It’s Time to Think About Predictive Analytics for Coding Compliance
For the period April 1, 2021 to September 30, 2021, the Office of Inspector General (OIG) issued 162 audit reports and 46 evaluation reports. They reported $787.29 million in expected audit recoveries, $1.17 billion in questioned costs, and $1.24 billion in potential savings. The expected investigated recoveries were pegged at $3 billion.
Those numbers are impressive indeed, but they come from one place. You.
Healthcare practices are under increased scrutiny from government agencies. CMS, RAC, MAC, and others – as well as from private payers who are increasingly using sophisticated auditing algorithms to uncover improper claims that can lead to them clawing back reimbursements you’ve received. Many providers are not leveraging the technological firepower necessary to assess their compliance risks to prevent these expensive takebacks.
Are you doing everything you can to monitor and assess your risk of a coding compliance audit?
Unfortunately for many practices, assessing compliance audit risk often involves benchmarking – an approach that reminds you where you have been but doesn’t tell you where you are going. While benchmarking can play a role in predicting modeling, you still need to properly integrate the data within predictive algorithms to push the models to the next level. If you want to look through the windshield rather than the rearview mirror when it comes to minimizing your audit risk, you should invest in real predictive analytics — as government and private payer regulators are doing.
Most common approaches to audit risk management
Organizations understand the need to monitor their compliance risk. However, there is a wide range of approaches being used – each with varying levels of sophistication and success.
Homegrown systems
Some smaller practices take this basic approach which employs a system of Excel spreadsheets and other APC audit worksheets. While this system may yield some rudimentary results, it is an awkward, resource-draining, time-consuming manual process with significant drawbacks.
A spreadsheet is not an integrated solution and presents an awkward process to get to every claim. Practices that use this approach need the help of the IT group to extract the claims for the specific doctors they want to review. These claims are then downloaded into a file and exported into a spreadsheet. Reviews are based on a small number of random claims with no clear way to identify which doctors to target. The chances of catching actual problem claims or doctors with issues using this probe-audit process are extremely low.
RAC tracking solutions
A Recovery Audit Contractor (RAC)-based process is only a slight upgrade from a homegrown system. The RAC technology was developed to provide access to your claims data with no sophisticated method for selecting which claims you should review. There is no precise targeting that tells you which claims or providers pose the most risk which ultimately doesn’t help you minimize that risk.
After getting access to the claims, the practice is still faced with the task of analyzing them. This system was built by companies with no real compliance or audit intelligence background and offers a limited method of uncovering true audit risk.
Vendor-developed workflow automation audit-based software
Implementing compliance audit-focused software from a vendor is clearly a step up from homegrown or RAC-tracking solutions, but there can still be issues that prevent gaining true visibility into overall compliance audit risk. There is an ability to identify problem providers, but the targeting mechanism is based on narrow, simplistic benchmarks that come exclusively from a small source of data provided by the vendor’s clients.
Using this software is often a time-consuming, multi-step process involving many “clicks” that can consume resources and delay the time to gather insights to assess the organization’s audit risk. Even after going through this rigorous process, the practice may still not gain visibility into all audit risks. The process shows only the claims that have been selected based on narrow logic. Although these vendors often claim to provide visibility to 100% of claims, the applications are in fact, only filtering outlier claims based on simplistic benchmarks that don’t go deep enough.
Artificial Intelligence/Machine Learning
This solution, based on the power of artificial intelligence/machine learning (AI/ML), is a major leap forward from the other audit-risk detection approaches. Rather than mining narrow client or vendor databases, an AI/ML approach is a true Big Data solution able to access millions of available claims. This predictive analytics solution requires less work while providing more accurate results.
Rather than relying on narrow benchmarks that provide limited results, an AI/ML solution targets 100% of providers and CPT codes that could most likely trigger an audit. Instead of benchmarks, this solution uses sophisticated algorithms that can quickly surface problems to predict audit risks for your practice.
Protecting yourself against expensive recoupments and intrusive, costly audits requires the same advanced technology that is being used by regulators. Isn’t it time you reviewed your audit-risk assessment approach to ensure you are implementing the most effective defense possible?
To learn more about how predictive analytics can help you minimize your coding compliance audit risk, download our White Paper.
By Frank Cohen, Director of Analytics and Business Intelligence