An association of health information managers, clinical coders, and clinical documentation specialists in Australia has released a guideline for adopting clinical coding AI solutions.
The guideline developed by the Health Information Management Association of Australia (HIMAA) emphasises a principle-based approach to taking up AI technologies to generate clinically coded data. It is meant for healthcare organisations, the clinical coding workforce, software companies, government agencies, users of coded data, and educators.
It outlines seven key considerations: governance, risk management, privacy and security, ethical and safe use, quality improvement, collaboration and partnership, and human expertise or human-in-the-loop.
WHY IT MATTERS
HIMAA recognises that AI offers “significant potential to enhance clinical documentation integrity, enable autonomous coding, support clinical coding audits, and improve health information management.”
It also sees how the role of clinical coders is evolving, “requiring new skills in AI oversight, exception handling, and algorithmic bias identification.”
“The adoption of this guideline is expected to lead to safer AI adoption by the sector, compliance with relevant legislation, standards and frameworks, and increased workforce readiness while confirming the role of the clinical coding workforce in AI environments,” HIMAA said, explaining its development of the AI adoption guideline.
The association urges health organisations to “go beyond technical deployment and consider the broader ecosystem of governance, workforce, and regulatory expectations.”
“Ultimately, the successful application of AI-driven clinical coding relies on balancing automation with human oversight, integrating governance with innovation, and ensuring compliance while optimising efficiency and accuracy,” HIMAA underscored.
Additionally, the association emphasised the need to invest in “scalable IT infrastructure,” which includes cloud-based processing and cybersecurity measures, to support AI’s long-term sustainability.
Meanwhile, the HIMAA guideline does not provide specific guidance on adopting AI for clinical documentation improvement, nor does it address technical specifications, IT infrastructure needs, interfacing specifications, AI product selection recommendations, or other related AI solutions such as ambient clinical documentation. Nonetheless, its principles may be applicable to these areas.
THE LARGER TREND
There are various benefits of implementing AI-assisted coding solutions; according to HIMAA, AI may help address coding workforce shortages and coding backlogs and accelerate the coding process. The technology may also be used to code a certain subset of data and perform routine or low-value tasks, so coders could work on higher-value tasks.
Northern Health in Victoria, for example, has recently transitioned to AI-assisted coding to mainly address its shrinking coding workforce. Te Whatu Ora Health New Zealand, meanwhile, is seeking to pilot an AI-assisted clinical coding solution for hospital admissions to enhance the speed and accuracy of delivering care.
However, HIMAA noted that there has yet to be published evidence of these potential benefits in the Australian healthcare context.
The association also noted common barriers to the uptake of AI coding solutions. At the early stage of adoption, these include integration challenges, investment and cost, lack of trust, risk appetite, change management capability, AI model readiness, workforce readiness and training, cybersecurity and data privacy risks, the lack of digitised medical records, and regulatory and governance considerations.
——————————————
