Maternity and Neonatal Equalities dashboard: Appendices

Appendix A: Presentation of the data

Disclosure rules

Disclosure rules for presenting data are essential to protect confidentiality, ensure legal compliance, maintain trust, and prevent the misinterpretation of information. These rules, often referred to as statistical disclosure control, minimise the risk that patients can be identified from aggregate, tabular data, especially when dealing with small populations. In line with other Maternity Services Data Set (MSDS) related products, NHS trust level data is either rounded or suppressed to avoid any disclosure risk. This means:

counts of zero are shown as zero
counts of 1-7 are rounded to 5
all other counts are rounded to the nearest 5
percentages (where applicable) are calculated from the rounded numerator and denominator
there is an exception for rare events, where rates per 1,000 are calculated instead of percentages. These rates are calculated from unrounded numbers but are not shown where the numerator is less than 8. This applies to the postpartum haemorrhage (PPH) and preterm birth metrics

Symbols

Symbols used in the downloadable Equalities MSDS data file:

- = zero / rounded to zero

: = not available

z = not applicable

* = cell value has been suppressed to ensure that the data remains non-disclosive (see above).

Confidence intervals

A confidence interval gives an indication of the likely error around an estimate that has been calculated from measurements based on a sample of the population. It indicates the range within which the true value for the population as a whole can be expected to lie, taking natural random variation into account.

For the MBRRACE-UK and CQC Maternity data, 95% confidence intervals are used, meaning if it were possible to repeat the same collection of data under the same conditions a number of times, we would expect 95% of the confidence intervals calculated in this way to contain the true population value for that estimate. It has not been possible to construct these within the publication timescales for the MSDS data.

Different reporting periods and intervals of data

Reporting periods and intervals differ because we are using a range of data sources with different geographical breakdowns. For consistency, all reporting periods start from 2022 to align with MSDS, as that is when data quality significantly improved following an uplift to the dataset in 2019. The time series goes up to the latest period available which varies across the data sources.

The MBRRACE-UK data is presented at national level which cover rare events. Data is presented annually or combined for 3 years to avoid small numbers when broken down by ethnicity and deprivation.

The CQC Maternity Survey data is presented at national and regional level. Due to small numbers it is not possible to present these at NHS trust level with the ethnicity and deprivation breakdowns. The information is presented annually broken down by ethnicity and deprivation. Where respondents have not answered the ethnicity question, their information has been removed from reporting. If the Lower Super Output Area (LSOA) could not be identified for a respondent, then this information has also been removed from reporting. The regional breakdowns are based on the location of the NHS trust providing the care so there is no unknown category. Different survey weighting methodologies have been applied at NHS trust level so will impact on regional counts, which will not sum to England level counts.

The Maternity Services Data Set data is presented at national and NHS trust level. MSDS is collected monthly but to avoid small numbers, data has been aggregated into 6 monthly periods or 12 monthly periods (with rolling periods where appropriate) for rare events, before it is broken down by ethnicity and deprivation. Further geographical breakdowns are possible if needed in a future iteration of the dashboard.

Ethnicity breakdowns

We have used the 5 broader ethnicity categories as used in the ONS 2001 Census to avoid small numbers and suppressing data. NHS trusts can use the information to drill down into their local data to investigate anything further.

Deprivation categories and breakdowns

Different deprivation categories and breakdowns have been used because we are using a range of data sources which have different deprivation categories applied to it.

The MBRRACE data is presented at national level. Baby’s deprivation categories for stillbirths and neonatal deaths use the proportion of families with low incomes in each local area. Mother’s deprivation categories for maternal deaths use 2019 English Indices of Deprivation. As these are rare events, the numbers are small so the data are presented in quintiles rather than deciles. It is expected that these will be updated in line with the 2025 English Indices of Deprivation for a later iteration.

The CQC Maternity Survey data is presented at national and regional level based on deciles as the numbers are larger.

The MSDS data is presented at NHS trust level based on quintiles from the 2019 English Indices of Deprivation as the numbers are smaller.

Appendix B: Limitations of the data

MSDS data

Missing submissions

Most NHS trusts submit monthly MSDS data within the submission window. However, if a NHS trust does not submit data for a given month, this may have a large impact on trust-level and national estimates.

Duplication issue

There are known duplication issues in MSDS for the number of bookings and the number of births and their related measures, leading to some double-counting in published figures.

Bookings

At the point of data submission, there are validation rules in place to ensure that there is only one booking appointment date recorded for each woman per NHS trust for a given reporting period. However, it is possible that a woman may have multiple booking appointments recorded for the same pregnancy if she has received care across different NHS trusts. This is due to the way the data set is designed. Once a first booking appointment date has been submitted by trust A, if subsequent care is needed in trust B, then trust B can only submit the care if they also submit a booking appointment date to mark the woman’s entry into their care. This subsequent booking date may actually be their first point of contact within trust B or an attempt to repeat the date of the first booking within trust A (if known, and subject to system arrangements and data validation rules). Therefore, in this scenario, 2 booking appointments are counted – one for each NHS trust.

