National Pregnancy in Diabetes Audit 2021 and 2022 (01 January 2021 to 31 December 2022)

The NPID audit monitors the quality of care for women with diabetes who become pregnant, against national standards (NICE CG63 superseded by NG3 in February 2015 and NICE Quality Standard for Diabetes*).

The NICE guideline was developed in response to the Confidential Enquiry into Maternal and Child Health (CEMACH) survey** of diabetes and pregnancy in 2002-03, which identified severe deficiencies in the provision of care before and during pregnancy for women with diabetes in England and Wales and poor outcomes against international standards.

The audit results will be relevant to women with diabetes who are pregnant or may become pregnant, and to all those involved in planning and delivering care to women with diabetes, such as:

• Diabetes and maternity services
• Primary care teams
• Specialist diabetes services
• Family planning services
• Local health boards (LHBs)
• Public health programmes
• Patient education programmes
• Diabetes and maternity networks
• Diabetes support groups and charities

The audit published data for the first time at service level in the 2015 report. Service level data is available as an interactive dashboard alongside the publication. Antenatal diabetes services will be able to use their data to identify areas of care to target with quality improvement work.

It is difficult to assess accurately the level of participation in the audit, as a definitive list of antenatal diabetes services is not available.

The 2019 National Maternity and Perinatal Audit (NMPA) Organisational Survey*** listed 164 maternity services in England and Wales with a multidisciplinary diabetes team. These covered 134 trusts. 182 services submitted to NPID in 2021 or 2022, from 133 trusts. There were 12 trusts which submitted data to NPID which do not appear in the NMPA survey. This may be due to organisational changes since 2019.

In England the data is collected under direction by NHS England. It is assumed that all women who are eligible for the audit, from all eligible sites, are entered. However this can not be verified as this information is not collected anywhere else, therefore case ascertainment cannot be calculated accurately. Welsh data is collected under an agreement with the Welsh Government. It is assumed that all women who are eligible for the audit, from all eligible sites, are entered. However this can not be verified as this information is not collected anywhere else, therefore case ascertainment cannot be calculated accurately.

Data has been published at the following geographic levels:

• National report – England and Wales 2021 and 2022 data
• Service level dashboard – service level, region level and England and Wales for 2020 to 2022

***http://www.maternityaudit.org.uk/Audit/Charting/Organisational

To minimise the burden of data collection and submission, the audit data is partly collected from antenatal diabetes services and partly obtained by linking the directly collected data to:

• National Diabetes Audit (NDA) data
• Hospital Episode Statistics (HES) data
• Patient Episode Database for Wales (PEDW) data
• Neonatal Research Database (NNRD)

All data collected by the NPID audit should already be stored in clinical records; no additional data should need to be collected.

The majority of the data collected by the NPID audit is entered into a custom-built secure online database by staff at antenatal diabetes services. As there is no facility to upload a pre-populated data file, data entry is manual which introduces a risk of data entry error. Several validation methods are used to mitigate this risk (see Validity section below).

The majority of the data fields are mandatory items, which means completeness is generally high, although a ‘Don’t Know’ response is available for many questions.

Linkage to the NDA core dataset is conducted to enrich the NPID dataset with information on ethnicity and deprivation, and also to provide additional details on diabetes type where this is missing in NPID.

Ethnicity and lower super output area (LSOA) was obtained for 97% of women in the NPID audit by linking to NDA data. Year of diabetes diagnosis was also obtained for 97% of women in NPID by linking to the earliest diagnosis year recoded in any year of NDA data. See the Methodology section for further details of the linkage process.

Diabetes type was added to the NPID audit collection for direct data entry from 01 January 2015. For 2014 records where diabetes type was not retrospectively entered directly to the NPID audit record, and for subsequent records where the diabetes type remained unknown in NPID, this has been obtained (where possible) by linking to NDA data.

Where it has not been possible to obtain these data items via linkage this may be because:

• The woman’s GP practice did not participate in the NDA
• The woman registered an opt-out so that her data was not included in the GP practice’s NDA submission
• The woman was diagnosed with diabetes after 31 March 2022
• These data items were left blank in the NDA submission

Following the NDA linkage almost all records (99.4%) had a known diabetes type in 2021-22.

