One out of every 143 U.S. patients who received a prescription for an opioid painkiller in 2008 obtained prescriptions from multiple physicians in a pattern that suggests misuse or abuse of the drugs, according to a recent estimate. While these “doctor shoppers” comprised only 0.7 percent of all patients with opioid prescriptions, they purchased almost 2 percent of all such prescriptions and 4 percent of the total amount of opioid drugs measured by weight.
Dr. Douglas McDonald and Kenneth Carlson of the research firm Abt Associates, Inc., of Cambridge, Massachusetts, derived national estimates from a massive dataset of 146.1 million records of opioid prescriptions dispensed by 76 percent of U.S. retail pharmacies in 2008. The prescriptions were for 9 different opioid drugs, including buprenorphine, methadone, and oxycodone, and were written by 908,000 prescribers for 48.4 million individual patients.
To identify elusive doctor shoppers from the vast majority of patients legitimately procuring opioids, the researchers sorted patients into groups based on their opioid purchasing patterns. This approach netted an “extreme” group of 135,000 likely doctor shoppers, who, on average, obtained 32 opioid prescriptions from 10 different physicians over a 10-month period.
The researchers’ analysis also revealed sizable gaps in prescription monitoring by both individual physicians and state-run monitoring programs. Mr. Carlson says the work will help set better thresholds for prescription drug monitoring programs (PMPs) to catch suspicious purchasing patterns and alert both physicians and pharmacies.
Extreme Opioid Purchasers
Text Description of Graphic
Mr. Carlson, a statistician, and Dr. McDonald, a sociologist, applied a statistical modeling approach to identify three distinct groups among the entire population of prescription buyers (see Figure 1). Patients in the first group, who made up the majority, had visited only one prescriber in the entire year; most were likely seeking relief for acute pain that resolved. Patients in the second group had obtained prescriptions from one to four prescribers, on average; they probably had chronic pain that was being treated by multiple physicians, including specialists.
The third group, however, contained the outlier patients that the researchers labeled “extreme.” These patients averaged 10 different prescribers throughout a 10-month period; moreover, when the researchers looked at only those who had paid for their prescriptions in cash, that average jumped to 15 prescribers per patient.
“Of course, we don’t have X-ray vision or trackers on these people to know for certain they are abusing or distributing opioids,” says Dr. McDonald. “But we can draw some inferences about who these people in this extreme group are.”
For instance, the age distribution of this group peaked between ages 26 and 35 (see Figure 2). This distribution corresponds to the age range with the highest prevalence of self-reported prescription drug abuse in the 2010 National Survey on Drug Use and Health by the U.S. Department of Health and Human Services.
Dr. McDonald and Mr. Carlson also calculated that this extreme group had purchased a total of 11.1 million grams of drug, which, when averaged per patient, was the equivalent of 109 milligrams of morphine per day for a whole year. More than 100 milligrams of morphine per day puts patients at high risk for overdose.
“For every day out of the year, that’s a lot of drug,” says Dr. McDonald. For example, the Centers for Medicare & Medicaid Services flags Medicare beneficiaries taking a daily dose of 120 milligrams or more of morphine equivalents for 90 days as at risk for overdose.
Text Description of Graphic
Improved Estimate of the Magnitude of Doctor Shopping
Abuse of opioid prescription drugs has spiked in recent decades, beginning in the 1990s with a general increase in prescribed painkillers and, specifically, with the availability of new synthetic opioid drugs, such as oxycodone. Between the mid-1990s and 2011, emergency room admissions for opioid misuse increased about tenfold. In the same period, oxycodone prescriptions more than doubled, and sales of the drug increased more than tenfold.
The increases in prescribing and in abuse of prescription opioids are clearly linked, but how much have drugs diverted by doctor shoppers fueled that abuse? “We know that doctor shopping happens. And physicians know that there are patients seeking these drugs for misuse,” says Dr. McDonald. “But we didn’t know how prevalent doctor shopping was.”
