Fraud, waste and abuse in health care claims: A bad situation worsened by the pandemic
Here's what benefits managers should know about detection and prevention.
Whether it is deliberate or accidental, it is hard to deny that fraud, waste, and abuse (FWA) are running rampant in health care claims and contributing to the rising cost of health care. FWA brings real consequences to corporate benefits managers who have a responsibility to their stakeholders, including both corporate employers and their employees who demand continued quality of care options while trying to keep down costs, prevent sharp increases in rates and other out-of-pocket expenses.
Related: Report: Overbilling and upcoding by hospitals cost Medicare $1 billion
FWA is a long-term problem evolving over time and impacting claims across all medical and pharmaceutical spend areas. It has long been estimated that one in six claims contains errors, and the COVID-19 pandemic has added new wrinkles to the problem. Here are some contributing factors as to why:
- Waste comes about through errors, such as unintentional one-offs that result from keying an incorrect value, chronic issues built on the misinterpretation of correct coding requirements, or the misuse of modifiers.
- Abuse tends to be more of a systemic issue that is inconsistent with sound fiscal, business, or medical practice in which, for example, coding to a higher level of service than actually provided or billing for medically unnecessary services like excessive treatments and over-utilization of diagnostic technologies.
- While it is relatively rare, outright fraud refers to more complex schemes used to obtain unauthorized or unearned benefit through intentional deception or a repeated pattern of billing behavior for services, procedures, or equipment that were never actually rendered.
Fortunately, applications of data analytics and data engineering can be applied to identify erroneous billings and payments. Corporate benefits management teams are increasing their focus on the detection of FWA in order to mimic the Payment Integrity (PI) and Special Investigations Unit (SIU) teams employed by insurers. Despite working diligently to detect these issues, new schemes continue to creep up regularly.
The COVID-19 pandemic has introduced rapid changes in regulations and policies to ensure access to patient care. The benefits of these changes come along with an increased exposure to FWA, allowing egregious offenders the opportunity to use creative means to take advantage of COVID-19 enhanced payments and loosened policies that enable FWA to pass through undetected. Examples of potential for FWA in pandemic period claims include billing for unnecessary, not rendered, or non-compliant services. Below are a specific examples of FWA challenges elevated by the COVID-19 pandemic:
- To encourage patients to seek necessary care, certain member cost-shares were waived, giving rise to the potential for overprescribing and overuse of telehealth services. Moreover, when FWA does occur, members are less likely to complain about aberrant billing if their share of the cost is zero, and payers will be less likely to notice when typical whistleblowers are silent.
- In response to the pandemic, telehealth (historically restricted to established patients in rural areas) was swiftly expanded to ensure access to care for patients who are new and established, urban and rural. Early indications are that while offering easy access to providers, telehealth comes with its share of aberrancies that include scheduling of “impossible” hours and an increased likelihood of unnecessary follow-up visits. It also affords the opportunity for coding of e-visit or virtual check-in claims to higher-level telehealth visits, as well as up-coding of evaluation and management (E&M) services to a higher level of service than actually rendered. Finally, certain unscrupulous telemarketers have persuaded elderly patients to share their personal health information, prescribed non-compliant treatments, and ordered unnecessary supplies.
- During the pandemic period, aberrant services levels abounded in many areas, including 1) excessive COVID-19 diagnosis (especially on claims resulting in enhanced COVID-19 payment levels), 2) early refills and disbursement of excess of prescribed quantities for pharmacy and durable medical equipment (DME), 3) extended visit durations and an increased number of visits for the various therapies, 4) extended lengths of stay for skilled nursing facilities (SNF) and home health, and 5) excessive lab testing including outlier levels of presumptive urine drug screens (that may or may not have been prescribed or medically necessary).
Since the beginning of the pandemic period, there is no doubt that billing and payment errors have been made at the expense of corporate employers, plan members and other stakeholders. The good news is the claims from the pandemic period are now rolling in sufficient quantities to be relevant and can be used to help put a stop to these practices, set new policy, and recoup losses. The challenge lies in putting data to use to identify the aberrant providers, erroneous claims, and incorrect payments.
