Automated clinical coding at Kettering General Hospital

Challenge

A key administrative task in the NHS is clinical coding. During this process, a patient’s diagnosis, treatment or reason for seeking medical attention is captured from a clinician’s notes and translated into codes. These codes must then be submitted by the care provider before they can be reimbursed by the NHS. Accurate clinical coding also helps the government monitor population health and clinical practices.

Kettering General Hospital (KGH), part of the University Hospitals of Northamptonshire NHS Group wanted help with their clinical coding process, as they struggled to recruit and retain experienced coders. The trust also had to balance throughput with accuracy, as processing higher volumes of codes could lead to lower quality.

Solution

Working with the Clinical Coding team at KGH, we began the UK's first Automated Clinical Coding project for Inpatients and Day patients, using AI to automatically assign ICD-10 and OPCS 4 codes to each episode of a patient’s care. Starting in Endoscopy, we built an automation program to identify, verify and submit clinical codes from patient notes.

Identifying and mitigating risks early on allowed us to define best practices for the automated clinical coding system, ensuring we delivered the needs of all stakeholders. We also established a process for iterative learning, where the AI could learn from previous examples. This reduced the number of exceptions — instances that the system cannot automatically process — and improved the speed and accuracy of coding. 

Impact

Using an automated clinical coding service in Endoscopy helped KGH to achieve several benefits. As well as releasing staff time equivalent to one full-time employee per year, with a saving of over £35k, the likelihood of human error was greatly reduced. Thanks to automation, staff roles now focus on validating and auditing complex cases, opening up more opportunities for them professionally and helping the NHS retain its employees. 

The flow of data within clinical coding processes was also improved, giving a more detailed picture of the care the trust provides, ensuring it is fully reimbursed for that care, and enabling accurate healthcare data to be provided to local government services.

As well as automating the end-to-end process in Endoscopy, we delivered a recommended roadmap for automation in other areas of the trust. This will help KGH gain more efficiencies, identify potential secondary conditions in patients, and improve its clinical outcomes.