Quality and Methodology Report 2021

1. Methodology background

Official statistic	No
Frequency	Annual
How compiled	Submitted by all HFEA licensed fertility clinics
Geographic coverage	United Kingdom (UK)
Register size	Approximately 100,000 cycles per annum
Document last revised	13 December 2023

2. About this quality and methodology report

This quality and methodology document contains information on the quality characteristics of the data as well as the methods used to create it. The information in this document will help you to:

• understand the strengths and limitations of the data

• learn about existing uses and users of the data

• reduce the risk of misusing data

• help you to decide suitable uses for the data

• understand the methods used to create the data

3. Important points

3.1. All licensed fertility clinics in the UK are required by law to provide information to the HFEA about treatments they carry out and their outcomes. We hold this information on our Register, which contains information about fertility patients, partners, donors, treatment and children born as a result of these treatments.

3.2. The information that we publish is a snapshot of Register data as at 19 April 2023. Results are published according to the year in which the cycle was started unless otherwise stated.

3.3. By this date, clinics were required to have submitted all records of 2021 treatments and outcomes, including births. However, due to the pressures of COVID-19 on clinics and a large-scale data migration project for our data Register, treatments and pregnancies from 2020-2021 and birth outcomes from 2019-2021 remain unvalidated. Data validation involves quality checking data submitted from licensed clinics across the UK to verify treatment, pregnancy and birth outcome data is correctly recorded on our data Register. Data validation is expected to be completed in summer 2024.

3.4. As our database consists of a live Register and we rely on accurate reporting from clinics, information may be subject to change and the figures published here may differ slightly to those published before or in the future.

3.5. This publication reports national data trends over time. Clinic-specific data is published on our website’s clinic search tool, Choose a Fertility Clinic, as well as in our State of the Sector reports.

4. Quality summary

We make information from the Register available to patients, clinic staff, media, researchers and the public to show how fertility treatment has changed over time. We publish annually data from our Register on treatment and outcomes, covering treatment type, age group, partner type, ethnicity, donor information, funding status, English region and UK nations. The data published is a snapshot of our Register data submitted by clinics and is subject to change.

5. Quality characteristics of UK Register data

5.1. Relevance

The Human Fertilisation and Embryology Act 1990 made it a legal requirement for every licensed fertility clinic in the UK to submit information on all fertility treatments and children born as a result to the HFEA. The HFEA holds a Register of these treatments since 1991, which is believed to be the largest database of fertility treatment in the world. We produce this statistical release on an annual basis with an underlying dataset. We do not hold detailed information on partner intrauterine insemination (IUI), surrogacy or treatments that takes place outside of licensed clinics which are not regulated by the HFEA.

5.2. Accuracy and reliability

We typically report on cycle level information rather than patient level information. We make an estimate of the number of patients but owing to multiple and sometimes inconsistent registrations at different clinics, cycles are the default reporting method.

For a typical cycle, clinics will initially report patient registration, cycle-related information and early outcomes (including pregnancies) in line with General Direction 0005. Clinics then have 12 months to subsequently report the outcome of cycles for which they reported a pregnancy.

The Register undergoes regular validation – this was last completed in December 2020. Treatment cycles performed on or before 31 December 2019 have been verified along with their pregnancy status. Live births have been verified for treatments that took place on or before 31 December 2018. One centre has been removed from the analysis due to data quality issues. Additionally, data on Northern Ireland is excluded when comparing information by nation, this is due to one clinic accounting for 60% of annual cycles in Northern Ireland experiencing data reporting errors. Data from Northern Ireland in 2020/21 has been excluded throughout the HFEA dashboard for this reason. Results are published according to the year in which the cycle was started. Three clinics are known as clinics that are not yet submitting via the new data submission system. These three clinics accounted for around 1,350 cycles in 2019.

