Lecture 6: Pragmatic Trials for comparative effectiveness and safety

Question 1

Premise for Pragmatic Trials

Answer 1

In the real world, there is always confounding by indication which can be addressed in design and/or analysis, but if there is a chance for strong confounding or lots of unmeasured confounders, a pragmatic trial is a real-world study wth baseline randomization

Question 2

When should LST be considered?

Answer 2

• When objective is to demonstrate comparative effectiveness/safety in the real world
• when randomization is required to avoid confounding
• when endpoints can be evaluated through FU
• If these considerations are not met, the choice you are left with is: Phase IV RCT, observational (non-randomized) study, or non-randomized prospective study with follow-up and endpoint collection

Question 3

To be considered “real world”, LST designs must:

Answer 3

• inclusion/exclusion must be minimal and focused on investigators interpretation of local label
• enroll investigators and subjects to reflect actual settings for the former and key patient demographics for the latter
• comparator should be active
• open label (patient/physician can discontinue randomized tX
• Study may need to be larger/longer than pooled clinical data

Question 4

To be considered successful, LST designs must:

Answer 4

• Success depends on sufficient enrollment and continuation over a long term
• requires informed consent and case-report forms to be short and easy to understand
• data collected based on routine practice