Automated clinical coding: what, why, and where we are?

Abstract

Clinical coding is the task of transforming medical information in a patient’s health records into structured codes so that they can be used for statistical analysis. This is a cognitive and time-consuming task that follows a standard process in order to achieve a high level of consistency. Clinical coding could potentially be supported by an automated system to improve the efficiency and accuracy of the process. We introduce the idea of automated clinical coding and summarise its challenges from the perspective of Artificial Intelligence (AI) and Natural Language Processing (NLP), based on the literature, our project experience over the past two and half years (late 2019–early 2022), and discussions with clinical coding experts in Scotland and the UK. Our research reveals the gaps between the current deep learning-based approach applied to clinical coding and the need for explainability and consistency in real-world practice. Knowledge-based methods that represent and reason the standard, explainable process of a task may need to be incorporated into deep learning-based methods for clinical coding. Automated clinical coding is a promising task for AI, despite the technical and organisational challenges. Coders are needed to be involved in the development process. There is much to achieve to develop and deploy an AI-based automated system to support coding in the next five years and beyond.

A survey on clinical natural language processing in the United Kingdom from 2007 to 2022

“…Advances in the above three technical objectives in particular ( Disease information and classification, Diagnostic Surveillance and Cohort Building for Epidemiological Studies ) offer a great opportunity for health systems to harness data from unstructured EHRs for better care. In addition, clinical NLP has great potential in (semi-)automated clinical coding for timely and more accurate auditing, surveillance and public health policing 141 . However, at the writing of this review, developments of automated coding are still in their infancy in the UK.…”

“…The deployment of text analytics capabilities with health systems has shown its great potential in facilitating more efficient and cost-effective clinical trials 81 , 147 . Another operational development is the use of clinical NLP models for facilitating efficient medical coding 141 : funded by NIHR recently as an AI Award, University College Hospital colleagues have been comparing 148 , 149 for automatically assigning ICD-10 codes for hospital admissions.…”

Automating the overburdened clinical coding system: challenges and next steps

“…This sizable market has stimulated the race to create the first widely adopted ACC model. Several large technology companies have already created semi-ACC systems, including Deloitte, Optum, and Capita 3 . Start-up AKASA recently created an ACC solution that outperformed human coders on the MIMIC-III dataset 17 .…”

“…Coders require months of training and can code around 60 cases per day. Even at this rate, cases pending coding can be backlogged by months 3 . Moreover, the manual coding process is prone to errors—accuracy ranges widely (50–98%; median of 80%) depending on the coder, diagnosis/service, patient complexity, etc 4 , 5 .…”

“…Given the language-based, pattern-heavy, data-driven nature of coding decisions, AI and NLP offer the promise of ACC to support coders. Dong et al recently analyzed the technical challenges of ACC and proposed future directions 3 . They also discuss the most accurate ACC system to date, which used a benchmark dataset (MIMIC-III) of US intensive care documents and ICD-9 codes 6 .…”