Previous estimates, relying on data from insurance claims or prescription drug monitoring programs in some states, have suggested that between 0.2 percent and 13 percent of all opioid patients are doctor shoppers, depending on the definition used. As Dr. McDonald points out, many doctor shoppers are likely to pay with cash to avoid a paper trail. So, to get a more reliable estimate, Dr. McDonald sought a national dataset that would include those cash payers.
He found it in a dataset that had been developed by the health information and services company IMS Health and recorded all of the opioid prescriptions dispensed during 2008 from 37,000 retail pharmacies across the United States, including cash purchases. The IMS Health dataset linked all the prescriptions dispensed to each individual patient by any of the pharmacies, a feature that made possible the study of doctor shopping.
Previous studies had largely relied on the professional judgment of physicians to set the parameters of what constitutes doctor shopping, which yielded widely differing estimates of its prevalence.
“We thought about it differently,” says Dr. McDonald. Rather than defining what would constitute doctor shopping from the beginning, he and Mr. Carlson asked, “What if we let the data sort itself out?” The two researchers came up with a mathematical model that did just that by grouping the patients according to their individual prescription patterns—namely, how many different prescribers they had over the course of the year.
Better Grounds for Red Flags
The researchers point out that limitations in their mathematical model likely restricted their estimates of doctor shopping. However, Dr. McDonald is confident that the analysis provides more detailed information about doctor shoppers than earlier studies and that it will help PMPs and physicians spot those rare patients who are likely to be abusing opioids.
“The results are giving us good information that the threshold for identifying doctor shopping is higher than use of three different prescribers. It is probably north of five or six different prescribers during a short period,” Dr. McDonald says. He notes that state-run PMPs that record scheduled drug prescriptions dispensed by pharmacies can now use these findings to improve their alert notifications to pharmacies and physicians to flag suspicious opioid drug purchases.
Going forward, Dr. McDonald and Mr. Carlson suggest that improvements in health care information technology should focus on improving the accessibility of PMPs. Currently, physicians can access these PMP databases to pull up a patient’s prescription history, but they often omit doing so because of time constraints and cumbersome access procedures. In addition, current PMP systems often do not detect doctor shoppers who hop state borders. Effectively addressing these limitations would mean linking all state PMP data and integrating prescription history data into patients’ electronic medical records.
“These doctor-shopping patients are clearly different, and they are exploiting the absence of good data management right now. The best way to improve the situation is to get better information in the hands of doctors—ideally, in their lap in advance of an appointment,” Dr. McDonald says. “Ultimately, health care providers are the front-line defense against prescription drug diversion.”
Such defenses are urgently needed because of the sharp rise in use of opioid prescription drugs in the United States, which has resulted in a dramatic increase in overdose deaths. According to the researchers, between 1999 and 2010 the number of deaths by opioid prescription drug use more than quadrupled to almost 17,000 per year, twice as common as deaths due to cocaine and heroin use combined. Moreover, prescription opioid abuse increasingly is seen to be a precursor to heroin use and addiction.
“Understanding the mechanisms of diversion of opioid drugs is important for intervention in doctor shopping,” says Moira O’Brien, health scientist administrator in the NIDA Division of Epidemiology, Services, and Prevention Research. “This study contributes to the small but growing body of studies aimed at estimating the magnitude of doctor shopping.”
Ms. O’Brien points out that the study shows that even among the groups of patients with legitimate chronic pain complaints, substantial numbers are getting opioid prescriptions from several prescribers. This signals potentially dangerous, uncoordinated care, and these patients may be at risk for medical complications or overdose. “These data really underscore the importance of PMPs making information accessible to providers about any opioid prescriptions a patient might have received.”
This study was supported by NIH grant DA028920.
McDonald, D.C. and Carlson, K.E. Estimating the prevalence of opioid diversion by “doctor shoppers” in the United States. PLOS ONE 8(7):e69241, 2013. Full text