Presentation 1

Title: The rise of the secure data environment: learnings from national and regional

Speakers: Rachel Denholm and Venexia Walker

Institution: University of Bristol

Abstract: The COVID-19 pandemic provided the opportunity for unprecedented access to
national electronic health record data for research in secure data environments. At the same
time, there has a been a push to develop regional secure data environments that contain the
granular information collected more locally. Jonathan Sterne, Rachel Denholm and Venexia
Walker co-lead a team of researchers at the University of Bristol who have worked within
these environments. During this seminar, Rachel and Venexia will share their learnings and
future directions for research using these resources.

Presentation 2

Title: A comparison of methods for externally validating the Kidney Donor Risk Index in the UK kidney transplant population

Speaker: Stephanie Riley

Institution: University of Plymouth

Abstract: Transplantation represents the optimal treatment for many patients with end-stage
kidney disease. The Kidney Donor Risk Index (KDRI) was developed to predict graft failure
following kidney transplantation. The survival process following transplantation consists of
semi-competing events, where recipient death precludes graft failure but not vice-versa. We
sought to externally validate the KDRI in the UK kidney transplant population, and assess
whether validation under a competing risks framework had an impact on predictive
performance. Additionally, we updated the KDRI using data from the United Kingdom to
explore whether this improved the predictive performance.

Using data from recipients of deceased donor single kidney-only transplants, held by NHS
Blood and Transplant, we externally validated the KDRI. Our outcomes of interest were one and five-year graft failure. Considering the semi-competing events, we modelled the
outcome in two ways: censoring the recipient at the time of death, and modelling death as a
competing event. Cox proportional hazard models were used to validate the KDRI when
censoring for death, and cause-specific Cox models were used to account for death as a
competing event. KDRI performance was assessed by discrimination, calibration, and overall
accuracy of predictions.


This event is organised by HDR UK South-West Training Lead, Yinghui Wei

Mailing list

To sign up to HDR UK South-West mailing list please complete this form