Nobody, at least me, thinks that its model is 100% accurate, due to a lot of unknown parameters (e.g. actual (today) event rate is more likely higher than the 5.2%, since it is the average during the study) these are the best, possible estimates.
An example of such an instance is time from event, to when the event is reported. Obviously events are not reported the instance they occur. I would expect there is a delay. How long, we do not know. Is it days, weeks, even months?
Serious Adverse Event and pregnancy should be reported within 24 hours of the knowledge of the occurrence. So, we could assume that it is true for the endpoint event also.
How often are these patients making office visits?
At Visit 1: evaluation of patient's eligibility in the study; Visit 1.1: Patients not eligible for participation in the study based on the study procedures on Visit 1 may possibly become eligible in the next 28 days: these patients may return at the discretion of the investigator for a second optional screening visit (for example, patients who are started on a statin at Visit 1, whose statin dose is changed at Visit 1, and/or needed to wash out non-statin lipid-altering medications.); Visit 2: randomization visit; Visit 3: Day 120 ± 10; Visit 4: Day 360 ± 10; Visit 5: Day 720 ± 10; Visit 6: Day 1,080 ± 10; Visit 7: Day 1,440 ± 10, Last Visit: Day 1,800 ± 10
What is the mechanism by which events are reported? Is there a consistency among all the sites in this regard? Is there a consistency among countries?
Events were recorded in the Electronic Data Capture (on a Case Report Form), inc. any supporting documentation (e.g., laboratory test reports, patient discharge summary, or autopsy reports)