Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

ORCID

N/A

Access Type

Open Access Thesis

Document Type

thesis

Degree Program

Public Health

Degree Type

Master of Science (M.S.)

Year Degree Awarded

2015

Month Degree Awarded

February

Abstract

Approximately 20% of congestive heart failure (CHF) patients are readmitted within 30 days of hospital discharge, a rate which may be affected by in-hospital and post-discharge care. Reducing this rate is important to hospitals, both to improve outcomes and to avoid reductions in Medicare reimbursement. Assessing outcomes within a short post-discharge window best measures the impact of the care, planning, and followup of that admission; but most research on the effects of changes in CHF care has measured outcomes over periods longer than 30 days, adding the unpredictable long-term course of CHF to the factors affecting the outcome. As well, almost no studies to date have included the appreciable effects of CHF comorbidities in their analyses.

This study addresses these needs by measuring rates of 30-day all-cause readmission, and by adjusting for comorbidities and demographic factors in our analysis.

We hypothesize that an improved CHF care protocol including both in-hospital and post-discharge components will reduce the risk of readmission, and may alter the rate of change of that risk.

We have analyzed as an interrupted time series data on 2764 discharges of CHF patients from a hospital that implemented such a change to assess the effect of the new protocol on the readmission risk and on the trend in that risk, comparing outcomes in the 22 months preceding introduction of the new protocol to those in the first 31 months of full implementation. Using multiple logistic regression, we have tested for an association between the new protocol and both the unadjusted risk of readmission, and that risk in a model including comorbidities and demographic factors as covariates.

Neither model found a statistically significant association between introduction of the protocol and log-odds of readmission (unadjusted p = 0.847, adjusted p = 0.755) or between introduction of the protocol and change in risk of readmission over time (unadjusted p = 0.437, adjusted p = 0.313).

These results, in comparison with other published results, can clarify what changes to care protocols have been shown to be effective. Further, post hoc power analysis of this study can inform study design for further research.

DOI

https://doi.org/10.7275/6374054

First Advisor

Penelope Pekow

Second Advisor

Nicholas Reich

Fourth Advisor

Brian Whitcomb

Included in

Biostatistics Commons

Share

COinS