Start Date

11-12-2016 12:00 AM

Description

We investigate the effect of unobserved patient health status on patient readmission rates and the impact of telehealth on patient health status. We develop a hidden Markov model to capture the evolving latent health status of a patient to model its impact on readmissions. We obtained a large, inpatient panel dataset of Congestive Heart Failure patient visits along with the American Hospital Association IT Supplement data. We find that telehealth exerts a positive impact only on patients in less healthy states, while this impact diminishes as patients’ health improves. Our results also show that less healthy patients tend to incur significantly higher readmission rates compared with healthier patients. These results suggest nonclinical factors such as patients latent health status can impact readmission significantly. Focusing solely on hospital readmission rates may yield myopic policies.

Share

COinS
 
Dec 11th, 12:00 AM

What Drives Patient Readmissions? A new Perspective from the Hidden Markov Model Analysis

We investigate the effect of unobserved patient health status on patient readmission rates and the impact of telehealth on patient health status. We develop a hidden Markov model to capture the evolving latent health status of a patient to model its impact on readmissions. We obtained a large, inpatient panel dataset of Congestive Heart Failure patient visits along with the American Hospital Association IT Supplement data. We find that telehealth exerts a positive impact only on patients in less healthy states, while this impact diminishes as patients’ health improves. Our results also show that less healthy patients tend to incur significantly higher readmission rates compared with healthier patients. These results suggest nonclinical factors such as patients latent health status can impact readmission significantly. Focusing solely on hospital readmission rates may yield myopic policies.