ART Registries and Codes of Practice: A Reference for Patients and Clinicians

ART registries and codes of practice are the two accountability instruments that govern assisted reproductive technology worldwide. A code of practice is the rulebook a fertility clinic must follow to operate legally in a given jurisdiction. A registry is the dataset of cycle outcomes those clinics are required to submit. Together, they define what fertility clinics must do and what they must report. This guide covers the major national instruments: HFEA (UK), RTAC (Australia/New Zealand), CDC (US), ANZARD, Q-IVF (Sweden), DIR (Germany), CARTR Plus (Canada), JSOG (Japan), REDLARA (Latin America), ESHRE EIM (Europe), and ICMART (global). What each measures, and what they miss.

What Are ART Registries and Codes of Practice?

When a fertility clinic reports a "success rate," that number comes from somewhere. Most national governments and professional bodies now require fertility clinics to submit their treatment data to a central registry. Some also require clinics to follow a legally binding code of practice that governs how they operate. Together, these two instruments form the accountability infrastructure of assisted reproductive technology (ART) worldwide.

A code of practice is the rulebook. It defines who is qualified to work in a fertility laboratory, what equipment standards a clinic must meet, how embryos must be handled and stored, what data must be collected, and what limits apply to treatments like donor insemination or multiple embryo transfer. Codes vary widely by jurisdiction. In some countries they carry statutory force. In others they are professional consensus documents with no enforcement mechanism.

A registry is the dataset. Clinics submit cycle-level records: how many oocytes were retrieved, how many embryos were transferred, whether a clinical pregnancy resulted, and whether a live birth occurred. Registries aggregate this into national summaries. They are the source of the "success rates" that appear in clinic marketing, media coverage, and patient decision-making.

Patients should care about both. Codes of practice tell you what a clinic is legally required to do. Registries tell you whether the aggregate outcomes across an entire country or region justify the clinical decisions being made.

But both instruments have limits. Sweden's Q-IVF and Australia and New Zealand's ANZARD do link successive cycles for the same patient and report some cumulative outcomes. Most other public-facing registry outputs do not. They count what happens in a given treatment episode rather than following a couple from first consultation to eventual outcome, whether that outcome is a live birth, a decision to stop treatment, or an undiagnosed condition that was never addressed. They report what technology produced. They do not report out-of-pocket cost, journey length, or whether the underlying cause of infertility was ever diagnosed.

Restorative Reproductive Medicine asks different questions. RRM clinicians start from diagnosis: what is causing this couple's infertility, and is there a treatable condition? Registry data is necessary for transparency in a field that handles profound decisions. But it is not sufficient for evaluating whether patients are being served well. Understanding what registries do and do not measure is the first step toward asking better questions.

How Registries Differ, and Why It Matters for the Numbers

Before reading the statistics in the rest of this article, it helps to know how the major ART registries differ on what they collect, what they omit, who is required to participate, who chose to, and which patient population reaches their data. The figures HFEA publishes about UK ART are not directly comparable to JSOG's figures about Japanese ART, because the two systems measure different patient populations under different governance, with different funding capture, and with different definitions of who counts as a tracked patient. Apparent per-cycle live birth rates can diverge between countries by 10 percentage points or more not because clinical outcomes differ, but because the underlying registry frame differs. The cards below establish the comparative lens; the per-registry detail sections that follow build on it.

