Prescription Opioid Use among Patients Seeking Treatment for Opioid Dependence

Abstract

Objective

This study was designed to assess non-medical prescription opioid use among a sample of opioid dependent participants.

Methods

A cross-sectional survey was conducted with a convenience sample of patients hospitalized for medical management of opioid withdrawal. We collected data related to participant demographics, socio-economic characteristics, the age of first opioid use, types of opioids preferred, and routes of administration. We also asked participants to describe how they first began using opioids and how their use progressed over time.

Results

Among the 75 participants, the mean age was 32 years (SD: ± 11, range: 18-70), 49 (65%) were men, 58 (77%) considered themselves to be “white,” 55 (74%) had a high school diploma or equivalent, and 39 (52%) were unemployed. All of these participants considered themselves to be “addicted.” Thirty-one (41%) felt that their addiction began with “legitimate prescriptions,” 24 (32%) with diverted prescription medications, and 20 (27%) with “street drugs” from illicit sources; however, 69 (92%) had reported purchasing opioids “off the street” at some point in time. Thirty-seven (49%) considered heroin to be their current preferred drug, and 43 (57%) had used drugs intravenously.

Conclusions

We found that many treatment-seeking opioid dependent patients first began using licit prescription drugs before obtaining opioids from illicit sources. Later, they purchased heroin, which they would come to prefer because it was less expensive and more effective than prescription drugs.

In the United States, recent recommendations to adequately treat pain have been associated with an increase in the number of prescriptions written for opioids.^{1, 2} Clinicians who prescribe these opioids often struggle with “a question of balance” between risk and benefit.³ On the one hand, if their prescribing practices are too liberal clinicians risk contributing to the problem of prescription drug abuse; while on the other hand, they may compromise the clinical benefit of adequate pain control for their patients if their practices are too conservative.

There are individual patients who have both clear objective evidence for a chronic pain disorder and a well-documented drug use disorder; however, many physicians do not feel prepared to effectively manage these complicated patients. For example, a survey of a nationally representative sample of 648 primary care physicians indicated that nearly one-half of them had difficulty talking to their patients about substance abuse.⁴ A similar study found that although most physicians (88%) reported that they ask patients about substance abuse, less (24%) do so at every visit and few (19%) do so annually.⁵ In another study of 161 primary care physicians, investigators found that 45% of physicians thought that they did not have adequate consultation and referral resources to assist them with patients who have chronic pain and “only 29% felt they had adequate consultation and referral resources in their communities to assist them with patients who might be abusing or selling opioid prescriptions.” ⁶

Although physical tolerance does develop, it has been believed that a drug use disorder was an uncommon adverse event when opioids are used for the treatment of “legitimate” chronic non-malignant pain. For example, one author concluded that “few legitimate, drug-naïve patients become addicted as a result of the intended use of OxyContin as an analgesic.”⁷ Similarly, in a 2002 statement to a committee of the United States Senate, the then Acting Director of the National Institute on Drug Abuse stated, “Reports of people becoming addicted to OxyContin, if used as prescribed, are rare.”⁸ However, it is now clear that some patients with chronic non-malignant pain who are treated with long-term opioids do indeed develop aberrant drug taking behaviors.⁹ For example, one author found that of 298 individuals who were admitted to a psychiatric facility in Appalachia for opioid dependence, 187 (63%) had problems related to OxyContin that was prescribed to them for chronic pain.¹⁰ Investigators have identified several patient characteristics associated with an increased risk for aberrant drug taking behavior: a prior history of a substance use disorder and younger age,¹¹ prior mental health problems and low educational achievement,¹² and a history of legal problems.¹³

Initial prescription opioid use may have multiple origins. Some purchase opioids from the illicit market (i.e., “on the streets”), others may have their first exposure to opioids as a result of receiving a prescription from a physician, and some may get these drugs from friends or family members. The 2006 National Survey on Drug Use and Health reports that, among persons aged 12 or older who used prescription opioids non-medically in the 12 months prior to the study, 70% reported that the source for their most recent use of these drugs was a friend or relative.¹⁴ On the other hand, another study found that the majority (84%) of 109 individuals seeking treatment for a substance use disorder stated that “they had legitimately been given a prescription for opioids for pain at some point from a physician.”¹⁵

