How we use sources, and what we left out

What counts as a source

A source needs four things: a publisher, a title, a year and a working link.

Preferred, in order. Professional society guidance, such as ACOG practice bulletins and AAP clinical reports. Government agencies, such as CDC, NIH, CMS and the Department of Education. Systematic reviews, particularly Cochrane. The original papers behind staging systems and grading scales. Federal regulations, quoted by section.

Not used. Other websites summarising the above. Marketing content. Anything that does not publish the underlying material.

Every citation on a page resolves to an entry in all sources, and the build fails if a citation points to a source that does not exist.

How numbers are handled

Nothing is adjusted. No figure is inflation-adjusted, averaged across sources, extrapolated or rounded by this site. A number appears the way its source published it.

The year is always stated. The CDC lifetime cost estimates are in 2003 dollars. Every page that uses them says so, every time, with three caveats attached. They are old. They are averages across all severities. And they count the individual's lost productivity but not a parent leaving work. ³

Figures that change annually are not restated. The SSI payment amount and the ABLE contribution limits change every year, so this site points to the agency that publishes the current figure rather than printing one that will be wrong. ^{4 5}

Where an agency has stopped publishing something, the site states that as a fact, without editorial comment about the agency.

What was left off the site

These are the things we wanted and could not source. Nothing here was guessed at and nothing here was published.

Filing deadlines for 41 of 51 jurisdictions. We printed a statute citation only where we could tie it to that state's own published code. For the rest, the state hub and the lookup tool say so plainly. This is the largest gap on the site. It is also the most consequential, because a missed deadline is permanent. ²

Per-state medical records fee caps and response deadlines. Federal rules apply everywhere and are on every relevant page. Thirty days, one 30 day extension, a reasonable cost-based fee, and no search or retrieval fee. ⁶ Some states set tighter limits. We could not verify all 51 against a primary source, so no per-state table is published.

Per-state births, cesarean rates and preterm rates. We did not republish 51 sets of figures we could not verify. Instead every state hub links to March of Dimes PeriStats and CDC WONDER. There you can pull the current number with its own source and year attached. ^{7 8}

Lifetime cost by condition and severity. No agency publishes it. The cost tool shows the CDC population anchors with their caveats. It then lists the cost drivers a life care planner would work through, rather than inventing a severity band. ³

Incidence figures for several conditions. These include kernicterus, newborn facial nerve palsy and newborn spinal cord injury. Where a rate could not be tied to a primary source, the page says the condition is uncommon or rare in words, rather than printing a number.

Hospital-level maternity data. Not mirrored. It would go stale silently, and neither CMS nor Leapfrog publishes a stable open feed of maternity measures. The hospital tool links to both and explains how to read a missing result. ^{9 10}

A GMFCS level distribution. We could not source a reliable breakdown of how children with cerebral palsy are distributed across the five levels, so none is published.

Numbers published but flagged for re-verification

A smaller category. These appear on the site with a citation, and they should be re-opened against the primary document before the site is treated as settled.

National vital statistics figures. These include the United States cesarean delivery rate and the preterm birth rate. They are cited to the National Vital Statistics System reports, which is the right source. They should be checked against the current annual report, because they move each year. ^{11 12}

Incidence figures drawn from single studies. These include the reported rate of brachial plexus birth injury and of persistent pulmonary hypertension. They are cited to the study that reported them, with the year. Read them as that study's finding, not as a current national rate. ^{13 14}

Flagging these publicly is the point. A site that lists only what it left out, and never what it is less sure of, tells you half the story.

If you find an error

Tell us. The corrections page explains how, and what happens next.

Corrections are published rather than quietly edited, because a page that silently changes is a page you cannot check.