Linkage of a national clinical liver transplant database with administrative hospital data: methods and validation

Abstract

Introduction: The UK Liver Transplant Audit (UKLTA) database contains clinical information on all liver transplants carried out in the UK. To expand its potential for research and service evaluation, we linked it to the Hospital Episode Statistics (HES), an administrative database of all admissions to English National Health Service (NHS) hospitals.

Materials and methods: In the UKLTA database, we identified the linkable records of first liver transplantation between April 1997 and March 2010. We linked UKLTA records to HES records on the basis of NHS number, gender, date of birth, and postcode, as well as procedure codes for liver transplantation and dates of transplant. In linked records, agreement of primary liver disease diagnoses according to both databases was expressed as a proportion of the linked records and using kappa statistic.

Results: There were 5,815 linkable records in the UKLTA database, of which 4,959 records were successfully linked with HES (85.3%). Among these, 4,922 records (99.3%) had at least one diagnosis coded in HES relevant to an indication for liver transplantation. The overall agreement of primary liver disease diagnoses between UKLTA data and HES was 77.8% (95% CI 76.6%-79.0%) with a kappa of 0.75 (0.74-0.76). Diagnostic agreement can be further improved by using broader groupings of clinically related diagnoses.

Conclusion: Linkage of clinical data and administrative hospital data provides a rich resource for the study of liver transplantation.