WellPoint's John Steffy: How to use social media to detect healthcare fraud

Social media sites can provide a wealth of information to aid in healthcare fraud investigations. But how do you wade through all that information--and interpret the results?

In fact, there's no substitute for manually reviewing sites, says John R. Steffy, a senior investigator in WellPoint's northeast special investigations unit.

But you might be surprised at the different kinds of fraud you can uncover through sites such as Facebook, Twitter, Pinterest and even MySpace. (Yes, MySpace still exisits.)

There's much more to it than finding photos of a person who is collecting disability payments crossing a marathon finish line. 

To learn how insurers can make the most of data mined from social media sites, FierceHealthPayer: AntiFraud spoke to Steffy, who specializes in investigating healthcare fraud by organized criminal groups, He previously worked for the Kings County district attorney's office in Brooklyn, New York.    

FierceHealthPayer: AntiFraud: What types of intelligence can fraud investigators gather through social media sites? Why is this intelligence valuable?

John Steffy: Social media sites are excellent places to gather open-source intelligence. A typical example is finding a photo of a person who's water skiing and may have submitted a claim for broken leg. Commercial payers and government investigators like this kind of intelligence because it acts as a checkmate on the fraudster.

But there's other valuable intelligence out there as well. Investigators will want to look at Facebook, Twitter, Foursquare, Pinterest and Myspace. People love to talk about where they are and what they're doing, and even our photos are geo-tagged these days. This becomes useful when a provider bills for services on dates when he or she may have been in Rome, for instance, or elsewhere on vacation. It's very powerful to confront a provider you may suspect of fraud with photos that indicate they were out of the country on the date they billed a particular payer for services.      

FHP:AF: How is social media intelligence most efficiently gathered?

Steffy: There's no substitute for manually reviewing social media available on the internet. This allows investigators to familiarize themselves with different sites and different types of user-generated content contributed to each site.

Each site has its own interface. It's important for an investigator to know the difference, for example, between LinkedIn--which is usually aimed at a professional audience and includes information that doesn't change often--and Twitter, where updates are brief and constantly changing. Twitter information might be small and truncated, but it's very up-to-date.

Investigators have to be careful with the use of data mining tools because they're built to return data. There's a distinction between data and intelligence.

Data have to be refined and sorted and sifted before you can consider them useful intelligence. When investigators manually review social media, they can give themselves a head start on the intelligence cycle by gathering data and sorting out the relevant bits of intelligence in one condensed step.

Data mining tools are often available from the same vendors that sell other investigative tools. But many social media data mining tools are meant to assist marketing teams.

Investigators also have to be careful when deciding which tools to use. There are obviously security concerns with placing any kind of software on a corporate or government computer or asset for investigative purposes. With pitfalls like that, manual review of social media is really the safest course to follow.         

FHP:AF: Does social media information have a use in commercial and government business fraud investigations?

Steffy: Investigators can--at the very least--use social media intelligence to augment or corroborate what they already know or evidence they've gained from other sources.

Further, investigators can use social media to build their knowledge of a profession or specialist. Let's say you know how an oncology practice works but want to learn more. Oncology practices will often post YouTube videos that detail the intricacies of how the profession functions.

Investigators should also be mindful of social media sites geared towards reviewing services provided by a profession. Yelp is very helpful here. You can use it to data mine for witnesses or glean information from online complaints you may not otherwise have known about.   

FHP:AF: Do criminals use social media to carry out or boost profits of healthcare fraud schemes?  If so, how?

Steffy: Criminals do use social media. Some groups, for example, try to steal data and information on beneficiaries. Often, criminals are able to hack partial data lists. Once they have these, criminal groups can avail themselves of different media to socially engineer additional data on victims. If criminals have the name, height and weight of an individual and perhaps their address, they can use social media to get a birth date and fill in other personal details. 

FHP:AF: In your opinion, is it a good investment for insurers to buy prepackaged social media data from external sources? Why or why not?

Steffy: That's an interesting and difficult question because every insurer SIU has a particular set of tools at their disposal, and one payer's SIU may not have the same tools as another's. Each SIU has to look at the tools they have and compare them with available social media tools and resources to see if there's a good fit for the company's particular needs.

And the answer to this question today may not be the same as the answer tomorrow, because technology in this field is constantly changing. Every day, there's a new social media site, and every day it seems there's a new program out there to troll for information in social media.

FHP:AF: Do you have any other tips for fraud investigators? 

Steffy: Many investigators think of social media as social networking sites such as Facebook and Myspace. But those are just one subset of what's available. You also want to keep in mind Yelp, LinkedIn and other sites geared toward the exchange of more professional, user-generated content. Those sites can be helpful in identifying potential witnesses, leads and targets. It's very easy, for example, to look at the Yelp site and do a search for "medical mill" or "pill mill." You'll end up with a list of results that you can typically narrow down by geographic area.

In addition to that, always keep in mind that social media are basically a self-disclosed network analysis. So anytime you have an investigation target, it's easy to use social media tools to augment or build on what you already know their network analysis is.   

Editor's Note: This interview has been edited for length and clarity.

Related Articles:
Nurse's firing over Instagram photo spurs social media debat
FDA's social media guidance: Better late than never
Why hospital leaders should use Twitter
How hospitals can harness social media to help people
Hospitals must address employee fraud reports with procedural fairness
How to make online reviews work for you
Traditional media reigns over social media for sharing health policy research

Suggested Articles

The HHS OIG is asking for an additional $23.7 million to support fraud oversight that has benefited from an emphasis on data analytics.

A New York surgeon was sentenced to 13 years in prison for fraud and more physician practice news from around the web.

A federal judge has ruled that the U.S. government’s remaining fraud case against UnitedHealth can move forward.