Digging into DEA data exposes sharp increase in use of pain medications

By Frank Bass
The Associated Press

It was safe to assume there was a story buried in the thousands of pages of controlled-substance data maintained by the U.S. Drug Enforcement Administration. I just didn't know what the story would be.

The DEA keeps the data as part of its Automation of Reports and Consolidated Orders System program (ARCOS), designed to keep an eye on more than 30 million annual transactions involving controlled substances. Manufacturers and distributors are required to file reports at least quarterly. Quarterly summaries by three-digit ZIP code are kept on its Web site. The three-digit ZIP codes are boundaries created by the U.S. Postal Service and the Census Bureau. They generally correspond with the first three numbers of the full ZIP code.

Technically speaking, putting the data into usable form wasn't too hard. It wasn't a huge amount of data, since there are only about 890 three-digit ZIP codes in the United States, and data are only available for eight years. We also were able to use numbers from the 2000 Census – the latest demographics available for three-digit ZIP areas – to figure out the social and economic characteristics of various ZIPs.

All of the DEA data were in PDF format, which required a little bit of work to convert into a usable format We used a program called PDF2XL to handle the conversion of thousands of pages into Excel spreadsheets. At $95, we thought it was a pretty good deal.

Once in Excel, the data required quite a bit of manipulation. I used the AutoFill function quite a bit and recorded a few macros to help me along. It was tedious work, taking about a month. Once I finished, I began examining the various Schedule II drugs listed for the latest year:

• DL-Amphetamine Base
• D-Amphetamine Base
• Methylphenidate
• Cocaine
• Codeine
• Buprenorphine
• Oxycodone
• Hydromorphone
• Hydrocodone
• Meperidine
• Methadone
• Morphine
• Fentanyl Base

We then looked at the available drugs for 1997 and created a list of drugs with data for both 1997 and 2005. Then, we looked at overall amounts. We threw out a few because they didn't seem to have much in common (the amphetamines, methylphenidate) and tossed a few more because of their potential other uses and small amounts (cocaine, buprenorphine, hydromorphone, methadone and fentanyl). That left us with five: codeine, hydrocodone, oxycodone, meperidine and morphine.

We began the actual analysis by tallying all five drugs, which confirmed that something unusual was happening. In 1997, about 49 million grams of the five drugs were sold at retail establishments around the nation. By 2005, that number had jumped to more than 90 million – an increase of 90 percent.

The obvious answer was, in a word, Oxycontin. In 1997, only about 4 million grams of oxycodone had been sold. That was less than 10 percent of the Big Five. By 2005, about 30 million grams had been sold – a third of the major painkillers. Even with the heavy oxycodone increase, my editors encouraged me to keep plugging away at the numbers. We figured that 90 million grams worked out to 9 billion milligrams – about 300mg per year for every man, woman and child in the United States. Unless you were a long-term chronic pain patient or Keith Richards, most experts agreed that was enough for a very bad overdose.

Use is spreading

I started looking for patterns. I used ArcView GIS software to map the percentage change in retail distribution across the country for oxycodone. There had been a 600 percent increase overall, which wasn't too surprising, given that it had really only been pushed since 1995. I found that the highest distribution in grams per 100,000 people was still mostly in Appalachia, but appeared to be moving west. There also were some isolated "hot spots" where distribution had increased in Fort Lauderdale, St. Louis, and Columbus, Ohio.

I looked at hydrocodone, used mostly in Vicodin and Lortab. It appeared to have replaced oxycodone as the painkiller of choice across much of Appalachia. Morphine sales had increased, too, mostly in Arizona and Nevada. Only codeine and meperidine distribution were down across the board, but there were a few more hot spots. Codeine use had increased in the suburbs, particularly around Nashville, Kansas City and Long Island, N.Y. So we had a story that didn't just involve "hillbilly heroin." People across the country were living in a world of pain.

After talking with my editors, we decided we wanted to focus on use, rather than abuse. I started speaking with DEA and National Institutes of Health officials, who were concerned about our writing about physician prosecutions. I also began chatting with a small underground movement of chronic pain advocates.

