By James C. Sherlock

The author has completed a detailed survey of individual Applied Behavior Analysis (ABA) providers in New Jersey licensed by Virginia Medicaid to provide services to children with autism.  He picked New Jersey for reasons familiar to readers of his work on nursing home chains. The survey revealed that most of those providers live in Lakewood and surrounding Ocean County.

In this case, record-keeping, the fundamental building block of government business, and oversight have failed to meet even this skeptic’s low expectations.  

The overall implication of the survey is that licensing out-of-state providers located far from Virginia’s borders to deliver ABA services in Virginia may not be worth the risk.  The other major failing shown by the full survey is that, while neither government tracks nor oversees ABA companies, they must do so.  While individual providers make errors in submissions, companies that employ them have been proven to be the primary source of fraud.

The implications of the full survey will be discussed in two parts: records and out-of-state providers.  This is about the effects of record inaccuracy.

The full survey will not be provided here to protect providers’ identities, but even the abridged version attached shows that individual providers failed at a very high rate to update their practice address records with the Commonwealth and the federal government.  Sometimes they updated one but not the other.  The full version shows that many of the women surveyed failed to update their maiden names upon marriage.

Widespread record inaccuracies effectively eliminate computerized analysis as an oversight tool. All other methods, such as those used to complete this survey, are far too manpower-intensive to be replicated at scale.  With millions of records and the inability to conduct computerized analysis, both federal and state governments are seriously hamstrung in basic oversight and in eliminating malpractice and fraud.    

The survey compared the records of those providers’ addresses maintained by:

• the Virginia Department of Health Professions (DHP) database of licensed individual providers who are subject to Board of Medicine oversight,
• the Centers for Medicare and Medicaid Services (CMS) National Provider Identifier (NPI) database, and
• public records of their current residences.

It revealed that the records held by federal and state government agencies responsible for oversight of autism services are so fundamentally flawed as to be useless for computer-based oversight and fraud detection.  

For example, almost 90% of ABA providers are women.  Most of them are young, reflecting the profession’s relative newness.  The full survey discussed here shows that when they get married, change their legal names, and move in with their spouse, many do not update their NPI records.  Eight of 73 women in the survey had married names and addresses that differed from their maiden names recorded in government records.  

Of what use are an NPI number and a Medicaid provider number to the federal and state governments if the names and addresses in those records are both out of date, or one is out of date, and the other is not?  Other methods of analysis, such as the one used to produce this survey, are too resource-intensive for widespread use. Record accuracy will need to be strictly enforced, with penalties that exceed current ones.

In this very small sample, the reconciliation of records is sufficiently complete for DMAS and the Attorney General’s Medicaid Fraud Control Unit to link the full version of this survey to individual providers’ Medicaid billing histories to determine whether anything warrants further investigation. 

But that will not address the fundamental problem of record inaccuracy defeating computer analysis.  Both governments will have to enforce much greater sanctions than are currently on the books to get their records right.

At that point, perhaps they can begin to ask about outcomes of therapy, not just pay the bills submitted.