Sequential co-enrolment in randomised trials in neonatal intensive care medicine

Area of Science:

• Medical research methodology
• Clinical trial design
• Biostatistics

Background:

• Patient recruitment for clinical trials is often limited in specialized settings like neonatal intensive care units.
• Sequential co-enrolment into multiple trials presents challenges, including the risk of contaminated results.
• Infants may have prior trial enrollment as fetuses, complicating new trial participation.

Purpose of the Study:

• To define requirements for estimating different treatment effects (estimands) in sequential clinical trials.
• To analyze how accounting for prior trial participation status and treatment impacts estimand interpretation.
• To provide guidance for researchers when co-enrolment is unavoidable.

Main Methods:

• Consideration of a scenario involving two sequential randomized clinical trials.
• Description of various estimands based on how prior trial information is incorporated.
• Analysis of how differences in available prior trial data affect interpretation and generalizability of results.

Main Results:

• Estimands vary in their consideration of previous trial participation and treatment status.
• Analytical results can differ in interpretation and generalizability due to prior trial data availability, unless treatment interactions are absent.
• Co-enrolment necessitates careful data collection on prior trial status.

Conclusions:

• Researchers must collect data on co-enrolment and prior treatment status when sequential trials are involved.
• Trial analysis plans should be adapted to mitigate risks associated with co-enrolment.
• Accurate estimation of meaningful treatment effects requires addressing potential biases from concurrent or prior trial participation.