Official Statistics

Complications from abortions in England: comparison of Abortion Notification System data and Hospital Episode Statistics 2017 to 2021

Published 23 November 2023

Applies to England

This release presents statistics on abortion complications in England using data from the Abortion Notification System (ANS) compared to data on hospital admissions for abortion complications from Hospital Episode Statistics (HES) for the years 2017 to 2021.

These are official statistics in development, which are produced in accordance with Statistics and Registration Service Act 2007 and the Code of Practice for Statistics and meet high standards of trustworthiness, quality and public value. Official statistics in development were previously called ‘experimental statistics’. This development relates to the Abortion Statistics for England and Wales series, in particular the data sources, methods, coverage and quality concerns of complications following abortions, to assess and improve the existing statistics. This publication is not part of our routine publications. A fuller description of methodology and limitations of these statistics is provided throughout the release.

Introduction

The Office for Health Improvement and Disparities (OHID) acknowledges that there are limitations with the abortion complications data that is collected using the HSA4 form. This publication explores whether Hospital Episode Statistics (HES) can be used as a supplementary source for data on abortion complications.

Abortions are generally very safe, and most women will not experience any complications. According to the National Institute for Health and Care Excellence (NICE), “abortion is a simple, safe and commonly performed procedure”. However, there is a small risk of complication following an abortion. See more information in NICE guidance for abortion care.

According to NHS information on abortion, possible complications include:

• heavy bleeding
• infection
• damage to the womb
• sepsis
• needing another procedure to remove parts of the pregnancy that have stayed in the womb, also known as evacuations of retained products of conception

The type of abortion complication and risk of occurrence differ depending on the method of abortion and the number of weeks gestation. See more information on the risks and types of abortion complications. See the ‘Complication rate by complication type’ section in the ‘Results’ chapter of this publication for a full break down of the complication types included in this analysis.

This publication does not include data on deaths from abortion. Data on deaths from abortion are collected by the Office for National Statistics.

Key points

The key points of this release are:

• abortion complications are recorded differently in HES compared to the ANS and there are different strengths and limitations associated with using either data source
• between 2017 and 2021, the abortion complication rate is higher when using HES compared to the ANS for abortions in England for residents of England
• the abortion complication rate is highest when including incomplete abortions in HES. Incomplete abortions are usually characterised by complications such as bleeding and abdominal pain

Background

The ANS is a database containing all HSA4 forms submitted by the terminating registered medical practitioners to the Chief Medical Officer (CMO) at the Department of Health and Social Care (DHSC). This is for abortions taking place in England and Wales only. The HSA4 forms collect information about the abortion and the woman that had the abortion, such as the method of the abortion and the gestation (weeks) of the abortion. The ANS is maintained by DHSC.

Hospital Episode Statistics (HES) is a database containing details of all admissions, accident and emergency (A&E) attendances and outpatient appointments at NHS hospitals in England. This publication uses inpatient admissions data only and does not include data on A&E attendances or outpatient appointments. Only HES records containing an abortion related complication in the primary or secondary diagnosis position have been included in this analysis. Each HES record contains a wide range of information about an individual patient admitted to an NHS hospital including clinical information about diagnoses, patient information and administrative data. This system is maintained by NHS England.

Differences in the data sources

Complications are recorded differently in HES compared to the ANS.

The reasons for these differences are as follows:

• there is no common definition between ANS and HES as to what constitutes an abortion complication. This analysis presents 2 different abortion complication rates using HES data to account for this. Both the ANS and HES complication rates do not include an evacuation of retained products of conception. The additional HES abortion complication rate includes complications from incomplete abortions which contains evacuation of retained products of conception
• the ANS collects data on complications that occur up until the time of discharge from the place of termination using the HSA4 form. Complications that occur after discharge, or after the HSA4 form has been sent, are unlikely to be recorded. HES collects information on complications where a patient has been admitted to an NHS hospital or NHS funded treatment centre. HES will record complications that occur after discharge from the place of termination
• there are differences in how the abortion complication is coded. The ANS has a pre-defined list of complications to choose from (see table 1 in the ‘Results’ chapter). There is a free text box for ‘other complications not listed’. HES admissions are coded by clinicians using the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10). See the annex for a full list of ICD-10 codes included in this analysis
• the ANS data can include less severe complications that do not require admission to hospital. In these cases, the complication might be treated by the abortion provider. For example, an allergic reaction to abortion medication taken in clinic might be treated with an antihistamine on site. HES will pick up more severe complications which require an admission to hospital
• the data from the ANS used in this analysis includes complications occurring from all legal abortions for residents of England that were provided by an approved clinic or hospital in England. This includes cases where the abortion medication has been administered at home. HES data used in this analysis includes admissions to NHS hospitals, and treatment centres funded by the NHS, for residents of England. This could include an admission to an NHS hospital for a complication occurring after an abortion that was performed outside of England, or that was performed illegally (without a signed HSA4 form)
• the date of an abortion complication in ANS is based on the date of termination, whereas in HES it is based on the date of admission to a hospital setting

