Dissertation/Thesis Abstract

Using the Gelberg-Andersen Behavioral Model for Vulnerable Populations to Assess the Factors Associated with the Timing and Patient Outcomes of Clinical Ethics Consultations
by Homan, Mary Elizabeth, D.P.H., The University of Oklahoma Health Sciences Center, 2017, 116; 10606249
Abstract (Summary)

The study question asked: Does the late timing of an ethics consultation predict adverse patient outcomes (excess length of stay, low realization rate) adjusting for patient characteristics? As such, the study aim was to apply the Gelberg-Andersen Behavioral Model for Vulnerable Populations to predict the two adverse patient outcomes (excess length of stay, low realization rate) as a function of predisposing, enabling, need, and health behavior variables of adult patients for whom an ethics consult was performed during the hospital stay.

Using hierarchical modeling in logistic regression combined with an epidemiologic approach to sequentially test for effect modification and confounding I concluded that predisposing variables such as being male, Black or African-American, Hispanic, American Indian/Alaska Native, or 65+ years old did not negatively affect a patient’s length of stay (Model 1). Enabling variables including living wills and designation of a health care proxy did not negatively affect a patient’s length of stay. Other enabling variables such as having insurance, including Medicaid, did not provide a protective effect on length of stay. Only the health behavior variable of consults after the first two hospital days negatively affected the odds of excess length of stay (p < .001). The odds of an excess length of stay (Model 1) are six times greater when a patient had a late ethics consultation compared to those patients who had an early ethics consultation (OR = 5.98; 95% CI [3.73, 9.61]). This finding clearly has implications for the practice. Earlier ethics consultation may reduce length of stay.

In Model 2, low realization rate, using the same epidemiological modeling strategy as excess length of stay, the predisposing variable of being Black or African-American (p = 0.43), the enabling variable of the presence of a health care proxy (p = 0.03), and the health behavior variable of a consultation after the first two days (p < .001) affected the odds of a low realization rate. The odds of a low realization rate are two times greater for Black patients compared to all other races (OR = 2.59; 95% CI [1.03, 6.54]). The odds of a low realization rate are half that of patients without a health care proxy than for those who had a health care proxy (OR = 0.53, 95% CI [0.30, 0.94]). The odds of a low realization rate if a consultation is conducted after two days is three times greater than if a consultation was conducted within the first two hospital days (OR = 2.99, 95% CI [1.87, 4.79]).

The findings from this study add to the ethics literature for two main reasons: 1. This was the largest study conducted at a Catholic health care organization and 2. I modeled a unique method of looking for predictors of two replicable and generalizable adverse patient outcomes. Stratifying by end-of-life cases versus non-end-of-life cases strengthened the argument that late ethics consultations affect both excess length of stay and low realization rate in the negative. Reductions in excess length of stay by one day would be considered clinically meaningful which represents a 12.5% reduction in median length of stay for this population.

Indexing (document details)
Advisor: Bratzler, Dale W.
Commitee: Fox, Mark D., Johnson, David L., Kinney, Sharyl, Raskob, Gary E.
School: The University of Oklahoma Health Sciences Center
Department: Allied Health Sciences
School Location: United States -- Oklahoma
Source: DAI-B 79/01(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Medical Ethics, Public health, Health care management
Keywords: Bioethics, Clinical ethics, Clinical ethics consultation, Ethics consultation, Gelberg-Andersen Model
Publication Number: 10606249
ISBN: 9780355153002
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest