In the treatment of chronic pain, there may be no area of greater controversy than the use of opioids. Changes in attitudes with respect to opioid use toward the end of the 20th century, and subsequent exponential increases in use, have been well documented.2,31,56,58 More recently, the burgeoning public health issue regarding opioid-related adverse events has perhaps been equally well documented, as the use of opioids in chronic pain brings with it marked potential for adverse events for the patient, including overdose, experience of physiological dependence and subsequent withdrawal, addiction, and negative impacts on functioning.2,6,38,56 Attention paid to the so-called “opioid epidemic” (eg, Refs. 19,32) has highlighted the need to clearly differentiate and identify the types of problematic prescription opioid use (eg, misuse, abuse, addiction) and discern their frequency in treated patients with chronic pain.
Attempts to calculate rates of problematic opioid use behavior have suffered from imprecise and poorly defined terminology. Two recent sets of expert consensus statements, one suggesting a framework for measuring abuse liability for use in trials of analgesics for those with chronic pain53 and the other a set of definitions for opioid-related adverse events,44 identified 8 loose and overlapping categories of problematic use, including misuse, abuse, addiction, aberrant use, dependence, nonmedical or nontherapeutic use, physical dependence, and psychological dependence (also see the review of Webster and Fine,60 who further define “pseudoaddiction”). The vagueness inherent in these definitions, areas of overlap among them, and their sometimes interchangeable use have made it difficult to determine exact rates and types of problematic opioid use. For example, the narrative review of Højsted and Sjøgren24 detailed the findings of 25 studies involving patients with chronic pain prescribed with opioids, concluding that the prevalence of problematic opioid use behavior ranged from 0% to 50%. Although this span was representative of the literature at the time, it was of questionable value for delineating the scope, impact, and prevalence of the problem or in facilitating informed clinical and policy decisions regarding the allocation of screening and treatment resources. Martell et al.,38 in their review of opioid use for low back pain, reported a similar range of current problematic opioid use (3% to 43%).
The purpose of this study was to perform an updated review of problematic opioid use in chronic pain using explicitly defined terms44,53 for rates of problematic use in the literature. We synthesized the data to clarify and calculate prevalence estimates to increase precision and utility. As a secondary set of analyses, we investigated whether variation in the rates of problematic opioid use were related to differences in study characteristics (ie, primary study purpose, study design, method of assessment, clinical setting).
2.1. Literature search strategy
We searched the clinical and scientific literature using Science Direct, Google Scholar, PubMed, and PsychINFO/PsycArticles databases for articles published between January 2000 and January 2013. We repeated the search in November 2013 to include articles published or accepted since January 2013. We used broad search terms to increase the probability of accurate identification of target articles (Table 1). We also reviewed reference lists to identify any articles that the initial search had missed.
2.2. Abstract screening
The abstracts of all studies identified in the literature search were read by 2 reviewers to assess eligibility for full-text data extraction. To be eligible for data extraction, studies met the following criteria: (1) only adult participants (ie, 18+ years of age), (2) sample composed of individuals with chronic noncancer pain (persistent pain lasting longer than 3 months), (3) participants were using opioids orally (to exclude studies of opioids delivered transdermally or through injection/intrathecal pump), (4) the abstract listed 1 or more of the following terms in reference to opioid use: abuse, misuse, dependence, addiction, or aberrant/problematic behavior, and (5) quantitative information was provided (as opposed to a commentary or qualitative review) regarding rates of problematic opioid use.
2.3. Full-text data extraction
Each study fitting the inclusion criteria was read in full by 2 members of the study team to extract and record data on a standardized data extraction form. The extracted information included participant demographics and pain details (ie, sample size, gender, age, pain duration, ethnicity, pain location), primary objective (eg, assessment of prevalence, medication safety/efficacy), study design (ie, cross-sectional/prospective/retrospective), study setting details, country of data collection), and method used to identify problematic opioid use (ie, structured/unstructured clinical interview, urine drug screen [UDS], chart review, clinical judgment, questionnaire).
