How DEA can get more out of its data

Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.

The Drug Enforcement Administration helps prevent the diversion of legal, prescription opioids. One of its requirements is the collection of industry-generated sales data. The Government Accountability Office has found ways the DEA could tighten up the quality and governance of the data. For more, Federal Drive with Tom Temin spoke to the GAO’s director of Homeland Security and Justice issues, Triana McNeil.

Interview transcript:

Tom Temin: Ms. McNeil, good to have you on.

Triana McNeil: Thank you for having me.

Tom Temin: First of all, tell us in more detail what the function of DEA is that you were looking at.

Triana McNeil: So DEA plays a key role in addressing diversion. That is, when legally produced drugs are illegally obtained for nonmedical use. Their role is to really prevent the diversion of controlled substances in certain prescription drugs like opioids. So we focused our efforts on DEA’s diversion control division.

Tom Temin: Got it. And so this is not just opioids, but anything that might be recreationally or somehow used outside of its prescribed purpose by the prescribed person.

Triana McNeil: Exactly.

Tom Temin: And I guess the question is, how do they even control that? Because if I get a prescription for something and I give it to my brother-in-law, how would they know?

Triana McNeil: That’s a really good question. So there are two main laws: the Controlled Substances Act from 1970, and the SUPPORT Act from 2018 that we focused on in our study. The Controlled Substances Act calls on DEA to maintain a closed system of distribution. The distribution is supposed to only occur between registered businesses that handle controlled substances. So in 2005 DEA began focusing its attention on wholesale distributors of prescription opioids that ship the drugs from manufacturers to pharmacies. So in our report we’ve got a graphic that shows how that process works and the flow of the prescription drugs, from the manufacturer to the distributor. The distributors then ship those to the pharmacies and the pharmacy then provides that to the patient. There’s so many different ways that those controlled substances can be diverted at all the different phases of that process. And that’s why it’s so important that DEA get a handle on all the entire flow of drugs in the process.

Tom Temin: And it sounds like this is mainly a data driven process. They derive the information of the flow through the data reported by these entities. What sets of data, what sorts of data does DEA need to be able to carry out this function?

Triana McNeil: So DEA has three main data systems that it uses. One is CSOS, the Controlled Substance Ordering System. This is where electronic orders are placed. Folks can also place orders manually with a paper form, so the CSOS is not complete. And then they’ve got [Automation of Reports and Consolidated Orders System]. That system pulls in all the data from CSOS plus the paper order forms. And it really monitors the flow, the sale and the dispensing of the controlled substances through disclosed system. This is historical information. Some of it is dated. Some can be many months up to a year old. Then the third data system is SORS. This is where be a received and stores suspicious order reports. This is more real-time, but each registrant can really define what they think a suspicious order is in their own terms. And that’s something we point out in the report.

Tom Temin: Yes, you found some issues with the data. What were the chief ones?

Triana McNeil: Some of the main issues that I have just outlined and I’ll just underscore in the CSOS, where orders for these drugs are placed in an electronic environment, it’s not a complete system because some people are still submitting paper order forms — some pharmacies, for example. Then you have ARCOS. Some of the data in there is very old. It’s a gold mine because it’s historical. It goes back many, many years, and it collects a lot of good pieces of detailed information. But some of it is dated, and then you’ve got SORS.

Tom Temin: And just tell us what that stands for?

Triana McNeil: SORS is the Suspicious Order Reports, that system collects all of the information that the registrants are flagging as, “I think this order might be suspicious, DEA. You may wanna look at that.” And the issue with that is all the different registrants, be it manufacturer, distributor pharmacy, they may define in different ways. So it can be like comparing apples and oranges.

Tom Temin: We’re speaking with Triana McNeil. She is GAO’s director of Homeland Security and Justice issues. And so basically, what they’re looking for is anomalies within that distribution system, on the assumption that not every distributor or pharmacy is on the up and up.

Triana McNeil: That’s what they should be looking for.

Tom Temin: Well, are they?

Triana McNeil: What we found was, there’s a number of opportunities that exist for DEA to improve its management and use of data. So the first point, DEA’s data systems are not designed to conduct real time analysis. And so we identified ways that DEA could use algorithms to identify, for example, unusual volumes of deleted transactions or drugs that were disposed of, rather than sold. And by doing this type of data, mining DEA can be more proactive, real-time and more quickly identify all the different types of questionable patterns nationwide, and then flag those registrants that need to be investigated.

Tom Temin: And what about the issue of data governance, too, because that came up in the report?

Triana McNeil: It did. We found that DEA lacked a data governance structure to manage all of that data that it collects and maintains. So, for example, right now in the newly created suspicious opioid order system SORS, that’s voluntary to submit that information electronically. So DEA has to manage both electronic and paper files whenever it gets any kind of information from its registrants that something looks suspicious. So it’s unclear who is going to centralize all that fragmented electronic and paper information. It’s also unclear us who’s going to analyze it. Does DEA have one staff person or 20 that will be mining that data, and then how are they going to do the analysis? And further, it’s unclear how the analysis will be used and if they’ll provide it to their field investigators or the industry as well. So a governance structure will help address the outstanding questions that we have and ensure that DEA’s diversion activities are targeted and timely so that they can stay ahead of any new trends. Because we really have to get a handle on this opioid crisis and the larger drug issue.

Tom Temin: I guess it’s fair to say they do have some effectiveness in anti-diversion. I mean, one piece of evidence is that opioid deaths are finally starting to fall again.

Triana McNeil: Yes, that is true, but have they fallen enough? I would say no.

Tom Temin: All right, and what are your principal recommendations specifically for the DEA?

Triana McNeil: So we made a number of recommendations. The two that I’ll highlight here, we asked them to develop and implement additional ways to use algorithms. They’ve got to be proactive and identify problematic drug transaction patterns. We also recommended that they establish and document a data governance structure to ensure that they are maximizing all of the information that the industry is providing to them. They agreed with both recommendations, and they said that they’re currently trying to implement them.

Tom Temin: Triana McNeil is director of Homeland Security and Justice issues at the Government Accountability Office. Thanks so much for joining me.

Triana McNeil: Thank you so much, Tom.

Tom Temin: We’lll post this interview along with a link to her report at Hear the Federal Drive on demand and on your device. Subscribe at Apple Podcasts or Podcastone.