Data sources

Abortion Notification System (ANS)

In accordance with the Abortion Act 1967 and the Abortion Regulations 1991, all abortions performed in England and Wales must be notified to the CMO within 14 days of the procedure. Any complication known to the abortion provider terminating the pregnancy that occurred up until the time of discharge from the place of termination should be recorded on the HSA4 form. The ANS is a database containing all HSA4 forms.

The HSA4 form explicitly states that ‘an evacuation of retained products of conception is not a complication’ therefore, these cases are not included in the ANS complication rate.

This data is published by OHID in the annual Abortion statistics for England and Wales publication.

Strengths of the data

There is a legal requirement for registered medical practitioners to provide a HSA4 form for every abortion they perform. This ensures that DHSC reporting accurately covers all legal abortions in England and Wales. This data has been collected since 1968, thus providing a long timeseries of data.

The HSA4 form collects other detailed information on the abortion, such as the:

• number of weeks gestation
• method of abortion
• woman’s date of birth

This is legally required in accordance with the Abortion Act 1967 and Abortion Regulations 1991. HSA4 forms missing this information are sent back to the terminating practitioner for completion. These factors can affect the risk of experiencing a complication following an abortion, therefore it is useful to have this information available when analysing abortion complications.

Limitations of the data

The HSA4 form is not designed to capture information on complications occurring after discharge from the place of termination. The guidance for filling in the complications section of the HSA4 form states “you must select any complications that occurred up until the time of discharge from the place of termination”. It is possible that providers may complete and send the form on the same day the termination occurred. Therefore, any complications that occur after discharge from the place of termination, or after the form has been submitted, may not be recorded.

Due to the above, complications may be less likely to be recorded for terminations where either both or one of the abortion medications is administered at home for early medical (EMA) abortions under 10 weeks’ gestation^{[footnote 1]}. This is because the woman might have a fully remote consultation and treatment. They might not be in contact with the abortion provider after they have been sent the pills, which is before the abortion has taken place. In this case, it is unlikely that the person filling in the HSA4 form will know if the woman experienced a complication following an abortion. Abortions where both abortion medications were administered at home account for 52% of all abortions in 2021.

If the terminating practitioner is informed of a complication after discharge, or after HSA4 form has been submitted, it is unlikely that the complication will be recorded in the ANS. Patients experiencing a complication may present at an NHS setting (such as the GP or A&E), therefore independent abortion providers may not be informed. If the abortion provider is informed, they would need to have documented the relevant HSA4 form identification number and contact DHSC to ask for the form to be returned to them or updated with the relevant information. In 2022 there was no evidence of this occurring.

Hospital Episode Statistics (HES)

HES is a database containing details of all admissions, A&E attendances and outpatient appointments at NHS hospitals in England.

The analysis presented in this publication uses inpatient admissions data only. A person can be admitted to hospital as an inpatient through an elective, emergency or other route. Unfortunately, we are not able to accurately capture abortion complications treated in A&E or in outpatient care due to data quality and coding issues.

Abortion complications have been quantified using HES admitted patient care records where an admission with a diagnosis code relating to an abortion complication is listed. See the annex for full list of ICD-10 codes used in this analysis. The diagnosis for the admission is coded by the hospital coding team using a discharge letter which is completed by a clinician. The patient may have multiple diagnoses, or experience multiple abortion complications, but these are still counted as a single admission.

Strengths of the data

HES data covers all integrated care boards (ICBs) and sub ICBs^{[footnote 2]} in England. This includes private patients treated in an NHS hospital and care delivered by treatment centres funded by the NHS.

At present, there is no commonly agreed definition of what should be included as an abortion complication used by both ANS and NHS England. The HES data includes additional information on incomplete abortion which includes evacuation of retained products of conception. This allows for the construction of an alternative abortion complication rate.

The coding for complications recorded in HES comes from the patient discharge letter which is filled in by a clinician. This means complications in HES have been verified by a medical practitioner.

In this analysis, complications that occur after discharge from the abortion provider can be captured by HES, provided that the woman is admitted to hospital as an inpatient. HES is not restricted by the patient’s abortion pathway.

