# Quantifying the impact of clinical coding in chronic kidney disease on risk of death and COVID-19 death

## TL;DR

This study shows that patients with chronic kidney disease who are properly coded in medical records have a lower risk of death and COVID-19 death compared to those who are not coded.

## Contribution

The study introduces evidence that clinical coding for CKD stages 3–4 is associated with reduced mortality and COVID-19 death risks.

## Key findings

- Coded CKD stages 3 and 4 were linked to significantly lower adjusted hazards of death and COVID-19 death compared to uncoded CKD.
- Propensity score matching and competing risk regression confirmed the association between clinical coding and lower mortality risks.
- CKD stage 5 could not be analyzed due to insufficient data on uncoded cases.

## Abstract

Patients with biochemical evidence of chronic kidney disease (CKD) without a diagnostic code (uncoded CKD) in primary care are at increased risk of death, acute kidney injury (AKI), and unplanned hospital care. Uncoded CKD is highly prevalent and there is no data to evaluate whether patients with uncoded CKD were at an increased risk of COVID-19 death. Aim: to assess whether patients with uncoded CKD stages 3–5 were at increased risk of death and COVID-19 deaths.

Descriptive and inferential analyses to measure adjusted hazard of death, and COVID-19 death in patients with CKD stages 3–5 from 2.85 million primary care patients in Greater Manchester, England. Sensitivity analyses using propensity score matching and competing risk regression.

Coded CKD stages 3 and 4 (versus uncoded) were associated with significantly lower adjusted hazards of death (HR 0.81, CIs 0.77–0.86, p=<0.0001; HR 0.45, CIs 0.34–0.60, p=<0.0001, respectively), and COVID-19 death (HR 0.74, CIs 0.55–0.99, p = 0.03; HR 0.55, CIs 0.30–0.99, p = 0.045, respectively). Descriptive analyses were conducted for patients with CKD stage 5 due to low numbers of patients with uncoded CKD stage 5, precluding survival analyses.

Our retrospective cohort study suggests that clinical coding is a digital intervention associated with a lower adjusted hazard of death and COVID-19 death in patients with CKD stages 3 and 4, and should be considered a key element in the organisation and delivery of care for people with CKD.