2.3.1. Coding of current opioid misuse, abuse, and addiction
Problematic use of opioids was categorized according to recent consensus statements published by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT)44 and Analgesic, Anesthetic, and Addiction Clinical Trials, Translations, Innovations, Opportunities, and Networks (ACTTION)53 panels, an overlapping group of experts with representation from private, public, and governmental domains. Based on these consensus statements, and the associated commentaries of Butler9 and Sullivan,54 the following definitions were used to categorize problematic use as misuse, abuse, or addiction.
- (1) Misuse: Opioid use contrary to the directed or prescribed pattern of use, regardless of the presence or absence of harm or adverse effects.
- (2) Abuse: Intentional use of the opioid for a nonmedical purpose, such as euphoria or altering one's state of consciousness.
- (3) Addiction: Pattern of continued use with experience of, or demonstrated potential for, harm (eg, “impaired control over drug use, compulsive use, continued use despite harm, and craving”).9(p2243)
Only these 3 terms were coded. Additional categorization of terms was not deemed appropriate because the terms defined in the consensus statements were either not relevant to the purposes of the present review (eg, diversion, intoxication, suicide-related use) or were not specific enough in their delineation of problematic use patterns (eg, aberrant opioid-related behaviors, nonmedical opioid use).
The following guidelines were used to code the type of problematic use. First, the percentage of study participants meeting criteria for each type of problematic use was extracted and recorded, where possible, to the tenths decimal place. A single percentage was recorded from studies that met the criteria for only a single type of problematic use, whereas studies that reported separately on more than 1 type of problematic use provided more than 1 estimate (eg, 1 for participants meeting criteria for misuse and 1 for participants meeting criteria for addiction). Second, when studies reported a range of values regarding the percentage of patients meeting criteria for 1 type of problematic use, a minimum and maximum value was recorded. Third, only current problematic opioid use was recorded; data were not used if a study reported only on historical or lifetime problematic opioid use. When insufficient or ambiguous information was provided in the published articles or available supplemental data, we contacted study authors for additional details.
When possible, rates of opioid misuse, abuse, and addiction were recorded directly from study text (eg, Refs. 5,36). When no specific rate was reported, a calculation was performed based on the number of patients meeting criteria for misuse, abuse, or addiction divided by the sample size (eg, Refs. 57,65). When multiple forms of behavior indicating the same type of problematic use were collapsed and reported in the study as a single value, the single value was recorded (eg, Ref. 17, where a percentage of 3.2% was presented as a combined value for various forms of opioid misuse). Finally, when the original study included a psychometric evaluation of a questionnaire and used nonquestionnaire data to evaluate issues of sensitivity and specificity (eg, in the identification of questionnaire cut-scores), then rates of problematic use from the nonquestionnaire data were recorded (eg, Ref. 21).
Each included study had at least 1 codeable percentage (with an upper limit of 6 if minimum and maximum values for misuse, abuse, and addiction were all reported). Categorization of problematic opioid use was performed independently by 2 reviewers (K.E.V and M.L.M) and, in the case of disagreement regarding categorization, a consensus was reached after discussion.
2.3.2. Rating of study quality
The quality of each study was rated using 8 of the 9 criteria used by Chou et al.13(p.146.e3) in their review of measures to predict and identify problematic drug-related behaviors. The first criterion of Chou et al., which determines whether the study evaluated test performance in a population other than the one used to derive the instrument (ie, derivation vs validation study), was coded but eventually discarded as it was deemed less useful in discriminating between high and low quality.
The remaining 8 criteria evaluate study sampling issues (eg, consecutive sample or random subset, proportion of missing data), adequate description of study methods (eg, sample and patterns of opioid prescription, criteria to identify problematic behavior), and potential influence of raters on identification of problematic behavior (eg, rater blinding regarding the identification of problematic use). Consistent with Chou et al., studies that met the majority of the criteria (5 or more) were regarded as higher quality.
2.4. Analytic plan
Extracted data were entered into SPSS (version 21; IBM Corporation). The primary variables of interest were average rates of misuse, abuse, and addiction across studies. Because a small number of studies reported these rates as a range of values, 2 sets of calculations were performed for each type of problematic use, a minimum and a maximum. When only a single value was recorded, that value was entered as both the minimum and maximum value as that ensured that both the minimum and maximum calculations included the complete set data. Although we expected minimum and maximum values to be close to one another, this method of calculation was deemed to make best use of all available data and allow equal weightings for each study's data.
