In today’s healthcare environment where reimbursement is increasingly tied to risk adjustment, providers and health plans might be unknowingly overestimating the health risks within member populations, which can lead to disruptions to operations and compliance penalties. Elevated risk scores can garner higher reimbursements, but overestimating risk scores may result in costly compliance penalties if the practice is revealed during an audit.
To avoid these potential penalties, organizations should reevaluate their risk adjustment practices to pinpoint problematic processes and identify where technology can assist their efforts. The investments in time and resources to address these issues will not only improve risk adjustment accuracy, but may also elevate the level of care that members receive.
Under risk-adjusted models, there is an incentive to capture and report – and potentially over-report – chronic conditions that can increase reimbursement. But improper coding of conditions can occur even if not intentional. Claims submitted by providers may contain coding errors that will persist if not reviewed and corrected by the health plan. One common coding error that improperly inflates risk scores is when acute conditions are reported year after year despite their resolution. For example, a patient who sees his primary care physician for an ear infection (or other condition) a year after his stroke occurred cannot be coded as an acute stroke. Such “upcoding” would result in inappropriate overpayments to the health plan.
The government’s primary tool for holding plans accountable for their risk scores are Risk Adjustment Data Validation (RADV) audits. A recent RADV audit found that Medicare overpaid for nearly half a sample of patients enrolled at a subsidiary of UnitedHealth Group, the nation’s largest Medicare Advantage insurer¹. While overpayment recoveries have been modest so far, the government is placing greater pressure on CMS to expand the use and aggressiveness of audits. Penalties for improper coding can be up to three times the government’s damages caused by the violator. The Civil Monetary Penalty (CMP) may range from $5,500 to $11,000 for each false claim. And lawsuits and negative media attention damage organizations’ reputations and hurt their ability to attract and retain members.
How Technology Can Help
Technology can help organizations improve their risk adjustment and coding compliance in several ways. For starters, I recommend the following:
- Use natural language processing (NLP) technology to improve coding accuracy and standardize coding across the team. NLP – which can be used to analyze free text in medical records – is proven to effectively increase coding accuracy. Organizations should use an NLP solution with general precision and recall metrics in the 90-plus percentage range that is capable of continuous learning and improvement. In addition, technology-driven workflow solutions can reduce discrepancies in coding and help standardize practices across the team. It’s important to note that NLP technology isn’t intended to replace human coders, but rather to supplement their specialized skills and improve their workflow.
- Leverage NLP to efficiently manage a rigorous quality assurance (QA) process. A typical QA process consists of managers or internal “auditors” reviewing a random sample of 5 to 10 percent of coder outputs to check that accuracy levels are maintained. Rather than relying on a sample set, NLP can enable a much more targeted and streamlined review process. NLP can process coded documents and identify charts where there is high likelihood of unsubstantiated codes. Internal auditors can then spend their time reviewing high-risk documents instead of selecting a small subset to review.
- Leverage coder feedback to better identify and target educational opportunities. Organizations should make use of coder feedback for targeted provider education. Technology-enabled coding platforms should allow coders to flag instances where diagnoses are not sufficiently documented in the chart but where the patient likely does have the condition. Analytics can then be run to identify specific providers that most frequently under-document and/or specific conditions that are most frequently under-documented. Making use of this type of coder feedback helps to ensure that providers are continually pushed to document and code accurately and fewer risk conditions slip through the cracks.
- Use analytics to flag conditions, providers, and members with high audit risk, and take corrective action. Analytics platforms should be able to systematically analyze historical data with regard to specific error patterns and compliance gaps that must be addressed. This should include benchmarking providers’ coding patterns to identify any potential outliers across the provider network. These providers can then be targeted for further education and training. Similarly, plans should use analytics to identify potentially high impact members with greater audit risk. Several scenarios can make members more prone to audit risk, including the following:
Coding activity flags:
Health condition flags:
- 7+ HCCs
- Vascular disease
- RAF change of more than 1.0 from prior year
- Diabetes with complications
- Few delete records
- Major depression
- Acute conditions from non-acute settings
Regulators are increasing their scrutiny on risk adjustment compliance, especially in the Medicare Advantage space. Organizations must be prepared for increased audit activity and potential lawsuits. It is crucial to develop strong programs to ensure ethical participation and avoid compliance penalties aided by sophisticated, modern technology. Given the large amounts of money at stake – and not to mention organizations’ reputations – it’s better to be safe than sorry.