcare institutions.