Utilising longitudinal biomarker data to improve risk prediction and follow-up scheduling in patients with Chronic Kidney Disease

Doctors currently guess how often kidney patients need blood tests, and often schedule them too frequently or not often enough. Chronic kidney disease affects roughly one in ten UK adults, and can silently progress to kidney failure requiring dialysis or transplant. Existing risk calculators only use a patient’s most recent test results, ignoring the rich history of past measurements that could reveal whether the disease is accelerating or stable. This project will build a statistical tool that uses a patient’s entire sequence of blood and urine test results to predict their future risk of kidney failure more accurately, and then recommend the optimal date for their next appointment. If successful, the tool could replace one-size-fits-all follow-up schedules with personalised intervals—some patients might safely wait six months, while others need checking every two weeks. For the NHS, this means fewer unnecessary appointments for stable patients and earlier detection of deterioration in those at high risk, reducing emergency dialysis starts and hospital admissions. The research is applied and data-driven, drawing on existing longitudinal datasets rather than new laboratory work, so results could be integrated into clinical software within a few years.