Limitations of the data

HES only covers NHS hospitals in England unlike data from the ANS which covers all hospitals and clinics in England and Wales. Any complications that are treated in hospitals in Wales, or in treatment centres in England not funded by the NHS, are not captured in the HES data used in this analysis.

Similarly, HES data may include treatment of a complication arising from an abortion that was not performed in England. This makes it difficult to analyse the complication rate from abortions performed in English hospitals.

There may be abortion complications that do not require admission to hospital but instead can be dealt with at the time of the abortion; for example, an allergic reaction to an abortion medication, resulting in swelling of the face, might be treated with an antihistamine by the abortion provider. This might be picked up using the HSA4 form but would not be counted using HES data.

There can be issues with the accuracy of the coding of the diagnosis. A clinician may use an incorrect ICD-10 code, mistype an ICD-10 code or leave out this information. This can result in an incorrect diagnosis being inputted into the database. There may also be differences in coding practices across different hospitals.

There can also be confusion around the medical terms used to describe an abortion. Miscarriages are often described as a ‘spontaneous abortion’ while abortions may be described as an ‘induced miscarriage’. This can result in the admission being given a code for an abortion complication when it should have been given a code for a miscarriage complication and vice versa.

The total number of abortions is required in order to create a denominator for the complication rate per 1,000 abortions. While this is possible with ANS data, there is no total number of abortions using HES data. This is because not every abortion requires admission to hospital. For this reason, we estimate the HES denominator using ANS data on legal terminations in England for residents of England.

Results

Methodology

ANS

For analysis using data from the ANS, the complication rate has been calculated by taking the number of legal abortions with at least one complication and dividing this by the total number of legal abortions in England for residents of England. This has been expressed as a rate per 1,000 abortions.

HES

For analysis using data from HES, the complication rates have been calculated by taking the number of admissions in NHS hospitals in England for residents of England with a diagnosis code relating to abortion complications and dividing this by the total number of legal abortions in England for residents of England (from the ANS). This has been expressed as a rate per 1,000 abortions.

Due to lack of comparable data in HES, we have used data from the ANS to estimate the group at risk (the number of abortions) which forms the denominator of the rates. To match the HES data used in this analysis as closely as possible, we have used only abortions in England for residents of England. We acknowledge that using this method may lead to biased estimates as the ANS estimate for the number of abortions may not exactly reflect the true group at risk. For example, HES may include complications from an abortion performed outside of England but treated in a hospital in England. This would not be included in the ANS estimate for the number of abortions which could lead to an overestimation of the rate of complications per 1,000 abortions. We expect the effect of this bias to be small.

This analysis presents 2 HES complication rates to account for the differences in what can be included as an abortion complication. One rate consists of ICD-10 codes for complete abortion with a complication (HES). The other rate consists of ICD-10 codes for complete and incomplete abortion with a complication (HES including incomplete abortion). Incomplete abortion includes retained products of conception. See the annex for a full list of ICD-10 codes used in each rate.

Complication rate

Figure 1 below shows that from 2017 to 2021, the ANS abortion complication rate is lower than the HES abortion complication rate and HES abortion complication rate including incomplete abortions.

The rate of complications is much higher when including incomplete abortions using HES (includes retained products of conception). This is likely due to incomplete abortions usually being characterised by complications such as bleeding and abdominal pain.

Prior to 2020, the ANS complication rate remained stable at around 1.6 to 1.7 per 1,000 abortions. In 2020, the ANS complication rate fell to 1.2 per 1,000 abortions before rising slightly to 1.4 in 2021. The HES abortion complication rate and the HES abortion complication rate including incomplete abortions both fluctuated between 2017 and 2021.

Figure 1: crude abortion complication rate by data source, England, 2017 to 2021

Complication rate by age

Figure 2 below shows that in 2021, the complication rate increases with age group. For all 3 abortion complication rates (ANS, HES and HES including incomplete abortion), the lowest abortion complication rate is for those aged under 20 and the highest rate is for those aged 35 and over.

For the ANS rate, this relationship holds for both medical and surgical abortions. Across all age groups, women are more likely to have a medical abortion than a surgical abortion (see table 2 in the ‘Abortion statistics 2021: data tables (revision)’ spreadsheet on the Abortion statistics for England and Wales: 2021 page). This analysis can’t be replicated using the HES data.

Figure 2: crude abortion complication rate by age group and data source, England, 2021

Complication rate by method and gestation

Data on method and gestation of the abortion is not available for this analysis of HES abortion complications rates, therefore this component of the analysis is limited to ANS data.

