Examination of factors associated with cancer survival among complex patients using competing risk model
Monday, November 4, 2013
: 2:30 p.m. - 2:50 p.m.
Background: Approaches to describe risk of disease include Cox proportional hazards regression; however, this method fails to account for the competing risk of death. Purpose: To determine significant predictors of breast cancer-specific-survival among female breast cancer (BCa) patients while allowing competing risk for other causes of death. Methods: 4220 BCa patients identified from Florida cancer registry between 2007 and 2010 were linked with their electronic medical records. Competing risk survival model was used for analysis. Results: Two-year cumulative BCa death and death from other causes were 4.83 and 2.89 percent respectively. Median follow-up for patients who died from BCa and from other causes were 569 and 642 days respectively. Multivariate competing risk model showed that increased chance of BCa death was associated with black race, triple negative status with effect increased over time, unknown biomarker receptor status, being unmarried, Medicare reliance, having poorly- or un-differentiated tumor grade, regional diagnosis stage, larger tumor size, and more positive nodes detected. Higher probability of death from other causes was associated with being diagnosed at an older age, being unmarried, Medicaid recipients or Medicare beneficiaries, having more comorbidity, moderately-differentiated tumor grade, poorly- or un-differentiated tumor grade, larger tumor size, not having chemotherapy, and having less lymph nodes examined. Temporal increase of detrimental triple negative effect on BCa survival appeared in competing risk model but was not observed in Cox model. Conclusions: Competing risk model may be a better tool in examining factors associated with cancer survival among patients with complex conditions.
Public health or related organizational policy, standards, or other guidelines
Demonstrate use of competing risk model
Compare factors identified by traditional cox model and competing risk model
Keyword(s): Women's Health, Cancer
Presenting author's disclosure statement:
Qualified on the content I am responsible for because: I am a biostatistician, and have been using survival techniques to analyze public health data for multiple funded grants.
Any relevant financial relationships? No
I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines,
and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed
in my presentation.