Research Methods
When one paper contains two trials, do not average the hours
The afamelanotide report gives a 6-month United States result and a 9-month European result. They are not one pooled afternoon.
Some papers report two protocols under one title. The afamelanotide trials in erythropoietic protoporphyria did that. Both used a 16 mg subcutaneous implant or placebo. They did not use the same number of implants, the same length of observation, or the same median hours of pain-free sun.
Do not pool by eye
The United States median at 6 months was 69.4 hours versus 40.8. The European median at 9 months was 6.0 hours versus 0.8. Averaging 69.4 and 6.0 produces a number neither trial measured. A methods note keeps each median next to its country, its duration, and its implant count.
Registries help
NCT01605136 and NCT00979745 are the registrations named with the paper. If a sentence cannot be tied to one of them, it is probably a blend. Blends are how a careful trial becomes a looser claim.