1,357 of 603,967 registered trials list a Philippine study site — 0.22% of the world's. Only 281 of them are Philippines-only. In all the rest the country supplies 5,217 of 166,220 study sites, or 3.1%, and 85.2% of the whole slice is run by industry.
A trial "in the Philippines" is usually a trial run from somewhere else with a Philippine site attached. 1,076 of the 1,357 run in more than one country, and the question worth asking of those is not whether the Philippines is in them — it is how much of them it is.
of multi-country trials give the Philippines under a tenth of their sites — 836 of 1,076. Across all of them it is 5,217 sites out of 166,220.
multi-country trials put half or more of their sites in the Philippines. That is the whole set in which the country is the principal place the study happens.
trials — 20.7% — list no country but the Philippines. Those are the domestic research record, and they are one trial in five.
Against the countries a sponsor would actually weigh it against, the Philippines places 10th of 11 — above Vietnam on 1,324, below everyone else.
Singapore runs 3.0× as many as the Philippines — 4,096 against 1,357 — with a small fraction of the population, and Thailand 4,329. Those are raw counts, and they are raw on purpose: every population series good enough to divide by is CC BY or stricter, and this site is built to public domain or CC0 only. So the comparison here is presence, not intensity, and the per-capita version of this chart is one you will have to do in your head.
Those 1,076 multi-country trials reach 141 other countries. The company the Philippines keeps in a trial protocol is mostly not its neighbours.
United States appears in 61.5% of them, which is what a global Phase 3 protocol looks like: a sponsor assembles sites across continents to enrol a target number of patients, and the Philippines is one of the places that target gets met. The list below it is Europe and East Asia, not South-East Asia — these trials are not regional studies that happen to include the Philippines.
Trial phases are a ladder. Phase 1 is first-in-human, a few dozen people, where the science is. Phase 3 is the large comparison against existing treatment, where the patients are. The Philippine distribution is not shaped like a research base.
723 trials are Phase 3 — the stage that needs large numbers of patients and comparatively little infrastructure.
trials are Phase 1, first-in-human. That is 1.6% of the slice, against 53.3% at Phase 3.
have a company as lead sponsor. 13 are led by the US National Institutes of Health. Public research funding barely appears.
5,217 Philippine study sites across 414 named places. 4,837 carry coordinates; 380 do not and are absent from the map below rather than guessed at.
Trials are counted by the year they started, from 1991 to the present. The busiest year on this chart is 2010.
Read the right-hand end carefully. A trial appears here on the year it started, so the last year or two is always still filling up — a study beginning this quarter may not be registered yet. The fall after 2010 is longer than that effect explains, but the final points are not yet a trend.
Four limits. The first is a deliberate absence rather than a gap.
The register is a US federal work and public domain, but its records are written by sponsors, and NLM's policy warns that contributed content may carry its own copyright. So the extraction requests structured fields only — no titles, summaries or descriptions. That prose was never fetched, never cached and is not in this repository, and a check fails the build if a column that could hold it ever appears.
The obvious next question is whether what gets studied here resembles what people here are ill with. Every disease-burden series is CC BY-NC-SA or stricter, so that comparison is not attempted rather than approximated. Diabetes Mellitus, Type 2 leads the conditions with 108 trials; what that means is left to the reader.
The conditions charted are NLM's own MeSH mapping, not the sponsors' free text, because the derived vocabulary is consistent and unambiguously public domain. NLM mapped nothing to 138 trials — 10.2% — and those are absent from the condition figures.
A status is only as true as its last update. 8 trials in this slice still say they are recruiting while also giving a completion date that has passed — a contradiction inside one record, found without consulting a calendar.
Every figure on this page traces to one of these, through a CSV in
data/ph-trials/. Each is checked against its source query on every
build.
All 1,357 registered trials listing a Philippine study location, with their status, type, phase, sponsor class, enrollment, dates and every study site in every country — Keyless: a plain GET, no account, no token, no form. Filtered server-side on filter.advanced=AREA[LocationCountry]Philippines, which the API resolves identically to the free-text location query. Paged at the maximum of 1,000; the count is read from the first page only, because the token pages do not carry it and taking it from the last page is how a paged extract quietly becomes however many rows arrived
The licence this page relies on, quoted in full: “Works produced by the U.S. government are not subject to copyright protection in the United States” and “may be freely used or reproduced without permission in the U.S.” — The same policy carries the caveat that shaped this whole pipeline: “you may encounter documents, illustrations, photographs, or other content contributed by or licensed from private individuals, companies, or organizations that may be protected by U.S. and international copyright laws.” Trial records are contributed content, so the extraction projects structured fields only and never requests the sponsor's prose
The 603,967 registered trials the Philippine slice is measured against — Read once per run so the denominator is the register's own figure on the day of the snapshot, rather than a number copied from documentation that has since moved
The condition vocabulary this page charts: NLM's own mapping of each record to Medical Subject Headings, and the 138 trials it mapped to nothing — Sponsors type their own condition strings and they are as consistent as free text always is. The derived MeSH terms are NLM's work rather than the sponsor's, which makes them both cleaner and unambiguously public domain. The 10.2% of trials NLM never mapped are published as a figure rather than dropped
Would have provided the human-readable description of what each trial is actually testing — Rejected on the licence, not on usefulness. These fields are written by the sponsor, and NLM's policy expressly warns that contributed content may carry its own copyright. The pipeline never requests them, so they are not in the cache and not in this repository — the field list in fetch_trials.py is the whole of the control, and a check asserts the published schema cannot grow one
Would have turned the regional comparison into trials per million people, which is the comparison a reader actually wants — CC BY 3.0 IGO and CC BY 4.0 respectively, and this site is built to public domain or CC0 only. So the regional table is raw counts, and the page says plainly that Singapore's threefold lead is not adjusted for the fact that it has a small fraction of the population
Would have supported the original question: whether what gets studied in the region matches what people there are actually ill with — CC BY-NC-SA 3.0 IGO — non-commercial and share-alike. That half of the question is unanswerable inside this licence rule, so it is not attempted rather than approximated, and the page states the gap instead of implying the trial mix reflects need
I wrote a companion post — “A Drug Trial “In The Philippines” Is Usually Somebody Else’s Trial” — in simple, everyday words.