An Exploration of Opioid Use in the United States
By Jason Fischel, MD FACEP
UMASS Amherst
Practicum for Public Practice
Kathryn Tracy, DrPH MPA
Summer 2016
Site Supervisor
David Jaslow, MD MPH
Abstract
The purpose of this manuscript was to explore the opioid epidemic in the United States with evidence from scientific literature to help elucidate important research questions. Several questions were addressed in this manuscript. Questions pertained to the scope of the actual public health problem in addition to its etiology. Other important aspects of the manuscript assess the role of the media in the daily discussion of opioid abuse, in addition to important resources that are available to those who suffer from opioid addiction. This manuscript is the culmination a of self guided evidence-based research review from the scientific literature.
There has been a tremendous amount of media attention in the past several years about the perceived rates of opioid abuse. Daily, the media covers stories that pertain to opioid abuse, deaths related to overdoses of heroin or other opioids, or the importance of first responders carrying the opioid antidote naloxone. Just recently, even The Wall Street Journal(Kamp, 2016) reported on opioid related deaths in Maine from the year prior. The number of deaths was “record setting.” It reported on the debate in Maine regarding the state’s discussion surrounding the increased availability and broad use of naloxone by laypersons. The topics of opioid and heroin addiction have become extremely political and mainstream. The goal of this manuscript is help refocus the discussion on opioid addiction and the scope of the public health problem based on scientific evidence.
In part, this topic would seem controversial because different sources quote many different statistics when discussing the depth of the public health problem surrounding opioid abuse and overdose. The Centers for Disease Control and Prevention report quite clearly that the trend in opioid abuse is alarming (CDC, 2016). In 2014, more people died from overdoses than in any other year. Over half of these overdoses involved an opioid. The CDC reports that opioid pain relievers are a driving force in the number of opioid related deaths. They further cite the age adjusted death rate from overdoses including any opioid has quadrupled since 1999 (CDC, 2016). In fact, in 2007, overdoses surpassed motor vehicle collisions as the leading cause of injury death (Beauchamp, 2014).
In 2010, 12.2 million people reported using pain relievers for nonmedical purposes in the past month for the first time (Meyer, 2014). According to the Substance Abuse and Mental Health Services Administration, after marijuana, nonmedical use of prescription pain relievers is the second most common type of drug use. It also reported the most commonly abused drugs included oxycodone, hydrocodone, methadone, morphine, and, codeine (Meyer, 2014). It is not surprising that the unintentional overdose death rate increased 124% between 1999 and 2007. It is thought that the majority of these deaths are related to opioid use.
While the statistics clearly point to a significant threat to the health of the population, part of the issue with this public health problem is clearly defining the problem. Rates of opioid abuse are commonly quoted. However, depending on the source, the quoted numbers seem variable. In addition, defining the terms abuse, misuse, dependency, and addiction is also important (Vowles, 2015). Without knowing the exact definition of a term a study is trying to quantify, significant confusion or misinterpretation of data can occur. One study attempted to define these terms and quantify the rate of opioid misuse in patients who receive the medication for chronic pain. This is important because patients who receive opioids for chronic pain seem to be at particularly high-risk group for opioid abuse.
The study was a systematic review by Vowles and colleagues in 2015 that was published in a mainstream peer reviewed journal. Other papers were used in order to calculate rates of concerning opioid use, however, the studies that were analyzed were rated for quality prior to inclusion in the analysis. As mentioned, previous studies had quoted problematic opioid use amongst patients with chronic pain between 0 to 50% of the time (Vowles, 2015). A similar number has been found in anther study. This huge variability has little utility in defining the scope of a public health problem. In this study, overall rates of abuse ranged from .08 to 81% and addiction from .7 to 34%. However, when only high quality studies were evaluated the rates for misuse were .2 to 56% and addiction from .7 to 23%. Misuse has a broad definition and includes disorganized use, underuse, overuse, and use in conjunction with other substances such as alcohol.
However, this study was criticized by another author for inclusion of studies that were, in fact, of low quality. Scholten and Henningfiled point out that the definition of opioid misuse in the above study is identical to “patient noncompliance.” Their work suggests opioid noncompliance is similar to the general patient population noncompliance rate with other medications after comparison with another meta analysis (Scholten, 2016). In addition, though Vowles reports rating the studies for quality, many of the studies were rejected by a Cochrane review because they were found to be low in quality. As such, the Cochrane review was unable to publish a meta analysis, but rather published a literature review. The Cochrane findings were in stark contrast to Vowles’s conclusion and found that opioid use for chronic pain does not seem to be a major risk for opioid dependence. Moreover, likely the most important finding is that there is a dearth of high quality research on this topic.
Few would contest that the rates of opioid use are increasing. Given this, it is not surprising that the rates of opioid sales are increasing. The number of outpatient retail prescriptions tripled over the last two decades with its peak in 2011 at 219 million (Compton, 2015). This begs examination of physician prescribing patterns. Certainly, the aforementioned rise in opioid use would ultimately not be possible without an increase in physician prescribing.
Physicians are, in fact, writing a far greater number of prescriptions for opioids. This is well documented in the literature. After all, pain is the number one complaint why patients present to the emergency department (Poon, 2014). One recent study attempted to analyze the trends in opioid prescribing from physicians practicing in emergency departments. The study, done by Mazer-Amirshahi and colleagues was published in 2014. It is a retrospective analysis of publicly available data from the National Hospital Ambulatory Care Survey from 2001 to 2010. The database is maintained by the Centers for Disease Control and Prevention. The study found opioid prescribing increased from 20.8% to 30% over the study period. This represents a relative increase of 49%. DEA schedule 2 narcotic use also increased from 7.6 to 14.5%, a relative increase of 90.8%. Upon discharge from the emergency department, oxycodone and hydrocodone had the greatest relative increases over the study period (Mazer-Amirshahi, 2014).
