This tutorial introduces you to CompARE, a web-tool for designing trials with composite endpoints.
2019-02-21
This tutorial introduces you to CompARE, a web-tool for designing trials with composite endpoints.
A composite endpoint consists of two or more outcomes combined in a unique endpoint. Patients who have experienced any one of the events specified by the components are considered to have experienced the composite outcome.
Rationale behind the use of Composite Endpoints:
Probability of observing the Composite Endpoint:
Effect measures for the Composite Endpoint:
Parameter Effect | Null hypothesis | Alternative hypothesis | |
---|---|---|---|
Risk difference | \(\delta_* = p_*^{(1)} - p_*^{(0)}\) | \(\delta_* = 0\) | \(\delta_* < 0\) |
Relative risk | \(\textrm{R}_* = p_*^{(1)}/p_*^{(0)}\) | \(\log( \textrm{R}_* ) = 0\) | \(\log( \textrm{R}_* ) < 0\) |
Odds ratio | \(\textrm{OR}_* = \frac{p_*^{(1)}/q_*^{(1)}}{p_*^{(0)}/q_*^{(0)}}\) | \(\log( \textrm{OR}_* ) = 0\) | \(\log( \textrm{OR}_* ) < 0\) |
Both the composite probabilities and effect depend on the composite components and on the association between them. To calculate the composite parameters and the sample size, you can use CompARE.
Consider the event probabilities under the control group \(0.12\) and \(0.10\) and the odds ratios \(0.75\) and \(0.65\) for the Endpoint 1 and 2, respectively.
Consider the event probabilities under the control group \(0.07\) and \(0.15\) and the risk ratios \(0.7\) and \(0.6\) for the Endpoint 1 and 2, respectively. Assume a correlation equal to \(0.1\).
Consider the event probabilities under the control group \(0.07\) and \(0.2\) and the risk ratios \(0.8\) and \(0.5\) for the Endpoint 1 and 2, respectively. Assume a correlation equal to \(0.2\).
Consider the event probabilities under the control group \(0,07\) and \(0.15\) and the odds ratios \(0.8\) and \(0.65\) for the Endpoint 1 and 2, respectively. Assume a correlation equal to \(0.1\).
Consider that the probability for the Endpoint is \(0.1\); the probability for the Endpoint 2 is \(0.15\). You expect a risk reduction for the Endpoint 1 of \(0.80\) and \(0.65\) for the Endpoint 2. Consider the risk difference as the measure for testing the treatment differences between groups.
Consider that the probability for the Endpoint 1 takes values between \(0.078\) and \(0.112\); the probability for the Endpoint 2 takes values between \(0.117\) and \(0.17\). You guess that the correlation between the composite components is weak and expect a relative risk for the Endpoint 1 of \(0.85\) and \(0.75\) for the Endpoint 2.
Reproduce the sample size calculation in the paper:
Cannon CP, Weintraub WS, Demopoulos LA, et al. Comparison of Early Invasive and Conservative Strategies in Patients with Unstable Coronary Syndromes Treated with the Glycoprotein IIb/IIIa Inhibitor Tirofiban. New England Journal of Medicine. 2001;344(25):1879-1887.