Different registries operate under different mandate architectures. HFEA is a statutory licensing register created by the 1990 Act; CDC NASS is mandatory federal surveillance under the FCSRCA; ANZARD is accreditation-conditional, with submission tied to RTAC accreditation that clinics need for Medicare rebate eligibility; JSOG operates mandatory professional-society reporting for its members; DIR, CARTR Plus, and the parallel US SART CORS are voluntary professional-society databases; Q-IVF is a voluntary national quality register with near-universal Swedish coverage; ESHRE EIM and REDLARA are voluntary multi-country aggregations; and ICMART is the global meta-registry that aggregates national submissions. Coverage is not a clean function of mandate: Q-IVF's voluntary structure achieves near-universal Swedish participation, REDLARA reports membership of more than 90 percent of regional centres, and DIR's 2024 Jahrbuch covers 139 voluntary member centres. Registries that capture both publicly funded and private self-pay cycles describe the national ART universe; registries that capture only private self-pay describe only a subset, and the country's average per-cycle rate can shift several percentage points depending on which population reaches the data. Patient-linked cumulative reporting (Q-IVF, ANZARD) allows cumulative-per-patient rates to be computed; cycle-level-only reporting does not. Publication lag determines how stale the published figures are when patients encounter them in clinic marketing materials. The single most common interpretive error in patient-facing IVF reporting is treating cycle-level rates from one frame as comparable to patient-level rates from another. The frames are not interchangeable.

National Codes of Practice

HFEA

Human Fertilisation and Embryology Authority (United Kingdom)

RTAC

Reproductive Technology Accreditation Committee (Australia and New Zealand)

ANZARD

Australia and New Zealand Assisted Reproduction Database

Q-IVF

Swedish National Quality Registry for Assisted Reproduction

CARTR Plus

Canadian Assisted Reproductive Technologies Register Plus

DIR

Deutsches IVF-Register (Germany)

JSOG

Japan Society of Obstetrics and Gynecology

ESHRE EIM

European Society of Human Reproduction and Embryology, European IVF Monitoring Consortium

ICMART

International Committee for Monitoring Assisted Reproductive Technologies

REDLARA / RLA

Red Latinoamericana de Reproduccion Asistida / Registro Latinoamericano de Reproduccion Asistida

What a Code of Practice Does

A code of practice defines the minimum standards a fertility clinic must meet to operate legally in a given jurisdiction. The content varies, but most codes address the same core concerns: staff qualifications, laboratory standards, infection control, data collection and submission requirements, consent procedures, limits on embryo transfer numbers, donor anonymity rules, and the handling of unused embryos. The scope and legal force of these codes differ substantially by country.

United Kingdom: HFEA Code of Practice (9th Edition, v9.4)

The Human Fertilisation and Embryology Authority (HFEA) Code of Practice is among the most comprehensive fertility regulatory codes currently in force. It is legally binding under the Human Fertilisation and Embryology Act 1990 and the 2008 amendment. Any UK clinic that licenses embryo creation, storage, or transfer must comply with it. The code runs to several hundred pages and is updated iteratively; the 9th edition was launched in January 2019 and is currently in force as version 9.4 from October 2023, reflecting updates across consent, welfare-of-the-child assessments, gamete storage time limits, and data submission standards.

The code says nothing about why a patient needed treatment. It governs the conduct of the procedure. It does not require a clinic to document whether an underlying diagnosis was ever sought, offered, or declined. Clinics can fully comply with the 9th edition while never investigating the reproductive health condition that brought a couple through their door.

Australia and New Zealand: RTAC Code of Practice (2024)

The Reproductive Technology Accreditation Committee (RTAC) Code of Practice governs fertility clinics in Australia and New Zealand under the Fertility Society of Australia and New Zealand (FSANZ). Unlike the HFEA framework, RTAC accreditation is a professional scheme, not statutory law. Its practical force is Medicare eligibility: RTAC accreditation is required for clinics to receive Medicare rebates. The 2024 edition covers clinical governance, laboratory quality, and patient safety. Data submission to ANZARD is a condition of accreditation.

United States: No Single National Code

The United States has no national code of practice for ART. The federal Fertility Clinic Success Rate and Certification Act (FCSRCA) of 1992 requires clinics to submit data to the CDC but establishes no clinical practice standards or enforcement mechanisms. FCSRCA is the only federal US statute on ART data collection. Major professional societies publish practice guidance that carries no legal force. State regulations vary widely. A US couple evaluating a fertility clinic cannot rely on a national regulatory standard to define what that clinic is required to do before recommending treatment.

