Every fertility clinic on earth advertises success, and almost none of them are lying — they're just choosing denominators, endpoints, and age bands with a marketer's eye. Per-transfer versus per-retrieval, pregnancy versus live birth, blended age bands, selected patient mixes: the arithmetic games are legal, standard, and entirely survivable once you know how to read the registries yourself. This guide teaches the skill — SART and CDC data for US clinics, HFEA for the UK, and the written-question method that substitutes for registries everywhere else, including abroad.
The Four Games and How to Catch Them
Game one: the endpoint. "Pregnancy rate" counts positive tests, including the meaningful fraction that end in early loss; "clinical pregnancy" counts heartbeats; only live birth counts babies. The gap between headline pregnancy and live-birth rates commonly runs 10–20 points, and clinics choosing the bigger number have already told you their communications policy. Game two: the denominator. Per-transfer rates exclude everyone whose cycle produced no transferable embryo — flattering for clinics that screen or select aggressively; per-retrieval and per-cycle-start rates include the whole funnel and read lower for identical quality. Neither is dishonest; comparing one clinic's per-transfer against another's per-retrieval is. Game three: the age band. Live-birth odds fall so steeply with age that any unbanded aggregate is noise — a clinic treating a younger mix will beat a better clinic treating an older mix on aggregates every time. Insist on your band, in every conversation, at every clinic, without exception. Game four: the patient mix. Clinics that decline hard cases post better numbers; clinics that accept them post honest ones — which is why an unusually high outlier rate should raise questions, not hopes, and why registry footnotes about patient selection are the most informative fine print in the field.
Reading SART and CDC Data (US)
US clinics report to SART (the professional society registry) and the CDC (mandated national reporting), both publicly searchable by clinic. The skill: use the live births per intended egg retrieval framing SART now emphasizes — it captures the whole funnel including freeze-all strategies — banded by age, and read the cumulative outcome tables showing odds across subsequent transfers from one retrieval. Check cycle volume (rates on 40 cycles bounce; rates on 900 mean something), note the percentage of cycles using donor eggs or PGT (both reshape numbers), and read the patient-mix characteristics both registries publish. What the registries can't do: rank clinics for you — they measure outcomes, not your fit — and both explicitly warn against league-table use. What they absolutely do: expose any gap between a clinic's marketing and its filings, which is the single fastest integrity test in American fertility medicine.
Reading HFEA Data (UK) — and the Registries' Limits
The UK's HFEA publishes clinic-level results with the most consumer-friendly presentation in the field — live-birth rates per embryo transferred, age-banded, with confidence intervals drawn so patients can see when apparent differences are statistical noise (most are). The HFEA's own guidance is the maturest sentence in success-rate literature: most UK clinics' results are statistically consistent with the national average, and choosing on service, location, cost, and trust is legitimate once a clinic clears the competence bar. That framing travels: everywhere on earth, the goal of reading success data isn't finding the magic clinic — it's excluding the poor ones and freeing yourself to choose among the competent on the factors that actually differ.
Why age bands are everything (illustrative registry-shaped rates)
Illustrative rates shaped on published registry patterns for own-egg blastocyst transfers — actual registry values vary by year and definition. The gradient is the point: any success number not banded for your age is marketing, not information.The Statistics That Protect You From Yourself
Two consumer-statistics habits complete the toolkit. Confidence intervals are not decoration: a 52% rate on 60 transfers and a 46% rate on 900 transfers are statistically indistinguishable, and the smaller program's edge is as likely noise as signal — HFEA draws the intervals for you; for SART-scale data, a rough rule serves: differences under five to eight points between mid-volume clinics rarely mean anything, and year-to-year swings at small programs mean even less. Chasing a two-point difference across an ocean is spreadsheet theater performed for an audience of one. Selection effects run both directions: a clinic with modest numbers may be the region's referral center for the hardest cases — the honest phrase for which is "better doctor, worse-looking table" — while a gleaming rate may be built on declining everyone over 38. The patient-mix disclosures, where registries publish them, convert this from suspicion to reading; where they don't, the written question's context-volunteering test does the same work.
And one framing correction that saves money and grief: success rates are a property of patients at clinics, not of clinics — your age, reserve, diagnosis, and embryo cohort travel with you to whichever program you choose, and they carry more of your personal probability than the clinic's aggregate does. The registry-reading skill exists to exclude bad programs and puncture inflated ones; it cannot find a clinic that transcends your biology, and marketing that implies otherwise has priced that implication in.
Volume, Access, and the Numbers Nobody Publishes
Two unlisted metrics belong in the same file as the rates. Volume is quality's quiet correlate — embryology is craft, craft compounds with repetition, and a program running a thousand cycles yearly has debugged failure modes a fifty-cycle program meets annually; volume appears in every registry and deserves reading alongside rates rather than after them. Access policy is the mirror-image disclosure: ask directly what patients the clinic declines (age caps, BMI cutoffs, reserve thresholds) — the answer simultaneously explains the published numbers and tells you whether you'll face the policy yourself at cycle two, when your own numbers have moved. Programs answer this plainly; the evasive answer is, as everywhere in this guide, itself the data.
The Written-Question Method — for Clinics Without Registries
Most of the world's clinics — including the international destinations our sister site ivfabroad.co maps — report to no public registry, which moves the burden to a structured written question: "For patients in my age band using [own/donor] eggs, what is your live birth rate per embryo transfer, and your cumulative live birth rate per retrieval, over the last two full years?" Every serious program tracks these internally; producing them in writing is trivial for the honest and revealing for the rest. Grade the answer on three axes: did they answer the actual question (live birth, your band, stated denominator), did they volunteer context (volume, patient mix, definitions), and did they resist the urge to convert the answer into a promise. Our live guide on how clinics abroad report rates extends this method, and the clinic-metrics guide covers the non-statistical vetting that surrounds it.
Donor-egg programs get one special note: because donor age replaces recipient age as the driver, donor-cycle success rates cluster high and differentiate clinics poorly — lab quality and matching practice matter, but a 60% donor-cycle rate is the category norm, not a distinction. Read donor numbers against donor benchmarks, not own-egg ones.
The Bottom Line
Success-rate literacy is one afternoon's work with a permanent payoff: learn the four games, pull the registry data where it exists, run the written question where it doesn't, and weight cumulative-per-retrieval numbers in your own band above everything a homepage says. Then remember the HFEA's grown-up truth — most competent clinics cluster, statistically — and spend your remaining diligence on the things that genuinely vary: laboratory leadership, communication quality, cost structure, and whether the physician answering your questions treats them as welcome. The number-reading gets you past the marketing; the fit — laboratory, communication, cost, trust — decides the rest, and it deserves the majority of your remaining attention.
This site covers what — protocols, medications, add-ons, and what the evidence says about treatment itself. For where — destinations, donor laws, costs, and trip logistics, country by country — our sister site covers the map.
Compare destinations at ivfabroad.co →