Despite the large problem of prescription opioid abuse, there is little known about the histories and trajectories of this problem. It is not clear whether these people intentionally seek out prescription drugs or their problem was the result of the unintended consequences from a physician's medical treatment of chronic pain. This knowledge gap makes it difficult for physicians to balance the conflicting medical views on limiting opioid use versus adequate pain control. The purpose of this study was to determine how a group of patients, who had been admitted for opioid detoxification, initiated a pattern of drug abuse and to describe how it progressed over time. We compared the characteristics of those who first used opioids with a “legitimate prescription” (i.e. licit use) with those who first used opioids diverted from a friend, a family member or from illegal purchases (i.e. illicit use).

Methods

We conducted a cross-sectional survey of patients hospitalized for the medical management of opioid withdrawal. The study protocol was approved by the human studies committee of the sponsoring university and the Medical Director for Research at the host hospital. Those who were invited to participate were given written information about the study and then verbal consent was obtained before proceeding to interview questions. We obtained only verbal consent because we did not collect any personal identifiers (name, address, etc.) for our permanent database. Our participants often were providing information about illegal activity. If they had to sign a consent form with their name, then they would actually increase their personal risks by participating in the study; their names would be potentially discoverable. By collecting only anonymous data, their risks associated with study participation were reduced. The human studies committee accepted this rationale. Participants were free to withdraw from the study at any time without prejudice.

Setting and Participants

Participants were recruited from an inpatient “detoxification unit” of an urban tertiary-care teaching hospital. This 18-bed ward is in a 550-bed public hospital that serves a regional population of approximately 1.5 million. It occupies a separate wing of the hospital, has 1,200 to 1,400 admissions per year, is the largest of 3 inpatient detoxification programs in the region and serves a socio-economically diverse group of adults aged 18 years or more. It is staffed by 4 physicians, 1 nurse practitioner, 6 chemical dependency counselors, and a varying number of nurses and other support staff. The goal of the staff is to refer every patient to some sort of aftercare counseling program (e.g., outpatient, inpatient, or residential) at the time of discharge.

Seventy-five participants were recruited from among those admitted to the detoxification unit between February 15, 2006 and August 2, 2007. All patients admitted to this unit had been evaluated in the Emergency Department by a certified chemical dependency counselor prior to admission and were required to meet DSM-IV criteria for opiate dependence. To be eligible, participants must have wished to become abstinent from opioids, been at least 18 years old, been able to understand spoken English, been able to provide informed consent, and had urine toxicology positive for opiates on the day of admission. There were no exclusion criteria other than patient refusal. This was a convenience sample, and patients were approached and recruited during times when the medical student interviewers (M.C.C. and C. E. K.) were available (e.g., evenings, weekends, and summer hours). Therefore, not all patients admitted during the study period were invited to participate.

Measures

We collected quantitative data that included: basic demographic (e.g., age, gender, and race/ethnicity) and socio-economic (e.g., education, employment, type of health insurance) characteristics, age of first opioid use, types of opioids preferred, routes of administration, history of other drug use (e.g., cocaine, marijuana and benzodiazepines), history of prior mental health treatment, and criminal history (i.e., number of prior arrests, number of misdemeanor and felony convictions, and time spent in jail or prison).

We also asked participants to describe how they first began using opioids, how their drug use progressed over time, and if they were ever questioned by a physician about their drug use. The responses to these open-ended questions were summarized by written notations on the data collection forms.

Data Analysis

Descriptive statistics were used to summarize socio-demographic and other baseline clinical characteristics of the participants. Participants who reported that their addiction began with opioids that were prescribed for them (i.e., licit use) were compared with those who traced the onset of their addiction to either diverted prescription medications or from “street drugs” (i.e., illicit drug use). The Fisher exact test was used for between group comparisons of categorical variables and the Student t-test was used for between group comparisons of continuous variables. Equal variances were assumed, except for number of prior arrests, misdemeanors and felonies, as these variables were not normally distributed.

