Every change that alters the substance of an article gets logged here, with the date and what changed. If you think something on this site is wrong, tell me and it will end up on this page.
12 September 2026: isotretinoin is a good practice statement, not a strong recommendation
What was wrong. Six passages across five articles said the 2024 AAD acne guideline “strongly recommends” isotretinoin. It does not. Isotretinoin is covered by recommendation 4.1, which the guideline lists as one of its five good practice statements, not one of its 18 graded recommendations. The pages affected were isotretinoin, cystic acne, teen acne, types of acne and acne treatment options.
Why it happened. Several widely read secondary summaries of the guideline list isotretinoin under “strong recommendations”, and the original drafts followed them instead of the guideline itself. The error was found by checking every guideline-strength claim on the site against the primary document, after the evidence map was built directly from it and disagreed with the articles.
What it does and does not mean. A good practice statement is not a weaker endorsement. It is the label a guideline panel uses when it considers the case clear enough that formal grading would add nothing. Isotretinoin remains the guideline’s answer for severe, scarring or treatment-resistant acne. The correction is about accuracy in describing the guideline, not a change in what the guideline advises. All six passages now say “good practice statement” and explain the term.
11 September 2026: first full fact-check and link check
Every source link on the site was tested, and a sample of the most load-bearing statistics in each of the 50 articles (trial results, prevalence figures, dosages and regulatory statements) was re-read against the page it cites. Of roughly 150 claims checked, 145 matched their source exactly. These are the ten that were corrected, clarified or softened:
- Maskne and acne mechanica: A meta-analysis result was described as an odds ratio. It is a relative risk (1.42, 95% CI 1.31–1.54). Corrected, and the confidence interval added.
- Back acne: A 2022 laboratory study was described as testing 70 C. acnes strains. The contact-time experiments used 30 representative isolates selected from that collection. Corrected in the text and in the infographic.
- How to start adapalene: Specific figures from a 2014 Japanese trial (100 patients; 100% vs 70% still on treatment at week 4) could not be verified because the paper is paywalled. The claim is now described in general terms and flagged as unconfirmed.
- Niacinamide for acne: A split-face study was described as reducing surface oil. In that 30-person arm the difference only approached statistical significance (p=0.055) and the sebum excretion rate did not change. Reworded.
- Post-inflammatory hyperpigmentation: Attributed Fitzpatrick skin types IV to VI to StatPearls; the source says types III to VI. Corrected.
- Acne scars vs dark spots: Said StatPearls lists sunscreen first among treatments for dark marks. The source frames daily UV protection as the backdrop to a stepwise plan whose first step is calming the inflammation. Reworded.
- Makeup that won’t clog pores: Cited product-replacement intervals (6–12 months for foundation) that appear only inside a graphic on the source page, which says there are no set rules. Replaced with the open-jar symbol guidance.
- Topical clindamycin: A resistance study was described as sampling 50 patients; the 30.6% figure comes from the 36 patients whose swabs grew C. acnes. Clarified.
- Should you pop pimples?: Cited the same US scarring cohort as another article using a different patient count. Reworded to “nearly 2,000” so the two articles agree.
- Site-wide: Every source link on the site (314 unique URLs) was tested; all resolve. Source counts and a “links checked” date now appear under each article’s byline.
No claim checked was contradicted by its source, and no dosage, safety or pregnancy claim needed changing. The method behind these checks is described in the Editorial Policy.
When a claim doesn’t match its source
Reading primary sources closely means we regularly find claims repeated with complete confidence across the internet that do not match the paper being cited. When that happens we do two things. We flag it on the page, in a box headed “Commonly repeated, and not quite right”, setting the common claim next to what the source actually says. And we link the source, so you can read it and decide for yourself rather than take our word for it.
What we do not do is chase it. We don’t write to publishers, we don’t name the sites carrying the claim, and we don’t set ourselves up as anybody’s fact-checker. Almost every discrepancy of this kind is honest: a detail added in a summary of a summary, a study described from its abstract, a finding that was true of a 1998 formulation and is no longer true of a 2021 one. Pointing at the primary source and letting it speak is more useful than an argument, and it is the part we can actually stand behind.
Our own mistakes are a different matter, and those we do own. Every one of them is logged on this page with the date and what changed. If you think one of our flags is itself wrong, please tell us, and we will correct the correction.
The flags currently live on the site sit inside the acne treatment evidence map and the ingredient combination checker.