How Can Vendors Research an SNF Market Responsibly With Public Data?

Define the market and research question before downloading data, use stable CMS facility identifiers, preserve dataset versions and field definitions, separate facility facts from calculated metrics, and publish uncertainty and exclusions. Use public records to describe facilities and documented events—not to infer purchase intent, budget, clinical need, culpability, or a decision-maker. Citations and ratings are not lead scores, sponsored rankings, or permission to make targeted claims about a facility.

Frame a question the data can actually answer

Start with a neutral question such as “How many active Medicare- or Medicaid-certified nursing homes are located in these counties as of the dataset date?” Define geography, certification status, observation period, unit of analysis, and exclusions before viewing outcomes.

Write down claims the project will not make. A deficiency record can establish that a public record exists under the dataset’s definitions. It does not establish present conditions, purchase intent, budget, contract timing, or that a product would have prevented the finding.

Define geography using named states, counties, ZIP codes, or a documented radius method.
Define whether the unit is a certified facility, campus, organization, owner record, or survey event.
Choose an as-of date and observation window appropriate to each dataset.
Predefine active-status handling, missing values, duplicates, and suppressed fields.

Build a reproducible facility spine

Use the CMS Certification Number or the dataset’s documented facility identifier as the primary join key. Names and addresses change and are poor substitutes. Keep raw downloads immutable, retain data dictionaries, and record the dataset page, retrieval date, file version or update date, and transformation code.

Join one-to-many tables deliberately. A facility can have multiple surveys, deficiencies, penalties, ownership records, or dates. Collapsing these to one row without a declared rule creates double counting and misleading recency.

Download only fields needed for the stated question from official CMS dataset pages.
Profile identifiers, date ranges, nulls, duplicates, category values, and outliers before analysis.
Document every filter, join, grouping, and derived field in plain language.
Spot-check calculated results against source rows and Care Compare facility pages.
Separate CMS-published values from vendor-calculated values in labels and exports.

Publish context, not a covert prospecting score

Report counts, rates, or trends with denominators, date windows, refresh dates, and known limitations. Ratings combine defined measures and should be described using current CMS methodology and source dates. Ownership fields may not reveal operational control or buying authority.

Do not call a facility “ready to buy” because it has citations, a rating change, or an ownership entry. Do not imply CMS endorsement. If a vendor publishes rankings, sponsorship and commercial relationships must be conspicuous, and payment must not alter a purportedly objective rating.

Use neutral labels such as “facilities with a public survey record in the selected period.”
Give users the source, as-of date, denominator, and method alongside each result.
Create a correction channel for identity, join, and calculation errors.
Set a refresh cadence that reflects source publication rather than claiming real-time status.
Review outreach language separately so descriptive data is not converted into unsupported need or intent claims.

Illustrative example: county market inventory — illustrative example

A vendor wants a reproducible inventory of certified nursing homes in three counties and a descriptive view of public survey activity. The example is a research method, not a sales ranking.

The analyst defines the counties, active-status rule, facility-level unit, survey observation window, and retrieval date before downloading data.
Provider Information becomes the facility spine keyed by CMS Certification Number. Survey records remain a separate one-to-many table.
The output reports facility count and the number with at least one published survey record in the defined window. It labels both the CMS fields and the analyst-derived flag.
A methods note lists dataset URLs, update dates, exclusions, missing-value treatment, joins, and the query used to avoid double counting.
The vendor does not rank “sales readiness,” infer buying authority from ownership, or tell facilities that a citation proves they need the vendor’s product.

A responsible market view is auditable, time-bounded, and modest about what public records can reveal.

Practical checklist

Research question, geography, unit of analysis, as-of date, and observation window are written first.
Official dataset pages, data dictionaries, retrieval dates, and source versions are retained.
CMS Certification Number or documented stable identifier anchors facility joins.
One-to-many surveys, deficiencies, penalties, and ownership records are not accidentally double-counted.
Filters, joins, exclusions, missing values, derived fields, and denominators are documented.
CMS-published fields and analyst-calculated metrics are clearly distinguished.
Results are spot-checked against source rows and current Care Compare pages.
Limitations, refresh date, and correction channel appear with published results.
No citation, rating, or ownership field is presented as evidence of purchase intent or buying authority.
Sponsorship cannot secretly influence rankings, and no CMS endorsement is implied.

Limitations

CMS datasets are published on different schedules and may describe different observation periods; a download is not a real-time operational picture.
Identifiers, ownership structures, survey histories, and rating methods require their dataset-specific definitions and may change.
Public regulatory data does not reveal purchase intent, budget, procurement authority, clinical need, or causation.
This guide is not legal advice about privacy, advertising, outreach, procurement, antitrust, or data licensing.

Official sources

CMS Provider Data Catalog — nursing homes — Official CMS nursing-home datasets, update information, and data dictionaries.
Provider Information dataset — Official CMS dataset page with facility identifiers, locations, ownership fields, ratings, and metadata.
Medicare Care Compare — nursing homes — Official Medicare facility search and comparison experience.

Useful next steps

Facility map and sales-visit planner — Research locations and organize a multi-stop driving route.
Resource center — Browse public facility research references.
Survey guide — Understand the survey records represented in public data.

Editorial update: 2026-09-17. Editorial standards and corrections.

Explore POCDesk — Free Nursing Home Compliance Tools

Official Sources

POCDesk analyzes official government data. Verify any facility's record or read the underlying regulations at these primary sources:

Medicare.gov Care Compare — CMS's official nursing home comparison tool
CMS Provider Data Catalog — the raw nursing home inspection & deficiency datasets
CMS Nursing Home Certification & Compliance — enforcement policy and the Special Focus Facility program
42 CFR Part 483 (eCFR) — the federal requirements of participation behind every F-Tag
National Long-Term Care Ombudsman Resource Center — free advocacy help for residents and families