Reading guide
Reading future observational results
Always ask who is counted, who is missing and what was measured.
Editorial draft; scientific and translation review pending.
Denominators before percentages
A future dashboard should state the eligible population, number contributing each outcome, follow-up window and missing data. Reports received are not necessarily unique people; unique people are not automatically a population denominator. Adverse-event reports alone cannot establish incidence.
Separate source confidence and outcome
A label-only exposure differs from material with analytical documentation. Patient-reported change differs from a clinician-recorded measurement. Retain those distinctions rather than collapse them into a single score. Report improvement, no change, worsening and loss to follow-up with uncertainty.
No live efficacy leaderboard
Do not turn early volunteer reports into “best peptide” rankings or automatically generated regimen advice. Public outputs should be approved aggregate analyses with disclosure controls, methods and version history. Sparse cells, rare conditions and combined attributes may identify people even when names are removed.
References
- 1.
- 2.