How we read the data
Every number on this site comes from one federal file. This page says exactly what we do to it, because a directory that will not show its working is just an opinion with a table.
The source
The CDC Assisted Reproductive Technology (ART) Success Rates, 2022 final release, published on data.cdc.gov. US fertility clinics are required to report their outcomes, so this is a census rather than a survey: 457 clinics across 51 states and territories, and 379,224 individual reported cells.
The headline figure
The number we lead with on every clinic page is precisely this: live births per embryo transfer, donor eggs, frozen embryo transfer, all patients, 2022. Nationally that is 45.1%.
We picked one measure rather than blending several. The CDC splits donor results by what was transferred - fresh eggs, frozen eggs, frozen embryos, donated embryos - and there is no per-cell denominator in the release, so any weighted composite across those would be invented arithmetic dressed up as rigour. Frozen embryo transfer is reported by 300 of 457 clinics, against 29 for fresh eggs, which makes it both the most comparable and the most representative single figure in current practice.
Three things in this file that produce false numbers
Each of these was found by reading rows. Each one silently breaks a site that misses it.
- Suppressed cells are shipped as 0.0. 234,039 of the 379,224 rows - 62% - carry a footnote explaining that the cell was withheld, while the value column reads zero. Published naively, most of this site would be fabricated zeros. We mark these and render “Not reported”. A suppressed cell is not a zero.
- Small-sample cells are fractions, and the numeric column lies. Where a cell has fewer than twenty transfers, the CDC writes the value as a fraction such as
"5/13"and sets the numeric column to 0.0. Trusting the numeric column publishes a 0% live birth rate for a clinic that achieved 5 births from 13 transfers. This affects 34,982 rows across 79 clinics. We parse the fraction and show it as “5 of 13 transfers” - never as a percentage, because a percentage from thirteen implies precision that is not there. - Clinic 9999 is not a clinic. It is the CDC’s own national aggregate. Left in a list it tops every ranking. We exclude it everywhere and use it only as the benchmark each clinic is compared against.
Eight clinics genuinely reported zero live births from between five and eight transfers. That is real data and we show it as “0 of 5 transfers”. It is not a 0% success rate and we will not display it as one.
What we deliberately do not do
- We do not rank clinics by percentage as the default. The default sort is transfer volume, because a high rate on a handful of cycles is noise.
- We do not adjust for case mix. Clinics with strict acceptance criteria report better numbers; the CDC data cannot correct for that and neither can we, so we say so on every relevant page rather than pretending the comparison is clean.
- We do not include 'Donated Embryos' in the donor-egg headline. That is embryo donation, which is a different thing from egg donation, and it is labelled as such in the detail.
- We do not accept payment to change, soften, reorder or remove any figure. There is no arrangement under which that is possible.
Corrections
If you run a clinic and believe something here misrepresents you, write to hello@donateegg.com. We will check it against the federal file the same week. Where the file itself is wrong we will say so on the page; we will not quietly overwrite the public record with a clinic’s own figures.
Ingest and derivation code is versioned. The data is refreshed when the CDC publishes a new final release, which is annual and runs roughly two years behind.