Editorial policy

Voxsanity publishes information about clinical trials, medicine approvals, and research funding, drawn from public government registries. This page sets out how we decide what to publish, what we will never do, and how you can hold us to it.

Last updated: July 2026

We would rather be checkable than reassuring. Everything below is written so that you could catch us breaking it.

Independence

What we show is never influenced by who funds us. Coverage and rankings are not for sale.

Voxsanity may take revenue from any source, including pharmaceutical companies, clinical trial sponsors, medical device manufacturers, and contract research organisations. We are not going to promise you that we will refuse their money.

We are promising something narrower and more useful: that their money buys nothing. It does not buy a place in our coverage. It does not buy a kinder description. It does not buy a higher position in a list. The revenue is not the problem. Influence over what you read is the problem, and that is what this policy rules out.

We have chosen this promise deliberately, over the easier one. A pledge never to take pharmaceutical money sounds stronger and is worth less: it is a promise about our bank account, not about your information, and a company that wanted to shape what you read would simply route around it. A promise about influence is the one that actually protects the reader, and it is the one we can be held to, because you can check the output.

How the content is made

Most of what you read here is generated automatically from public government data. Each night, a batch process pulls the latest records from the registries we use, including ClinicalTrials.gov, the FDA through openFDA, NIH Reporter, and the Pharmaceutical Benefits Scheme, and an AI language model rewrites the clinical source text into plain English. The plain English version is stored separately from the original wording, and the original is always preserved and linked, so any summary can be checked against its source.

AI-generated text is labelled as such on every page it appears, directly above the text, so you always know what you are reading. These summaries are reviewed against the source on a sample basis, and they can still contain mistakes, which is why the original registry wording is always one click away. How we use AI is set out in full in our terms of use.

A smaller amount of content is written by hand rather than generated: the "state of access" notes on some condition pages, and pages like this one. Hand-written and machine-generated content are held to the same rules below. Neither is for sale.

What coverage means

Coverage means which conditions get a page on this site, and what appears on it.

Voxsanity currently covers 103 conditions. That list was chosen on public health grounds: burden of disease in Australia, how many people are affected, and whether people with that condition have few options and little plain English information. It was also constrained by whether the public registries we depend on actually hold usable data for that condition.

When we expand the list, the same test applies. A condition is added because people in Australia need the information. Never because someone asked us to add it, and never because someone offered to pay for it to be added. We will not sell a place on the list, and we will not sell an absence from it.

The trials, medicines, and research figures shown on a condition page are whatever the public registries say. We do not curate that set to anyone's advantage.

What "not for sale" rules out, specifically

To be concrete, all of the following are things Voxsanity will not do:

  • No paid placement in search results, rankings, or listings. Ordering and inclusion are set by defined technical criteria applied to the data: a trial's recruiting status, phase, and recency, and whether a medicine has a reviewed government mapping of its PBS listing to that condition. Nobody can pay to appear higher, or to appear at all.
  • No paid removal, suppression, or softening. We will not remove a trial, alter a medicine's record, or reword a condition's data because a sponsor, a manufacturer, or anyone else asked us to. Corrections are made because something is factually wrong, and for no other reason. If a company disputes a fact, it uses the same corrections process as everyone else, and the record continues to say what the registry says.
  • No favourable treatment for sponsors we already list. Voxsanity publishes sponsor pages that aggregate publicly registered trial activity by the organisation running the trials. This is factual disclosure, not paid placement, and it stays that way. A sponsor on this site has not paid to be there, cannot pay to change what it says, and cannot pay to be taken down. Being listed buys nothing. If a sponsor also becomes a commercial partner, we disclose that on the page, and it still buys nothing.
  • No advance preview for anyone commercial. No paying party gets to see, approve, or influence what we publish before it is published.

How we will judge any future revenue

Voxsanity intends to be a sustainable business, and may add sponsorships, advertising, verification badges, or other paid features, including ones bought by companies in the health sector. Every one of them has to pass the same standing test before it launches:

Does it change what we show, how we describe it, or the order we show it in? If yes, we do not build it.

A paid feature that is clearly separated from the data, clearly labelled as paid, and has no influence over coverage, descriptions, or rankings can pass that test. A feature that lets someone buy visibility, buy a better description, or buy their way out of an unflattering fact cannot pass it, and no amount of labelling repairs that.

When a paid feature does launch, this page will say what it is, who pays for it, and what it does not buy.

Independence from our data sources

Voxsanity depends on public data providers, including the Pharmaceutical Benefits Scheme, the Therapeutic Goods Administration, ClinicalTrials.gov, and the NIH, and may hold formal or informal relationships with some of them, including licence terms for reusing their data. None of those relationships gives a provider any say over how their data appears here. We use a source because it is accurate and useful, not because of any partnership, and a provider cannot obtain a softer description, a better position, or the removal of an unflattering record by virtue of the relationship. Where a licence requires us to reproduce a provider's wording unchanged, as the PBS licence does, we do that and say so; that is a constraint on us, not influence for them.

Who is accountable, and how to challenge us

Final editorial authority rests with Samuel Gyimah, founder of Voxsanity Pty Ltd (ABN 82 700 348 867). There is no editorial board and no medical review panel; saying otherwise would be exactly the kind of reassuring fiction this page exists to avoid. This policy binds his own decisions, not only those of some hypothetical future advertiser. He cannot buy or sell a place in the coverage, a kinder description, or a higher ranking either, and if he ever did, the output would show it and you could catch it.

Corrections work the same way regardless of who benefits. We do not fix flattering errors faster than unflattering ones, and we do not treat a correction that embarrasses a listed company any differently from one that embarrasses us. Accuracy is the only test, and the corrected record says what the registry says, not what anyone would prefer it to say.

If you believe we have broken this policy, that something is on this site because someone paid for it, or missing because someone paid for it to be missing, then report it. Every trial, condition, and medicine page carries a "Report an issue" link that reaches us directly. Use it, or the contact page, and tell us what you think we got wrong and why. We aim to respond within 48 hours.

We would rather hear it from you than not hear it at all.

Related: what our data does and does not cover, and our methodology.