Clinical trials · off ClinicalTrials.gov’s own register

Which studies are recruiting,
and how near the nearest site is.

ClinicalTrials.gov is the public register of clinical research, kept by the U.S. National Library of Medicine: every registered study, from a one-clinic survey to a drug trial run at a thousand hospitals, as its sponsor entered it. Give this page a condition and a place. It asks the register directly, from your browser, which studies are recruiting, and lays each one out with its phase, sponsor, enrolment and dates — and the listed sites nearest you, measured from the register’s own coordinates.

It is a way into the register, not advice. A listing is not a recommendation — the register says itself that listing a study does not mean the U.S. government has evaluated it — and whether a trial suits you is a question for a clinician who knows you. Every result links to its record on ClinicalTrials.gov, which is the source: read it there before you contact anyone.

Search the register Reading a record

Read live from ClinicalTrials.gov · asked straight from your browser, no relay · nothing you type is kept here

The search

A condition, a place, and who is recruiting.

Recruiting studies only, until you tick more. A place with a distance is measured from the register’s own coordinates for that place; “Use my location” asks your browser instead and sends only a rounded point. Enter searches from any field.

Reading a record

What the words on a record mean.

Every label on a result is the register’s own field, in plain words. These are what the register means by them.

Status

Recruiting: At least one site is enrolling participants now.
Not yet recruiting: Registered and approved to start; nobody enrolled yet.
By invitation: Enrolling by invitation: participants come from a group the researchers chose in advance, so you cannot volunteer.
Active, not recruiting: Under way, closed to new people.
Completed: Finished as planned. Results, if posted, are on the record.
Suspended, terminated, withdrawn: Paused, stopped early, or stopped before its first participant.
Unknown status: It said it was still going, its completion date has passed, and nobody has verified the record in two years.

Each site carries its own status as well: a study can be recruiting in one city and closed in another. The site lines on a result show both.

Phase

Early phase 1: Once called phase 0: very limited exposure in people, with no intent to treat or diagnose.
Phase 1: First studies in people: how the body handles a drug, side effects as the dose rises. May enrol healthy volunteers.
Phase 2: Does it work for this condition, and what are the common short-term side effects.
Phase 3: Larger trials once there is early evidence, to weigh the benefit against the risk.
Phase 4: After approval: more on safety, effectiveness and the best use.
Not applicable: No FDA-defined phase — device and behavioural trials, among others.

A combined label (Phase 1/2, Phase 2/3) is one study spanning both. Observational studies have no phase: they follow people and assign nothing.

Dates, numbers, who

Start, primary completion, completion: As the sponsor gives them. Estimated until they happen, actual after; a record that says neither is printed as it stands.
Enrolment: The target (estimated) or the final count (actual) of participants.
Who can join: Sex, the age range and whether healthy volunteers are accepted — the summary only. The full eligibility criteria are on the record.
Last updated: The day the register last posted a change the sponsor made. An old date on a recruiting study is worth a phone call before a journey.

Method

Where every word comes from, and where yours go.

The source

ClinicalTrials.gov, run by the National Library of Medicine at the U.S. National Institutes of Health. Sponsors and investigators enter their own studies; the register checks the entries for apparent errors and gaps, not for their science. This page reads its public, keyless API, version 2, straight from your browser — the register answers any site (its CORS header is *), so there is no relay, no cache and no copy on labs.llc.

clinicaltrials.gov/api/v2/studies — the search: query.cond, query.term, query.locn, filter.geo, filter.overallStatus, filter.advanced (phase, kind, age), countTotal and pageToken, twenty studies a page.
clinicaltrials.gov/api/v2/version — the register’s data timestamp, printed under the results as given (it names no time zone).
clinicaltrials.gov/api/v2/stats/size — how many studies the register holds, printed at the top.

Re-use of the register’s data is governed by its own terms and conditions (clinicaltrials.gov/about-site/terms-conditions). Every result here names the register, links to its record and carries the register’s own last-update date.

What is fetched, and what is worked out here

Fetched, printed as the register holds it: every title, condition, sponsor, enrolment figure, date, site, site status and summary, and the count of matches. Status and phase codes are turned into plain words; the codes are the register’s. The order of the studies is the register’s — its best match, or its last-update date if you ask for that.

Worked out here: the distance from your point to each site, along the great circle (haversine, Earth radius 3,958.8 miles), from the coordinates the register gives each site; the order of the nearest sites inside one record; and, when you type a place with a distance, the point to measure from. For that the page asks the register for sites in the place you typed and takes the coordinates it gives the best-matching, most-listed one — other places of the same name are offered beside it. When no site it returns spells what you typed (“NYC”, “Philly”), the city nearly all of those studies share is the register’s own reading of it, and the page says so; when they share none, the place is searched by name and the page says the distance was not applied. A country or a state is searched by name, never shrunk to a point.

The limits

The register is what sponsors enter, and it lags: a site shown as recruiting may have closed last week. The distance filter is the register’s own — a study qualifies when any of its sites is inside the distance — and a site with no coordinates cannot be measured, so it is listed without a distance. The register places each site at its city’s coordinates, not its street address, so every distance runs to the site’s city, not to its door: a site in the very city you searched from reads 0 miles, whichever side of town it is on. The words you type go through the register’s search engine, which widens them to synonyms, so a result may name your condition only among its keywords. Twenty studies are drawn at a time; the count above them is the register’s.

If the register cannot be reached, or refuses a search, the page says so in words — the register’s own message where it sends one — and offers to try again. Nothing is filled in while it is away.

What happens to what you type

Your search goes from your browser to clinicaltrials.gov and nowhere else: labs.llc has no relay here, logs nothing and stores nothing, and the request carries no referrer. The register sees the search and your IP address, as a visit to its own site would, and its privacy policy applies there.

“Use my location” happens only when you press it, and your browser asks you first. The point is rounded to two decimal places — about a kilometre — before it leaves the page, and it goes only to the register, inside the search.

The page’s address holds no search unless you copy a link to one. A link that carries a condition (?cond=, the form a link from a symptom checker takes) is read once and then taken out of the address bar; the site’s analytics tag reads the address as the page loads, so a condition in a ?cond= link can reach it. The same link written #cond= is never sent to any server. Kept in your browser: whether you want the notice cards, and — for the one reload that switching editions causes — the search on screen, deleted as the page comes back.

Questions and corrections: labs@labs.llc.