A Markov model to evaluate hospital readmission

Area of Science:

• Medical Statistics
• Public Health
• Epidemiology

Background:

• Analysis of non-fatal recurring events is common in chronic-degenerative disease research.
• Chronic Obstructive Pulmonary Disease (COPD) and Respiratory Failure (RF) are significant chronic-degenerative conditions.

Purpose of the Study:

• To estimate the probability of hospital readmission for patients diagnosed with COPD or RF.
• To identify factors influencing readmission rates in these patient groups.

Main Methods:

• Utilized Markov Chain modeling to analyze repeated hospital admissions.
• Employed the Nelson-Aalen estimator to estimate transition probabilities between states.
• Analyzed Puglia Region hospital discharge data from 1998-2005 for patients with COPD or RF codes.

Main Results:

• Patients diagnosed with RF showed increased readmission probability (OR = 1.618 first transition, 1.279 second transition).
• Patients with COPD or RF as the principal diagnosis at first admission had higher readmission odds (OR = 1.615 first transition, 1.193 second transition).
• Clinical severity and discharge ward did not significantly affect readmission probability.

Conclusions:

• Readmission time is influenced by the initial pathology's severity.
• For severe COPD or RF cases, effective management post-initial admission can reduce readmission risk, with conditions potentially becoming secondary diagnoses.