Putting data to work
Utilizing data analytics to address these issues relies on multiple data sources that are complex and increasing in size. Claims data is the most important area to master from a data engineering perspective. While claims data is transactional in nature, coupling provider contracts, covered benefits, regulatory requirements, and global payment policies adds a new dimension to understanding claim payment inaccuracies.
These additional data sources are often unstructured and depend on technological innovation to match policies or contractual terms to claims payments. Getting the data into usable formats is step one; after that, the information presents a sizable logic puzzle that has to be completed in order to unearth billing and payment issues.
One of the first lines of attack in the identification of overpayments is to utilize rule-based algorithms for global payment policies and correct coding requirements. Customized algorithms can match these policies to claims data to identify inaccurate payments and other issues with adjudicated claims. Essentially, there are thousands and thousands of custom “what-if” decisions utilizing global payment policies as a source of truth. This scalable solution uses logic and market intelligence to identify low-hanging fruit and to claw back overpayments.
More challenging are provider contracts. Because provider contracts are typically complex, pulling relevant data points in a scalable fashion presents a challenge. Optical character recognition (OCR) is a valuable methodology to pull relevant contract terms in near real-time and translate those terms into a consistent searchable format. Customized OCR platforms can train on multiple contracts over time and be built to increase accuracy in pulling relevant terms, to ensure that the correct billing codes are used on all claims and that the contracted rate is adhered to.
While the two examples above present post-pay and retroactive views of payment integrity, the future of health care will rely on pre-pay determination of the potential for inaccurate payments. This will allow for proactive measures to understand contracts, claim systems, and operational aspects to adjudicate claims. For example, time-series forecasting can apply historical information on payments (e.g., realized vs. erroneous) to spot trends and root causes as to why certain types of claims are erroneously paid.
The ability to forecast one to three months ahead will be important to understand which types of claims are likely to be paid in error. In addition, the application of anomaly detection and outlier analytics to identify the potential for previously unknown patterns or conditions and to detect individual and aggregated abnormal patterns compared to peer groups. Finally, predictive analytics can be used to detect complex patterns and either predict the likelihood of individual claims that will be erroneously paid, catch abnormalities on global edit payments, or cluster providers into groups for education and outreach based on claims history.
Finally, artificial intelligence (AI) can be used to efficiently extract information from high frequency data sets, discover data gaps and automate assurance of data quality, and profile and score providers and claims. AI coupled with machine learning (ML) drives algorithms that learn from business decisions and feedback which fuel working efficiencies and improve payment accuracy, reducing false positives and the overall cost of claims administration.
Another benefit of applying leading-edge technology to FWA is that it can help overcome the tendency of the two functions (PI and the SIU) that typically work on these types of claims issues to “silo off.” The PI team is tasked with ferreting out waste and abuse, while fraud is the province of the SIU. When analytics identifies a provider that seems to be engaged in FWA, a review of medical records or a more comprehensive investigation can be undertaken that includes both PI and the SIU.
Enhancing internal expertise
While data analytics solutions do tremendous things, the tools required for the detection for FWA in claims data are not available in an “off the shelf” solution or service that can simply be purchased and put to work. For corporate benefits managers to have confidence that provider billing errors and payer payment errors have been properly denied by the administrative services only (ASO) or third-party administrator (TPA), it requires access to comprehensive claims data and the contractual ability to “audit” more than a small sample of that data.
Additionally, it requires top-notch analytics tools and techniques that are combined with automation, machine learning, custom configuration, and manual review and intervention to analyze the claims data. Also, the employment of human capital in the form of tech geniuses as well as skilled, experienced payment integrity specialists and fraud investigators, remains an important part of the solution to this vexing challenge. Finally, detection of FWA requires robust internal business analytics solutions that depend on knowledgeable resources across IT, PI, and the SIU to maximize actionable findings. If any individual facet is missing, the overall process will fall short of the maximum potential, limiting quality outcomes and financial benefits to both the corporate employers and their employees.
Maria Turner (mturner@aarete.com) is a managing director at global consultancy AArete, where she is a leader in the Healthcare Payment Intelligence™ practice.
Read more: