Improving coding and primary care management for people with Chronic Kidney Disease: an observational controlled study in east London.

Background
Evidence from the UK national chronic kidney disease (CKD) audit in primary care shows that diagnostic coding in the electronic health record for CKD averages 70% with wide practice variation. Coding is associated with improvements to risk factor management, and there is evidence that CKD cases coded in primary care have lower rates of unplanned hospital admission.
Aim
To increase the diagnostic coding of CKD (stages 3-5) and key aspects of primary care management, including blood pressure to target and prescription of lipid lowering medication to reduce cardiovascular disease (CVD) risk.
Design and Setting
Prospective, controlled, cross sectional study set in four clinical commissioning groups (CCGs) in east London.
Method
Interventions to improve coding formed part of a larger system change to the delivery of renal services in both primary and secondary care in east London. These included dashboards with key performance indicators, searches to enable practices to identify un-coded cases and monitor progress, and safety tools to identify cases with a falling estimated glomerular filtration rate (eGFR). This was supplemented with practice-based facilitation on using the quality improvement (QI) tools, and with education sessions. Additional renal specific clinical facilitation was provided for practices in the lowest decile of CKD coding.
Quarterly anonymised data on CKD coding, blood pressure values and statin prescriptions were extracted from practice computer systems for a one year period pre and post the start of the intervention.
Results
All three intervention CCGs showed significant coding improvement over a one year period following the intervention (regression for post intervention trend p<0.001). The CCG with highest coding rates increased from 76% to 90% of CKD cases coded, the lowest coding CCG increased from 52% to 76%. The comparison CCG showed no change in coding rates.
Combined data from all practices in the intervention CCGs showed a significant increase in the proportion of cases with blood pressure achieving target levels (difference in proportion p<0.001) over the two year study period. Differences in statin prescribing were not significant.
Conclusions
Clinically important improvements to coding and management of CKD in primary care can be achieved by QI interventions which use shared data to track and monitor change supported by practice based facilitation. Alignment of clinical and CCG priorities and the provision of clinical targets, financial incentives and educational resource were additional important elements of the intervention.