Queueing network model for obstetric patient flow in a hospital.

Area of Science:

• Operations Research
• Healthcare Management
• Biostatistics

Background:

• Hospital patient flow is complex and challenging to manage.
• Accurate prediction of patient load is crucial for effective hospital operations and capacity planning.
• Obstetric patients offer a suitable cohort for probabilistic flow analysis due to random admissions.

Purpose of the Study:

• To develop a queueing network model for hospital patient flow.
• To analyze and predict the distribution of obstetric patients across different hospital wards.
• To validate the model's accuracy against real-world hospital data.

Main Methods:

• Utilized a two-year log of patient movements from the University of Tsukuba Hospital.
• Applied queueing theory, specifically M/G/∞ and M/M/m queue models.
• Analyzed patient admission rates and length of stay histograms for probabilistic flow modeling.

Main Results:

• Confirmed Little's Law for patient flow within individual wards.
• The proposed queueing network model accurately predicted the probability distribution of patients per ward.
• Model predictions showed strong agreement with observed patient data.

Conclusions:

• The developed queueing network model provides a reliable method for predicting patient distribution in hospital wards.
• This approach can significantly enhance hospital capacity planning and resource allocation.
• The model's simplicity and accuracy make it a valuable tool for healthcare management.