DiscoverBlood & CancerHazard ratio? P value?
Hazard ratio? P value?

Hazard ratio? P value?

Update: 2019-05-23


Episode 17:

David L. Streiner, PhD, of McMaster University, Hamilton, Ont., and the University of Toronto, joins Blood & Cancer host David Henry, MD, of Pennsylvania Hospital, Philadelphia, to talk hazard ratios and P values as they examine the clinical relevance of findings from a phase 3 trial.


In Clinical Correlation, Ilana Yurkiewicz, MD, of Stanford (Calif.) University, talks about how to balance family versus patient preferences.


Show notes

By Emily Bryer, DO, resident in the department of internal medicine, University of Pennsylvania, Philadelphia 

  • A phase 3 trial is a randomized, controlled trial testing a new intervention against placebo or treatment as usual.
  • Randomization maximizes the chances that the groups are equivalent but does not guarantee it.
  • With randomization, you also are accounting for variables that are unknown and/or cannot be controlled for. 

Phase 3 trial discussed by Dr. Henry and Dr. Streiner:

  • AURELIA trial: Bevacizumab plus chemotherapy vs. chemotherapy alone for platinum-resistant ovarian cancer.
  • Inclusion criteria: Ovarian cancer that has progressed on a platinum-based therapy.
  • Randomization: 361 patients randomized 1:1 to receive bevacizumab plus chemotherapy versus chemotherapy alone.
  • Primary endpoint: Progression-free survival (PFS).
  • Main outcome: PFS had a hazard ratio 0.48 (95% confidence interval, 0.38-0.60).
    • A hazard ratio of 0.48 means that patients in the experimental group had half the risk of experiencing a bad outcome (progression) than patients in the comparison group did.
    • The hazard ratio includes a confidence interval (CI) at the end of the value because it is an estimate. The CI represents where the true hazard will fall 95% of the time. If 1.0 is included in the range, then the result is not statistically significant, and the events could have happened by chance.
  • An ad hoc analysis is conducted at the end of the study.
    • It is not a prespecified statistical idea.
    • It is thought provoking and hypothesis generating but not conclusive.


AURELIA trial: J Clin Oncol. 2014;32(13):1302-8.









Hazard ratio? P value?

Hazard ratio? P value?