What is a new-user design and why is it used in observational cohort studies?

• A. Use time-varying exposure
• B. Exclude new users
• C. Randomize
• D. Include only new initiators at cohort entry to avoid prevalent-user bias

Answer: (D)

Starting follow-up at the moment a person begins a treatment and including only those who are new initiators at cohort entry is the key idea. This approach avoids prevalent-user bias, which happens when you include people who have already been using the drug before the study starts. Those ongoing users often differ in important ways: they may have tolerated the drug well, had time to adapt, or already experienced early events that influenced whether they stay on therapy. As a result, comparing new users to others (or to non-users) can distort estimates of safety or effectiveness because the early risk period and reasons for continuing or stopping treatment are not equally represented.

By enrolling only new initiators, you ensure that exposure starts at the same point for everyone, baseline characteristics are measured around initiation, and early events are captured consistently. This makes the comparison more like a randomized trial in terms of timing and baseline assessment, reducing biases tied to prior exposure and survivorship. Time-varying exposure and randomization are separate concepts and aren’t what defines this design; the defining feature is restricting the cohort to individuals at the moment they start the treatment to avoid biases from past use. Excluding new initiators would reintroduce bias rather than prevent it.