Data and linkage

 

Obtaining access to linked data has been a recurrent challenge when evaluating integrated care initiatives like Principia’s enhanced support. By their very nature, these initiatives involve practitioners working together across organisational boundaries, but the data on hospital, general practice, community and social care are held separately and are not usually linked. To conduct this analysis, we needed to find a way to follow individuals as they moved between care homes and hospitals, while maintaining patient confidentiality. The situation was complicated by the need to include residents who paid for their care home stay themselves, without local authority support, since they are still eligible to receive the enhanced support, yet might not feature in local authority databases.

The Improvement Analytics Unit has access to pseudonymised data from the Secondary Uses Service (SUS) – a national, person-level database that is closely related to the widely-used Hospital Episode Statistics (HES). SUS contains information on A&E attendances, inpatient admissions and outpatient appointments that are funded by the NHS in England, but unfortunately it does not record accurately whether an individual resides in a care home. Therefore, the unit needed a new database containing information on the residents of care homes, and a method of linking that data to SUS.

The care home resident database was assembled by a team at the Arden & Greater East Midlands Data Services for Commissioners Regional Office (Arden & GEM DSCRO). The process began with extracts from the National Health Applications and Infrastructure Services (NHAIS) database, which contains information on all registrations with general practices in England. The data in question were monthly snapshots, containing a list of individuals registered at each general practice, including their residential address and NHS number. The addresses from these NHAIS extracts were cross-referenced with a database of care homes obtained from the Care Quality Commission (CQC). This enabled the Arden & GEM DSCRO to construct a merged database containing individuals who moved into care homes, based on whether they changed their recorded address to one that matched the location of a care home. The database covered care home stays that began between August 2014 and July 2016. It included the month during which the individual moved into the care home and (where applicable) the month during which they left and the reason for leaving (ie, whether they had moved elsewhere or died). The merged database also contained some limited information about the care home itself, such as whether it offered nursing care in addition to residential care, and how many beds it contained.

The database of care home stays was pseudonymised before it was transferred to the Improvement Analytics Unit. This meant that names, addresses and full dates of birth were removed, and the NHS number replaced by a linkage key. That key was used to link the database of care home stays to the SUS data that the Improvement Analytics Unit already held. The linked care home and hospital data were analysed within an accredited secure data environment based at the Health Foundation. The overall approach to information governance was scrutinised by the project Oversight Group and information governance experts at NHS Digital. The Data Access Advisory Group at NHS Digital recommended that the data be provided for this project. At no point did the Improvement Analytics Unit have access to identifiable data. Throughout, the minimum amount of data was used.

The process resulted in a truly unique database, containing the hospital histories of every care home resident in the selected areas. At a time when information about the quality of care offered to care home residents can be lacking, the database provides valuable insights. To give one example, Figure 1 shows how often care home residents were admitted to hospital each year for certain conditions. The rationale of looking at admissions is that, for some conditions, such as fractures and sprains, admissions are often preventable (see Box 2). In other cases, such as pneumonia, admission rates might be reduced by making improvements to the care offered. Therefore, the linked data sets can be revealing as to the quality of care offered to care home residents.

Figure 1: Rates of hospital admissions for care home residents per year

Box 2: Conditions for which we considered emergency admissions to be potentially avoidable

The analysis included conditions that are often manageable, treatable or preventable in community settings without the need to go to hospital, as well as those that may be caused by poor care or neglect. These conditions were:

  • acute lower respiratory tract infections, such as acute bronchitis
  • chronic lower respiratory tract infections, such as emphysema
  • diabetes
  • food and drink issues, such as abnormal weight loss and poor intake of food and water, possibly due to neglect
  • fractures and sprains
  • intestinal infections
  • pneumonia
  • pneumonitis (inflammation of lung tissue) caused by inhaled food or liquid
  • pressure sores
  • urinary tract infections.

To calculate the number of potentially avoidable emergency admissions, the Improvement Analytics Unit counted admissions with one of these conditions as the primary diagnosis for the first consultant episode of the hospital spell. Note, however, that this list of conditions was originally intended to be applied to the general population aged 65 or over, rather than to care home residents. Within the timeframe of this study, it was not possible to validate the appropriateness of these conditions for the care home population. Sometimes individuals need to be admitted to hospital for these conditions, regardless of the quality of the care offered. The metric is not perfect but we would expect the enhanced support to have greater impact on admissions for these conditions than others.


Pseudonymised data sets have been stripped of identifiable fields, such as name, full date of birth and address. However, a unique person identifier (such as NHS number) has been replaced with a random identifier. The scrambled version of that field is used as the ‘key’ to link data sets together. For this analysis, the scrambled version of the NHS number was used to link together hospital records for the same individual over time.

Except for December 2014 (see technical appendix for details: www.health.org.uk/publication/impact-enhanced-support-rushcliffe).

§ For example, rather than using national data, the work was restricted to certain local areas (see ‘Selecting a comparison group’).

Previous Next