Evidence-based health information: it is not medical advice and does not replace an eye examination.Information, not medical advice. Disclaimer

Our method

How we choose studies, how we report on them, how we use artificial intelligence and how we correct mistakes.

Per gli occhi rests on a simple idea: there is plenty of opinion in circulation about eye health, and very few numbers. We start from the numbers, that is, from published studies, and we report them in a way that anyone can understand and check.

Where our sources come from

Every article starts from a search of PubMed, the database of the US National Library of Medicine that collects the world’s biomedical literature, and from the guidelines of scientific societies and health institutions (the World Health Organization, European ophthalmology societies, national health agencies).

The details of each citation (authors, title, journal, year, DOI) are not copied out by hand: a program reads them directly from PubMed. The same program checks that every study cited really exists and that it has not been retracted, that is, withdrawn by the journal. A retracted study cannot be cited.

Which studies count most

Not all studies carry the same weight. In order, we give most weight to:

  1. systematic reviews and meta-analyses, which bring together and compare all the studies on a given question (Cochrane reviews are the benchmark);
  2. guidelines from scientific societies, which translate the evidence into recommendations for doctors;
  3. randomised controlled trials, in which people are assigned at random to a treatment or to a comparison;
  4. observational studies, which follow groups of people over time and find associations, not necessarily cause and effect;
  5. laboratory and animal studies, which are useful for understanding mechanisms but not for drawing conclusions about people.

In every article we say what type of study we are describing, how many people it involved, for how long, and what its limitations are. We prefer absolute risks (‘from 9 to 4 people in 100’) to relative ones (‘50% lower’), which on their own can mislead.

What we do not do

  • We do not give personal medical advice, make diagnoses, or recommend treatments or doses.
  • We do not name brands of medicines, eye drops, supplements, lenses or devices; we mention active ingredients only when describing a study.
  • We do not publish recovery testimonials or statements attributed to doctors or experts.
  • We do not use photographs of diseased eyes, of patients or of doctors: our cover images are graphic illustrations of irises drawn by a computer program, a different one for every article, in the colours of its section. They are neither photographs nor realistic images generated by artificial intelligence.

Artificial intelligence

We use artificial intelligence tools to prepare and translate our texts, starting from the selected studies. We say so openly, in every article, right below the headline:

  • until an article has been checked by a qualified person, it carries the label ‘AI-generated text’;
  • once every statement has been checked against the studies cited, it shows the name of the person who checked it and the date.

Whatever the status of an article, the same automated checks apply: sources come from PubMed or from a closed list of institutional sources, every number in brackets in the text must match a source, every source must be cited in the text, and a list of banned wording (promises of a cure, dosages, brand names, superlatives) blocks publication. They remain tools, however: they do not replace review by a person, which is our aim for every article.

Languages

The magazine is published in Italian, English, German, French and Spanish. The versions in languages other than Italian are translations of the same article, with the same sources and the same structure; study titles remain in their original language of publication, which is almost always English. We have chosen a European perspective: no health rules that apply in only one country, and the emergency number is 112.

Corrections and updates

If an article contains an error, we correct it and state at the end of the article what has changed and when. We periodically review our articles in the light of new studies; the date of the last update is shown below the headline. You can report an error by writing to redazione@pergliocchi.com or through the Contact page.

Who is responsible for the content

The magazine is published by Publilia di Alessandro Ruggiero, which holds editorial responsibility for it. The rules on advertising and sponsorship are on the Independence page; the limits of our content are set out in the Disclaimer.