Science Notes
Five files. Fifteen studies. Zero borrowed certainty.
Open a folder to inspect the question, study design, finding, and limit together. The pair label is an editorial index; the medicine answers to its sources.
Open evidence files
- Weight Is Not the Whole OutcomeThree trials, three different endpoints: preventing diabetes, changing risk factors, and testing cardiovascular events directly.
- Three Questions, Three Kinds of EvidenceA sham-controlled procedure trial, a patient-level treatment synthesis, and a study of trial methods belong in the same chart—but never in the same evidence box.
- The Screen Is Not the InterventionDigital care can mean a managed pathway, a telephone service, or a communication technique. The chart gets clearer when those parts stop borrowing one another's results.
- Inside mRNA EvidenceOne file for trial synthesis, one for lymph-node biology, one for observed safety. Different lenses; no single one gets the whole microscope.
- The Record Is Not the Whole PatientCoverage changes several outcomes at once. Spending can mislabel need. Diagnosis counts can rise when the system looks harder.
Company identities and market context live in FIVE PAIRS and MARKET ROUNDS. Source handling and project boundaries live in SOURCES & RISKS.