Computer assisted coding solutions adopted at many health systems across North America are not built for specialized clinical populations, like pediatrics. Learn how they are built, why they do not work for a pediatric context, and what the alternative is.
Computer assisted coding (CAC) software is a tool that reads clinical documentation and suggests applicable codes for approval by a medical coder. These tools can help health providers mitigate the increasing number of diagnosis and procedure codes that coders must use, and can further enable the workforce to focus on complex cases and clinical improvements.
As health information management (HIM) professionals seek to boost coding productivity, improve data quality and reduce denials, more providers are adopting CAC software. However, many of the CAC solutions that health systems are implementing rely on technology that is not built to meet the needs of their unique clinical setting.
CAC was readily adopted during the transition from ICD-9 to ICD-10 in 2015, which dramatically increased the number of diagnosis and procedure codes and the complexity of the coding. However, most CAC vendors today still rely on legacy technology that was built using rules and keyword-based natural language processing (NLP). These tools work by identifying specific terms in the physician notes and matching that verbiage with specific codes. In other words, they are built to search for terms but do not account for the clinical context of the notes. For example, if the clinical documentation said the patient had a “heart attack”, the software would suggest the respective code. However, if the documentation mentioned that the heart attack was ruled out or still being investigated in another location of the chart, the CAC would struggle to modify its prediction of this code according to this context.
The challenge with this legacy model of building a CAC system is that to function well in every clinical setting, tens of millions of rules would need to be established. Achieving this feat is impossible. As a result, these solutions have made concessions to fit generic contexts as best as possible. While the technology behind legacy CAC solutions still falls short in a generic context, in a specialized setting, these shortcomings can have a significant impact on coding workflows.
Pediatric hospitals’ unique patient population differentiates their clinical documentation from that of other health providers. Physicians may use dissimilar terms, the coding rules vary, and consequently, different diagnosis and procedure codes are relevant for their individual contexts. What works well in the adult setting may not be valid in pediatrics. As a result, legacy CAC solutions that have been built to meet the coding demands of adult medical specialties will not operate well in a pediatric context.
In our conversations with children’s hospitals across North America, the same issues were raised again and again: the legacy CAC solutions that they used were not tuned for their unique patient population. As a result, the rules-based (and even machine learning) technology often provided incorrect suggestions that did account for the clinical context and conflicted with pediatric coding guidelines. The tools also suggested adult codes where pediatric alternatives exist.
It is clear that these CAC solutions do not function as intended. For instance, if the clinical notes reported that a child’s mother smoked, the tool would identify “smokes” and suggest a code that the baby smokes as well. For a tool intended to streamline and speed up medical coding workflows, these recurring errors and inappropriate suggestions fail to help HIM professionals and health providers achieve their goals. Worse yet, they can also lead to increased denials if coders start incorrectly accepting autosuggestions.
Improper solutions perpetuate the operational efficiency issues that existed pre-CAC: extreme backlog, very expensive overtime bills, and low-quality work-life balance for employees. Not to mention the negative impact of miscoding that can reduce reimbursements or the continued inability to use AI for important secondary use cases, like population health management, clinical research, and wastage in revenue cycle management. To address these challenges, health systems cannot continue to rely on the same tools that they always have.
Health systems are different and to meet these varied coding needs an improved tool is necessary. That is why Semantic Health developed an AI-powered concurrent coding and auditing platform. Where CAC tools function based on a long series of rules, the AI-powered platform uses deep learning and natural language processes to accurately digest clinical information and suggest medical codes based on the context. To do this successfully across specialized providers, the algorithm is trained on data from each context, like that of pediatric health providers. Our model is tuned based on how coding is done in a pediatric setting using pediatric-specific documentation nuances and coding directives in our predictions to make them more accurate. This drives better coding outcomes across care settings and offers the full benefits of assisted coding to pediatric providers.
The Semantic Health Information Platform also accounts for the drawbacks of disconnected coding workflows associated with CAC systems that require medical coders to consult multiple sources to accomplish their work. In our platform, all insights are presented to medical coders in a single, intuitive user interface to improve coding efficiency and accuracy. We also highlight the clinical evidence that supports every code suggestion so that coders do not have to review the full documentation and notes can be checked more efficiently. We simplify the manual coding process so that everyone can focus on what’s most important: ensuring, delivering, and documenting top-quality care for those who need it.
With an AI medical coding tool, health systems can fix their current processes to unlock key benefits:
We welcome any questions and are happy to discuss the unique needs of your health system. Reach out to email@example.com to learn more.
Semantic Health helps hospitals and health systems unlock the true value of their unstructured clinical data. Our intelligent medical coding and auditing platform uses artificial intelligence and deep learning to streamline medical coding & auditing concurrent with patient admission, improve documentation quality, optimize reimbursements, and enable real-time access to coded data for secondary analysis.