Historical Note: UK Multi-Centre Self-Audit (1993)

In the early 1990s, the Medical Research Council and the Royal College of Obstetricians and Gynaecologists jointly examined outcomes from IVF programmes across multiple UK centres. Published in 1993, this was an early effort at systematic self-audit. It sat alongside the statutory HFEA register, which had been established by the 1990 Act and was already collecting cycle-level data through HFEA licensing from 1991. The audit work helped normalize the transparency culture that the statutory register operationalized.

National Registries

What the Registries Cover

Each major national registry collects cycle-level data from accredited fertility clinics. Most report annually, with a two-to-three-year lag between treatment year and publication. The figures below reflect the most current data in the RRM Academy research library.

United Kingdom: HFEA Register

The HFEA operates both the regulatory code (above) and the national data register. Clinic-by-clinic outcome data is searchable on the HFEA website. Annual trend reports are published covering all licensed treatments in the UK. The HFEA register is among the most accessible national datasets globally, with patient-facing tools that allow clinic comparison.

What it measures well: cycle counts by age band, live birth rate per embryo transfer, multiple birth rates, and trends in single embryo transfer adoption. What it largely does not report: cumulative outcomes per patient over multiple cycles, patient drop-out, or time elapsed from first consultation to live birth.

Sweden: Q-IVF (Annual Since 2007)

Sweden's Q-IVF National Quality Registry for Assisted Reproduction is one of the methodologically strongest national registries in the world. It publishes annual reports covering treatment outcomes from the prior year; the most recent report in the RRM Academy library covers treatment year 2023 (Q-IVF et al., 2025).

Q-IVF is notable for its attention to cumulative outcomes. Sweden was among the first countries to routinely track and publish patient-level cumulative live birth rates across multiple treatment cycles, not only single-cycle results. This makes Swedish registry data particularly valuable for comparing with restorative cumulative outcome data.

Japan: JSOG ART Summary

The Japan Society of Obstetrics and Gynecology (JSOG) Ethics Committee publishes an annual ART summary. Japan operates one of the highest-volume ART programmes globally, performing over 450,000 cycles per year in recent reports. The most recent JSOG summary in the RRM Academy library covers treatment year 2023.

Japanese registry data is strong on cycle volume and age stratification. It documents, among other metrics, the extraordinary scale of frozen embryo transfer cycles in Japan, which now far exceed fresh transfers. As with most registries, patient-level cumulative data and drop-out rates are not the primary reporting focus.

Australia and New Zealand: ANZARD

The Australia and New Zealand Assisted Reproduction Database (ANZARD), maintained by the National Perinatal Epidemiology and Statistics Unit at UNSW Sydney, publishes annual reports. The most recent in the RRM Academy library covers treatment year 2023.

ANZARD reports clinic-level and national aggregate data on IVF, ICSI, frozen embryo transfer, and donor cycles. It is notable for its reporting on multiple birth rates and its tracking of the shift toward elective single embryo transfer, which Australian national policy has strongly encouraged.

Canada: CARTR Plus

The Canadian Assisted Reproductive Technologies Register Plus (CARTR Plus), operated by the Canadian Fertility and Andrology Society (CFAS), publishes annual reports. The most recent in the RRM Academy library covers treatment year 2024.

Canadian ART regulation rests on a federal-plus-provincial split. The federal Assisted Human Reproduction Act (2004) prohibits human cloning, commercial surrogacy, and several other practices; Health Canada enforces sperm and ovum safety provisions (section 10, in force February 2020) and reimbursement rules (section 12, in force June 2020). Most clinical-practice regulation sits with the provinces following the 2010 Supreme Court of Canada decision in Reference re Assisted Human Reproduction Act, which narrowed federal jurisdiction over medical practice. CARTR Plus is a voluntary registry operated by the Canadian Fertility and Andrology Society; participation is broad among accredited clinics but not legally mandated.

Germany: DIR Jahrbuch

The Deutsches IVF-Register (DIR) publishes an annual Jahrbuch (yearbook) drawing on data from German fertility centres that are DIR members. The 2024 edition covers 139 member centres. The most recent report in the RRM Academy library is the DIR Jahrbuch 2024, published in 2025. DIR also publishes a patient-facing Sonderausgabe (patient edition) of its annual data.

