Treatments effects from randomized trials and propensity score analyses were similar in similar populations in an example from cardiac surgery.