The first analytic step involved the calculation of unweighted raw means and SDs for rates of misuse, abuse, and addiction across all included studies. In addition, we calculated a number of weighted means, including weighting for raw sample size and log-transformed sample size. The log transformation was performed to address the large variability in sample size and apparent exponentiation of the sample size distribution within the largest studies. In addition, a Winsorizing procedure was performed for studies with sample sizes of greater than 1334 participants, which was the point at which outliers were identified within stem-and-leaf plots; there was also evidence of a bimodal distribution at this cut-point. For the analyses using the Winsorized sample size data, samples size for all studies with greater than 1334 participants were set to 1334, the value of the next largest sample size.
In addition to the analyses involving weightings by sample size, weighted means were calculated for study quality. Furthermore, means for studies of high and low quality were evaluated separately. Finally, a weighted interaction term of log-transformed sample size and quality rating was calculated using standardized scores (z-scores).
As a secondary set of analyses, differences in rates of problematic use were assessed in relation to primary study purpose (ie, Was the assessment of prevalence of misuse, abuse, and addiction the primary aim?), study design (ie, retrospective, cross-sectional, prospective), method of identification (eg, questionnaire, structured/semistructured interview, chart review, UDS), and clinical setting (eg, primary care, pain clinic). A series of analyses of variance (ANOVAs) was used to analyze for differences in rates of problematic use based on these study characteristics.
3.1. Search results
Figure 1 displays the flow of information throughout the different phases of the search in a manner consistent with the PRISMA statement.41 The search yielded a total of 311 records for screening after the exclusion of 46 nonempirical papers, such as reviews, letters, and commentaries. An additional 9 records were identified in the updated, November 2013, search, yielding a total of 320 records for screening.
Kappa values indicated an acceptable level of agreement among raters, range κ = 0.79 to 0.91. All articles that had a discrepant rating after this stage of evaluation were retained for full-text data review.
A total of 78 articles were retained for full-text review. Of these, 40 were excluded for the reasons outlined in Figure 1. Therefore, 38 articles were used in data synthesis.
3.2. Characteristics of selected studies
Individual study characteristics are located in Table 2. The majority of studies, 35 (92%), reported on either misuse or addiction, whereas the remaining 3 studies reported on both. In total, 29 studies (76%) reported on rates of misuse and 12 (32%) on rates of addiction. Abuse was reported in only a single study, that of Banta-Green et al.,5 as this was the only study that reported specifically on participant intention. Therefore, no further calculations of abuse prevalence were performed.
Generally, considerable variability regarding study characteristics was apparent. Sample size, for example, ranged from 63 to 938,586 participants. Quality ratings ranged from 0 to 8. Sample size and quality ratings were significantly and negatively correlated with one another, r = −0.36, P < 0.05. There was also variability in reporting basic demographic and pain-related information. Specifically, 77.5% of studies reported on participant sex, 70.0% provided some information on age (with 15.0% providing nonnumeric information that could not be extracted—eg, “most patients fell into the 35- to 50-year-old range”), 47.5% provided information of any kind on participant ethnicity, and only 22.5% provided information on education. Regarding pain-related information, a minority of studies provided information on pain location (42.5%), or information on pain duration (37.5%).
3.3. Rates of opioid misuse and addiction
Overall, sizeable variability in rates of both misuse and addiction was indicated across reviewed studies. Rates of abuse ranged from 0.08% to 81.0% and addiction rates ranged from 0.7% to 34.1% across all studies. For high-quality studies (n = 13 for misuse and 10 for addiction), misuse rates ranged from 2.0% to 56.3% and addiction rates from 0.7% to 23.0%. For low-quality studies (n = 16 for misuse and 2 for addiction), misuse rates ranged from 0.08% to 81.0% and for addiction from 8.4% to 34.1%.