German registry data is particularly useful for its detailed reporting on embryo culture methods, freeze-all protocols, and elective single embryo transfer trends across a large, mature national programme.

Latin America: RLA / REDLARA

The Red Latinoamericana de Reproduccion Asistida (REDLARA) publishes the Latin American Registry of Assisted Reproduction (RLA). The most recent report in the RRM Academy library covers treatment year 2021.

REDLARA aggregates data from member clinics across Latin America and reports membership of more than 90 percent of centres performing ART in the region. Participation is voluntary, but uptake is high. National regulatory frameworks vary widely, which makes REDLARA's regional aggregation one of the few sources of cross-border comparative data for Latin American ART.

United States: CDC ART National Summary

The Centers for Disease Control and Prevention (CDC) publishes annual ART Fertility Clinic and National Summary Reports under the FCSRCA mandate. The CDC report covers both clinic-specific outcomes and national aggregate data. Publication is typically two to three years behind the treatment year. The most recent report in the RRM Academy library covers treatment year 2019.

The CDC report is the only mandatory national ART data collection in the United States. It provides clinic-by-clinic live birth rates by age group, a powerful consumer transparency tool. It does not provide cumulative per-patient outcomes, out-of-pocket cost data, or diagnosis information.

Europe-Wide: ESHRE EIM

The ESHRE European IVF Monitoring (EIM) Consortium aggregates national registry data from across Europe and publishes annual reports in Human Reproduction. The RRM Academy research library indexes the EIM series back to its treatment-year 2013 analysis, with the most recent being De Geyter et al. (2025), reporting on ART activity in Europe in treatment year 2020. ESHRE EIM is a meta-registry: it aggregates what national registries submit rather than collecting clinic data directly, so it inherits each contributing registry's methodological constraints.

Global: ICMART World Report

ICMART publishes world reports on global ART practice, drawing on national registry data from dozens of countries. Publication lags several years due to the complexity of international data aggregation.

What the Registries Measure Well, and What They Miss

ART registries measure one thing well: what happens during a funded treatment cycle. They count retrievals, transfers, and live births, aggregate that data by age band, and let patients compare clinics. That is useful. It is also almost exactly where the usefulness ends.

The registries' real structural function is to serve the industry that populates them. Independent analyses of US national ART data have documented this dynamic from inside the field: Gleicher and colleagues (2019, Human Reproduction Open) traced a worldwide decline in IVF birth rates from a 2001-2002 peak and attributed part of the trajectory to clinic practice changes the registry framework continued to count favourably, and Kushnir and colleagues (2018, BMJ Open) showed that US national ART live birth rates vary considerably across stimulation protocols in ways the headline registry rates do not surface. The data categories registries collect are the same categories that make IVF look as effective as possible. The measures that would reveal costs, attrition, long-term harm, and diagnostic shortcuts are absent.

What the registries deliberately do not measure

Cumulative live birth rate per patient. The most important question for a couple is not whether a single cycle produced a live birth. It is whether they ever reach a live birth across however many cycles they can sustain. Most major registries do not track this. Malizia et al. (2009, NEJM) followed 6,164 patients through 14,248 cycles at a single large US center. After six cycles, the cumulative live birth rate was 72% under optimistic assumptions (non-returners perform as well as continuers) and 51% under conservative assumptions (non-returners achieve no live births). A 21 percentage-point gap driven entirely by what happens to patients who stop. Age widens and narrows that gap substantially: women under 35 reached 86% optimistic and 65% conservative; women 40 and older reached 42% and 23%, respectively. Smith et al. (2015, JAMA) analyzed 156,947 UK women across 257,398 IVF cycles. The first-cycle live birth rate was 29.5%. The prognosis-adjusted cumulative live birth rate reached 65.3% by the sixth cycle. That gap between one cycle and six cycles reflects the treatment course many couples actually require, not an edge case. ESHRE treats cumulative live birth rate as a primary effectiveness outcome in its ART guidance, which the registry response to per-cycle reporting has been slower to match. The per-cycle headline persists in clinic marketing: Wilkinson and colleagues (2017, BMJ Open), in a systematic review of UK private and NHS fertility-clinic websites, documented how direct-to-consumer success-rate advertising on clinic pages diverged from the HFEA registry data the same clinics were required to submit.

