What are key sample size considerations for a cohort study?

• A. The number of researchers and budget.
• B. Expected incidence, expected effect size, desired power, alpha level, follow-up duration, and anticipated loss to follow-up.
• C. Only expected effect size and p-value.
• D. The number of exposures measured concurrently.

Answer: (B)

Planning the sample size for a cohort study hinges on how many outcome events you need to detect an association with enough precision, over the time you will follow participants, while accounting for people who drop out. The expected incidence tells you how often the outcome occurs in your population, which affects how many people you must enroll to observe enough events, especially if the outcome is rare. The expected effect size is the magnitude of the association you want to be able to detect (for example, a specific relative risk or hazard ratio); smaller effects require a larger sample to have enough power. The desired power is the probability of detecting a true effect if it exists (commonly 80% or 90%), and the alpha level is the probability of a false-positive finding you’re willing to accept (often 0.05); higher power or stricter alpha increases the needed sample size. The follow-up duration matters because longer follow-up usually yields more events, boosting power, but it can also increase loss to follow-up. Anticipated loss to follow-up must be considered to ensure you still have enough participants and events after attrition.

The other options don’t fit these planning needs as directly. Feasibility aspects like budget and the number of researchers affect how large a study can practically be, but they aren’t the statistical inputs used to calculate the necessary sample size. Relying only on the expected effect size and a p-value ignores the role of power and the significance threshold in study planning. The number of exposures measured concurrently is about design complexity and data collection, not the statistical size needed to detect an association.