Table 3 and Table 4 display means, SDs, and 95% CI calculations for misuse and addiction, respectively. Regarding the calculation methods used to evaluate average rates of misuse and addiction, most means (excluding means calculated by raw sample size weighting and low-quality studies) were within 8% of one another for misuse and within 3% of one another for addiction. Specifically, rates of misuse ranged from a minimum of 21.7% for the mean weighted by the Winsorized sample size to a maximum of 29.3% for the unweighted mean. Rates of addiction ranged from a minimum of 7.8% for the mean weighted by Winsorized sample size to a maximum of 11.7% for the unweighted mean. Calculation of 95% CI indicated an overall range across all methods of mean calculation of 12.9% to 37.5% for misuse and 3.2% to 17.3% for addiction.
Two mean calculation methods yielded means that were markedly different from the rest including means weighted by raw sample size and means of low-quality studies. Means weighted by raw sample size were approximately 69% for misuse and approximately 4% for addiction. For low-quality studies, means were approximately 32% for misuse and 23% for addiction. The 95% CI calculated for these 2 methods were also noticeably broader than those calculated using the other methods, overall range of 16.5% to 76.5% for misuse and 0.8% to 39.2% for addiction.
3.4. Comparisons of study design, diagnostic method, and clinical setting
As noted, because the studies identified for data extraction were quite varied regarding their characteristics, we examined rates of misuse and addiction across studies regarding primary study purpose, study design, assessment method used to identify problematic behavior, and clinical setting. For each of these 4 variables, 4 ANOVAs were conducted (minimum/maximum; misuse/addiction). A Bonferroni correction was used for all pairwise comparisons to help control against the commission of a type I error.
Across all analyses, results indicated only 1 significant difference in relation to study characteristics. Specifically, mean unweighted rates of opioid addiction were lower in the 7 studies that identified the assessment of prevalence as the primary study objective, minimum/maximum mean = 5.5%/6.2% (SD = 4.6%/6.2%; 95% CI = 2.1%-10.8%), in comparison with 5 studies for which prevalence assessment was a secondary objective, minimum/maximum mean = 18.4%/19.4% (SD = 10.7%/9.4%; 95% CI = 9.0%-27.6%), all F > 8.3, all P < 0.02.
For opioid misuse, 11 studies (37.9%) identified the assessment of prevalence as the primary study aim and 18 studies (62.1%) as a secondary aim. No significant differences were indicated in average rate of misuse across studies, all F< 2.7, all P> 0.11.
All other comparisons did not indicate any significant differences in relation to the additional study characteristics evaluated. Specific findings are detailed in the following paragraphs and descriptive information is provided in Table 5.
Regarding study design, of the 38 studies reviewed, 39.5% were prospective, 34.2% were cross-sectional, and 26.3% were retrospective. No significant differences were indicated by any of the analyses comparing rates of misuse and addiction with design, all F< 1.0, all P> 0.37.
The assessment method used also varied substantially across studies with the majority, 64.9%, using only a single assessment method (questionnaire: 21.6%, clinical judgment/chart review: 21.6%; structured/semistructured interview: 13.5%; UDS: 8.1%). The remaining 35.1% of studies used a UDS plus at least 1 other method, which were coded as a single assessment category (ie, UDS plus at least one other method). The misuse comparisons were nonsignificant, all F< 0.71, all P> 0.59. For addiction, although comparisons of questionnaire and structured/semistructured interviews with multiple assessment methods reached a traditional level of significance, P < 0.05, the follow-up Bonferroni-controlled pairwise comparisons were not significant.
Finally, for evaluations involving clinical setting, 52.6% of studies involved data collected within a specialty chronic pain clinic with an additional 26.3% of data collected in primary care. Of the remaining studies, the clinical setting from which the data were collected was not clearly identified (eg, clinical trials registry; toxicology laboratory). Given the diversity in clinical setting, comparisons used only data from pain clinics and primary care. Consistent with the other analyses of study characteristics, no significant differences were indicated, all F< 0.52, P> 0.49.