Figure 3 below shows that in 2021, the ANS abortion complication rate increases with the number of weeks gestation of the abortion, independent of the method of the abortion. The total complication rate is 0.4 per 1,000 abortions where gestation is 2 to 9 weeks and rises to 23.8 where gestation is 20 and over weeks.

The medical complication rate is higher than the surgical complication rate for all gestation groups other than 2 to 9 weeks. However, the total surgical complication rate (for all gestations) is higher than the total medical rate at 3.9 complications per 1,000 abortions compared to 1.1. This is because the complication rate for medical abortions performed at 2 to 9 weeks gestation is very low (0.3 per 1,000 abortions) and this type of abortion makes up 97% of all medical abortions.

Figure 3: crude abortion complication rate by gestation (in weeks) group and method, England, 2021

Complication rate by complication type

Tables 1 and 2 show that in the ANS and HES, the most common abortion complication is haemorrhage (65.8%) and delayed or excessive haemorrhage (82.0%) respectively.

Please note that the complication types are not directly comparable because the options listed in ANS are not comparable with the ICD-10 codes used in HES.

Table 1: complication type using the ANS, England, 2021

Complication type	Percentage of total complications
Haemorrhage	65.8%
Sepsis	11.4%
Cervical tear	11.1%
Other	6.3%
Uterine perforation	5.4%

Table 2: complication type using HES, England, 2021

Complication type	Percentage of total complications
Delayed or excessive haemorrhage	81.4%
Other and unspecified	10.6%
Genital tract and pelvic infection	8.0%
Embolism	Not included, to protect patient confidentiality

Coverage of complications

If an event requiring additional medical attention occurs after an abortion, the complication may or may not be captured in either the ANS or HES. There are also other events that are not captured in either ANS or HES. For this reason, the complication rates presented in this analysis are incomplete because they do not include those complications diagnosed at other sites such as the GP, 111 or A&E. Some examples are listed in the table below.

Table 3: coverage of complications using HES and the ANS (through HSA4 form), England

Scenario	Included in HES admissions data	Included in ANS
The complication occurs in the abortion clinic, clinical assessment determines treatment is needed and the patient is admitted to hospital	Yes	Yes
The complication occurs in the abortion clinic, clinical assessment determines treatment is needed and treatment is given by abortion provider	No	Yes
The complication occurs outside of the abortion clinic, after the patient has been discharged, and the patient is admitted to hospital. If the abortion provider is informed and the HSA4 has not yet been submitted	Yes	Yes
The complication occurs outside of the abortion clinic, after the patient has been discharged, and the patient is admitted to hospital. The abortion provider is informed but the HSA4 form has already been submitted	Yes	No
The complication occurs outside of the clinic, after the patient has been discharged, and the patient attends their GP and is referred on to another treatment provider. The abortion provider is not informed	No	No
The complication occurs outside of the clinic, after the patient has been discharged, and the patient calls an ambulance but is not admitted to hospital. The abortion provider is not informed	No	No
The complication occurs outside of the clinic, after the patient has been discharged, and the patient calls 111 but is not admitted to hospital. The abortion provider is not informed	No	No
The complication occurs outside of the clinic, after the patient has been discharged, and the patient attends A and E but is not admitted to hospital. The abortion provider is not informed	No	No
The complication occurs, after the patient has been discharged, and the patient is admitted to hospital but they do not disclose that an abortion took place	No	No

Glossary

Evacuation of retained products of conception

A procedure to remove parts of the pregnancy that have stayed in the womb.

Chief Medical Officer (CMO)

The CMO is the most senior government advisor on health matters in the UK. The CMO advises government on public health issues. This extends to recommending policy changes affecting the law governing abortion and advising doctors who perform abortions, regarding the interpretation of that law. There are 4 in total, each one advising either His Majesty’s Government (CMO for England and medical adviser to the United Kingdom government), the Northern Ireland Executive, the Scottish Government or the Welsh Government.

Medical abortion

Medical abortion involves taking 2 different medicines to end the pregnancy. The medicines are prescribed by the hospital or clinic and are usually taken 1 or 2 days apart.

On 30 March 2022, Parliament voted to amend the Abortion Act 1967 to allow eligible women in England and Wales to take one or both pills for early medical abortion up to 10-weeks at home, following a telephone or e-consultation with a clinician.

In 2021, 87% of abortions were medical abortions and abortions where both medicines are taken at home accounted for 52% of all abortions.