Patient drop-out. Registries do not capture couples who stop. They count only those who returned. The attrition is documented and substantial in specific cohorts. Sharma et al. (2002, Fertil Steril) analyzed 2,056 patients undergoing 2,708 IVF cycles at a single UK center and found that 64% of patients did not continue after the first attempt. Continuation was more likely in women 35 or younger (39% continued) than in women over 35 (27% continued), and more likely when five or more oocytes were retrieved (40% versus 23% when fewer were obtained). The predictors of dropout are the same predictors of poor prognosis. Pearson et al. (2009, Fertil Steril) found, in 2,245 women, that the point at which the first cycle failed predicted discontinuation: women whose first cycle ended in a chemical pregnancy or early miscarriage were more likely to stop entirely than those whose cycle ended before embryo transfer. The emotional weight of a pregnancy that failed, not a cycle that simply did not work, is a documented driver of attrition. Neither registry captures this. What registries report as a completion rate is a selection artifact. The success figures that reach patients assume the dropout never happened.

Out-of-pocket cost per live birth. No major national registry publishes cost data alongside outcome data. CDC national data show live birth rates of approximately 37 to 38 percent per ART cycle, with substantial variation by age. At $12,000 to $15,000 per cycle in a self-pay market, the arithmetic for couples requiring multiple cycles reaches $30,000 to $45,000 before medication costs and failed frozen transfers. Chambers et al. (2014, Fertil Steril) analyzed ART utilization across multiple countries and found that consumer cost as a proportion of disposable income independently predicts both access and the number of embryos transferred per cycle: a 1 percentage-point drop in cost relative to income predicts a 3.2% increase in utilization. Countries where ART is more expensive relative to income transfer more embryos per cycle, accepting higher multiple-pregnancy risk because cycles are rationed. No registry pairs this cost context with the success figures it publishes.

Adverse outcomes. De Geyter et al. (2025, Human Reproduction), reporting ESHRE EIM data on 923,318 ART cycles across 41 European countries in 2020, found that 13.1% of ART singleton deliveries were preterm (before 37 weeks), with 2.4% very preterm (before 32 weeks) and 0.8% extremely preterm (before 28 weeks). Singleton pregnancies are the intended outcome of single-embryo transfer policy, so their preterm rate is not an artifact of twin or triplet gestations. OHSS affects 1 to 5 percent of fresh cycles and up to 10 percent of high responders; late-onset cases after discharge are undercaptured in registry data. Hart and Wijs (2022, Frontiers in Reproductive Health) review longer-term outcomes in ART-conceived offspring from childhood to adolescence across cardiovascular, metabolic, neurodevelopmental, and psychological dimensions; the major ART registries themselves do not maintain longitudinal follow-up of offspring beyond perinatal records.

Whether the underlying cause of infertility was ever diagnosed. A registry record shows a cycle was initiated and an embryo was transferred. It does not show whether a systematic diagnostic workup preceded the transfer. The 2023 ESHRE evidence-based guideline on unexplained infertility defines the diagnostic category that captures couples proceeding to treatment without an identified specific cause; it is itself a recognised and substantial proportion of the ART population. By contrast, Sanchez-Mendez et al. (2025) found that 1,310 couples entering a NaProTechnology program received a mean of 2.5 diagnoses per couple. The ability to identify specific etiological factors in nearly all couples reflects a diagnostic approach the registry record cannot capture. The registry records the cycle. Whether the underlying condition was ever found is not part of the dataset.