Accurate identification and enumeration of problematic opioid use in those with chronic pain is important. Our review evaluated the current state of the literature regarding rates of opioid misuse, abuse, and addiction in chronic pain. The results are concordant with previous work in many ways. Chiefly, the substantial variability in studies evaluating problematic opioid use remains apparent as there were many designs used, methods of identification used, and study settings examined. The range of rates of problematic use was even broader than that has been reported in previous work24,38 with rates ranging from 0.08%52 to 81%.62
We took several steps within the review to address this expected variability. First, we coded for specific types of problematic use by adopting the definitions offered by the IMMPACT and ACTTION groups.44,53 In the order of severity, these types were: misuse (use not in accordance with prescribed directions, regardless of the presence or absence of harm resulting from use), abuse (intentional use for a nonmedical purpose), and addiction (use demonstrated harm or high potential for harm). In total, 38 articles were included in the full review, with 76% providing information on misuse and 32% providing information on addiction. Only a single study reported on abuse. Although the rates of misuse encompassed the entire range documented (ie, 0.08%-81%), the range for rates of addiction was somewhat more constrained, 0.7% to 34.1%.
Second, we calculated several weighted means and also separate means for high- and low-quality studies, with the overall goal of determining whether a subset of these scores would provide a degree of confidence with the rates identified. With the exception of means weighted by sample size and means for low-quality studies, which were particularly different than other means calculated, there appeared a level of concordance across the majority of mean calculations. On average, misuse was documented in approximately 1 of 4 or 5 patients (actual mean percentage range: 21.7%-29.3%) and addiction in approximately 1 of 10 or 11 patients (actual mean percentage range: 7.8%-11.7%). Perhaps the 2 most robust calculation methods were the sample size by study-quality interaction term and the mean of the high-quality studies only. For these 2 methods, rates of misuse ranged from 23.6% to 24.9% and rates of addiction from 8.8% to 10.7%. Furthermore, the observed SD for the high-quality studies was approximately half of that observed for the low-quality studies and two-thirds of that observed across all other calculations, suggesting a lesser degree of variability among these studies, and therefore perhaps bolstering confidence to some degree in the accuracy of these values.
Third and finally, we examined whether differences in study results could be at least partially explained by variability in the study methods that were used. Almost all comparisons based on study characteristics indicated a lack of significant differences regarding rates of abuse and addiction across different study designs, methods of assessment, and clinical settings (specialty pain clinic vs primary care). Only a single statistically significant difference was indicated between studies with a primary purpose of assessing addiction prevalence and those that assessed it as a secondary purpose such that lower rates of addiction were indicated in studies specifically designed to assess prevalence. As these analyses were likely underpowered because small cell sizes and the ranges analyzed were broad, these results ought to be interpreted cautiously, and we include them here to primarily provide information of potential use to future studies in this area.
We can make several recommendations for future studies of problematic opioid use in chronic pain. First, studies must specify the relevant demographic and pain-related details. At a minimum, we suggest that these include gender, age, and ethnicity, as well as pain location and duration. These details were included in a surprisingly small number of studies despite their demonstrated relevance in treatment response and role in the potential for problematic opioid use.24 The inclusion of measures of pain intensity and interference would likely provide valuable additional information. Second, there is likely a benefit to be found in specifying type of problematic use that is being assessed and specifically designing studies to evaluate prevalence as a primary objective. Such specification may aid in decreasing variability across studies regarding rates of problematic use and perhaps also have the added benefit of allowing for greater precision in the language used in relation to patterns of opioid use in chronic pain. Third, at present, there is no clear gold standard for use in the identification of misuse, abuse, and addiction.48 Perhaps the most thorough method is the Aberrant Drug Behavior Index (ADBI) used by Butler et al.10 The ADBI involves a triangulation approach consisting of self-reported patterns of opioid use evaluated by a structured interview, physician-reported patterns of use, and a UDS. A positive ADBI, indicating the presence of problematic opioid use, consisted of either a positive rating on the structured clinical interview or positive ratings on both the physician report and UDS. In this review, this triangulation method was coded as indicating misuse, but it seems feasible to modify it so that it also provides information regarding abuse and addiction.