Onset of labour and mode of delivery were obtained for 83% of pregnancies ending in 2020 and 84% of pregnancies ending in 2021. The data was linked to the relevant financial years of HES, allowing for up to 7 days difference in pregnancy end date between the NPID audit data and the HES data.

The relative timing of the NPID audit dates and HES release dates means that it is not yet possible to link the 2022 NPID audit data.

However, onset of labour was missing or coded as ‘not known’ for 23% of linked HES pregnancies and the delivery method was missing or coded as ‘not known’ for 16% of linked HES pregnancies.

Although 84% of 2020 and 89% of 2021 pregnancies could be linked to a hospital episode record in the PEDW data, onset of labour was missing or coded as ‘not known’ for 54% of linked pregnancies and the delivery method was missing or coded as ‘not known’ for 46% of linked pregnancies.

Data on admissions to neonatal care are was obtained for the first time in the 2020 reporting year and provided again for 2022 following linkage to the Neonatal Research Database (NNRD). Further information is available using the following link: https://www.imperial.ac.uk/neonatal-data-analysis-unit/neonatal-data-analysis-unit/. Around 30% of nearly 15,000 NNRD records which could be linked to NPID were not recorded in NPID as having an admission to Special Care Baby Unit (SCBU) or Neonatal Intensive Care Unit (NICU). Conversely, there were a number of live born infants in NPID who were recorded as being admitted to SCBU or NICU and who did not appear in the NNRD data.

The data entry system validates certain data items such as height, weight and HbA1c values at the point of data entry, so that records can only be submitted where values are within the valid range.

For questions with categorical response options (e.g. pregnancy outcome), users must choose a response option from a drop down list so only valid data values can be entered.

As the audit records several dates in each pregnancy, date entry errors may make a record internally inconsistent. For example, the first contact with the antenatal diabetes team may be recorded as after the pregnancy end date or the first HbA1c measurement may be recorded at a date before the woman was pregnant.

The number of such errors has been reduced by:

• Providing an annual data quality report back to the submitting antenatal diabetes service for them to review and correct the errors identified
• Introducing validation within the data entry system that cross-checks the dates at the point of data entry and flags errors which must be resolved before the record can be submitted (as of 1 January 2015)
• Making available reports within the online system that flag certain unlikely combinations to those entering the data so these can be checked prior to submission deadlines
• Allowing each service to download their own raw data from the data entry system for review
• Providing management reports for each submitting organisation through the Clinical Audit Platform (CAP) to reproduce some of the measures used in the NPID report so that the impact of missing data on calculated measures can be seen

Some other issues that have been identified as the dataset accumulates over time are:

• Second pregnancies for the same woman being entered incorrectly i.e. as twins with different pregnancy end dates, rather than 2 separate pregnancies
• Inconsistent diabetes type information for different pregnancies for the same woman

There are also a small number of medically implausible combinations of values present in the data, such as women recorded as having type 1 diabetes not on insulin. As it is not possible to correct these, they have been left in the data and flagged as a data quality issue where used in the report. In the 2014-2022 dataset, diabetes type has been reset from type 1 to unknown where there is no record in any year in NPID or NDA data of insulin use.

Following the data quality review period, a final dataset was extracted.

Where outstanding errors remained, the data was cleaned by NHS England.

Where possible, the majority of the data from each record was used and only the invalid dates removed. For date errors, pregnancy end date was assumed correct and dates inconsistent with this were set to missing.

A small number of errors that were flagged on the data quality reports remained uncorrected and could not be resolved by cleaning rules, such as:

• All other data completed but ‘Alive at 28 days’ left blank

• Fundamental errors such as the same outcome data recorded against 2 different pregnancies

771 records with such unresolved errors were excluded from the final cleaned dataset for 2014-2022. 31% of these records were excluded because the ‘Alive at 28 days’ field had been left blank.