Participant comments, in response to the open-ended questions, were summarized and then grouped into categories of similar themes. Each category was labeled with a direct quote from one of the participants, and those participants with similar comments were counted as positive responders within that category.

Power analysis was based on a minimum sample size of 30 participants in each group, a two-sided alpha of 0.05, and a reference group proportion of 0.40. We calculated the power to detect a moderate, clinically important effect size (i.e., an odds ratio of 1.5 or greater) to be 81%.

Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) Version 14.0 (SPSS, Inc., Chicago, IL, 2005) and Statistical Analysis System (SAS) Version 9.1.2 (SAS Institute Inc., Cary, NC, 2004).

Results

Participant Characteristics

The characteristics of the sample are summarized in Table 1. The typical participant was about 30 years old and white. Over half were male, about half had at least a high school education, and about half were unemployed. All of these participants considered themselves to be “addicted” to opioids. During the study period of February 2006 to August 2007 there were 1850 admissions to the inpatient detoxification unit. Unfortunately, there is not information about how many of these admissions were for detoxification from alcohol, opioids, or other drugs (e.g., benzodiazepines); however, the clinical staff of this unit estimated that the proportion for these groups is approximately 40%, 50% and 10% respectively. Using these estimates, the 75 participants included in the study sample would represent about 8% of the approximately 925 admitted for opioid detoxification. Of these 75 participants, 31 (41%) indicated that their addiction began with their own licit prescription, 24 (32%) from diverted prescription medications, and 20 (27%) from “street drugs.” However, over 90% had purchased drugs “off the street” at some point in time, nearly two-thirds of the participants considered heroin to be their preferred drug, and over half had used drugs intravenously.

Table 1. Participant Characteristics.

Characteristic	Mean (±SD) or N (%)
Current Age (years)	31.5 (± 10.9)
Gender (Male)	49/75 (65%)
Race (self-identifies as “white”)	58/75 (77%)
Education
No high school diploma	32/74 (43%)
High school diploma	11/74 (15%)
Some college	23/74 (31%)
Post-secondary degree	8/74 (11%)
Health Insurance
Self-pay (no insurance)	21/74 (28%)
Commercial Insurance	24/74 (32%)
Private Pay	13/74 (18%)
Medicare	3/74 (4%)
Unmanaged Medicaid	10/74 (14%)
Managed Medicaid	3/74 (4%)
Employment
Unemployed	39/75 (52%)
Full-time	19/75 (25%)
Part-time	12/75 (16%)
Disabled	4/75 (5%)
Retired	1/75 (1.3%)
Prior mental health treatment (yes)	34/75 (45%)
Age of first opioid use (years)	19.2 (±5.9)
Number of prior arrests	6.5 (± 11.8)
Number of misdemeanors convictions	1.7 (± 3.1)
Number of felonies convictions	0.8 (± 2.4)
Jail time (months)	16.3 (±51.4)

Participant Comments

Some comments were gathered from all 75 participants. Many participants (49%), indicated that their drug use started because of a social situation. They alluded to “pill parties” as a common source. One person noted “[type of opioid] was handed to me by my friend, this guy I know, someone who was at the party.” Another person said “kids are using it like Viagra.” Some participants also mentioned how common it was to have prescription drugs available in high schools. One participant noted that opioids were even available “at the prom.” Some of the participants noted that top athletes at these schools would start using opioids to “make it through a game” as a way to deal with the pain from a prior injury. Later, these same athletes would use opioids to “get high” on the weekends and during the off season. During interviews common themes arose such as “I used them to feel normal,” “It helped take away my emotional pain and stress,” or “ [name of opioid] made me feel like a better person.”

When asked about the origin of these opioids, many said their “parent's prescriptions.” One participant, who admitted to being a drug dealer, suggested that young people are wary of heroin and other street drugs; they prefer the pharmaceutical grade opioids because “they know they're pure.” Another noted, “The best way to get opiates is to look for the dying person who will give [them] up.”

