🔔 CMS Coverage Related-Document Graph

One structured edge per filed related-document link between coverage policies in the official CMS Medicare Coverage Database — the source policy (LCD/Article) and its typed related endpoints (a related Article, a related LCD, or a related contractor). A graph over ~12,299 filed edges, keyed and deduplicated on source_type:source_id:version:related_num. Built for coverage-policy, compliance, and knowledge-graph teams mapping how CMS policies reference one another.

Differentiation: the cross-references between LCDs and Articles are scattered across two tables in two CMS ZIPs; this graph joins them into one flat, deduplicated, typed edge. Missing endpoints are kept as explicit nulls (a related-Article edge with no related-LCD keeps a null r_lcd_id), never dropped or guessed, and no clinical relationship is inferred beyond the filed link — this is a document graph, not a coverage determination.

Output is one cms_coverage_document_edge row per result; billing is pay-per-event, the value event being one coverage document edge (a $0.02 start fee per run, then $0.02 per coverage document edge). Source: downloads.cms.gov, www.cms.gov.

No login, no API key and no CAPTCHA solving are involved: the source is read logged-out with an identified contact User-Agent.

📊 Sample Output

Real rows from run nJbbFraIEUgDyPw2x on build 0.1.7 (2026-09-19), the same input as the Quick start below — every value is as the source published it (emails masked, long text shortened):

source_type	source_id	source_version	related_num	edge_type	r_contractor_id
LCD	32553	20	2	->article	228
LCD	32553	21	2	->article	228
LCD	33252	29	1	->article	368
LCD	33252	29	3	->article	368
LCD	33261	40	1	->article	368
LCD	33261	40	2	->article	368
LCD	33261	40	3	->lcd	368
LCD	33267	17	1	->article	368

The run finished with the status message: NORMAL: 50 unique rows (0 duplicates collapsed). billable=50

✅ What you get

Each row is flat JSON with these fields (from the dataset schema and the sample run; a field the source does not publish for a given row is null):

• record_type (string/null) — e.g. cms_coverage_document_edge
• source_type (string/null) — e.g. LCD
• source_id (string/null) — e.g. 32553
• source_version (string/null) — e.g. 20
• related_num (string/null) — e.g. 2
• edge_type (string/null) — e.g. ->article
• r_article_id (string/null) — e.g. 56424
• r_article_version (string/null) — e.g. 3
• r_lcd_id (string/null) — e.g. 38927
• r_lcd_version (string/null) — e.g. 11
• r_contractor_id (string/null) — e.g. 228
• source (string/null) — e.g. cms-coverage-related-document-graph
• release_date (string/null) — null in every sample row
• source_url (string/null) — e.g. https://www.cms.gov/medicare-coverage-database/
• observed_at (string/null) — e.g. 2026-09-19T00:20:11Z
• license (string/null) — e.g. CMS Medicare Coverage Database — U.S. Government public data; CPT is copyright A
• attribution (string/null) — e.g. U.S. Centers for Medicare & Medicaid Services, Medicare Coverage Database (https
• terminal (string/null) — null in every sample row
• reported_total (string/integer/null) — null in every sample row
• raw_seen (string/integer/null) — null in every sample row
• duplicates (string/integer/null) — null in every sample row
• unique_found (string/integer/null) — null in every sample row
• delivered (string/integer/null) — null in every sample row
• scope (string/null) — null in every sample row
• note (string/null) — null in every sample row

✅ What you get

• record_type (string/null) — e.g. cms_coverage_document_edge
• source_type (string/null) — e.g. LCD
• source_id (string/null) — e.g. 32553
• source_version (string/null) — e.g. 20
• related_num (string/null) — e.g. 2
• edge_type (string/null) — e.g. ->article
• r_article_id (string/null) — e.g. 56424
• r_article_version (string/null) — e.g. 3
• r_lcd_id (string/null) — e.g. 38927
• r_lcd_version (string/null) — e.g. 11
• r_contractor_id (string/null) — e.g. 228
• source (string/null) — e.g. cms-coverage-related-document-graph
• release_date (string/null) — null in every sample row
• source_url (string/null) — e.g. https://www.cms.gov/medicare-coverage-database/
• observed_at (string/null) — e.g. 2026-09-19T00:06:34Z
• license (string/null) — e.g. CMS Medicare Coverage Database — U.S. Government public data; CPT is copyright A
• attribution (string/null) — e.g. U.S. Centers for Medicare & Medicaid Services, Medicare Coverage Database (https
• terminal (string/null) — null in every sample row
• reported_total (string/integer/null) — null in every sample row
• raw_seen (string/integer/null) — null in every sample row
• duplicates (string/integer/null) — null in every sample row
• unique_found (string/integer/null) — null in every sample row
• delivered (string/integer/null) — null in every sample row
• scope (string/null) — null in every sample row
• note (string/null) — null in every sample row

