How do you find out which programs offer the gold standard in opioid addiction treatment?

Over the past few years of reporting on opioid addiction in America, I've often interviewed people who were doing well in recovery programs who told me, “I’m on medication-assisted treatment now, but before this, I was in a program where they taught me: That's not real recovery.”

Medication-assisted treatment (MAT) is a method of treating opioid addiction in which patients go to talk therapy and take one of three Food and Drug Administration-approved medicines — methadone, buprenorphine or naltrexone. The existing research shows MAT with methadone or buprenorphine keeps people in treatment for longer, reduces cheating with illicit opioids, and prevents overdoses and deaths. The federal and California state governments have created various policies and programs to try to encourage addiction treatment centers to use MAT. But folks on the ground kept telling me that treatment programs remained reluctant, often out of a belief that people who take methadone and buprenorphine are just substituting one drug for another and aren’t truly overcoming their addictions. Methadone and buprenorphine are both opioids themselves, just like painkiller pills and heroin, but they’re designed to reduce cravings and aid with recovery.

So I wanted to know: How prevalent is this attitude? What are the chances that, if I were to seek opioid addiction treatment for myself or a loved one, I'd run into a facility that didn’t offer the best possible treatment, that might not even tell me about the best possible treatment? Could there be other reasons, such as financial constraints, that keep facilities from offering MAT? I could gather all the anecdotes I wanted, but I didn't think I’d get the real picture unless I was able to get systematic data on addiction treatment centers.

I decided to start by answering these questions for California. The state makes tons of opioid-use statistics publicly available, but it doesn’t keep a list of which centers offer or support MAT. The federal government’s Behavioral Health Treatment Services Locator does keep track of non-MAT facilities, but because it depends on a voluntary survey, sent out just once a year, its database is incomplete and can be out of date. (During the training sessions for the Center for Health Journalism's 2018 Data Fellowship, our instructors had told us always to check how a dataset is made and how often it's updated. These tips were a huge help for me in understanding the strengths and weaknesses of the Behavioral Health Treatment Services Locator.)

So I embarked on cobbling together the data myself. Here are the lessons I learned.

Even three minutes on the phone makes a big difference.

To verify that my list of California non-MAT facilities, which I'd downloaded from the Behavioral Health Treatment Services Locator, was accurate, another reporter and I called all 150 phone numbers on the list. On average, my calls took only about three minutes. I identified myself as a reporter; asked whoever picked up if their facility offered MAT; and if not, why not. My initial intention was just to verify the Treatment Services Locator data. I didn’t realize I was also gathering information that would become critical later, when I ran into other data roadblocks. More on that below.

My calls also taught me to verify voluntary, survey-based datasets. It turned out that 14 of those 150 facilities actually did offer MAT. In addition, 11 of the phone numbers had some problem: They were disconnected or rang endlessly over three separate tries. At one number, a man told me he was not interested in donating to my cause and hung up on me.

I did recognize pretty quickly that I was getting more information from my calls than anticipated, and that a few themes came up repeatedly. Many places said they did not offer MAT but did refer to other MAT facilities. Many noted they weren't licensed to dispense medication. And, of course, staffers at about a third of the facilities talked about their philosophical opposition to MAT. My coworker and I took notes on our calls in a shared Google spreadsheet. Afterward, I used the sheet to classify each of the calls into buckets: offer_mat, refer_to_mat_or_offsite, not_licensed_to_dispense_medicine, philosophically_opposed. The buckets let me quickly pull numbers on the proportions of facilities that fit into each category.

Sometimes data can’t answer your questions

I had one big blank spot in my data. At the time I downloaded datasets from the state and federal governments, the fed listed 1,183 “substance abuse” facilities for all of California. The state’s records, however, indicated it had licensed 1,889 drug and alcohol programs. From my calls, I knew the MAT status of the 1,183. What about the remaining 706 facilities? California's logs didn't indicate anything about MAT. My magazine, Pacific Standard, didn't have the people-power to call 706 numbers.

At first I thought I could deduce whether those hundreds of facilities offered MAT by using billing data from Medicare and Medi-Cal, which is California's version of Medicaid. Since these datasets are very large, much larger than what Excel can handle, I learned a little SQL to be able to analyze it. To start, I got a ton of help from the mentor assigned to me by the Center for Health Journalism's Data Fellowship. With a little persistence, however, it seems possible to learn SQL from online resources as well. Googling your SQL questions really does work.

Unfortunately, my Medicare/Medi-Cal idea didn't pan out. I needed to find associations in those datasets that would point me to possible clinicians who treated opioid addictions without MAT. The associations I found were too loose, however. Clinicians who might treat addiction looked too much like clinicians who didn't treat addiction, in terms of billing. Even in huge datasets, there just might not be enough information to answer specific questions.

I had to rely on the data I already had. I ended up using my Treatment Services Locator calls as a proxy for the state-licensed facilities I couldn't get MAT info on. For example, since in my calls I found a significant minority of facilities were philosophically opposed to MAT, I could surmise there could be hundreds more like them, among the licensed California facilities for which I had no MAT information. You can see more about how I dealt with imperfect information in my story, in the subsection “Getting Past ‘Abstinence Only.’”

When I first found my Medicare/Medi-Cal idea wasn't working, I really worried that my story would have this big gap and wouldn’t seem thorough. Later, I realized the data I already had was telling me more than I thought. I just had to analyze it correctly, and explain the uncertainties that remained.

In many ways, I'm still frustrated there is no database anywhere that can accurately tell searchers whether an addiction treatment facility supports MAT, before they pick up the phone. Just imagine being a desperate patient, parent, or spouse, seeking treatment now. You'd want that information upfront. Instead, you have to call and ask. Still, I hope my story illuminates for policymakers the barriers to more facilities offering this life-saving treatment. For people with addictions and their families, I hope it shows you can and should ask about MAT. Any facility should be able to tell you how they approach it, in less than three minutes.