What the registries do measure, and what that measurement actually means

Clinic-level transparency. Statutory registries operated by regulators (HFEA in the United Kingdom, CDC under FCSRCA in the United States, RTAC-linked ANZARD in Australia and New Zealand) require clinics to submit outcome data that is then published. Society-run registries (CARTR Plus in Canada, REDLARA in Latin America, JSOG in Japan) operate as voluntary professional schemes. Before registries existed, success claims were unaudited. That ended. What registries did not end is patient selection bias and the absence of independent verification: society-run registries operate without external audit, governed by the same societies whose member clinics fund them. CDC includes medical-director attestation and Westat validation visits but the audit is not independent of the reporting infrastructure. The data are real. The governance of the data is partly self-referential.

Age stratification. All major registries report outcomes by age band. This matters clinically, because ovarian reserve and egg quality decline in ways that affect per-cycle rates substantially. Age stratification makes the per-cycle figure more interpretable. It does not address dropout, cost, harm underreporting, or diagnostic bypass.

Initiated cycle reporting. Most registries count all initiated cycles, including cancellations before retrieval. This is more conservative than counting only completed transfers. A clinic reporting 30% live birth per initiated cycle includes its cancellations in that number. Technically accurate by registry standards. Whether it informs the patient decision it is used to support is a different question.

Why this matters for couples

The figures couples encounter when evaluating fertility care come almost entirely from registry infrastructure. Those figures are real. They are also systematically incomplete in ways that favor the appearance of efficacy over the reality of the patient journey. A couple working from registry-published figures alone cannot determine how many couples who started the same protocol eventually stopped, what the total cost to live birth was, whether an underlying condition was ever diagnosed, or what adverse events occurred along the way. The registry was not designed to answer those questions.

How RRM Academy Uses These Documents

RRM Academy maintains these registry and code-of-practice records in the research library because primary sources belong to patients and clinicians, not to the institutions that generated them.

RRM Academy treats these registries as the most favorable possible self-portrait of the IVF industry. That is not rhetorical. It is methodologically accurate. Registries report what the industry chose to measure, in the metrics the industry selected, verified by the professional societies that clinic membership fees support. They exclude attrition. They exclude cost. They exclude long-term harms. They exclude diagnostic bypass. What they publish is the best case the industry has made about itself, under conditions of self-governance. That upper bound is where the comparison work begins.

Registry data supplies the cycle-level denominators, age-stratified benchmarks, and per-transfer rates the industry uses to represent its own efficacy. The IVF Clinical Ledger methodology treats those numbers as that upper bound, then asks what happens when attrition, cost, and diagnosis are added back. Boyle et al. (2025, Journal of Restorative Reproductive Medicine) conducted a retrospective clinic-based analysis of RRM treatment outcomes referenced against publicly available IVF registry data for 2019, reporting a 41% crude live birth rate in the RRM group. That comparison only becomes analytically sound when the IVF benchmark is well-characterized, which is why HFEA, Q-IVF, ANZARD, and CARTR Plus reports are indexed here. Sanchez-Mendez et al. (2025), in a cohort of 1,310 couples followed at a NaProTechnology centre in Spain over five years, reported a crude take-home baby rate of 35.3% and an adjusted cumulative rate of 62.1% (95% CI 58.8-65.4) when accounting for duration of active participation in the program. Age-stratified adjusted rates ranged from 83.7% in women under 30 to 24.4% in women over 40. Registry data does not produce patient-level cumulative outcomes across a multi-year program course. But registry per-cycle rates, age-stratified and adjusted for realistic dropout, can be used to estimate what cumulative IVF outcomes look like over the same time horizon. The honest comparison requires that adjustment.

The deeper methodological treatment of this question is in progress. RRM Academy's ART registry comparison research is developing a patient-level cost-outcome accounting framework that registries do not currently provide: cumulative cost per live birth, attrition-adjusted patient-level outcomes, and the diagnostic context registries do not record. The registry documents indexed here are the foundational sources it builds on.

All Registry and Code Records

The 14 canonical instruments in this guide are indexed in the RRM Academy Research Library. Each entry links to the primary source record.

Frequently Asked Questions

This content is for educational and reference purposes only and does not constitute medical advice. Registry and code-of-practice information reflects publicly available primary sources at the time of writing. For clinical questions about evaluating fertility care options, consult an RRM clinician or qualified healthcare provider.