⚙️ Sample inputs

1. Quick start — the Store example (this is what the sample above came from)

{
  "maxRecords": 50
}

The sample run charged exactly: 4 × $0.02 apify-actor-start + 50 × $0.02 coverage_document_edge = $1.08 on the Free tier — every delivered row was billed.

2. A smaller, narrowed run

{
  "maxRecords": 5
}

Caps the run at 5 rows — about $0.12 on the Free tier ($0.02 start + 5 × $0.02).

3. A full-size run

{
  "maxRecords": 2500
}

Up to 2500 rows (the schema default for maxRecords) — about $50.02 on the Free tier ($0.02 start + 2500 × $0.02) if the source has that many.

4. A filtered run

{
  "maxRecords": 50,
  "fieldMatch": {
    "edge_type": "->article"
  }
}

Keeps only rows whose edge_type contains ->article (a value from the sample run above), matched case-insensitively after the source is read and before the maxRecords cap, so the cap counts matches and only delivered matches are billed. searchText does the same across every column. Zero matches end the run as EMPTY_SELECTION with no row charged — if the source has moved on since that sample run, pick a value you see in your own first run. The source is still read in full — this narrows what you receive and pay for, not what the source serves.

🧾 JSON sample record

One real record from run nJbbFraIEUgDyPw2x, exactly as it lands in the dataset (emails masked, long text shortened):

{
  "record_type": "cms_coverage_document_edge",
  "source_type": "LCD",
  "source_id": "32553",
  "source_version": "20",
  "related_num": "2",
  "edge_type": "->article",
  "r_article_id": "56424",
  "r_article_version": "3",
  "r_lcd_id": null,
  "r_lcd_version": null,
  "r_contractor_id": "228",
  "source": "cms-coverage-related-document-graph",
  "release_date": null,
  "source_url": "https://www.cms.gov/medicare-coverage-database/",
  "observed_at": "2026-09-19T00:20:11Z",
  "license": "CMS Medicare Coverage Database — U.S. Government public data; CPT is copyright American Medical Association (CPT/HCPCS descriptors and free-text descriptions are excluded).",
  "attribution": "U.S. Centers for Medicare & Medicaid Services, Medicare Coverage Database (https://www.cms.gov/medicare-coverage-database/)"
}

🔧 How it works

Source. The actor reads downloads.cms.gov, www.cms.gov — endpoints: https://downloads.cms.gov/medicare-coverage-database/downloads/exports/current_article.zip, https://downloads.cms.gov/medicare-coverage-database/downloads/exports/current_lcd.zip, https://www.cms.gov/medicare-coverage-database/. Public pages and feeds only; nothing behind a login.

Transport. Plain HTTPS from the Apify platform, no proxy. robots.txt is read first and a disallowed path is never fetched. Every request carries an identified contact User-Agent. Pacing: CRAWL_DELAY=0.0.

Terminal states. A run ends NORMAL, CAPPED (your cap was reached), PARTIAL (something was withheld and the message says what), GENUINE_EMPTY (the source was read and truly had nothing in scope) or BLOCKED (the source refused or changed shape — the run FAILS loud and bills nothing). A zero-row run is never reported as a silent success.

Charging. Each coverage document edge is charged at the moment it is pushed (coverage_document_edge); a row that fails to charge is not delivered, so the dataset count always equals the charged count.

🔧 How it works

Source

The official current_lcd.zip and current_article.zip from downloads.cms.gov (the lcd_related_documents.csv and article_related_documents.csv members — the .mdb is not read), streamed with an early cap and header layout canaries that fail loud (unbilled) on shell body, header drift, or zero-parse.

Compliance & licence

• Only the filed edge endpoints are emitted; descriptor/free-text columns are excluded by position where present. Every edge carries the CMS licence (license) and attribution (attribution).

What is not done. No login, no cookie or CAPTCHA bypass, no private or personal-account data, no browser automation.

