Unplanned dialysis starters: an audit of primary care management using linked electronic health record data.

Problem

In the UK, the majority of people with chronic kidney disease (CKD) stages 3-5 are managed in primary care. In this setting progression to end stage renal failure (ESRF) is uncommon. The identification of progressive CKD and timely referral to nephrology services requires regular review and is supported by NICE guidance. In east London the rates of ESRF and numbers on renal replacement therapy (RRT)/dialysis are higher than the England average, with up to 30% being late presentations and referrals. Late presentation is associated with excess morbidity and mortality. Planned RRT with patient education, modality choice and access preparation through hospital low clearance clinics leads to better patient experience and outcomes.This study aims to compare patient characteristics and primary care management between those with unplanned dialysis starts and those with planned RRT using linked primary and secondary care records.

Approach

The audit cohort comprised people starting renal replacement therapy at Barts Health NHS Trust renal unit identified retrospectively over 3 years, searching in August 2017. Each case was classified into planned, unplanned or excluded. Exclusion criteria included transfers from other hospitals, dialysis independent acute kidney failure and those with previous kidney transplants. Cases registered with a GP practice in inner east London CCGs (City and Hackney, Newham and Tower Hamlets) were pseudonymised and linked to the primary care electronic health record via the North and East London Commissioning Support Unit (NELCSU). Demographics, GP consultations and clinical values in previous two years, co-morbidities and prescribing in the previous six months were extracted from the primary care records using the RRT start date as the index date.

Findings

Of the 999 RRT cases identified from September 2014, 292 (29%) were classified as unplanned starters; 463 (46%) as planned and 244 (24%) were excluded. There were statistically significantly more deaths, as a crude proportion, in the unplanned starter group (p 0.01). 465 cases from the inner London CCGs were linked to primary care data. The proportions of linked cases in the unplanned, planned and excluded groups were similar to the whole sample. Further analysis will include comparative data for each group and using a logistic regression model we aim to identify predictors of unplanned starters from the primary care data.

Consequences

The identification of preventable factors in primary care settings for unplanned dialysis starters will contribute to efforts to prevent or delay the onset of ESRF.