Treatment data and pregnancy rates for 2020-2021 and birth rates for 2019-2021 are preliminary due to delayed data validation and are likely to be underestimated as a result of missing data. A manual data quality exercise was carried out between January-March 2023 to explore the preliminary data and compare the data to previously validated data to determine the extent that this may be underestimated. It is estimated that birth outcomes in 2021 are likely at least 20% below the true value (estimated based on 2018 validated birth rates). Pregnancy values are expected to be more representative for reported treatments. Birth rates are affected by missing birth outcome data in recent years, however we have increased the reliability of birth rates provided in our publications by removing cycles where an embryo transfer occurred but no outcome data was recorded (e.g., embryo transfer with pregnancy data but no birth data). Birth rates per embryo transferred are typically three percentage points lower than the comparable pregnancy rates following data validation.

In the HFEA dashboard, a small number of cycles have been excluded due to data quality issues. Cycles have been removed where: sperm/egg source is unclear, pregnancy or births were recorded but embryo transfer data was missing, partner type was not recorded, and a very small number of cycles were excluded due to errors in clinic cycle information. The data provided in the HFEA dashboard may not match those provided in publications due to filters and suppression used. For example, IVF cycles in the HFEA dashboard will be slightly higher than in publications as we have not restricted data to only include cycles begun with the intention of immediate treatment.

5.2.1. New data Register

Data used to create publications comes from our Register data. A large-scale data migration project provided an opportunity to ensure that data is reported more accurately and reliably. As a result, some data may differ from previous publications. These changes are applied not just to the latest year, but to all previous years, ensuring comparability. The key changes are outlined below:

1. Partner type. This refers to any partners of a patient undergoing treatment. In previous publications, cycles where a patient had indicated they had a partner, but provided no details of a partner, were recorded as having a male partner. This is now more accurately classified based on specific registration details and egg/sperm source data, indicating that the number of total cycles with single patients was higher than previously reported (approximately doubled from 4,119 to 8,205 cycles in 2019).
2. Egg share cycles. This occurs when a patient who is already undergoing IVF treatment donates some of their eggs to the treatment clinic. Previously, cycles were classified as egg share even if no eggs were donated or if only egg donation occurred and no treatment took place. These cycles have now been reclassified as standard treatment cycles if no eggs were donated, or egg donations if no treatment occurred. This improved accuracy indicates that the number of egg share cycles are lower than reported in previous years (for example, in 2000 there was a 9% decrease between old and new reporting databases from 300 to 273 cycles).
3. Surrogacy. This is defined as a process of a patient carrying a baby on behalf of another person or family. Due to the way the data was stored, surrogacy with own eggs was provided in data reports within egg donation cycles. Following improvements in data classification, surrogacy using own eggs can now be estimated and are recorded separately to egg donor cycles.
4. Egg and/or embryo storage. These cycles were previously defined by the intention that eggs and/or embryos are stored and no embryos are transferred to the patient at the time the cycle takes place. These cycles are now more accurately classified based on specific treatment details, meaning some cycles previously considered ‘treatment’ or ‘donation’ have been corrected as egg or embryo storage cycles. This improved reporting indicates that the number of egg and/or embryo storage cycles and the total number of eggs and embryos stored are higher than previously reported.
5. Total donated embryos. In previous publications there was an error in this field, meaning that the data provided listed the total number of donated embryos or donor eggs and donor sperm used, rather than the total number of embryos that were donated. In this report this has been corrected for current and previous years to be the total number of embryos that were donated for use by other patients in that year.
6. Ethnic group classifications. There are changes in ethnic group numbers due to reclassification of groupings based on the 2021 Census for England and Wales. In particular, a decrease in Other ethnic groups (-14%, from 3,330 to 2,867 cycles) and small increases among Asian (+3%, from 9,770 to 10,084) and Black (+9%, from 2,137 to 2,326) ethnic groups. Ethnicity was the variable with the largest proportion of missing values - 10% of cycles in 2019 had no ethnicity recorded.
   Ethnicity data submitted from clinics changed in 2021. Until September 2021, each clinic submitted ethnicity data as in broad categories of ethnicity through patient registration forms. Since October 2021, clinics select one of 16 pre-defined ethnic groups based on the ethnicity data provided in the Registration form by patients and submit the data to HFEA. The pre-defined ethnic groups are similar to the ethnic groups in the 2021 Census for England and Wales. However, the pre-defined ethnic groups from registration forms do not include a subcategory within the Other ethnic group ethnicity for Arab ethnicities.
   The HFEA data collection system contains one question on ethnicity, while the 2021 Census for England and Wales asked for respondent’s ethnicity in two stages. Ethnic groups provided by clinics were categorised manually by the Research and Intelligence team following submission from clinics to best match the Census data. This was done into one of the five aggregated ethnic groups based on which aggregated ethnic group was chosen by the most respondents in the Census. For example, in the Census individuals who identified as “Chinese” most commonly placed themselves in the “Asian, Asian British or Asian Welsh” category. As such, all Chinese patients have been placed in the “Asian” category in our publications.
7. IVF and ICSI cycles. In previous publications all frozen embryo transfer cycles were classified as ‘frozen embryo IVF cycles’ whereas in this publication, and future publications, we are able to separately classify ‘frozen embryo IVF cycles’ and ‘frozen embryo IVF with ICSI cycles’. This results in the number of IVF with ICSI cycles being higher than previously reported.

