To err is human. Every person, no matter how talented or diligent, makes mistakes. Placed under stress or faced with disorganized or hard-to-locate information, people are even more likely to make errors. In most circumstances, the consequences of making the wrong choice or missing a key piece of information are minor. In healthcare, however, human error can have a significant impact on patient safety, organizational revenue, and regulatory compliance.

For health insurers and provider systems that support Medicare Advantage (MA) and Commercial Exchange plans, risk adjustment is a critical operational activity—and one that’s particularly susceptible to mistakes. Let’s explore why human error is such a risk for risk adjustment, and how healthcare organizations can manage these risks and gain a clearer view of their population’s health.

Why risk adjustment coding is prone to errors

HCC coding for risk adjustment programs is a complex, labor-intensive process. There are a variety of factors that introduce opportunities for overlooked information, incorrect codes, and inconsistencies. A few of the leading factors that cause errors are:

• Large volumes of data. Reviewing charts for even a small subset of members involves a ton of data. Let’s say a plan identifies a target list of 1,000 patients for HCC coding. If each patient has an average of 10 charts, and each chart has an average of 30 pages, that quickly adds up to 300,000 pages to review. Even when the labor is spread between a team of people, that’s a substantial amount of work.
• Constrained deadlines. For MA plans, annual Sweeps deadlines happen in January, March, and September. For Commercial Exchange plans, audits must be complete by May. Even though these deadlines happen annually, getting medical charts from providers in a timely fashion can be a challenge. Consequently, coding often happens on tight deadlines where quick review passes may introduce coding errors or overlooked codes.
• Multistep processes. Risk adjustment relies on documents coded by physicians, claims coded by back office personnel, and HCC codes entered by medical coders. At any stage in the process, an individual contributor can make a mistake that impacts the rest of the chain. For example, if a provider discusses a condition with a patient in their office, but doesn’t clearly document it or define a treatment plan, downstream coders can’t submit a HCC code for the patient.
• Human judgement. While certain aspects of healthcare follow clear procedures and rule sets, others are more subjective. Risk adjustment is governed by regulations from CMS, but how healthcare organizations interpret these rules and apply related coding guidelines is subjective. Differences in human judgment, even when reviewing the same set of information, can cause discrepancies and omissions during the HCC coding process.

The consequences of HCC coding errors

The healthcare mistakes that make headlines usually involve improper patient care or compliance and security breaches. While risk adjustment may seem like a purely administrative activity, coding errors can have a big impact on costs, compliance, and quality of care—all of which impact consumers, too.

Understanding risk at a population level can have substantial financial implications for health plans and at-risk provider systems. If a patient population is affected by conditions that aren’t documented or coded, healthcare organizations don’t get paid for them (or in the case of ACA risk adjustment, they have to pay out more money to other plans in their risk pool). Additionally, if patients have conditions that aren’t being managed, healthcare organizations may be on the hook for high-cost emergency bills or ineffective testing and treatments. All these additional costs get passed through to patients in the form of premiums, copays, and coinsurance.

From a compliance perspective, coding errors can be incredibly costly as well. Healthcare organizations participating in MA are under more scrutiny than ever. In the wake of MA plans upcoding conditions to receive larger payments, CMS is on a mission to reclaim $1 billion in overpayments by 2020 through annual audits. If plans submit unsupported HCC codes, they’re likely to get hit with substantial fines and reputational damage.

HCC coding errors don’t just have a monetary impact—they can also affect care quality at an individual level. Let’s say a patient is diabetic and has hypertension, but their primary care doctor only records diabetes in their charts. Not only will this patient receive treatment that doesn’t account for comorbid conditions—their health plan also won’t have the data to inform care management activities for hypertension. As a result, both the provider and the plan will get dinged on their HEDIS and MA Star quality ratings, and the patient’s overall health may worsen over time.

How to mitigate the risk of human error in RA

Reducing the margin of error for risk adjustment is critical for healthcare organizations who want to provide high quality care and remain financially competitive. Here are a few of the common approaches to mitigating risk during the coding process:

• Hire more coders. Organizations are hiring more coders to divvy up the onerous work of manually reviewing and coding charts. In the U.S. alone, there were over 206K medical coders in the U.S. in 2016, and that number is projected to grow 13% by 2026. Having a larger pool of coders reduces the review load placed on any one individual, preventing errors that happen due to mental fatigue. However, this adds to the cost of performing HCC coding in-house or with an external vendor, and introduces additional opportunities for human error throughout the review chain.
• Perform additional rounds of QA. To ensure coding decisions follow guidelines and catch any missed coding opportunities, many organizations hire coders to perform QA on coded charts—sometimes three or four rounds of additional review. This also adds to the overall coding cost, and doesn’t capture every chart.
• Retrieve charts earlier and more often. Requesting patient charts well ahead of code submission deadlines, and getting smaller batches of charts more often, can reduce the deadline pressure on coders and leave more time for QA. However, this process adds to the provider’s administrative burden and can cause abrasion if requests are too frequent. There are also guidelines governing how early plans can request charts, so timelines are still somewhat constrained.
• Use technology to support the coding process. Healthcare organizations are increasingly turning to AI-powered coding solutions to augment their risk adjustment process. Starting with machine learning models to identify information in charts that may indicate relevant HCCs, medical coders can focus their time on reviewing potential coding opportunities and flagging patients who may need clinical follow-ups based on their suspected conditions. Here at Apixio, we’re big believers in the power of technology: Our clients find 27% more codes than coders performing manual reviews, and review charts 4-6 times faster than coders working alone.

Taking on patient risk as a health plan or provider system is, well, risky. Transforming patient risk—which involves messy, ambiguous data and a chain of human decision-making—into a neat set of codes to submit to CMS and HHS is no easy task. Because this process is driven by human activity, mistakes will inevitably happen. The best way to protect your organization is to invest in the right resources, processes, and technology to make sure mistakes are caught before code submission.

Learn how Apixio can help your organization reduce the risk in risk adjustment with our AI-powered HCC coding and auditing solutions.