💰 Pricing example

Event	Free	Bronze	Silver	Gold
Actor Start (apify-actor-start)	$0.02	$0.02	$0.02	$0.02
Coverage Document Edge (coverage_document_edge)	$0.02	$0.018	$0.016	$0.0134

Worked at the live Free-tier price:

• 8 coverage document edges: $0.02 start + 8 × $0.02 = $0.18
• 25 coverage document edges: $0.02 start + 25 × $0.02 = $0.52
• 2500 coverage document edges: $0.02 start + 2500 × $0.02 = $50.02

A run that delivers zero rows charges the $0.02 start fee only. A BLOCKED run (source refused) fails loud and charges no value event. The start fee is charged once per GB of run memory; the default run memory is 4096 MB.

Yield on the sample run: NORMAL: 50 unique rows (0 duplicates collapsed). billable=50. maxRecords is a hard ceiling on what is delivered and billed, never a target.

⚖️ Legal & ToS

This actor reads public, logged-out pages and feeds published by downloads.cms.gov, www.cms.gov. It collects only what the source publishes to any visitor, keeps to the source's robots rules (checked on every run), identifies itself with a contact User-Agent, and does not access accounts, private data or anything behind authentication. Use the output in line with the source's terms and your local law; the intended use is B2B research and monitoring.

❓ FAQ

Q: Do I need an API key or a login?
A: No. The source (downloads.cms.gov, www.cms.gov) is read logged-out; the input schema has no key field and the actor carries no secrets.

Q: Why did my run return 0 rows?
A: Read the run's status message. GENUINE_EMPTY means the source was read and had nothing in scope for your input; BLOCKED means the source refused and the run failed without billing a value event — retry later or narrow the input. A zero-row run bills the start fee only.

Q: How many rows can one run return?
A: Up to maxRecords (default 2500). Raise the cap for a bigger run; you pay per delivered row.

Q: How fresh is the data?
A: Every run reads the source live at run time; nothing is cached between runs. Put it on a schedule for a continuous feed.

Q: What formats can I export?
A: The dataset downloads as JSON, CSV, Excel, XML or RSS from the run's Dataset tab or the Apify API, and any run can push to a webhook or integration.

Q: How is this different from the other NexGen Watch actors actors?
A: Same output shape and billing model; this one covers downloads.cms.gov, www.cms.gov. The siblings under Related Actors cover the other sources or slices — run several on one schedule for a combined feed.

Q: Are there rate limits?
A: The actor paces itself against the source (CRAWL_DELAY=0.0) and honours its robots rules; there is no per-buyer limit beyond your Apify plan's concurrency.

🆘 Troubleshooting

• Run FAILED with BLOCKED → the source refused the request or changed its page shape → nothing was billed beyond the start fee; retry after a while, and if it persists open an Issue with the run id.
• Status says CAPPED → your cap (maxRecords) was reached → raise it for a bigger run.
• Input validation error on start → a field is outside the schema's allowed values → start from the Quick start block and change one field at a time.
• Run TIMED-OUT → a very wide request on a slow day → raise the run timeout in Run options or narrow the input; what was delivered before the timeout is still in the dataset.

🔗 Related Actors

• Chicago Business Owners Register — One record per Chicago business owner/officer statutory-register row (account, DBA name, owner name parts, title) from the City's public Business Own…
• CMS Article Response-to-Comment Ledger — One structured record per filed public comment-and-response on a CMS Article — the article id and version, the response-to-comment number, the (HTML-…
• CMS CLIA Laboratory Registry — One clia_lab_record per CLIA-certified laboratory: CLIA number, facility name, full address, certification date, termination status, and lab classifi…
• CMS Dialysis Facilities Scraper — A clean, structured roster of Medicare-certified dialysis facilities in the United States, sourced live from the official CMS provider-data API (http…
• CMS HCPCS Coverage Policy Index — One structured policy_code_link per HCPCS code ↔ coverage-policy relation — across BOTH Local Coverage Determinations and Articles in the official CM…
• CMS Hospice Facilities Scraper — A clean, structured roster of Medicare-certified hospice facilities in the United States, sourced live from the official CMS provider-data API (https…
• 🏢 About NexGenData — NexGen Watch is NexGenData's fleet of more than 300 public data actors built on official sources, pay-per-result. Browse the catalog at apify.com/nexgenwatch.

⭐ Found this useful?

If this actor saved you a manual check, a quick review on the Apify Store helps other teams find it. Feature request or a source that changed? Open it from the Issues tab — every one is read.