# Follow-up codebook (written after the registered analysis; exploratory)

Each record is a randomized placebo-controlled vaccine trial whose registry record mentions a crossover, unblinding, an open-label period, an extension, or a booster. You see its arms, the time frame and description of its adverse-event tables, every adverse-event group with its description, and which groups the study compares (`comparisons`: vaccine groups against placebo groups), with no counts.

Decide whether, in each comparison, the vaccine groups and the placebo groups report adverse events over the same follow-up for each participant. Code the trial (all its comparisons together; if they differ, take the least favourable):

- `equal`: by what the record says, both sides' groups cover the same period: the blinded placebo-controlled period only, or the same fixed time after each participant's doses, with any crossover, open-label or extension period reported in other groups or not at all.
- `vaccine_longer`: the vaccine groups' reporting runs on (through an open-label period, to the end of the study) while the placebo groups' reporting stops when placebo recipients were unblinded, crossed over to the vaccine, or left the placebo-controlled period.
- `placebo_longer`: the reverse.
- `unclear`: the record doesn't say enough to tell, or says things that conflict.

Decide from the record alone; don't look anything up. Write a JSON array, one object per trial in the order given: {"nct": "...", "follow_up": "equal|vaccine_longer|placebo_longer|unclear", "reason": "a sentence quoting or pointing to what in the record decides it"}.