Their stories were compelling. Many Americans weren't able to get appropriate amounts of painkillers because physicians were reluctant to prescribe large amounts that could bring them to the attention of DEA investigators. More than 100 physicians had been prosecuted over a four-year period. Many were guilty of running "pill mills," but we reviewed several plea bargains and trial transcripts that showed many had pleaded guilty to different charges because the alternatives were worse.

The physicians were suffering, but so were patients and families. One source called the day after her husband had died. She had driven him hundreds of miles every month to a pain specialist who was prescribing grams of Oxycontin for pain from a connective tissue disorder. When he died, she feared that an autopsy with a full toxicology test would be done, and the physician would be charged with murder. She had his body cremated hours later.

I flew to Myrtle Beach, where eight physicians had been rounded up in a notorious 2001 indictment. One of the physicians had worked at a pain clinic for barely two months and resigned in disgust; she was threatened with a 100-year jail sentence. Her attorney told her to accept a plea bargain and cooperate with federal investigators. "The first hog gets the best slops," he told her. She fought the case in court, and wound up with a two-year sentence. The U.S. Attorney's office that prosecuted her ultimately refused to comment on the case.

Finally, I talked to a U.S. Marine veteran of the first Persian Gulf War. He was taking large amounts of Oxycontin every day, the legacy of a couple of helicopter crashes. The first crash had nearly killed him, but he shook it off. The second one also was bad, and he had to retire on disability. For years after his discharge, he said he was shuttled from one Veterans Administration hospital to another. Finally, someone began prescribing opiates for him. He hated the idea, but said, "Everything changed."

Explaining why

We had the human elements. Next, we needed to figure out why the tide had turned on pain management. One trend was obvious; the highest increases in painkiller use correlated with an older population. Baby boomers are getting older, and, well … let's just concede that maybe the "baby" part of the generational tag sometimes applies.

The second trend was a little harder to find, but we felt stupid once we figured it out. Opened a magazine lately? Watched a television show? Gone to a doctor's office and not tripped over a pharmaceutical sales representative or two? Direct-to-consumer advertising by pharmaceutical companies has almost tripled over the past decade, from $11 billion to $30 billion.

No one could pin down exactly when the third reason for the increase had occurred, but most experts pointed to the early 1990s, when some influential physicians began making headway against the "no pain, no gain" mentality that had prevailed. It had been a long, slow process, but most physicians seemed willing to address the issue of patient pain, rather than dismiss it as a helpful, if occasionally annoying, symptom of another problem.

The final challenge was writing a balanced story. There are a few semantics that are important in writing about opiates:

• Be sure to understand the difference between dependence and addiction. One expert put it this way: "People who are addicted to opiates will rob a liquor store for their painkiller. People who are dependent on painkiller can't make it to the liquor store."

• The biggest quantity of diverted drugs doesn't come from pharmacy burglaries, street corners or pill mills. Most people who abuse them start at home. Ever share a codeine tablet with a family member? Or worse, grab a Vicodin from one of your spouse's old prescriptions? Come out of the house slowly, with your hands up.

• Simply because the DEA has labeled these drugs as "controlled substances" doesn't mean they're dangerous, per se. According to one study, nearly 400 people died in a year from using Oxycontin. But let's put that figure in perspective. About 7,200 people died that same year from using non-steroidal anti-inflammatory drugs (NSAIDs). I can't (and no one else can, either) provide the number of lives that have been wrecked by opiates, but be very, very cautious writing about single opiate overdoses.

We shared the numbers with our state bureaus, which resulted in separate sidebars in Arizona, Connecticut, Florida, Idaho, Indiana, Kentucky, Maryland, Missouri, Montana, Nebraska, New York, North Dakota, Ohio, Oklahoma, South Dakota, Tennessee, Washington and Wyoming. Other bureaus dropped state-specific figures into the main story and moved it on their state wires.

We alerted our members to the story in advance and provided a Web page with 52 zipped files for their use. Each state had spreadsheets showing distribution in each three-digit zip for all five drugs and line charts showing the year-to-year changes in distribution for each drug, as well as five separate maps showing the three-digit ZIP boundaries and change in distribution over the eight-year period.

The story was on the front page of more than five dozen U.S. newspapers, widely aired on local television and blogged about at great length.

---

Frank Bass is a reporter with the AP's Washington-based investigations team.