5.3. Coherence and comparability

Although there is no standard definition of an IVF cycle, the HFEA counts a cycle as an embryo transfer, or equivalent stage if no embryos are created. Any remaining frozen embryos used subsequently are counted as new cycles. This may differ from a clinical commissioning group’s definition of a cycle.

We recommend that readers always view the latest version of Fertility Trends, which includes data going back to 1991. This is because inclusions and exclusions may change year on year as we improve our methodology. For example, pre-2018 we did not have the same exclusions on birth rates that we do now, and in 2019 we took the decision to exclude under 18’s and over 50’s when showing age-related data. Within the HFEA dashboard, only individuals aged between 18 and 65 were included. In the current Register, there are approximately 7,000 records out of approximately 1,640,555 records with no age recorded and they are filtered out for any age-related data. Any exclusions or inclusions are applied not just to the latest year, but to all previous years, ensuring comparability within publications.

Licensed clinics have been required to provide Register information to the HFEA since 1 August 1991. Clinics report data based on our collection forms and the information requirements have changed over time. The Human Fertilisation and Embryology Act 2008 contained provisions enabling two females in a same-sex couple to register a birth from 1 September 2009 onwards. Where possible, historical data is provided for comparison. HFEA data is typically presented by the year in which the treatment cycle started, not the year in which a resulting birth was reported, unless otherwise specified. Other data providers, such as the Office for National Statistics (ONS), publish birth rates according to the year the child was born. There are different ways to account for the outcomes of treatment. Our live birth data counts all births where one or more babies were born showing some sign of life, including those who sadly go on to die within the first month of life (neonatal deaths).

Our multiple birth data counts only births where two or more babies were born alive, including those where one or more of the babies died within the first month of life. Still births – where a baby is born after 24 weeks gestation showing no signs of life – are not included in either live birth or multiple birth counts in the period covered by the reports due to the way clinic birth rates are currently reported. This means that a multiple pregnancy which results in the birth of one live baby and one stillborn baby is not counted within our data as a multiple birth. The ONS however, classes a multiple birth as a pregnancy resulting in the birth of more than one baby, whether alive or stillborn.

Clinical improvements have led to increased chances of a live birth for most patients since 1991 as freezing techniques have improved with the introduction of vitrification in lieu of slow freezing.

There is a large degree of comparability in birth statistics between countries within the UK; the regions included in the reports are based on the location of the clinic rather than patient’s residency. Internationally, we provide data to ESHRE each year following data validation to allow publication of UK fertility treatment and outcome figures alongside those of other European countries.

