Perfect HCC Risk Adjustment Scores: ForeSee Medical

Perfect HCC Risk Adjustment Scores: ForeSee Medical

For around 15 years, Medicare Advantage (MA) coding has been hampered by the administrative burden related to the HCC-RA model but now with new Artificial Intelligence (AI) tools, relief has arrived!

ForeSee Medical is a company that uses AI-related sciences such as Natural Language Processing (NLP) and Machine Learning (ML) to improve the workflow related to the HCC-RA model. Their disease detection algorithms and machine-learned natural language processing help to perfect the Risk Adjustment Factor (RAF scores) and reduce the time it takes to code clinical notes. ForeSee Medical uses NLP to discover diseases from text notes(like PDFs), and data (lab results, vital signs, medication, and problem lists) already stored in provider EHR systems.

On March 28, 2003, CMS announced the proposed final version of the CMS-HCC Risk Adjustment model (HCC-RA) for use in payment for MA patients with implementation on January 1st, 2004. The adoption of HCC-RA fundamentally changed the way payers and providers that participate in MA plans are compensated. The model closely ties reimbursement to accurate documentation and coding of a population’s disease burden. In fact, CMS has a scoring system to track disease burden, the RAF score, the higher the score the higher the reimbursement. HCC-RA commands a lot of attention because payers and providers need CMS to afford them the proper financial resources required to care for their patients. The model has become widely accepted and now CMS and other payers are using that reimbursement model in many other programs such as Accountable Care Organizations (ACOs), the Affordable Care Act (ACA), Primary Care First, and Geographic Contracting.

The HCC-RA model requires ICD coding skills that many physicians have not been trained to perform, and even for professional coders it requires an enormous amount of time per patient. In addition to coding their own patient encounters, primary care providers now have the additional burden of discovering and coding patient disease in consult notes, radiology reports, and lengthy hospital discharge summaries. Dr. Seth Flam, Co-Founder, and CEO state that “the labor required to perfect RAF scores has increased the administrative burden related to caring for MA patients and has proven to be an overwhelming drain on medical group resources”.

HCC-RA is mostly performed retrospectively. Since payers don’t always trust the coding skills of providers, they typically review the claims submitted by a medical group and then request that numerous charts are faxed for clinical review to assure that RAF scores are accurate. This retrospective chart review process adds to the expense and administrative burden for participants in HCC-RA healthcare delivery models.

ForeSee Medical, does the reading for providers using NLP, extracting clinical events and linking them to the original documentation. Since providers don’t all “speak the same way”, ForeSee software has their natural language processing (NLP) engine on a machine learning platform to accommodate the wide variation of medical notation terminology. Provider clinical notes (unstructured data) are “read” by the NLP engine, it then extracts the patient chart information, codifies it (structured data), and creates a longitudinal patient record. Every longitudinal patient record is matched against ForeSee Medical’s proprietary set of disease detection algorithms to accurately identify the patient’s disease burden.

ForeSee Medical is focused on discovering diseases and conditions that may not have been coded during the current year or even the prior year. Some software systems rely completely on the “recapture” of ICD codes that were used in a prior year to assist in documenting patient disease burden for the HCC-RA model. ForeSee Medical’s novel approach, focused on disease discovery, is more comprehensive and results in more accurate RAF scoring. In fact, if providers rely on recapture alone, they will inevitably under-document disease related to new patients or patients with progressive diseases like diabetes thereby underreporting RAF scores. Since the HCC-RA model relies on precise documentation of every patient's disease,under-documentation will result in suboptimal reimbursement. Disease discovery harnesses the power of the ForeSee Medical artificial intelligence engine and a complex set of proprietary algorithms, to discover diseases that HCC code recaptures alone simply never will. In fact, in a study of a well-performing organization after patient diagnoses were documented at annual wellness visits, the number of new disease suspects dwarfed recapture suspects by around a 4:1 ratio.

ForeSee Medical supports the old-fashioned retrospective workflow, but the good news is that the software also supports prospective (1 or 2 days before a patient encounter) and concurrent (at the point of care) workflows. The prospective model makes it easy for coders to help the physician plan for the upcoming encounter. Clinical decision support built into the medical group’s EHR is updated in real-time so it’s available and accurate when providers need it most - at the point of care, reducing the hassle related to HCC-RA.

ForeSee Medical’s HCC-RA software is expertly designed to help risk-bearing organizations thrive in the various CMS value-based risk-adjusted programs. Their proprietary disease detection algorithms and machine-learned natural language processing rationalize patient data across the healthcare system. Dr. Sol Lizerbram, Co-Founder and Executive Chairman state that “ForeSee’s novel use of Artificial Intelligence helps to perfect RAF scores and provide clinical decision support at the point of care by seamlessly integrating with modern EHRs”. We conclude the ForeSee software has significantly reduced the “hassle factor” associated with the CMS HCC Risk Adjustment Model.