Japan Medical Facilities (MHLW Open Data)

Every registered medical facility in Japan — hospitals, clinics, dental clinics, maternity homes, pharmacies — as English-schema JSON.

This Actor delivers Japan's official nationwide medical-facility register, published as open data by the Ministry of Health, Labour and Welfare (医療情報ネット) through the e-Gov data portal: roughly 245,000 facilities with names (kanji, kana, romaji where published), facility type, prefecture, address, coordinates, homepage, bed counts by ward type, and clinical departments. One run makes at most two upstream requests — the rest is offline processing of the official snapshot.

Unofficial tool. Not affiliated with or endorsed by MHLW or the Government of Japan. The e-Gov dataset page states the data is freely reusable under the portal terms; every record embeds the source, the dataset link, and the retrieval timestamp.

---

Who is this for?

• Healthcare market entry & sales teams — medical device makers, pharma, health-IT vendors building target lists of Japanese hospitals by prefecture, bed capacity, or department mix.
• Research & analytics — health-services researchers and consultants who need the full facility universe with coordinates for geographic analysis.
• Data & AI teams — a clean facility master for entity resolution, mapping products, or RAG pipelines — without parsing five differently-shaped Japanese CSVs inside a 27 MB ZIP.

What you get

• Five facility types in one schema: hospitals, medical clinics, dental clinics, maternity homes, pharmacies — filterable by type and prefecture.
• Bed counts by ward type (general, long-term care, psychiatric, tuberculosis, infectious disease) for hospitals and clinics.
• Clinical departments joined from the register's speciality table (hospitals / clinics / dental), deduplicated by department code.
• Honest sentinels handled: the upstream's "0.0 / 0.0" not-geocoded coordinates (~15k rows) become null instead of placing facilities in the Gulf of Guinea.

Sample output (dataset item)

{
  "facilityId": "1810000000",
  "name": "○○総合病院",
  "nameKana": "マルマルソウゴウビョウイン",
  "facilityType": { "key": "hospital", "ja": "病院", "en": "Hospital" },
  "prefecture": { "code": "18", "nameJa": "福井県", "nameEn": "Fukui" },
  "address": "福井県福井市○○町1-1",
  "latitude": 36.0641,
  "longitude": 136.2196,
  "homepageUrl": "https://www.example-hospital.jp/",
  "bedCounts": { "general": 240, "psychiatric": 0, "total": 240 },
  "departments": [
    { "code": "01", "nameJa": "内科" },
    { "code": "10", "nameJa": "小児科" }
  ],
  "snapshotDate": "2025-12-01",
  "source": "医療情報ネット（厚生労働省）",
  "sourceUrl": "https://data.e-gov.go.jp/data/dataset/iryou_teikyouseido_mhlw",
  "license": "e-Gov data portal terms (free reuse per dataset page)",
  "retrievedAt": "2026-08-25T09:12:44Z"
}

Input example

{
  "facilityTypes": ["hospital", "clinic"],
  "prefecture": "Tokyo",
  "includeDepartments": true,
  "maxItems": 1000
}

prefecture accepts JIS codes, Japanese names, or English names. maxItems goes up to 200,000 for full-country pulls.

Pricing

Pay per result — see the pricing tab. A whole-prefecture hospital list is typically a few hundred records.

FAQ

Where does the data come from? From the nationwide 医療情報ネット open-data snapshot published by MHLW on the e-Gov data portal. The Actor automatically detects the newest snapshot (or you can pin one via datasetZipUrl), and snapshotDate on every record tells you which one you got.

How fresh is the data? The government publishes this register twice a year (June and December snapshots). That's the honest cadence — running the Actor daily won't produce newer data, so schedule accordingly. For a facility master, twice-yearly is what the official source offers; no vendor selling you "live" Japanese facility data is getting it from this register any faster.

Is this an official government service? No. This is an unofficial tool, built and maintained independently, using officially published open data.

Does the data contain personal information? The published facility CSVs carry no person-name, phone, or e-mail columns — and this Actor enforces that structurally: fields are extracted by whitelist, name-shaped columns are excluded by pattern even if a future snapshot adds them, and an automated legal test asserts that no phone-, e-mail-, or personal-name-shaped strings appear anywhere in the output. Addresses are facility addresses as officially published.

Why is there no phone number for facilities? Because the official open data doesn't include one. We won't enrich from other sources and pretend it came from the register.

Any resource notes? A full nationwide run over clinics with departments processes ~150 MB+ of decompressed CSV; use a run with 2 GB+ memory on Apify for country-scale pulls. Prefecture-scoped runs are light. Upstream load is minimal by design: at most 2 requests per run, throttled and backoff-compliant, no rate-limit evasion.

Search terms this Actor answers

Japan hospital database · Japan medical facilities list · Japan clinics dataset · Japan pharmacy database · Japanese healthcare facilities data · MHLW open data English · hospital bed counts Japan · Japan healthcare market data

More Japan public-data Actors by the same developer

• Japan Government Tenders & Awards — tender notices and award results from the kkj.go.jp portal.
• Japan Care Facilities — long-term care providers across 35 service types (MHLW open data, CC BY).
• Japan Government Subsidies & Grants — open subsidy calls from the official jGrants API.
• Japan Renewable Energy FIT/FIP Facilities — METI-certified renewable installations, privacy-safe schema.
• Japan Financial Licenses & Registrations — FSA-licensed banks, securities firms, crypto exchanges and more.
• Japan Construction & Real-Estate Licenses — licensed contractors and brokers from MLIT's register.
• Japan Trade Statistics — monthly import/export flows by HS code and partner country, from Japan Customs.

---

Data source: 医療情報ネット（厚生労働省）open data via the e-Gov data portal — https://data.e-gov.go.jp/data/dataset/iryou_teikyouseido_mhlw . Reuse per the portal terms; source is stated in every record.