Integrated Care Board mortality method

Introduction

This section describes the approach used by the Office for Health Improvement and Disparities (OHID) for the production of mortality indicators at Integrated Care Board (ICB) level, for example in the Public Health Outcomes Framework (PHOF).

Integrated care boards (ICBs) have geographical boundary definitions, which are being refined each year. They can be defined precisely at the lower layer super output area (LSOA) level, so starting with record-level mortality data, it is possible to calculate precise mortality rates for each ICB. The majority of ICBs are coterminous with lower tier local authorities (LTLAs) (either singly or groups of LTLAs). In some cases, however, ICBs are not coterminous, leading to the possibility of slivers, that is to say, areas with small populations that can be identified by differencing LTLA and ICB values. In these cases it is incumbent on statistics producers to ensure that no potentially disclosive small counts can be derived from the data published.

Method

While ICB boundaries are increasingly aligning with LTLA boundaries, they are not yet established and continue to change. For this reason, a pragmatic approach has been taken to making sure that the risk of disclosing small counts by differencing is minimised.

Where ICBs’ boundaries are coterminous with LTLAs, the ICB values are calculated normally, from record-level data.

Where the ICB boundaries and LTLA boundaries do not align precisely, the approach taken depends on the extent of the difference:

1. Where the boundaries are very different, and the ‘slivers’ between the different boundaries are in fact not slivers but substantial populations, there is no risk of small numbers being disclosed. In these cases, the ICB values are calculated normally, from record-level data. The local authorities which cross ICB boundaries and for which values have been calculated precisely and independently are:

• East Suffolk (NHS Norfolk and Waveney ICB and NHS Suffolk and North East Essex ICB)
• North Yorkshire (NHS Humber and North Yorkshire ICB and NHS West Yorkshire ICB)
• Waverley (NHS Frimley ICB and NHS Surrey Heartlands ICB)
• Westmorland and Furness (NHS Lancashire and South Cumbria ICB and NHS North East and North Cumbria ICB)

2. Where the differences between the boundaries are small, creating potentially disclosive slivers, the ICB boundaries for this analysis have been aligned to the LTLA boundaries by reallocating LSOAs to the appropriate neighbouring ICB. Because the populations that have been reassigned to neighbouring ICBs are very small compared with ICB populations, the effect on the indicator values for ICBs is also very small.

The cut-off used to decide which of the above approaches applies is a population of 40,000. This was chosen because it is approximately the population of Rutland, the smallest LTLA for which these indicators are routinely produced. City of London and Isles of Scilly are smaller and are combined with Hackney and Cornwall respectively to avoid disclosure.

Details of affected ICBs

The following ICBs (2022 definitions) have been aligned to LTLA boundaries (2023 definitions) for the purpose of producing mortality indicators as set out above. The populations quoted are ONS 2022 rebased mid-year estimates.