Fertility Trends is our annual publication on national and regional trends in fertility treatments and live birth rates. Delayed data validation due to impact from the COVID-19 pandemic and a large-scale data migration project moving our 30-year fertility Register to a new system means Fertility Trends 2021 and Ethnic diversity in fertility treatment in 2021 reports contain preliminary treatment and pregnancy data for 2020 and 2021 and outcome data for 2019, 2020 and 2021. We are currently upgrading our systems to enable us to better use our data for patients, healthcare professionals and researchers and expect this work to be completed by summer 2024. Clinic-level information is published on the HFEA Choose a Fertility Clinic, website and may differ due to different reporting periods and processes.

In the HFEA dashboard, figures less than or equal to 7 have been suppressed and figures greater than 7 have been rounded to the nearest 5. This greater level of suppression than our publications was needed, as the filter options within dashboards increase the risk of data being combined to identify individuals. This approach is in line with NHS Digital dashboards.

5.4. Accessibility and clarity

Our recommended format for accessible content is a combination of HTML web pages for narrative, charts and graphs (including alt-text), with data being provided in usable formats such as Excel spreadsheets. Underlying datasets provide more detailed statistics. Alternative formats can be requested for all tables/figures in our reports. The HFEA dashboard has undergone an accessibility review, and improvements are ongoing. An update with these changes is expected in summer 2024.

5.5. Timeliness and punctuality

A snapshot of Register data is extracted each year and we aim to publish this data in Fertility Trends in May/June of each year. There is a two-year time delay in the data on which we report e.g. in 2023, we report on treatments that took place in 2021, allowing 12 months for birth outcomes to occur and to be reported by clinics to the HFEA. Due to the impact of COVID-19, a large-scale data migration project and delayed data validation, a Fertility Trends report was not published in 2022. However, an interim report looking at treatments for 2020 using provisional HFEA data was published in May 2022 – Impact of COVID-19 on fertility treatment 2020.

6. Concepts and definitions

6.1. Cycles and treatment cycles

The term ‘cycle’ covers the range of reasons patients undergo fertility treatment: with the intention of becoming pregnant as soon as possible (most patients), fertility preservation (a small but growing number of patients), or with the intention of donating eggs or embryos.

The term ‘treatment cycle’ or ‘treatment now’ includes only those cycles where the patient recorded on their registration form that they intended to try to become pregnant as part of their treatment (IVF, DI and egg sharing cycles are always treatment cycles).

This distinction is important because it has an impact on the birth rates we report. Throughout the reports, we use ‘treatment cycles’ to assess outcomes. Birth rates are calculated for those patients that intended to become pregnant to avoid underestimating the live birth rate.

6.2. Birth rates and pregnancy rates

The calculation of birth and pregnancy rates includes only cycles intended for treatment now and only cycles where the relating outcome is recorded on the Register. Pre-implantation genetic diagnosis (PGT-M/SR) and pre-implantation genetic testing for aneuploidy (PGT-A), which involves checking the chromosomes of embryos prior to treatment, have been excluded.

To give patients a clear idea of their own chance of birth or pregnancy, we present overall live birth and pregnancy rates with patient eggs only, excluding donor eggs and surrogacy to mitigate for the fact that donor eggs are likely to be frozen at a younger age and may artificially inflate the chances of a live birth rates. Fresh and frozen embryo transfer IVF cycles have been presented separately where possible. Pregnancies are counted as any cycle where foetal heart pulsations are seen.

We use per embryo transferred to measure IVF success rates as we believe it is the most appropriate measure of a clinic’s practices and success. This ensures multiple embryo transfers are not favoured in calculating birth rates. The data for DI and IVF birth rates per treatment cycle are available in the underlying data tables for completeness. This measure takes into account the outcomes of patients who do not reach the embryo transfer stage. There are many reasons why patients may stop treatment before the embryo transfer stage: for example, over or under reaction to the stimulation drugs, a failure to successfully fertilise any collected eggs, or other external factors in the patient’s life.

