What if there was a way to detect fraud for SUD facilities?

The following is coauthored by Cecille Avila (@cecilleavila), Austin Frakt (@afrakt), and Melissa Garrido (@garridomelissa).

Millions of Americans struggle with substance use disorder, with estimates suggesting as many as 1 in 13 people needed treatment in 2018. Between high demand for services and lack of regulation, this is an area of health care already rife with predatory behavior. Substance use disorder fraud was a significant problem before the COVID-19 pandemic, and it could potentially get even worse.

According to the Centers for Disease Control and Prevention (CDC), more individuals reported an increase in substance use during the COVID-19 pandemic, and overdose deaths have gone up significantly. Now that vaccination rates are increasing and social distancing measures are being lifted throughout the United States, well-intentioned individuals or concerned family members may view rehabilitation facilities or treatment centers as a logical next step to address some of these issues.

But it’s not as easy as Googling “substance use disorder rehab facility near me” to find treatment centers that can provide appropriate — or even genuine — treatment. The amount of fraudulent behavior that has occurred in the past decade can make it feel as if there are equal numbers of results for substance use disorder treatment facilities and horror stories.

In some cases, individuals have been sent out-of-state as part of elaborate and profitable patient brokering schemes that do not actually connect them to treatment. Some treatment facilities have profited off patients by overbilling for unnecessary drug tests. Treatment facilities with a known history of violations have continued to operate, resulting in the deaths of patients. In Florida, a man running sober homes was found to have coerced residents into prostitution.

Using news articles and publicly available legal filings, it might be possible to identify potential warning flags for facilities by looking at their websites or asking certain informational questions, such as:

  1. What is included on a facility’s website? References to accreditation can indicate a facility has met certain standards set by an outside organization (i.e., the Joint Commission).
  2. What’s not included on a facility’s website? Does it omit what kind of treatment plans are provided?
  3. What treatments or services does the facility provide? Medication assisted treatment is evidence-based, but equine therapy is less so.
  4. Who is going to provide treatment? For example, are staff members named and licensed?
  5. Are there clear policies in case of emergency (i.e., overdose)?
  6. Is recovery guaranteed? (This is not possible for a facility to guarantee.)
  7. Is there any financial assistance offered to help offset paying for care at the facility? A sliding scale financial assistance payment model can be helpful but offers of being paid (in cash or in kind) should be viewed with skepticism.

The list goes on, and it would be impossible for individuals to try and find answers for every facility in question in a reasonable amount of time. Perhaps what would be helpful is if there were an automated way — even a resource — that could guide people toward more trustworthy providers who offer appropriate addiction treatment.

How? One potential approach could be to use insurance claims data to identify patterns that suggest fraudulent activity. For example, suspicious behavior might be detected through unusual frequencies of procedures (i.e., drug tests, including urinalysis), unusual numbers of patient encounters per specific providers, potentially unnecessary treatments, unbundled lab tests (instead of billing under a single code), duplicate or multiple billing, or upcoding (billing for more expensive procedures).

However, flagging a single suspicious incident, such as high frequencies of tests or procedures, may reflect an overly cautious provider and not a fraudulent one. But perhaps a provider with a high rate of suspicious activities relative to its peers is more likely to be fraudulent. If nothing else, it could assist insurers, regulators, and law enforcement in focusing resources to scrutinize providers more likely to be engaged in illegal fraud.

While using claims data to detect fraud can be helpful, it’s not a perfect system. Claims data depends on what providers submit, so providers can get away with potentially problematic behavior if they don’t bill insurers for it. Patient kickbacks, in which individuals receive both indirect and direct incentives to refer patients to certain facilities, is a significant problem in this industry and does not generate claims data. Facilities that benefit from patient kickbacks without raising any other flags could remain undetected.

One challenge: An individual payor will only see or experience a subset of providers’ behavior. Perhaps a pattern of fraud can only be reliably detected by tracking providers across multiple payors.

Unfortunately, pooling data across insurers is difficult. Some data sets that do so (e.g., all-payer claims) do not identify providers, rendering them useless for the task. But a few resources exist that have data elements up to the task, such as those provided by FAIR Health or the Health Care Cost Institute, among others.

Of course, accessing data is just the first step. The next is to develop and validate an algorithm to identify more likely fraudulent providers. Doing so could help reduce unethical provider behavior.

While the COVID-19 pandemic dominated the news, the opioid epidemic did not fade away. Being able to connect individuals to appropriate and effective services is necessary to prevent even more harm from being done and without concern about falling victim to fraud.

Research for this piece was supported by Arnold Ventures.

Hidden information below


Email Address*