Case Study: Cardio Toxicity

Question 1

Hypotheses

Answer 1

Adjuvant CHEMO associated with an increased risk of cardiotoxicity in elderly BC patients

• Adjuvant RT is associated with an increased risk of cardiotoxicity in elderly BC patients
• Adujuvant CHEMO and RT synergistically interact to produce an increased risk of cardiotoxicity

Question 2

Study Period

Answer 2

12 months before index date (BC diagnosis), 6 months treatment, then post treatment, then censor after 10 years from pre-treatment

Question 3

Study Cohort

Answer 3

Started with 152,200 patients
Some excluded (largest exclusions were: Diagnosed prior to 1992, HMO membership, Missing Medicare coverage)
Ended with 31,748

Question 4

Exposure measurement

Answer 4

Surgery (mastectomy, BCS)
RT (LSRT)
CHEMO

Question 5

Outcomes

Answer 5

Heart Disease and Survival

Question 6

Confounders

Answer 6

Age, Race, Year of cohort study, Marital status, geography, hormonal status, comorbidities

Question 7

Required sample size

Answer 7

5500 exposed

Question 8

Statistical Analyses

Answer 8

-Describe study cohort and subgroup (univariate and bivariate analyses)
-Evaluate propensity to receive treatment (Logistic regression)
-Determined odds of experiencing cardiac events (Logistic regression)
Determined rate of cardiac events (Cox proportional hazards model)
-Estimated cardiac-event free survival (Kaplan-Meier)

Question 9

What is a Propensity study?

Answer 9

Attempts to estimate the effect of the treatment by accounting for covariates that predict receiving the treatment.

Also attempts to reduce bias due to confounding by variables that could be found in an estimate of the treatment effect obtained from simply comparing outcomes among units that received treatment versus those that did not

Question 9

What is a Propensity study?

Answer 9

Attempts to estimate the effect of the treatment by accounting for covariates that predict receiving the treatment.

Also attempts to reduce bias due to confounding by variables that could be found in an estimate of the treatment effect obtained from simply comparing outcomes among units that received treatment versus those that did not