Participants were asked to explain to the investigators why they first tried opioids. Thirty-seven participants (49%) reported first trying opioids because they were curious and/or someone they were with had opioids (i.e., situational). Another 15 (20%) reported first using opioids after a surgical or dental procedure (e.g., appendectomy, tooth extraction), 14 (19%) to treat a painful condition (e.g., non-traumatic back pain), and 9 participants (12%) reported their first use was after an injury (e.g., ankle sprain, broken bone). Table 2 shows opioid use history for those who first used opioid licitly versus those who first used opioids illicitly.

Table 2. Summary of the Qualitative Results.

Characteristic	Licit Use N (%)	Illicit Use N (%)	Pair-wise P-value	Group P-value
“Why did you first try opioids?”				< .001
Situational	0/25 (0%)	32/43 (74%)	<0.001
Pain	21/25 (84%)	10/43 (23%)	<0.001
After Injury	2/25 (8%)	1/43 (2%)	0.550
After Surgery	2/25 (8%)	0/43 (0%)	0.132
“Why did you continue using opioids?”				< .001
I liked the feeling	6/24 (25%)	33/41 (81%)	<0.001
Pain control	18/24 (75%)	8/41 (19%)	<0.001
Current Preferred Opioid				.024
Hydrocodone	18/31 (58%)	34/44 (77%)	.126
Oxycodone	26/31 (84%)	39/44 (89%)	.732
Heroin	9/31 (29%)	28/44 (64%)	.005
Fentanyl	1/31 (3%)	1/44 (2%)	1.00
Other	3/31 (10%)	0/44 (0%)	.067

Those who reported that they first tried opioids to treat pain were more likely than the others to have reported obtaining their first opioid through a licit prescription. Whereas, participants who reported using opioids for the first time because they were curious or because they were in a situational circumstance were more likely to have obtained their first opioid illicitly (Table 2).

When participants were asked, “Why did you continue using opioids?” after their initial exposure to opioids (regardless if licit or illicit), 39 (52%) reported that they continued using opioids because they “liked the feeling” and 26 (35%) because they wanted “to control pain.” Participants who initially used opioids illicitly were more likely to continue using opioids because they “liked the feeling” opioids gave them, whereas those who started using opioids licitly were more likely to continue using opioids to “control pain” (Table 2). Several participants noted that the use of diverted prescription drugs is expensive and that they turned to intranasal or intravenous heroin because it was less expensive and more effective than oral prescription medication.

Participants were also asked to report if any doctor had ever inquired about a substance use problem before giving them a prescription for opioids. Of the 53 participants who answered this question, 39 (74%) reported that their prescribing doctor never questioned them about this; however, 44 out of 67 participants (66%) stated that at some point they asked their doctor for help with their substance use problem.

Comparison Between Initially Licit Users and Illicit Users

Comparisons between those who reported obtaining their first opioid through their own prescriptions (i.e., licit use) and those who reported obtaining their first opioid from a diverted prescription or from the street (i.e., illicit use) are shown in Table 3. In sum, as compared to those who used illicit opioids, those participants who reported obtaining their first opioid from a licit source (i.e., their own prescription) were about 5 years older, more likely to have a college degree and more likely to have health insurance. There was also a difference in the use of other drugs among these two groups. Those who reported that the origin of their first opioid came from a licit source were less likely to have ever used marijuana. In addition, first-time licit users were less likely than first-time illicit users to have used drugs via an intranasal or intravenous route. First-time licit users were also less likely to have past legal problems, prior arrests, misdemeanor convictions, and felony convictions than first-time illicit users. As shown in Table 2, they were also less likely to report heroin as their current drug of choice.

Table 3. Comparisons between first-time licit users and first-time illicit users.

