Birth Injury Marketing
Data · method

The hospitals that never say if they deliver

By Kevin Schwaner, Founder; Chief Marketing OfficerUpdated

Of 4,658 hospitals in the CMS maternal care file, 2,357 carry an affirmative indicator for inpatient labour and delivery and 1,330 affirmatively do not. The remaining 971 carry neither. Every county-level count of delivering hospitals, including ours, is therefore a floor rather than a census, and any map built from it understates coverage by an unknown amount.

What the file actually contains

Obstetric service indicator across the CMS maternal care file.
IndicatorHospitalsShare
Affirmative: provides inpatient labour and delivery2,35750.6%
Affirmative: does not1,33028.6%
No indicator either way97120.8%
Obstetric indicator, as a share of the file. Confirmed, confirmed-negative and unreported hospitals in the CMS maternal care file. published withheld or not availableObstetric indicator4,658 hospitals in the file2,3571,330971
Figure 1: a fifth of the file answers neither way. Any count of delivering hospitals built from it is a floor.

Why this matters to anyone quoting the number

Obstetric access reporting leans heavily on counts of delivering hospitals. A county with no confirmed delivering hospital is routinely described as an obstetric desert. Some of them are. Others contain a facility that delivers and has not populated the field.

The error runs one way. Because a blank is not an affirmative, every count of confirmed delivering hospitals is at or below the true number, and every count of counties without one is at or above the true number. Nobody can say by how much without going hospital by hospital, and we have not done that.

How we handle it, and what that costs.
ChoiceConsequence
Count only affirmative indicatorsDelivering-hospital counts are floors. Obstetric-desert counts are ceilings.
Report the unreported separately on every pageEvery state page names the number of hospitals with unknown status alongside the confirmed count.
Never imputeNo hospital is assigned a status it did not report, and no county figure is filled in from a neighbour.
How a blank field becomes a wrong map. An unreported obstetric indicator, if read as a negative, converts a delivering hospital into an obstetric desert. Field is blankno indicator either way Read as "does not deliver"county becomes a desert Read as "unknown"count published as a floor We do the secondand label it on every page
Figure 2: the same blank, two readings. Only one of them can be defended, and the other produces a more dramatic map.

What we are not claiming

This is not an accusation of bad data. The file is a directory with a service field, not a survey with a response obligation. A blank is a blank.

And it is not a reason to discard the file. It is the best national source that exists for this question. It just has to be quoted as a floor, which is how it appears everywhere on this site.

Clinical measures and legal measures are never joined on this site. Adverse-event rates describe medicine; litigation describes a legal process. Where both appear together it is to describe a market, not to suggest that one predicts the other.

Sources

  1. CMS Care Compare: Maternal Health · Hospital: PC-02, PC-07a, PC-07b, SM-7; 2024 reporting period
  2. CMS Care Compare: Footnote Crosswalk: definitions for suppression codes

Cite this analysis

Free to reproduce with attribution, including commercially. Charts may be embedded as published.

Birth Injury Marketing. “The hospitals that never say if they deliver.” 25 August 2026. https://birthinjurymarketing.com/insights/what-cms-does-not-tell-you-about-delivering-hospitals/

Journalists and researchers: the underlying data is already published as CSV at /data/, free and without conditions, no form. Need a different cut, by state, county, hospital or year? Ask, and I will also tell you what the data cannot support.

More

Data · analysis

Where births and legal capacity are furthest apart

Joining county births to Offices-of-Lawyers establishments produces a market map nobody has assembled, and a measure much weaker than it looks.

Data · analysis

Where the birth happened is not where the family lives

More than a thousand counties record births and have no hospital confirmed to deliver them, which puts residence and venue in different places.

Data · analysis

Where the births actually are, by county

The distribution of U.S. births is far more concentrated than most geographic media plans assume, and the tail is longer than it looks.

Practice economics

What a slow callback costs you here

A familiar marketing rule, worked through for a category where one inquiry can cost more than a month of another practice area.

Practice economics

The counties your advertising cannot reach, and what does

The counties with no lawyer office at all are not a media opportunity. They are a referral problem, and a different service line answers them.

Economics

Cost per lead can fall while cost per signed case rises

In birth injury both movements are real at the same time, and only one of them matters. Here is the mechanism, and how to tell whether it is happening to you.

Operations

Why more spend stops producing more birth injury cases

Signed case volume is limited by how many inquiries a firm can competently review, not by how many it receives. Above that ceiling, additional spend buys a queue.

All articles → · The data behind them → · The publications →

Authorship & review

Author
Kevin Schwaner, Founder; Chief Marketing Officer. Credentials.
Last reviewed
Corrections
Found an error? Tell me and it goes in the dated corrections log, whether or not it flatters me.