Data is not currently cleansed to identify and only count the first booking appointment as it can be complex to identify this as distinct from any subsequent booking appointments. This is partly due to the ability of trust B to either record a new booking date or replicate the original one, and with the potential for trust B to submit booking appointment information into MSDS before trust A. This could happen if there are issues with their system or data submissions, alongside other complexities in the data.

For this scenario, for any England figures calculated and presented in this dashboard, a woman is only counted once in relation to having had a booking appointment within the given time period, when booking information has been submitted by both trust A and trust B. If trust A and trust B submit booking information for a woman with any different information (about her demographics or care), then the woman will be counted twice in the England figures across the 2 relevant categories. For example, where a woman has an ethnicity recorded as Asian or Asian British by trust A but Mixed by trust B, then the woman would be counted once in the overall total for bookings but twice across the ethnic group breakdowns for bookings in England.

Therefore, the sum of NHS trust level data related to booking appointments will not equal England figures.

Births

A separate duplication issue occurs with the counts for births and of babies born. This is because a mother could receive her antenatal care and deliver her baby with trust A. After delivering her baby with trust A, she could be transferred to trust B for postnatal care for herself and baby. She would therefore have a new booking appointment recorded for Trust B but also potentially another record of the birth and baby. For example, if she gave birth at trust A and then transferred to trust B afterwards, for trust B to record her and her baby’s postnatal care, they may need to replicate the birth and baby records first to be able to do this. Therefore, 2 births are counted – one for each NHS trust.

For this scenario, for any England figures calculated and presented in this dashboard, a woman is only counted once in relation to giving birth within the given time period, when birth information has been submitted by both trust A and trust B. We have been made aware that this can occur more frequently. For example, where neighbouring trusts share care for the same mother and baby with a shared maternity information system that automatically synchronises updates across all the trusts concerned with their care.

If trust A and trust B submit birth information for a woman with any different information (about her demographics or care), then like the booking appointments, the woman will be counted once in the overall birth counts for England but twice in the England figures across the 2 ethnic or deprivation categories.

In both the bookings and births duplication scenarios, if more trusts get involved with any aspect of the care (besides trust A and trust B), then this increases duplication further.

We are investigating how to deduplicate these records and identify the true birthing trust, which is primarily reliant on the accurate recording of the site at which the birth occurred (the recording of this has significantly improved over the last year). This is to ensure that both trusts recording the birth agree exactly at which trust the birth occurred at, and we can then identify that as the correct birth and baby record for reporting purposes.

Appendix C: Definitions

Data definitions: table for downloadable data files

Data found in the Maternity and Neonatal Equalities dashboard are also made available in a comma-separated variables (csv) file to allow users to carry out their own analyses.

The data definitions table below supports the downloadable csv data files. Not every data field is contained in each file, this is set out in the first column of the data definitions table below.

Which data files contain the column	Column name	Column description	Data type
All	ReportingPeriod	The time period for which data is presented (such as in single year, or for example, January 2025 to June 2025 format)	String
All	ReportingPeriodLength	The length of the reporting period for the data presented, in months	String
All	OrgLevel	The specific level of organisation or geography for the data presented, such as provider (NHS Trust), region or England	String
All	OrgCode	The NHS Organisation Data Service code for the named organisation	String
All	OrgName	The name of the organisation	String
All	DataSource	The data source used for the metric, such as Maternity Services Data Set (MSDS)	String
All	MetricCategory	What is being counted for the metric, such as women or babies	String
All	MetricName	The metric name – further information on each MSDS metric can be found in the metric construction tab in the dashboard, and CQC and MBRRACE-UK provide further information about definitions on their respective webpages	String
MBRRACE-UK and CQC	MetricType	Where applicable: whether the numbers given are for the total cohort, or are split by ethnicity or deprivation (quintile or decile)	String
All	MetricBreakdown1	Any breakdown of the main metric into subcategories with their own values, such as by different ethnic groups or modes of birth	String
All	MetricBreakdown2	Any further breakdown of the metric subcategory, for example, different modes of birth within each ethnic group	String
All	Denominator	The total population of people or events which were relevant for the metric being calculated	Numeric
All	Numerator	The count of people or events from the denominator, which met the specific criteria for the metric	Numeric
MSDS and CQC	Percentage	Where applicable: the reported percentage for the metric	Numeric
MSDS and MBRRACE-UK	RatePer1000	Where applicable: the reported rate per 1,000 for the metric	Numeric
MBRRACE-UK only	RatePer100000	Where applicable: the reported rate per 100,000 for the metric	Numeric
MSDS only	DQResult	Where applicable: a Pass/Fail flag to indicate if the trust’s data for the metric passed a data quality check. Values are then only shown for the metric where the data quality checked was Passed. Currently applies to rates of Preterm birth and Postpartum Haemorrhage from MSDS	String
MBRRACE-UK and CQC	CILowerLimit	Where applicable: the minimum value in a calculated range (confidence interval) where the value is likely to fall, with 95% certainty	Numeric
MBRRACE-UK and CQC	CIUpperLimit	Where applicable: the maximum value in a calculated range (confidence interval) where the value is likely to fall, with 95% certainty	Numeric