Methodology

Data Sources

Every claim in Nous Vita points at a document. The library behind it is 382 documents from 126 distinct publishers, and the tables on this page are generated from that library each time the site is built rather than written by hand, so they cannot quietly fall out of date. We do not use marketing materials, manufacturer claims, or affiliate content as evidence.

How this was built, in plain terms

Most interaction checkers are a list of pairs. Somebody wrote down that this drug and that supplement do not mix, and the tool looks the pair up. Such a tool can only ever know what someone remembered to write down, and it has nothing to say about the combination nobody thought to check. Nous Vita is built the other way round, in five layers.

1. 1

   We work out what you actually took

   You type a brand name, a nickname, or a spelling nobody agrees on. We resolve it to a substance — the molecule or the plant extract — so that fish oil, omega 3 and EPA/DHA all land in the same place. The registry holds 357 of them. Anything we cannot resolve is simply not checked, which is why every clear result on this site tells you our lists are incomplete rather than telling you that you are fine.

2. 2

   Every substance joins the families it belongs to

   A substance is never only its own name. Warfarin is a vitamin K antagonist, an anticoagulant, and a drug with a narrow margin between the dose that works and the dose that harms — all at once, and each of those memberships carries different consequences. We maintain 61 such families, and every substance declares which ones it belongs to. This is why the tool can react to a medicine nobody ever wrote a pairwise entry for: it does not need to recognise the product, only the family.

3. 3

   We model the harm, not the pair

   The thing that hurts you is bleeding, or a potassium level that disturbs your heart rhythm, or liver injury. It is not an abstract 'interaction'. So the harm is the root of the model, and each harm records the separate routes that lead to it. A supplement can raise bleeding risk entirely on its own, or only alongside one particular kind of blood thinner, and those are different facts resting on different evidence. We show you which one applies to you instead of collapsing both into a single warning that fits neither.

4. 4

   Every claim carries its document

   Each of the 75 curated relationships in the engine points at one or more documents in the library, and records the sentence in that document that supports it. The number of them resting on nothing at all is 0. That is not a promise, it is a measurement: an automated check recomputes it and refuses to pass if it is anything other than zero, and the same check runs over the 33 guidance rules and 61 families. A confident warning with nothing behind it is worse than no warning.

5. 5

   All of it runs on your device

   The registry, the families, the rules and the evidence library are compiled into the page itself when the site is built. Your regimen is evaluated in your browser, by code that arrived with the page. Nothing you enter is transmitted anywhere, because there is nowhere for it to go.

Evidence Grading

Every supplement recommendation and safety claim carries an evidence grade. We use a three tier system adapted from standard evidence-based medicine frameworks.

How the Safety Checker Works

When you enter supplements and medications into Nous Vita, the Safety Checker performs two types of analysis, both running entirely in your browser.

First, it looks at each substance on its own. Some supplements carry a documented risk with no second ingredient required: a known likelihood of liver injury, drawn from the NIH LiverTox database and reported with the category it assigns; or a harm such as bleeding that the literature attributes to the supplement by itself. A standalone risk is reported as a standalone risk, and is not disguised as an interaction.

Second, it checks each substance against every other thing you take and against the conditions in your profile. This does not work by looking your pair up in a table. Each substance belongs to families, and our rules are written about families, so a rule about vitamin K antagonists reaches every vitamin K antagonist we know of. An interaction is reported when published evidence documents a clinically meaningful effect: altered absorption, metabolism or clearance; additive or opposing effects on the body; or documented adverse events. The checker reports the mechanism, the direction of the effect, and who said so.

Nous Vita works like an application on your device. Once the site has loaded, you can disconnect from the internet and everything still works: the safety checker, the regimen builder, the tracking dashboard. The interaction database is part of the site itself, loaded once at build time, not fetched from a server while you use it. No internet connection needed. In the future, if we offer optional encrypted cloud backup, that is the only feature that would need a connection, and it would be opt-in only.

Research Pipeline

We maintain an adversarial research pipeline to keep the knowledge base current and honest. New evidence is evaluated by multiple independent reviewers who are incentivized to find weaknesses in the evidence, not just support for existing conclusions. We explicitly look for studies that contradict our current assessments. When the evidence shifts, we update our recommendations and publish the rationale for the change. The pipeline includes systematic monitoring of PubMed for new supplement drug interaction publications, regular review of NIH LiverTox updates, and periodic re evaluation of existing evidence grades.

Limitations

Nous Vita is an informational tool, not a substitute for clinical judgment. Our analysis has important limitations that users and clinicians should understand.

The interaction database covers published, documented interactions. It does not cover theoretical interactions that lack published evidence, interactions with rare or newly approved medications that have limited post marketing data, or individual patient factors such as genetics, liver function, age, or polypharmacy burden. A result of no known interactions does not guarantee safety. The absence of evidence is not evidence of absence.

Our evidence library reflects what has been published and indexed. Rare or novel reactions may not be documented anywhere yet, and no source we hold accounts for individual susceptibility. This is sharpest for liver injury: the NIH LiverTox database is authoritative about the substances it has assessed, and silent about the rest. An absent entry means not assessed, and nothing more than that.

Supplement quality and composition vary by manufacturer. Nous Vita evaluates the substance, not specific brands. A product may contain fillers, contaminants, or different doses than labeled.

Updates and Versioning

The interaction database and evidence grades are updated as new research is published. Major updates are versioned and changelogs are published. The NIH LiverTox data is refreshed when NIDDK releases updates. Users can verify the last update date on this page.