Markov chain modelling for geriatric patient care.

Area of Science:

• Geriatric Medicine
• Biostatistics
• Health Services Research

Background:

• Geriatric patient care pathways are complex.
• Understanding factors influencing care duration is crucial for resource allocation and patient management.
• Markov chain modeling offers a robust framework for analyzing time-dependent processes.

Purpose of the Study:

• To demonstrate the applicability of Markov chain modeling to geriatric patient care data.
• To utilize these models for assessing the impact of covariates on patient care trajectories.
• To interpret the influence of patient age and admission year on care patterns.

Main Methods:

• Phase-type distributions were fitted using maximum likelihood estimation.
• Analysis of events ending patient care periods to estimate event probability dependence on care phases.
• Inclusion of patient age at admission and year of admission as covariates.

Main Results:

• Significant differential effects of covariates (age, admission year) on model parameters were identified.
• Interpretations of covariate effects provided insights into care pathway influences.
• The fitted models demonstrated the impact of specific patient characteristics on care duration.

Conclusions:

• Phase-type distribution models are suitable for describing geriatric patient times spent in care.
• The ordered phases in the models reflect an interpretable structure related to increasing care intensity.
• The study validates Markov chain modeling for analyzing complex healthcare utilization patterns in the elderly.