Text

theses

Context: Among cancer survivors, multi-morbidity is common, with the majority of survivors co-managing at least one additional condition alongside their cancer care needs. A co-diagnosis of diabetes, a heterogeneous group of metabolic disorders, occurs in 16-29% of all cancer survivors. Both diabetes and cancer place considerable stress on the individuals and populations affected by these conditions. Research has identified multiple factors associated with better outcomes for specific cancers and diabetes when considered separately; however, there is a need to develop a holistic understanding of the impact of co-managing these conditions. Additionally, health disparities for cancer and diabetes among specific racial and socioeconomic groups for mortality and several health-related outcomes have persisted for decades. This study was guided by Link and Phelan’s theory of ‘fundamental causes,’ a social epidemiological theory that aims to identify the fundamental social causes of health inequalities by exploring four key relationships between socioeconomic status and health outcomes. According to the theory of ‘fundamental causes’ theory socioeconomic disparities in health are rooted in unequal resource distribution which positions those with more resources in a better position to adopt behaviors that will benefit their health. Guided by this theoretical framework, this dissertation examines the impact of social factors on co-managing diabetes and cancer on multiple risk factors and disease outcomes. The overall goal of this dissertation is to examine the impact of social factors and diabetic status on multiple outcomes in a nationally representative sample of cancer survivors, as articulated in the three specific aims below. Specific Aims: The aims of this dissertation were to: (1) identify factors associated with the co-occurrence of diabetes and cancer and factors known to be associated with higher risks for poor outcomes (i.e., obesity and physical inactivity); (2) examine differences in self-assessed health (i.e., proxy for mortality), physical and mental quality of life between diabetic cancer survivors and non-diabetic survivors; and (3) examine differences in multiple psychosocial outcomes for diabetic cancer survivors and non-diabetic cancer survivors. Design, Setting and Patients: All aims of this study utilize data from the U.S. Medical Expenditures Panel Survey Household Component (MEPS-HC) and the supplemental self-administered survey “Experiences with Cancer” collected from 2011-2012. The study sample is restricted to non-institutionalized civilians who reported having a diagnosis of cancer during their lifetime who completed the supplemental questionnaire. Cases of non-melanoma skin cancers were excluded. Study subjects were stratified based on a comorbid diagnosis of diabetes compared to non-diabetic cancer survivors. Descriptive, cross-sectional analyses were conducted to characterize the study populations for all three aims. For Aim 1, a series of nested multivariate logistic regressions were constructed to examine predictors associated with co-occurring diabetes and cancer. Bivariate analyses explored significance of cancer site and co-morbid chronic conditions. Chi-square analyses were analyzed to identify differences between groups on lifestyle and preventive health behaviors. Multivariate logistic regression models were specified to examine the impact of social factors on obesity, physical inactivity and likelihood of complex multi-morbidity. For Aim 2, a series of nested, multivariate logistic regressions were constructed to examine predictors of increased mortality, measured by proxy using self-assessed health. Additionally, a hierarchical multivariate regression model was specified to examine the impact of social factors and diabetic status on physical and mental functioning. Physical and mental functioning were measured using the subscales for the SF-12. For Aim 3, multivariate regression models were specified to assess factors associated with psychosocial variables including: non-specific psychological distress measured by the Kessler Index; depression measured by the Patient Health Questionnaire-2; mental health functioning measured by the SF-12 mental health composite; and cancer related worries measured by two single-item measures assessing chance of recurrence and fears about recurrence. For a subset of the study sample, longer-term cancer survivors, a logistic model was specified to identify factors that are associated with clinically relevant levels of psychological distress. Results: Aim 1 of this dissertation showed that the following are risk factors for the co-occurrence of diabetes and cancer: (1) survivors with a high school education had 1.63 greater odds [CI 1.03-2.50, p=0.04] and those with less than a high school education had 3.15 [CI: 1.75-5.69, p<0.001] greater odds than survivors with bachelor’s degrees or higher; (2) African American/black survivors had 2.05 greater odds [CI: 1.18-3.55, p=0.001] than Caucasian/White survivors; and (3) survivors aged 55-less than 65 had 12.28 greater odds [CI: 160-94,28], aged 65-less than 75 had 10.19 greater odds [CI:1.31-79.39, p=0.03]; aged 75 and older had 7.79 greater odds [CI: 0.97-62.22, p=0.05] than survivors aged 18-39 years old. Female survivors had 37% lower odds [CI: 0.41-0.97, P=0.04] of diabetic status compared to male survivors; and survivors 10 or more years out from treatment had a 43% lower odds [0.33-0.98, P=0.04] of diabetic status compared to recently diagnosed survivors. Diabetic cancer survivors had 48% lower odds of meeting physical activity recommendations [CI: 0.34-0.80, p=0.003] and 73% lower odds of meeting weight recommendations [CI: 0.16-0.48, p<0.001] compared to non-diabetic cancer survivors, controlling for social and cancer related variables. Diabetic cancer survivors had 186% [CI: 1.84-4.45, p<0.001] greater odds of having five or more chronic conditions compared to non-diabetic survivors. Examinations of differences in physical health (Aim 2) revealed that being diabetic was associated with 1.71% greater odds of reporting poor/fair health holding all else constant (p<0.001) compared to non-diabetic survivors; however, no differences in mental health were found. The results of Aim 3 demonstrated that diabetic status was not significantly associated with any of the psychosocial outcomes examined; however, physical functioning was associated with lower non-specific distress (B=-0.14, p<0.001) and depression (B=-0.05, p<0.001), and increased mental health functioning (B=0.155, p<0.001). Conclusions: The research in this dissertation confirmed that race and educational disparities exist in the prevalence of co-morbid diabetes among cancer survivors and the mechanisms underlying these disparities warrants further study. Further, diabetic cancer survivors were less likely to engage in protective health behaviors (i.e., meeting physical activity recommendations and maintaining healthy weight) and were at higher risk for complex multi-morbidity compared to non-diabetic survivors. Diabetic cancer survivors had poorer health and poorer health related quality of life compared to non-diabetic cancer survivors. While diabetic status alone was not shown to be associated with psychosocial health outcomes compared to non-diabetic cancer survivors, poorer physical health was shown to be related to poorer psychosocial adjustments. Research to identify individual and contextual level barriers are needed develop prevention strategies that may help mitigate the excess risk of the co-management of diabetes and cancer. Future research is needed to distinguish the components of the diabetic cancer survivorship phenomenon that are issues of inequity versus issues of disparate outcomes based on biological differences associated with race/ethnicity.

ETD_7977

Ph.D.

Includes bibliographical references

by Denalee O'Malley

doi:10.7282/T35X2CVG

ETD doctoral

The author owns the copyright to this work.