Mixed time series approaches for forecasting the daily number of hospital blood collections

Area of Science:

• Healthcare Management
• Biostatistics
• Time Series Analysis

Background:

• Accurate prediction of hospital blood collection numbers is crucial for efficient resource management and patient care.
• Existing methods may not fully capture the complex dynamics influencing blood collection volumes.

Purpose of the Study:

• To develop and validate novel methods for predicting the daily number of hospital blood collections.
• To enhance the accuracy of blood collection forecasting in a large hospital setting.

Main Methods:

• Analysis of time series characteristics of daily blood collection and outpatient numbers.
• Application of Autoregressive Integrated Moving Average (ARIMA) and linear regression models.
• Development of a combined prediction model integrating mixed time series approaches.

Main Results:

• The combined prediction model demonstrated superior accuracy compared to individual models.
• The model effectively captured the characteristics influencing blood collection numbers.
• Findings provide a basis for optimizing blood collection management strategies.

Conclusions:

• A combined prediction model using mixed time series offers a more accurate approach to forecasting hospital blood collections.
• This method accounts for both internal and external factors affecting collection volumes.
• The developed model presents a significant advancement in predicting blood collection numbers.