Date of Graduation


Document Type

Thesis (MS)

Program Affiliation

Biomedical Sciences

Degree Name

Masters of Science (MS)

Advisor/Committee Chair

Sharon Giordano, MD, MPH

Committee Member

Holly M Holmes, MD, M.Sc

Committee Member

Carlos H. Barcenas, MD, M.Sc.

Committee Member

Scott B. Cantor, Ph.D.

Committee Member

Gary E. Gallick, MS, Ph.D

Committee Member

Robert C. Bast Jr, MD


Our objective was to determine predictors of potentially inappropriate medication (PIM) use and its impact on outcomes (including ER visits, hospitalization, all cause death, and composite of three) in breast and colorectal cancer patients receiving chemotherapy. We used data from the SEER database linked to Medicare claims. Our cohort included patients ≥ 66 years diagnosed with of Stage II/III breast or colorectal cancer between 7/1/2007-12/31/2009. Baseline PIM was defined using the Drugs to Avoid in the Elderly list (DAE) or Beers criteria. Univariate and multivariable logistic regression were used to determine the associations of baseline PIMs with different covariates. Event-free survival (EFS) was defined from the initiation of chemotherapy to outcome, and estimated using the KM method. Cox proportional hazards modeling was used to determine the association of baseline PIMs with EFS. The final analysis included 1595 breast and 1528 colorectal patients. The frequency of baseline PIM was 22.2% (DAE) and 27.6% (Beers) in the breast cohort, and 15.5% (DAE) and 24.8% (Beers) in the colorectal cohort. Baseline PIM was associated with younger age, baseline ≥5 medications, and female gender. In the breast cohort, 37.5% patients had at least one composite outcome. One-year EFS rate was 49%, 62%, 96%, and 45% for ER, hospitalization, death, and composite respectively. Variables associated with increased risk of the composite outcome included baseline ≥5 medications, advanced stage, higher comorbidity, and baseline ER/hospitalization. Baseline PIM using DAE was associated with increased risk of death in the breast cohort, HR 2.31 (95% CI 1.07-4.96). 45% of patients in the colorectal cohort had at least one composite outcome. One-year EFS rate was 42%, 54%, 91%, and 38% respectively. Variables associated with an increased risk of the composite outcome in colorectal patients included baseline ≥ 5 medications, older age, female gender, higher comorbidity. In the time-to-event analysis, we found no association between baseline PIM and most outcomes in either group, aside from baseline PIM using DAE and death in the breast cohort during chemotherapy. Baseline ≥5 medications was associated with increased risks of adverse outcomes in both. Our findings require further prospective confirmation but call into doubt the need to reduce PIM in older patients during chemotherapy.


polypharmacy, geriatric, adjuvant chemotherapy, breast cancer, colorectal cancer, Beers Criteria, health outcomes