Surgical abortion

Surgical abortion involves an operation to remove the pregnancy from the womb. There are 2 types of surgical abortion:

• vacuum or suction aspiration can be used up to 14 weeks, where the pregnancy is removed using suction
• dilatation and evacuation can be used after 14 weeks, where the pregnancy is removed using forceps

Incomplete abortion

Where some pregnancy tissue has remained in the womb. This requires a procedure known as an evacuation of retained products of conceptions.

International Statistical Classification of Diseases and Related Health Problems (ICD-10)

ICD-10 is a comprehensive classification of causes of morbidity and mortality and is published by the World Health Organization (WHO).

Crude rate

The number of complications in a specified population per year, divided by the total number of abortions in that population. 

Feedback

OHID is inviting views on abortion statistics for England and Wales. Your feedback will help us understand the use and the value of our products, improve our outputs and ensure our reporting better meets our audiences’ needs.

It should take no more than 5 minutes to complete the user engagement survey - all responses are anonymous. 

Our statistical practice is regulated by the Office for Statistics Regulation (OSR).

OSR sets the standards of trustworthiness, quality and value in the Code of Practice for Statistics that all producers of official statistics should adhere to.

If you have any comments about how we meet these standards, you can contact:

• the Abortion Statistics team directly at [email protected]
• OSR by emailing [email protected] or through the OSR website

You are also welcome to email the Abortion Statistics team directly if you have any feedback on this publication and other abortion statistics.

Annex: ICD-10 codes

ICD-10 codes used in this analysis

The following ICD-10 codes have been used in this analysis for both complete and incomplete abortions (note: incomplete abortion contains evacuation of retained products of conception):

For the HES complication rate (complete abortions only):

• O045: medical abortion - complete or unspecified, complicated by genital tract and pelvic infection ​
• O046: medical abortion - complete or unspecified, complicated by delayed or excessive haemorrhage ​
• O047: medical abortion - complete or unspecified, complicated by embolism ​
• O048: medical abortion - complete or unspecified, with other and unspecified complication

For the HES complication rate including incomplete abortions:

• O040: medical abortion - incomplete, complicated by genital tract and pelvic infection ​
• O041: medical abortion - incomplete, complicated by delayed or excessive haemorrhage ​
• O042: medical abortion - incomplete, complicated by embolism ​
• O043: medical abortion - incomplete, with other and unspecified complications ​
• O045: medical abortion - complete or unspecified, complicated by genital tract and pelvic infection ​
• O046: medical abortion - complete or unspecified, complicated by delayed or excessive haemorrhage ​
• O047: medical abortion - complete or unspecified, complicated by embolism ​
• O048: medical abortion - complete or unspecified, with other and unspecified complication

ICD-10 codes considered but not used in this analysis

The following ICD-10 codes were considered but have not been used in this analysis:

• O030: spontaneous abortion - incomplete, complicated by genital tract and pelvic infection
• O031: spontaneous abortion - incomplete, complicated by delayed or excessive haemorrhage
• O032: spontaneous abortion - incomplete, complicated by embolism
• O033: spontaneous abortion - incomplete, with other and unspecified complications
• O034: spontaneous abortion - incomplete, without complication
• O035: spontaneous abortion - complete or unspecified, complicated by genital tract and pelvic infection
• O036: spontaneous abortion - complete or unspecified, complicated by delayed or excessive haemorrhage
• O037: spontaneous abortion - complete or unspecified, complicated by embolism
• O038: spontaneous abortion - complete or unspecified, with other and unspecified complications
• O039: spontaneous abortion - complete or unspecified, without complication
• O044: medical abortion - incomplete, without complication
• O049: medical abortion - complete or unspecified, without complication
• O070: failed medical abortion, complicated by genital tract and pelvic infection
• O071: failed medical abortion, complicated by delayed or excessive haemorrhage
• O072: failed medical abortion, complicated by embolism
• O073: failed medical abortion, with other and unspecified complications
• O074: failed medical abortion, without complication
• O075: other and unspecified failed attempted abortion, complicated by genital tract and pelvic infection
• O076: other and unspecified failed attempted abortion, complicated by delayed or excessive haemorrhage
• O078: other and unspecified failed attempted abortion, with other and unspecified complications
• O079: other and unspecified failed attempted abortion, without complication

1. On 30 March 2022, Parliament voted to amend the Abortion Act 1967 to allow eligible women in England and Wales to take one or both pills for early medical abortion up to 10 weeks at home, following a telephone or e-consultation with a clinician. The legislation came into force on 30 August 2022, as part of the Health and Care Act 2022. ↩

2. ICBs were previously clinical commissioning croups (CCG) ↩