The results of this review have 2 key implications. First, misuse and addiction do seem to be distinct patterns of problematic opioid use, at least based on the definitions used here and the differences in observed mean rates between them. Second, misuse seems more common than addiction. Several types of misuse were identified within studies and included underuse, erratic or disorganized use, inappropriate use (eg, to manage symptoms of anxiety or other sorts of distress), use in conjunction with alcohol or illegal substances (eg, marijuana), and, of course, overuse. If it is accurate that approximately 1 in 4 patients on opioids display patterns of opioid misuse, but not addiction, then perhaps more efficient targeting of treatment resources would be of benefit. Some forms of misuse, for example, may be readily addressed through relatively low-intensity methods such as education or frequent follow-up visits. One prominent example of a fairly low-intensity intervention is that of Jamison et al.,28 who held monthly meetings with patients deemed to be at “high risk” of opioid misuse. These meetings were a combination of motivational approaches, opioid education, and opioid use monitoring, including a UDS, held monthly over the course of 6 months. At the conclusion of the study period, the documented rates of aberrant behavior was low and comparable to rates documented for another group of patients, who were deemed to be of “low risk” of opioid misuse at the onset of the study. These findings suggest that there are alternatives available to providers who treat high-risk patients beyond simply not prescribing the medications at all. A more recent study from the same group37 further highlights a potential key role of cravings in opioid misuse, which presents another option for intervention given that the substance abuse literature already provides effective interventions directed at altering the impact of drug cravings more generally, and these could perhaps be readily adapted to problematic opioid use.7,63,64
The results of this review have several limitations. The most obvious is the degree of variability within this literature. In spite of our attempts to minimize the impact of this variability, the range of misuse and addiction was incredibly broad, as were measures of dispersion. Furthermore, there are other potential sources of variability in findings that were not possible to code and extract in a uniform manner. These include duration of opioid use, history of nonopioid substance misuse, abuse, or addiction, dosage levels and frequency of use, as well as health care system variables, such as frequency of prescription reviews, drug testing, or use of opioid “contracts.” These sources of variability will likely continue to cloud our ability to make precise estimates. There is clearly room here for a series of carefully controlled studies where sources of variability are held constant, or as constant as possible, to more clearly illuminate prevalence rates of problematic opioid use in individuals with chronic pain.
There was 1 curious finding that we have not yet emphasized. The overwhelming majority of studies within this review took place in the United States. Only 3 of the 38 studies took place in other countries, which suggests that this issue is of both high interest and is perhaps a problem that is somehow uniquely relevant to the US. The latter interpretation is supported by the finding of Manchikanti et al.31 indicating that the US population, which represents approximately 5% of the Earth's population, consumed approximately 80% of the global supply of prescribed opioids in the first decade of this century. This is an intriguing issue and although there are likely many factors involved, neither the abundance of opioids prescribed for the treatment of chronic pain nor the large proportion of studies of problematic opioid use seem to have helpfully diminished the prevalence, impact, or cost of chronic pain in the US since the explosion in opioid use for chronic pain.22
One final, related, comment on the use of opioids in chronic pain seems appropriate. In short, it is not clear whether the risks of opioid use outweigh the potential for benefit. The efficacy of opioids and their suitability for the long-term management of chronic pain still remain very much in question3,4,13,51,54,55 and while this uncertainty in effectiveness is well established, it stands in somewhat stark contrast to the clinical reality of chronic pain treatment, where rates of prescriptions have skyrocketed such that opioids are now among the most frequently prescribed medications. What does seem clear, however, is that the rapid increase in opioid use has had what Sullivan54 referred to as “unintended” consequences that, for at least some patients, require an additional form of intervention to curtail patterns of problematic use and potential for harm. We are not certain whether the benefits derived from opioids, which are rather unclear based on the extant literature, compensate for this additional burden to patients and health care systems.
Conflict of interest statement
The authors have no conflicts of interest to declare.
This research was supported by a grant from the Center for Health Policy at the Robert Wood Johnson Foundation to the first and last authors.
The authors thank Dr Robert Valdez for his insightful and helpful comments on a previous version of this report.
. Adams EH, Breiner S, Cicero TJ, Geller A, Inciardi JA, Schnoll SH, Senay EC, Woody GE. A comparison of the abuse liability of tramadol, NSAIDs, and hydrocodone in patients with chronic pain. J Pain Symptom Manage 2006;31:465–76.
