In an ideal world, translating what occurred during the patient’s visit to a correct bill would be a simple task. Documents and other clinical data would stream in from physician notes and other sources, an automated system would recognize the appropriate information, structured or otherwise, translate the chart into the correct codes and seamlessly send the information along for billing. If this error-free system were to exist, your organization would receive all the money it deserves and each patient’s bill would reflect his or her care experience exactly.
Since coding using workflow software became the norm, this type of utopia has been just that – a far off fantasy. Computer Assisted Coding was meant to increase productivity and reduce coding errors – yet many of those promises have fallen short.
This raises the question: Is CAC really the solution to solving coding errors?
The promise of CAC vs the reality of CAC
Since CAC was first introduced to the health care industry, it has been touted as the answer to several systemic issues like coder productivity, financial protection and even as a tool for improving CDI programs. In 2013, the Healthcare Information and Management Systems Society and DAK Systems Consulting predicted that CAC could give suggestions at point of documentation, thus making codes more accurate from the beginning of care. HIMSS also explained that, when integrated with an EMR system, CAC could improve data abstraction and improve reporting functions.
“CAC alone has a lower recall and precision rate.”
While some of these promises have been fulfilled in part, CAC solutions still leave much to be desired. As ICD-10 Watch explains, CAC hasn’t done much for coding accuracy or the productivity of coders. CAC may improve coding speed and hasn’t been shown to increase coding error rates, but those are hardly gold standards of success. In fact, a study by the Cleveland Clinic found that CAC alone – that is, without the assistance of a credentialed, human coder – had a lower precision rate.
Another important factor that must be considered is the initial data that enters the system. If that data is inaccurate or articulated poorly, the computer will inevitably have trouble making sense of it, and thus produce code suggestions that are also inaccurate. Computer interpretation of human language input is still far from perfect, and without the guiding hand of an actual human, mistakes are likely to crop up.
How to leapfrog CAC with AI-assisted audits
In five years, it’s likely that many advanced organizations will have done away with current iterations of CAC in favor of a largely automated coding system. Imagine your vehicle today – it gets you from place to place and has safety features to protect you from harm. But if at some future point your car could be controlled by an artificial intelligence (AI) capable of reacting to everyday hazards at lightning-fast speed and requiring intervention only in highly specialized instances, it might be safer. Even then, this new reality doesn’t mean that the government is suddenly able to remove traffic signs, stop lights and guardrails. The same logic applies to automated coding and AI-assisted audits.
“Audits pre-billing can protect the revenue mid-cycle.”
Here’s how the future may look: An advanced automated coding system codes charts without the need for human intervention. In most cases, the computer isn’t just assisting, it’s in the driver’s seat.
For highly complex cases, a credentialed human coder makes adjustments as needed. Before any charts are sent to billing, an AI audit solution checks the codes using a combination of human expert-generated rules, along with advanced machine-learning techniques, calibrated to the level of risk the organization is willing to accept. Coders make adjustments as needed and audited codes are sent to billing. This system achieves the next stage of productivity and coders are elevated to focusing on exceptionally unique cases or cases with significant impact on organizational revenue.
This future may sound like a pipe dream, but it’s actually just around the corner.
Why your organization needs to focus on code quality
Today, your organization can already benefit from AI-assisted code audits. Looking Glass® eValuator™ from Streamline Health is an AI-assisted code auditing tool that enhances the revenue mid-cycle in two ways: a) by giving human coders real-time suggestions that merge clinical scenarios with best practice coding logic to improve code accuracy and lower compliance risk and b) selecting the best charts for auditors to audit, reducing the administrative burden to complete pre-bill audits comprehensively.
As everyone knows, the quality of your codes directly affects revenue. Mistakes during the revenue mid-cycle can leave money on the table or introduce legal risk. Post-bill audits simply can’t do much to improve the quality of your codes or your coders because feedback loops happen too late. On the other hand, pre-bill audits foster a culture of real-time accountability and may significantly lower your organization’s risk. Quality coding protects revenue.
With automated coding and AI assisted audits, the ideal future of revenue protection is within reach. CAC hasn’t solved the real problem; while it may speed up the revenue mid-cycle, it hasn’t done much for code accuracy. Streamline Health’s eValuator™ tool is the way forward to a future of quality codes free of human error. The journey toward accurate coding begins today.