Three sources, every page

Every state-specific page on Caregiving Atlas is built from three classes of source, and we cite each one:

  1. Primary statutes and regulations. The actual text of state laws, federal CMS rules, IRS publications, and Social Security Administration guidance. We link to the statute, not to a third-party summary of it.
  2. Government and academic data.CMS hospital transparency files, BLS occupational data, Genworth’s Cost of Care survey, peer-reviewed journal articles. Numbers in our content come from these sources with the citation attached.
  3. Credentialed reviewer input.A practicing professional in the relevant state — an elder-law attorney, a CFP, a nurse case manager, or a CPA — reads the draft, flags errors, and signs off. Their name appears above the fold on every page they review.

When sources disagree — and they sometimes do, especially on Medicaid eligibility edge cases — we name the disagreement and explain which interpretation we’ve adopted and why.

Source hierarchy

Not all citations are equal. When sources of different rank conflict, we follow the higher-ranked one and footnote the conflict. The hierarchy:

  1. Statutes— the actual text of state or federal law (e.g., 42 U.S.C. § 1396, Cal. Welf. & Inst. Code § 14005.62). We link to the legislative site, not a third-party summary.
  2. Regulations— codified rules implementing the statute (e.g., 42 C.F.R., Florida Administrative Code). Where state rules diverge from federal framework, we cite the state rule and explain the divergence.
  3. Agency guidance and binding letters— CMS All-State Medicaid Director Letters, IRS published rulings, State Medicaid Manual provisions, DHCS All-County Welfare Directors Letters. These bind agency staff and therefore practitioners.
  4. Peer-reviewed research— for empirical claims about caregiver burden, dementia prevalence, cost of care. Journals with a stated peer-review process; we name the journal in the footnote.
  5. Government data releases— CMS Hospital Compare, BLS Occupational Employment Statistics, Census ACS, Genworth Cost of Care survey. We cite the dataset and release date.

What we don’t cite as primary sources: general news reporting (we use it for breaking changes only, and then confirm against the underlying statute or guidance); industry trade publications (often paid placements); Wikipedia; AI-generated summaries of any kind; competitor caregiver sites; lawyer-marketing content farms.

Who reviews

Our reviewers are real people with real credentials. We list their names, professional licenses, and state of practice on their reviewer profile pages. We don’t take anonymous reviews. We don’t accept payment from reviewers in exchange for inclusion.

Reviewers are paid a flat per-page fee for their time. They don’t get a cut of any downstream revenue from the page they reviewed. That separation is deliberate — we don’t want a reviewer’s judgment biased by which page will likely earn the most affiliate revenue.

Reviewer credentialing process

Different topics demand different credentials. We don’t let a CPA review Medicaid eligibility rules; we don’t let an elder-law attorney review tax-treatment tradeoffs. What we look for, per topic:

How we vet a prospective reviewer:

  1. License verification.We check the state bar, state board of accountancy, state nursing board, or certification body directly — not the reviewer’s self-report. Lapsed or restricted licenses disqualify.
  2. Disciplinary-history check. Public disciplinary records via the state regulator. Anything unresolved is a hard no.
  3. Sample-page review.We ask the prospective reviewer to mark up a published page in their area. We don’t pay for this and we don’t commit either way before seeing the work. The point is to verify the reviewer reads with the rigor we need, not to test trivia.
  4. Conflicts-of-interest disclosure.Every reviewer signs a written disclosure naming any active cases, business relationships, paid speaking arrangements, or expert-witness work that could affect their independence on the topics they’ll review. The disclosure appears on their reviewer profile page verbatim — we don’t paraphrase it.
  5. Per-page conflict re-check. Before signing off on a specific page, the reviewer re-confirms in writing that they have no conflict on the specific subject matter of that page (e.g., a Florida elder-law attorney reviewing a Florida Medicaid page confirms no active client matter depends on the contested interpretation discussed).

Correction policy

We get things wrong sometimes. When we do, we want to know, and we want the record to show it.

What triggers a correction:

How to flag one: Cite or correct this page (linked at the bottom of every editorial page). Tell us the page, the specific text or footnote, and what you think should change. If you’re flagging based on a primary source, include the citation.

Our SLA:

Where corrections are logged: we maintain a running corrections record in the repository (docs/corrections.md). It lists the date, affected page, what changed, and what triggered the change. The log is open-source-style and survives editorial turnover.

What we don’t do

The clearest way to describe a business model is sometimes to name the things it excludes. Caregiving Atlas does not, will not, and cannot:

Update cadence

Caregiving rules change. Medicaid eligibility thresholds adjust annually. State legislatures pass new statutes every session. Medicare AEP rules and Advantage plan availability reset each fall. To keep up:

Our affiliate policy, in plain English

Specifically:

Editorial independence

The reviewer signing off on a page does not get to edit our recommendations to favor their own services or jurisdiction. The advertiser whose product we discuss does not see the article before publication and does not have copy approval. When we change a recommendation, it’s because the evidence changed, not because a partnership did.

If we get something wrong, we want to know. Corrections are dated and visible at the top of the affected page. Cite or correct any page using the link at the bottom of every article.

What we are not

Caregiving Atlas is not:

Tooling that keeps us honest

A trust commitment is only as good as the mechanism that enforces it. We maintain a small set of automated checks inside the repository so editorial drift is caught at PR time, not after a reader writes in:

Why we publish this page

Most of the caregiving information online is published by businesses that profit from your decisions. Some of it is good; much of it is incentive-distorted in ways the reader can’t see. By laying out exactly how we work — including which decisions earn us money and which don’t — we let you weigh our guidance against our incentives directly.

We’re betting that a thoughtful 50-year-old caring for an 82-year-old parent at 11 p.m. on a Tuesday would rather read a sourced page from a brand that’s honest about its model than an unsourced page from a brand that isn’t. If we’ve done this right, you’ll notice the difference within a paragraph.