Text

theses

Chemotherapy for non-hematologic cancers is primarily administered and managed in the outpatient setting, and little is known about factors associated with hospitalizations for patients experiencing treatment-related adverse events so severe as to require inpatient care. This retrospective analysis conducted within the SEER-Medicare linked dataset in the non-metastatic lung and colorectal cancer populations revealed predictors related to the likelihood of initial unplanned hospitalization, as well as those predictive of the number of hospitalizations experienced. The tumor types were selected to allow study among two of the most frequently admitted solid tumors identified in the literature, from a nationally validated, population-based dataset comprised of patients over age 65, a group that is underrepresented in clinical trials. Factors including patient age, sex, race, marital status, degree of residential urbanization, median income, educational level, cancer type, stage, receipt of radiation therapy and comorbidities were studied and considered as predictive factors. Two separate tumor-based cohorts, lung (n = 2457) and colorectal cancer (n = 1485), were constructed and analyzed in parallel. Patient eligibility included those age greater than 65 years at the time of diagnosis, non-metastatic lung or colorectal cancer as their first malignant primary tumor, uninterrupted Medicare Part A and B coverage with no HMO enrollment, and those who received intravenous chemotherapy at least one time prior to experiencing a cancer-related, non-surgical hospitalization. The cohorts were analyzed using advanced statistical models that accounted for the potential within-region effects of geography at the SEER registry level. Decreasing age, non-white race, lower rates of high school graduation, higher median income, degree of urbanization, receipt of radiation therapy and number of comorbidities were significant predictors of the likelihood of an initial unplanned hospitalization for lung cancer. Non-white race, receipt of radiation therapy, degree of urbanization and number of comorbidities were factors associated with an increased number of hospitalizations. For colorectal cancer, female sex, decreasing age, higher rates of high school graduation, lower median income, degree of urbanization and number of comorbidities were significant predictors of initial unplanned hospitalizations. Non-white race, receipt of radiation therapy, degree of urbanization and number of comorbidities were factors associated with increased number of unplanned hospitalizations.

ETD_5062

Ph.D.

Includes bibliographical references

Includes vita

by Kristen Lee Fessele

doi:10.7282/T33J39ZZ

ETD doctoral

The author owns the copyright to this work.