Characteristic	Licit Use Mean (±SD) or N (%)	Illicit Use Mean (±SD) or N (%)	Pair-wise P-value	Group P- value
Current Age (years)	34.7 (± 11.8)	29.2 (± 9.8)		0.031
Gender (Male)	16/31 (52%)	33/44 (75%)		0.049
Race (White)	26/31 (84%)	32/44 (73%)		0.401
Education				0.018
No high school diploma	8/31 (26%)	24/43 (56%)	0.017
High school diploma	4/31 (13%)	7/43 (16%)	0.752
Some college	12/31 (39%)	11/43 (26%)	0.309
Post-secondary degree	7/31 (22%)	1/43 (2%)	0.008
Health Insurance				0.003
Self-pay (no insurance)	3/30 (10%)	18/44 (41%)	0.004
Commercial Insurance	10/30 (33%)	14/44 (32%)	1.00
Private Pay	8/30 (27%)	5/44 (11%)	0.122
Medicare	3/30 (10%)	0/44 (0%)	0.063
Unmanaged Medicaid	3/30 (10%)	7/44 (16%)	0.731
Managed Medicaid	3/30 (10%)	0/44 (0%)	0.063
Employment				0.566
Unemployed	17/31 (54%)	22/44 (50%)	0.815
Full-time	7/31 (23%)	12/44 (27%)	0.789
Part-time	3/31 (10%)	9/44 (21%)	0.338
Disabled	3/31 (10%)	1/44 (2%)	0.330
Retired	1/31 (3%)	0/44 (0%)	0.413
Routes of administration
Prior subcutaneous use (“skin-popped”)	6/31 (19%)	19/44 (43%)		0.046
Prior intranasal use (“snorted”)	17/31 (55%)	38/44 (86%)		0.003
Prior intravenous use (“shot up”)	13/31 (42%)	30/44 (68%)		0.003
Ever used cocaine	24/31 (77%)	41/44 (93%)		0.082
Ever used marijuana	27/31 (87%)	44/44 (100%)		0.026
Ever used benzodiazepines	19/31 (61%)	32/44 (73%)		0.324
Prior mental health treatment (yes)	17/31 (55%)	17/44 (39%)		0.480
Number of prior arrests	2.1 (± 3.8)	9.7 (± 14.4)		0.002
Number of misdemeanors convictions	0.7 (± 1.7)	2.5 (± 3.7)		0.006
Number of felonies convictions	0.16 (± .5)	1.2 (± 3.0)		0.027
Age of first opioid use (years)	20.5 (± 7.0)	18.3 (± 4.8)		0.117

Discussion

In sum, we found that most of these patients, who were hospitalized for the medical management of opioid withdrawal, first began using prescription drugs that were either from their own prescription or that had been diverted to them through another prescription. Participants who initially obtained opioids from illicit sources preferred heroin to a greater extent than patients who were receiving prescription opioids licitly. Many then came to prefer heroin because it was less expensive and more effective than prescription opioids. This suggests that there is a progressive nature to opioid abuse and that prescription opioids can be the gateway to illicit drug addiction and that individuals who use prescription opioids illicitly may be at greater risk for subsequent heroin use than those who use prescription licitly.

Those who first used opioids from licit sources were different from those who first used opioids from illicit sources; they were older, less likely to be male, more likely to have a college degree, less likely to use non-oral routes for drug administration, and had fewer prior arrests and criminal convictions. This suggests that there might be individual characteristics that predict the initiation of drug use or conversely, it may be that how drug use is initiated is predictive of the progression of the subsequent social consequences.

There are several limitations of this study. First, the participants were recruited from one program in one city, which limits the generalizability of the findings. Second, the participants were not selected randomly, and therefore, there may have been some sampling bias or the findings may not be a valid representation of all those who were admitted to this facility. Third, there may have been some interviewer bias. Although the 2 interviewers made every attempt to ask questions and record the responses consistently, there is always the potential for some inconsistency between the 2 interviewers. Finally, the participants may not have been able to provide accurate answers due to recall bias. However, we found that this group of participants seemed open about their past drug use and seemed willing to cooperate with the medical student interviewers.

Conclusions

Many treatment-seeking opioid dependent patients first began using licit prescription drugs before obtaining opioids from illicit sources. Later they purchased heroin, which they would come to prefer because it was less expensive and more effective than prescription drugs; however patients who begin taking opioids from the illicit market tend to prefer heroin to a greater extent.