. Atluri S, Sudarshan G, Manchikanti L. Assessment of the trends in medical use and misuse
of opioids analgesics from 2004 to 2011. Pain Physician 2014;17:E119–28.
. Ballantyne JC. Is lack of evidence the problem? J Pain 2010;11:830–2.
. Ballantyne JC, Shin NS. Efficacy of opioids for chronic pain: a review of the evidence. Clin J Pain 2008;24:469–78.
. Banta-Green CJ, Merrill JO, Doyle SR, Boudreau DM, Calsyn DA. Opioid use behaviors, mental health and pain—development of a typology of chronic pain patients. Drug Alcohol Depend 2009;104:34–42.
. Birnbaum HG, White AG, Schiller M, Waldman T, Cleveland JM, Roland CL. Societal costs of prescription opioid abuse, dependence, and misuse
in the United States. Pain Med 2011;12:657–67.
. Bowen S, Chawla N, Marlatt GA. Mindfulness-based relapse prevention for addictive behaviors: a clinician's guide. New York, NY: Guilford Press, 2010. p. 179.
. Brown J, Setnik B, Lee K, Wase L, Roland CL, Cleveland JM, Siegel S, Katz N. Assessment, stratification, and monitoring of the risk for prescription opioid misuse
and abuse in the primary care setting. J Opioid Manag 2011;7:467–83.
. Butler S. The IMMPACT factor or IMMPACT strikes again! Pain 2013;154:2243–4.
. Butler SF, Budman SH, Fanciullo GJ, Jamison RN. Cross validation of the Current Opioid Misuse
Measure (COMM) to monitor chronic pain patients on opioid therapy. Clin J Pain 2010;26:770–6.
. Butler SF, Budman SH, Fernandez K, Jamison RN. Validation of a screener and opioid assessment measure for patients with chronic pain. Pain 2004;112:65–75.
. Chelminski PR, Ives TJ, Felix KM, Prakken SD, Miller TM, Perhac JS, Malone RM, Bryant ME, DeWalt DA, Pignone MP. A primary care, multi-disciplinary disease management program for opioid-treated patients with chronic non-cancer pain and a high burden of psychiatric comorbidity. BMC Health Serv Res 2005;5:3.
. Chou R, Fanciullo GJ, Fine PG, Miaskowski C, Passik SD, Portenoy RK. Opioids for chronic noncancer pain: prediction and identification of aberrant drug-related behaviors: a review of the evidence for an American Pain Society and American Academy of Pain Medicine clinical practice guideline. J Pain 2009;10:131–46.
. Compton PA, Wu SM, Schieffer B, Pham Q, Naliboff BD. Introduction of a self-report version of the Prescription Drug Use Questionnaire and relationship to medication agreement non-compliance. J Pain Symptom Manage 2008;36:383–95.
. Couto JE, Romney MC, Leider HL, Sharma S, Goldfarb NI. High rates of inappropriate drug use in the chronic pain population. Popul Health Manag 2009;12:185–90.
. Cowan DT, Wilson-Barnett J. A survey of chronic noncancer pain patients prescribed opioid analgesics. Pain Med 2003;4:340–51.
. Edlund MJ, Martin BC, Fan M, Devries A, Braden JB, Sullivan MD. Risks of opioid abuse and dependence among recipients of chronic opioid therapy: results from the TROUP study. Drug Alcohol Depend 2010;112:90–8.
. Edlund MJ, Sullivan M, Steffick D, Harris KM, Wells KB. Do users of regularly prescribed opioids have higher rates of substance use problems than nonusers? Pain Med 2007;8:647–56.
. Federal Drug Administration. Attention prescribers: FDA seeks your help in curtailing the US opioid epidemic. 2013:1–5. Available: http://www.fda.gov/Drugs/DrugSafety/InformationbyDrugClass/ucm330614.htm
. Accessed 13 May, 2014.
. Fleming MF, Balousek SL, Klessig CL, Mundt MP, Brown DD. Substance use disorders in a primary care sample receiving daily opioid therapy. J Pain 2007;8:573–82.
