What do these two terms mean – HCC and AI?
In recent years, most people in the healthcare industry have heard of HCCs (Hierarchical Condition Categories). HCCs are conditions and diseases that are grouped together into categories by body system or disease process to measure the health status of a patient. In medical coding, this is known as Risk Adjustment. The current Center for Medicare & Medicaid Services (CMS) HCC model includes nearly 10,000 ICD-10 codes that map to 79 HCC Categories.
CMS utilizes HCCs to reimburse Medicare Advantage plans based on the health status of their enrollees. The plan pays for the predicted healthcare costs of a patient or population of patients. Costs are estimated based on the demographical profile and the health status of the patient. The data that is utilized to determine health status is based on the diagnoses codes from billed claims and medical records which are collected by physician offices, hospital inpatient visits, and outpatient settings.
In order to categorize HCC data, we need to look at qualitative text data that is usually present in medical coding and survey methodologies.1 HCCs are dependent upon the assigned risk scores from ICD-10 coding and demographic information and in order to determine those costs a RAF (Risk Adjustment Factor) score is calculated for that patient.
Insurance companies use algorithms to predict the cost of care for patients. For example, a patient with few serious health conditions could be expected to have average medical costs for a given period of time. However, a patient with multiple chronic conditions would be expected to have higher utilization of services and a corresponding higher rate of health care costs. 2
How does AI factor into HCCs?
The term AI (Artificial Intelligence) is becoming a more commonplace and recognized term. AI utilizes machine learning to simulate human processes. Using , AI makes human processes more efficient. As AI relates to HCCs, machine learning natural language processing, disease detection algorithms, and suggestion scoring are terms for the technical processes that are utilized to identify otherwise unfound data in a patient’s history that help assist in determining the RAF score.
At the end of the day, many companies are limited by time and resources when analyzing and optimizing RAF scores. An established patient may have a new diagnosis for this current year, but it is possible that the prior year’s diagnoses have progressed and should be coordinated to a new non-progressed diagnoses and be recaptured. This over-time look at the patient’s history creates opportunities that would have been previously missed.
Natural Language Processing is one of the must-have tools for disease discovery. By leveraging the natural language processing for chart abstraction, we improve the efficiency of identifying underlying disease burden 3
The Future of HCCs
HCCs are the future of better health management. Medical coding professionals should be sure they stay up to speed on the latest advancements in coding in order to ensure that the level and intensity of the services provided to patients is aligned with the reimbursement. CMS has some great HCC software that ensures appropriate diagnosis codes are reported for the patient. Companies that utilize natural language processing software have introduced new tools that can discover insights previously locked in unstructured data like surveys in the form of PDFs and text notes. 4
As we move forward, let’s consider the complete picture of the risk adjustment factors and how the HCC score reduces the need to request additional medical details and audits of claims. If you want to know more about reimbursements and HCCs or you currently participate in a value-based reimbursement model. Let us know and we can help you get a handle on high-quality care and sustainable profits.