Mortality Following Widowhood: The Role of Prior Spousal Health Open Access

Heller, Debra Ann (2014)

Permanent URL:


Previous studies have demonstrated that widowhood is associated with increased mortality risk. Although prior research suggests that the context of the predeceased spouse's death may affect this association, information is limited regarding how the rapidity of the decedent's health decline affects the survival of the bereaved spouse. The goal of this study was to combine two methods - group-based trajectory modeling and survival analysis - to identify decedents' end-of-life morbidity trajectories and to examine their association with post-widowhood survival in bereaved spouses.

Subjects included 9,967 married couples enrolled in the Pennsylvania Pharmaceutical Assistance Contract for the Elderly (PACE) Program. Using the predeceased spouse's death date as an index date, predeceased and bereaved spouses' morbidity trajectories in the prior year were evaluated for three morbidity measures: the Combined Comorbidity Score, inpatient hospitalized days, and ambulatory visits. Kaplan-Meier and Cox proportional hazards models were used to evaluate associations between morbidity patterns and post-widowhood survival over three years.

Multiple trajectories were identified for each predeceased morbidity measure, including six patterns for Combined Comorbidity, four for inpatient days, and six for ambulatory visits. Among hospice users, stable low or late onset predeceased Combined Comorbidity trajectories were associated with elevated mortality rates in the bereaved, relative to chronic high morbidity (HR=1.47 and 1.62, respectively); no effect was apparent in non-hospice users. Relative to stable medium ambulatory visits, chronic high predeceased visits were associated with a lower mortality rate in the bereaved (HR=0.67; 95% CI: 0.48, 0.92), while a stable zero visit pattern was associated with a higher rate (HR=1.32; 95% CI: 1.14, 1.53). The effects of spousal morbidity on survival were neither confounded with nor modified by age, sex, race, or place of death. However, for Combined Comorbidity and ambulatory visits, the predeceased morbidity trajectory was confounded with the widowed subject's own morbidity trajectory.

These results demonstrate the utility of group-based trajectory modeling for describing end-of-life health decline. However, the impact of spousal morbidity trajectory on post-widowhood survival was not consistent across measures, and was confounded with subjects' own morbidity. More research is needed to examine the complex pathways through which spousal illness trajectories affect post-widowhood mortality.

Table of Contents


Rationale (1)

Problem Statement (1)

Purpose Statement (5)

Research Questions (5)

Significance Statement (7)

Definition of Key Terms (8)


Introduction (10)

Marital Status and Mortality (10)

Bereavement and Mortality (14)

Early Cohort Studies (14)
Gender, Age, and Duration of Bereavement (16)
Cause-Specific Mortality of Bereaved Spouses (19)
Studies Addressing Shared Environmental Effects (20)
Meta Analyses of Bereavement and Mortality (21)

Impact of Bereavement on Other Health Measures (22)

Context of the Predeceased Spouse's Death (25)

Health Conditions of the Predeceased Spouse (25)
Expectedness of Death (29)
Caregiving Burden (31)
The Importance of Place of Death (34)
Use of Hospice Services (36)

End-of-Life Health Trajectories (36)

Summary of Current Problem and Study Relevance (40)


Introduction (41)

Population and Sample (42)

The PACE Program (42)
Widowed Cohort (43)

Research Design (44)

Procedures (44)

Instruments (45)

Data Files (46)
Computed Measures (48)

Plans for Data Analysis (51)

Phase 1: Health Trajectory Analysis (51)
Phase 2: Survival Analysis (58)
Study Limitations and Delimitations (66)


Health Trajectory Analysis (70)

Combined Comorbidity Score Trajectories (70)
Inpatient Hospitalization Days (75)
Ambulatory Visits (78)

Survival Analysis (81)

Collinearity Assessment (81)
Crude Mortality Risks and Rates (82)
Comparison of Mortality Rates to Other PACE Data (84)
Kaplan-Meier Analysis Results (85)
Cox Proportional Hazards Modeling (86)


Introduction (92)

Summary of Study (93)

Rationale and Significance (93)
Study Sample and Research Questions (95)
Methodology Used (96)

Conclusions, Implications, and Recommendations (97)

TABLES (109)



Appendix A: Emory University Institutional Review Board Letter (191)

Appendix B: SAS Code (193)

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
  • English
Research Field
Committee Chair / Thesis Advisor
Committee Members
Partnering Agencies
Last modified

Primary PDF

Supplemental Files