Artificial Intelligence, Natural Language Processing, & Experienced Coders – Accurate Risk-Adjusted Payments
Artificial Intelligence, Natural Language Processing, & Experienced Coders – Accurate Risk-Adjusted Payments
On average, organizations managed 9.70 petabytes (PB) of data in 2018; this is an explosive growth of 569 percent compared to the 1.45 PB managed two years prior (Dell Technologies, 2019). Healthcare organizations are no different. In fact, according to HIT Infrastructure, healthcare organizations have seen a health data growth rate of 878% since 2016, reaching 8.41PB on average by 2018 (2019). As you can imagine, it is physically impossible for a human to read through and analyze the tremendous amount of data collected. Healthcare organizations are rapidly turning to Artificial Intelligence (AI) and Natural Language Processing (NLP). This helps them collect, transform, understand, and drive decisions in a timely manner.
Diagnoses codes are not enough
Diagnoses codes are not enough to provide a complete picture of a patient’s health. This lack of information leads to inaccurate calculation of patient’s risk scores, and potentially, are the cause for lower reimbursements. AI or Natural Language Processing (NLP) helps capture appropriate HCC codes for more accurate reimbursement. In general terms, the healthcare industry is relying more on AI or NLP to aid with ‘unstructured’ data, like patient’s clinical notes. AI and NLP collect and process this data much faster than a human ever could. This means extracting patient chart information, codifying it (converting it unto structured data), and creating a longitudinal patient record. This allows organizations to drive insights and improve decision making.
Powerful Engine that learns from Coders
While NLP and/or AI act as a massive collector and filter for accurate HCC diagnoses, as HCC consultants, we have seen countless times, the need for a certified coder to review the documentation of those encounters “flagged” by these powerful engines that may have the appropriate documentation to substantiate an HCC diagnosis. The end goal of AI and medical coding is to accurately identify whether it is a reportable HCC diagnosis. An experienced coder can utilize this filtered information and confirm the engine’s findings per the documentation and M.E.A.T. criteria, as well as to discard machine-driven mistakes; an example of this is a document mentioning a medical term only as past medical history. This review and decisions made by the coder help in training these engine (machine learning) algorithms. This allows the algorithm improve over a period of time, and become more accurate in time.
As consultants, we must always strive to bridge the gap between the machine’s capabilities, and the human (experience) factor. This can positively improve coder’s efficiency as well as the risk adjustment department’s effectiveness, allowing health plans to take care of patients in a holistic manner while obtaining true reimbursements (Healthpayer Intelligence, 2019).
This article was co-authored by Karen Youmans, President/CEO of YES HIM Consulting.