There is no doubt that emergency physicians prescribe a lot of opioids. Several studies have tried to quantify how often patients leaving the emergency department are given opioid prescriptions. One study performed a retrospective analysis of 19 emergency departments for a one-week period in October 2012 (Hoppe, 2015). There were 27,516 patients seen. 19,321 were discharged. 3,284 patients received an opioid on discharge. The average age of the patients who received an opioid prescription was 41. Oxycodone and hydrocodone were by far the most common opioids prescribed on discharge. The median number of pills prescribed was 15 (Hoppe, 2015).
Interestingly, a study in 2016 was performed to assess characteristics of emergency physicians and their prescribing habits (Varney, 2016). The concept of the study is interesting and begs further investigation. However, this particular paper is not further described as the study participants were all based out of a single military hospital and very homogeneous in make up. A larger study including multiple centers with a much larger number of physicians would be an beneficial area for further study.
It is not just emergency physicians who are prescribing lots of opioids. One study looked at prescriptions for opioids from a prescription database (Vector one National database from SDI Health) for 2009 by specialty. This database receives prescription information from more than 30,000 of the 62,132 pharmacies in the country. This represents approximately half of the prescriptions filled in the US annually. Most prescriptions were for hydrocodone and oxycodone containing products. 28.8% of the prescriptions were from primary care providers (general practitioner, family medicine, osteopathic physicians). Internists wrote 14.6% followed by dentists at 6.1%. Orthopedists wrote 7.7%. For the 10-19 year age group, emergency physicians wrote the third most prescriptions at 12.3% (Volkow, 2011). Of note, it was interesting that osteopathic physicians as a group were lumped in primary care providers. This is obviously an error in methodology as osteopathic physicians specialize in many different areas of medicine.
While it is clear that emergency physicians are amongst a group of physicians who prescribe a significant amount of opioids, it is important to decipher if the increased prescribing has any relation to addiction. One study from 2015 published in the Annals of Emergency Medicine attempted to clarify this. The study authors tracked patients who were considered opioid naïve by the state prescription monitoring program to see if they received an additional script within 425 days of the index emergency physician prescription (Hoppe, 2015). The data did show those who received and filled a prescription were 10% more likely to have filled a second prescription within that time period. Though an interesting finding, the study had many limitations. Chiefly, it was only performed at one center. This limits external validity. The baseline characteristics between those who received an opioid prescription and those who did not, were statistically different. In addition, the study did not take into account who gave and why the second script was given. There is a fundamental disregard for confounding (Hoppe, 2015).
Butler et al attempted a similar study later that year. She and her team approached patients in their emergency department who were identified as using heroin or non-medicinal use of opioids and surveyed them about their first experience with opioids (Butler, 2015). Of the 122 patients approached, 59 completed the full survey. 59% of these patient reported their initial exposure to opioids occurred from a legitimate medical prescription. Of those patients, only 29% of them received that prescription from an emergency department (Butler, 2015). Importantly, 80% of these patients reported non-opioid substance abuse or treatment prior to the opioid exposure. This study had numerous limitations. The most notable limitations were the small sample size and convenience sample that was used. In addition, it was subject to considerable recall bias. However, similar to the Hoppe 2015 study, it focuses on an area that raises important questions about potential origins of opioid addiction and the role of iatrogenic opioid addiction after a short course of opioid use.
It would seem the origins of the increased prescribing of opioids is not hard to trace. In the early 80’s, reports emerged in the scientific literature that suggested opioids had a low potential of iatrogenic addiction (Beauchamp, 2014). This was supported in a consensus paper published in 1997 by the American Academy of Pain Medicine and The American Pain Society. The consensus paper urged clinicians to more strongly assess and treat patients’ pain. It identified chronic pain, and makes mention that the chronic pain may even derive from conditions other than cancer. It further points to the huge social cost associated with pain and urges “dialogue” between clinicians and regulators so as to not interfere with appropriate management of pain.
At the same time, the drug industry began a strong marketing push. Between 1995 and 2006, Purdue Pharma funded 20,000 pain related educational programs that supported using opioids for noncancer pain (Kolodny, 2015). According to a government report, in 2001, oxycontin was the most prescribed brand named narcotic for moderate to severe pain (US GAO, 2013). Purdue Pharma was even giving financial support to the Joint Commission. Purdue and the Joint Commission made an agreement that Purdue was the only drug company allowed to distribute certain educational materials and an educational book about pain management. This book was even available for purchase on the Joint Commission’s website (US GAO, 2013).
Then, in 1995, the president of the American Pain Society presented a new campaign for the aggressive pain management. It was called “Pain is the fifth vital sign” (Kolodny, 2015). Though this sounds fairly innocuous, it has a very deep meaning. To practicing physicians, the classic four vital signs (heart rate, blood pressure, respiratory rate, and temperature) are exactly as the name describes, vital. To suggest that an assessment of pain be a new, fifth vital sign is actually tremendously significant. The Joint Commission soon adopted this as well. In fact, the Joint Commission had proposed patient’s rights standards in 2000 that mandated patients being involved in their care, including the realm of pain management, amongst other topics. They further mandated that pain be recognized and addressed appropriately (US Veterans Administration, 2000). These proposed standards were adopted in 2001.