Progressive Kidney Disease Identification Study - ProKIDNI Study

  • Research type

    Research Study

  • Full title

    The use of machine learning to identify patients with rapidly declining chronic kidney disease

  • IRAS ID

    306690

  • Contact name

    Umanath (Nitin) Bhandary

  • Contact email

    nitin.bhandary@royalberkshire.nhs.uk

  • Sponsor organisation

    Royal Berkshire NHS Foundation Trust

  • Duration of Study in the UK

    0 years, 11 months, 29 days

  • Research summary

    Chronic kidney disease (CKD) is a condition characterised by a reduction in kidney function. Earlier detection of CKD allows for modification of risk factors to either slow or prevent decline and allows for timely referral to specialist renal services by primary care. For those patients in whom renal function declines rapidly, earlier counselling and patient education is associated with improved health outcomes and patient choice with respect to options such as home dialysis treatments and pre-emptive transplantation.
    There has been an increasing amount of interest in developing tools to help with prognosis of CKD, as well as predicting the likelihood and time of requiring renal replacement. Our findings suggest an overabundance of traditional approaches, with an increased focus on effect sizes and hazard ratios, while at the same time a general lack in attempts to train state-of-the-art machine-learning algorithms (MLAs) that could automate the process of identifying patients at risk of developing end stage kidney disease (ESKD).
    We have designed a suitable methodology for the study. We propose running a pilot study, which will focus on developing a set of MLAs using available laboratory parameters routinely collected by renal physicians at RBH.

    The MLAs could subsequently be used to identify patients at greatest risk for ESKD. These patients could be selected for early referral to specialist renal services and intensive counselling and education (if already under the care of a renal unit). The preliminary outcomes of the pilot study will lead to the development of machine learning architecture, with the prospect of following up with larger-scale data collection and analytics. The model predictions will be validated by clinicians.

    The model that we propose, aims to identify a predicted date for onset of ESKD, which could be more useful to clinicians than other risk modelling tools when planning patient care.

  • REC name

    South Central - Berkshire Research Ethics Committee

  • REC reference

    22/SC/0038

  • Date of REC Opinion

    24 Feb 2022

  • REC opinion

    Favourable Opinion