All the birth and pregnancy rates we quote in these reports are for one full calendar year. They are calculated as follows:

1. Birth/pregnancy rate per embryo transferred: the number of live birth/pregnancy occurrences divided by the sum of embryos transferred in that year
2. Birth/pregnancy rate per treatment cycle: the number of live birth/pregnancy occurrences divided by the number of treatment cycles started

6.3. Multiple pregnancy and birth rates

The calculation of multiple birth rate includes only cycles intended for treatment now and only cycles where relating outcome is recorded on the Register. In contrast to birth and pregnancy rates, it includes PGT-M/SR, PGT-A, surrogacy cycles and all egg sources.

1. Multiple pregnancy rate: the total number of multiple pregnancies divided by the total number of pregnancies
2. Multiple birth rate: the total number of live multiple births divided by the total number of live births

6.4. Reciprocal IVF

Reciprocal IVF, or shared motherhood, is a procedure used by female same-sex couples, in which one partner has their eggs collected and used to create an embryo, and then the other partner has the embryo transferred and carries the pregnancy. The HFEA does not collect data on where this has happened, and so it has been approximated as when the date of birth of the egg donor and partner match, and when at least one of the following is satisfied:

1. The egg donor’s surname and the partner’s surname match
2. The egg donor’s surname at birth and the partner’s surname match
3. The egg donor’s surname and the partner’s surname at birth match
4. The egg donor’s surname at birth and the partner’s surname at birth match

Due to the way the data is stored on the Register, for instance not having an identifier for reciprocal IVF cycles recorded, this data has previously been provided in data reports within egg donation cycles. In the new reporting database, reciprocal IVF using own eggs can now be estimated and are recorded separately to egg donor cycles. It is also worth noting that reciprocal IVF represented a large proportion of donor egg and donor sperm cycles in previous reporting.

6.5. Distinction between donor embryos and separately donated sperm and eggs

Due to the way data is collected in clinics, we cannot easily distinguish between a patient using a donated embryo, and a patient using separately donated sperm and eggs in their treatment. As such, these have been combined in previous HFEA publications. This distinction is discussed in the Trends in egg, sperm and embryo report, for which treatments using explicitly donated embryos were identified by:

1. The sperm and eggs both being sourced from donors.
2. The egg donor’s date of birth and surname matching with the date of birth and either current or birth surname of a patient who has donated embryos.
3. The sperm donor’s date of birth and surname matching with the date of birth and either current or birth surname of the partner of the patient in the above.
4. The treatment not being for surrogacy.

We emphasise that this is a minimum estimate from our records, and the true values may differ.

6.6. Donor registrations

The HFEA dashboard contains information on new sperm and egg donors. This relates to the date of the first time an individual registered as a sperm or an egg donor. Registrations are used rather than donations, as HFEA Register does not record information on the number of sperm donation cycles. Some donors register once, while others register several times, and so only their first registration is used to avoid overcounting. We note that when an embryo is donated, this is recorded as separate sperm and egg donor registrations by the HFEA.

7. Methods used to produce reports using UK Register data

The figures supplied in these reports are from our data warehouse containing Register data as at 19 April 2023. A snapshot of data is pulled using Microsoft SQL and fed into Tableau where data analysis and visualisation can take place.

The information that we publish is a snapshot of data provided to us by licensed clinics at a particular time. Guidance note 32 in the HFEA Code of Practice sets out the legal basis and requirements which govern our interaction with licensed clinics and third-party providers. We work closely with clinics and third-party systems to ensure the importance and guidance around submission of Register data is understood through stakeholder groups, workshops and sharing good practice.

We use additional quality assurance processes, including: manually validating data submissions; carrying out regular quality assurance checks on data through the inspection process; publishing non-compliances with data quality issues in inspection reports; where relevant, reviewing quality reports and targeting clinics for audit where irregular data has been submitted. In publications, we suppress values under five and any calculations with numerators and denominators below five are also supressed.