Different antenatal diabetes services may be involved in a woman’s care during a pregnancy, for example if the woman moves house or needs to be transferred to a more specialist service.

For 2014 data, NPID audit data was linked only to the service submitting the data, which may have provided care throughout the pregnancy, only for the beginning, or only for the end.

From 2015, the NPID audit has recorded ‘Booking Hospital’ and ‘Delivery Hospital’ for each pregnancy.

This publication includes pregnancies ending on or before 31 December 2022.

The submission deadline for this data was 17 February 2023, with data quality amendments permitted up to 31 March 2023.

The time lag to publication is 10 months from the end of the audit period and 7 months after the end of the data quality amendment period.

Due to the relative timing of the NPID audit cycle and HES and PEDW data releases, onset of labour and mode of delivery data in this publication covers 2020 and 2021 NPID audit data linked to 2019/2020, 2020/2021 and 2021/2022 HES and PEDW years.

The report is published as a HTML webpage. Tables and charts used in the report are also provided in Excel format in the detailed analysis file.

Service level reporting is in an interactive dashboard format and local data users can filter for their own report, and compare against national performance and other services.

A methodology paper and glossary of terms is also available in the report.

The same data items have been collected over the 10 years of the audit with improvements made to data quality through minor collection changes and the addition of validation rules as explained in the validity section of this document.

The number of women recorded in the NPID audit increased each year between 2013 and 2019. In 2020 the number of services submitting, and the number of women included was lower compared to 2019. This is unsurprising given pressures of the global pandemic. In fact, it is a testament to the service providers that the number of women/pregnancies was only slightly reduced. In 2021 and 2022, the number of women and services increased to the largest number recorded for the audit to date.

Table 1: Number of women and services per year between 2014 to 2022

The increase in the number of women being recorded in the audit is likely to be due partly to more complete annual data submitted by units as the audit has become more established, and partly due to an increase in type 2 diabetes amongst women of child bearing age.

Differences in audit results over time may be due to changes in participation and data submission rather than changes in service delivery.

Other than the NPID audit, the last national survey of pregnancy in women with pre-existing diabetes was the CEMACH survey in 2002-2003.

There have been some regional pregnancy in diabetes surveys since then but no up to date national information is available outside the NPID audit.

Comparisons with the CEMACH survey, particularly where data for all diabetes types was reported together, may show differences which are in part due to the increase in the proportion of women with type 2 diabetes in the NPID audit compared to CEMACH.

The NPID Audit is supported by an advisory group including patient representatives, Diabetes UK, clinicians, researchers and analysts. The advisory group provides input to the design of the analysis and the report as well as the direction and development of the audit.

The advisory group sought feedback in summer 2015 regarding the process of participating in the audit and the way the data was reported from services participating in the NPID audit. The most common request was for service level data to be produced. Due to the small numbers of pregnancies in women with diabetes at individual service level, data for 3 years was aggregated. Service level reports were published for the first time alongside the 2015 report, and have been repeated in subsequent publications, including a larger set of data items. In the 2022 reporting period, an interactive service level dashboard replaced the service level report.

Your feedback on this publication is welcome and may be sent to [email protected] (please include ‘National Pregnancy in Diabetes Audit’ in the subject line). Alternatively you can call our contact centre on 0300 303 5678.

Patient identifiable data is collected by the NPID audit and obtained by linkage to other data sources. This information is held securely and with restricted access. Identifiable data is only released by NHS England where the release meets NHS England Information Governance procedures.

Women can withdraw their consent at any time and ask for their information to be removed by asking their antenatal diabetes service to inform NHS England or by contacting NHS England directly.

Although in England NPID data is collected under direction, and in Wales under an agreement with the Welsh government, women are able to opt-out of the data collection at any time. Patient information leaflets and posters are available in all clinics in England and Wales and there is detailed information on who to contact to opt out.

Data in the NPID reports have been subject to disclosure control to prevent identification of individuals. The disclosure control methodology is documented in the methodology page available from
https://digital.nhs.uk/data-and-information/publications/statistical/national-pregnancy-in-diabetes-audit/2022.