A patient searches 'clinical trials for IgA nephropathy' at 11 PM. The first 4 results look like neutral search tools. Several of them aren't. They're recruitment funnels dressed up as finders, where sponsor dollars decide which trials rank first and a form at the bottom hands her email to a contract research organization before she closes the laptop. The approved drugs she could ask her nephrologist about on Monday morning, the 5 that have come to market in the last 3 years, do not come up at all.
That is how many of the trial finders that rank highest in Google results work. Trial Friend was built on the premise that patients deserve a different kind of clinical trial finder, and this post is a public record of what that actually means once you get past the marketing copy.
Most trial tools were built to solve the sponsor's problem
Only 47% of clinical studies complete enrollment on time. 53% run past their planned timeline, and one in six takes more than twice as long as originally planned. On top of that, 11% of investigative sites in a given trial fail to enroll a single patient (Tufts Center for the Study of Drug Development). Pharmaceutical sponsors spend billions a year trying to close that gap, and the recruitment industry built to serve them has grown into a multi-billion dollar business of its own.
A tool built to solve that problem ranks trials by how badly a sponsor needs enrollees, not by how well the trial matches a patient's situation. It collects contact information and passes it to the sponsor or a CRO. It shows the trial that paid for the top slot ahead of the one closest to a patient's house, or the one with the most mature science. None of that is illegal. A lot of it is not even wrong from a sponsor's perspective. It is working as designed.
The design just isn't the patient's.
What that looks like when you are the patient
You see a short list of trials that feels curated but isn't. A page about your disease that reads like a brochure for one specific drug. A symptom checker that concludes, helpfully, that you should talk to your doctor about a trial conveniently enrolling this quarter. No warning that a medication you already take would disqualify you from half the trials on screen. No mention that an already-approved drug might be a better first conversation with your doctor.
Two rare disease foundation leaders I spoke with this month raised the same question inside the first 10 minutes of each call. Are you collecting patient data? Who do you sell it to? The speed with which that came up was not an accident. Foundation leaders have watched trusted-looking resources erode their members' trust before. The skepticism is earned.
The data problem is not hypothetical
In 2017, iCarbonX, a Chinese genomics firm backed by Tencent, took a majority stake in PatientsLikeMe, one of the largest patient-reported outcomes networks in the United States, a platform with more than 600,000 members sharing symptoms, medications, and lab values. Two years later, the U.S. Committee on Foreign Investment forced iCarbonX to divest on national security grounds, and UnitedHealth Group bought PatientsLikeMe instead (Forbes, 2019). A network originally pitched to patients as a way to share their experience with other patients had moved through a foreign data firm and a U.S. health insurer in 24 months. None of the underlying members got a meaningful say in any of it.
That case made headlines. It is not the only one. Research published in Nature Communications in 2019 found that 99.98% of Americans can be re-identified from a dataset using just 15 demographic attributes, even when the data has been 'anonymized' in the way most privacy policies promise (Rocher et al., 2019). For rare disease patients, where a diagnosis plus a ZIP code narrows the search to a handful of people in the country, the privacy math gets worse in a hurry.
What patient-first actually means here
The phrase 'patient-first' has been said about almost every patient-facing tool on the internet. It costs nothing to say. These are the operational commitments behind Trial Friend that make it concrete.
Sponsorship is labeled, and editorial is untouched
Every disease page covers all approved treatments and all active trials, regardless of which company makes them. Sponsors and advertisers cannot buy placement, rewording, or edits inside the editorial content on a disease page. Any paid placement, whether a sponsored resource, an advertisement, or a partner-provided educational asset, is visually distinct and labeled 'Sponsored' or 'Advertisement.' Approved drugs with a 20-year track record sit alongside the newest investigational therapy in the same list, at the same level of detail. If an approved drug is the better first step before a patient considers a trial, the page says so plainly.
No trial ranking by sponsor
When a patient filters trials by location, phase, or recruitment status, the list sorts by those objective criteria. No sponsor is paying to appear first. A small academic trial 40 miles from the patient's house ranks ahead of an industry-sponsored trial 400 miles away if that is what the filter asks for.
A medication conflict checker that tells patients when to stay put
Many users type in the drug they currently take before searching trials. Trial Friend flags every trial for which that medication is a disqualifying exclusion, along with the specific washout period required. If a patient is on Vanrafia for IgA nephropathy and a trial demands a 12-week washout, the checker says so. Plenty of tools will happily show a patient a trial she cannot realistically join. Saving her the phone call is the point.
No data sales, no sponsor handoffs, no retargeting pixels
Email addresses submitted for trial alerts are used to send trial alerts from Trial Friend. No resale to sponsors. No list rental to CROs. No retargeting pixels on disease pages to follow patients around the internet with ads for a specific drug. If a patient opts in to partner communications, Trial Friend sends those emails directly; the subscriber's address is never shared with the partner. A patient's interest in researching her disease is not a signal we monetize. The full rules live in our privacy policy.
Citations to primary sources on every AI answer
Each disease page has a chatbot trained only on that disease, grounded in ClinicalTrials.gov, PubMed, openFDA, and NIH sources. Every answer cites where the information came from, so a patient or caregiver can click through to the original study or FDA document. No talking points. No ghost-written sponsor copy. If the site does not know something, it says so.
Why we are writing this down
Every cold outreach email to a foundation in the last month has led, sooner or later, to the same conversation about data, trust, and incentives. Answering that conversation well one call at a time is possible. Answering it at scale requires a public record a foundation leader can read before the call, share with a skeptical board member, and cite when they tell their members about us.
There is a longer-term reason too. A tool that serves patients honestly will eventually need to work with sponsors, CROs, and payers to keep running. When those conversations happen, the commitments in this post are the line. Trial Friend will work with a pharmaceutical sponsor on a trial awareness program if the program is grounded in public data, transparent about sponsorship, and useful even to patients whose current answer is 'no, this trial is not right for me.' Trial Friend will not sell email addresses, hand off patient contact details to sponsors, or let sponsor dollars shape how a disease page is written. Writing those lines down makes them easier to hold to.
“Every free health tool has a revenue stream. Sometimes that stream is the user.”
Trial Friend, April 2026
Patients searching for trials at 11 PM are not a market segment to us. They are the reason this exists. The easiest way to prove that is not to write a longer mission statement. It is to publish the lines we will not cross and then behave accordingly, every time. Trial Friend is an independent, patient-first clinical trial finder.