The indicators on this site have been calculated using individual patient Hospital Episode Statistics (HES). HES is a data ‘warehouse’ containing patient-level data on hospital admissions for all NHS trusts in England. HES is also designed for use in research and service evaluation. HES data were made available by the NHS Health and Social Care Information Centre (HSCIC)(Copyright 2012, reused with the permission of the HSCIC. All rights reserved)
How we selected our sample
The indicators are based on deliveries occurring between April 2013 and March 2014. With one exception (unplanned neonatal readmissions within 28 days), all statistics are based on the number of women giving birth (‘deliveries’) rather than the number of babies born.
After removing duplicate records, we restricted the sample to women aged between 15 and 45 with singleton, term deliveries. Concentrating on this group allows attention to be focused on the group of women whose maternity care is most affected by clinical uncertainty and which varies the most between providers. Most indicators are then split by parity as this has a major influence on pregnancy and delivery outcomes. Additional exclusions have been applied to each indicator to focus on the appropriate group of women ‘at risk’ of the outcome of interest, as detailed here.
Data is presented at trust level because, at present, there's no reliable way to separate out the information for the labour ward and alongside or freestanding midwifery-led units.
This site presents indicators for each English NHS trust that met our minimum data standards. More information about the data quality checks that we carried out is available to download here.
Adjusting for risk factors
When presenting figures for individual NHS trusts, indicators must take into account differences between the women who give birth at each organisation. The clinical and demographic characteristics of women (case-mix) can affect both the demands placed on the service and the outcomes of care. Risk-adjustment was performed using a multivariate regression model that removed the effect of age, ethnicity, level of socio-economic deprivation and clinical risk factors. You can read more about the risk adjustment process here.