Inpatient Flow Distribution Patterns at Shanghai Hospitals.

Area of Science:

• Health Services Research
• Urban Planning
• Biostatistics

Background:

• Understanding patient distribution patterns is crucial for optimizing healthcare resource allocation.
• Empirical studies utilizing patient flow data are essential for developing accurate distribution models.
• Existing models require validation and refinement across diverse healthcare settings.

Purpose of the Study:

• To demonstrate an inpatient flow distribution model for various medical institution levels and specialties in Shanghai.
• To compare the effectiveness of four different distribution models (power, negative exponential, Gaussian, log-logistic) for simulating patient flow.
• To identify the most applicable model for predicting inpatient distribution in a major urban healthcare system.

Main Methods:

• Utilized patient flow data from Shanghai to analyze inpatient distribution.
• Applied four mathematical models: power, negative exponential, Gaussian, and log-logistic.
• Assessed model fitting using adjusted coefficient of determination (R²) and Akaike information criterion (AIC).

Main Results:

• The log-logistic function model demonstrated the best simulation effect and applicability across most hospitals in Shanghai.
• The distance-decay parameter (β) in the log-logistic model varied by hospital type, with an overall estimate of 1.67.
• Model simulations were less satisfactory for tumor, children's, and maternity hospitals, indicating a need for specialty-specific adjustments.

Conclusions:

• The log-logistic model is effective for simulating inpatient distribution in many Shanghai hospitals.
• Attenuation coefficient models are valid for Shanghai's healthcare system.
• Further specialized analysis is required to develop accurate inpatient models for specific medical specialties.