Contested Science

Every hub in this collection sorts claims into three bins rather than two camps: settled (multiple independent lines of evidence converge — randomised trials, cohorts, mechanism — and further debate is performance), genuinely open (real scientists disagree using real data, and honest uncertainty is the expert position), and unsupported (tested and failed, or asserted without testable content). The three-bin habit is the single most protective reading skill in health, because almost every manufactured controversy works by smuggling claims between bins — presenting the settled as open (vaccines and autism) or the unsupported as merely suppressed (ivermectin for COVID). The disagreement chart on each hub makes the sorting explicit and scores it.

The collection's second commitment is explaining belief without condescension, because contested claims run on genuine fuel: medicine's own failures (dismissed symptoms behind the chronic-Lyme and mould movements; the opioid catastrophe behind pharma distrust), real institutional misconduct (suppressed trial data in the antidepressant record, conflicts of interest in guideline history), premature consensus enforcement (the lab-leak gatekeeping episode), and the honest gap between population evidence and individual experience. Treating believers as fools misreads the mechanism and loses the argument; the hubs name the establishment failure inside each controversy because that failure is usually where the controversy came from.

Once seen, the patterns repeat across every topic in the collection: lead-time and survival-statistic illusions (screening debates), dose-blindness — the poison-is-in-the-dose error and its inverse (chemical scares, potency-blind cannabis reassurance), single-study weaponisation against systematic review (ivermectin, glyphosate), mechanism-as-proof — plausible biology standing in for outcome evidence (peptides, supplements, mould toxicity), and the regret asymmetry — harms of action counted while harms of inaction stay invisible (vaccine hesitancy, screening refusal, treatment abandonment). The hubs teach these as transferable tools: master them on any one topic and every future controversy arrives half-read.

A collection like this fails if it defends institutions instead of evidence. The record it documents includes: guideline panels with industry entanglement, safety signals dismissed for years (benzodiazepine dependence, antidepressant withdrawal), overdiagnosis engines built by well-meaning screening advocacy, communication that hid uncertainty until the hiding collapsed, and journals amplifying fragile findings. The usable lesson is not cynicism but calibration: institutional consensus is strong evidence when it rests on public data and survives adversarial review, weak when it rests on authority and conflict-shielded processes — and the hubs show, case by case, how to tell which kind is in front of you.

The transferable checklist the collection trains: (1) What bin does each specific claim sit in — settled, open, unsupported — and who benefits from moving it? (2) What is the exposure and the dose, versus the studied one? (3) Is the evidence outcomes in humans, or mechanism plus anecdote? (4) What does the best systematic review say, and what would change its mind? (5) Where are the absolute numbers behind the relative ones? (6) What is the alternative to the criticised option, priced honestly — including doing nothing? (7) And who is selling what — a question that applies with equal force to supplement marketers, plaintiff experts, industry scientists and attention-economy influencers. No checklist replaces expertise; this one reliably detects its absence.

These hubs exist because the space between medical journals and viral posts is where most health decisions actually get made — and that space is served mainly by people with something to sell. Each hub aims to be the page a clinician could hand a worried patient, a journalist could background from, and a sceptical reader could stress-test: sourced to primary literature, honest about uncertainty, explicit about where its own judgments sit on the evidence scale, and finished with a practical bottom line rather than a shrug. Corrections and challenges are part of the design — the editorial contact is on every page, and claims are only as good as their sources, which are linked so readers can check.