. Fleming MF, Davis J, Passik SD. Reported lifetime aberrant drug-taking behaviors are predictive of current substance use and mental health problems in primary care patients. Pain Med 2008;9:1098–106.
. Gaskin DJ, Richard P. The economic costs of pain in the United States. J Pain 2012;13:715–24.
. Højsted J, Nielsen PR, Guldstrand SK, Frich L, Sjøgren P. Classification and identification of opioid addiction in chronic pain patients. Eur J Pain 2010;14:1014–20.
. Højsted J, Sjøgren P. Addiction to opioids in chronic pain patients: a literature review. Eur J Pain 2007;11:490–518.
. Ives TJ, Chelminski PR, Hammett-Stabler CA, Malone RM, Perhac JS, Potisek NM, Shilliday BB, DeWalt DA, Pignone MP. Predictors of opioid misuse
in patients with chronic pain: a prospective cohort study. BMC Health Serv Res 2006;6:46.
. Jamison RN, Butler SF, Budman SH, Edwards RR, Wasan AD. Gender differences in risk factors for aberrant prescription opioid use. J Pain 2010;11:312–20.
. Jamison RN, Link CL, Marceau LD. Do pain patients at high risk for substance misuse
experience more pain? A longitudinal outcomes study. Pain Med 2009;10:1084–94.
. Jamison RN, Ross EL, Michna E, Chen LQ, Holcomb C, Wasan AD. Substance misuse
treatment for high-risk chronic pain patients on opioid therapy: a randomized trial. Pain 2010;150:390–400.
. Katz NP, Sherburne S, Beach M, Rose RJ, Vielguth J, Bradley J, Fanciullo GJ. Behavioral monitoring and urine toxicology testing in patients receiving long-term opioid therapy. Anesth Analg 2003;97:1097–102.
. Manchikanti L, Cash KA, Damron KS, Manchukonda R, Pampati V, McManus CD. Controlled substance abuse and illicit drug use in chronic pain patients: an evaluation of multiple variables. Pain Physician 2006;9:215–26.
. Manchikanti L, Fellows B, Ailinani H, Pampati V. Therapeutic use, abuse, and nonmedical use of opioids: a ten-year perspective. Pain Physician 2010;13:401–35.
. Manchikanti L, Helm S, Fellows B, Janata JW, Pampati V, Grider JS, Boswell M V. Opioid epidemic in the United States. Pain Physician 2012;15:ES9–E38.
. Manchikanti L, Manchukonda R, Damron KS, Brandon D, McManus CD, Cash K. Does adherence monitoring reduce controlled substance abuse in chronic pain patients? Pain Physician 2006;9:57–60.
. Manchikanti L, Manchukonda R, Pampati V, Damron KS. Evaluation of abuse of prescription and illicit drugs in chronic pain patients receiving short-acting (hydrocodone) or long-acting (methadone) opioids. Pain Physician 2005;8:257–61.
. Manchikanti L, Pampati V, Damron KS, Beyer CD, Barnhill RC, Fellows B. Prevalence of prescription drug abuse and dependency in patients with chronic pain in western Kentucky. J Ky Med Assoc 2003;101:511–17.
. Manchikanti L, Pampati V, Damron KS, Fellows BM, Barnhill RC, Beyer C. Prevalence of opioid abuse in interventional pain medicine practice settings: a randomized clinical evaluation. Pain Physician 2001;4:358–65.
. Martel MO, Dolman AJ, Edwards RR, Jamison RN, Wasan AD. The association between negative affect and prescription opioid misuse
in patients with chronic pain: the mediating role of opioid craving. J Pain 2014;15:90–100.
. Martell B, O'Connor P, Kerns R, Becker W, Morales K, Kosten T, Fiellin D. Opioid treatment for chronic back pain: Prevalence, efficacy, and association with addiction. Ann Intern Med 2007;146:187–92.
. Meltzer EC, Rybin D, Meshesha LZ, Saitz R, Samet JH, Rubens SL, Liebschutz JM. Aberrant drug-related behaviors: Unsystematic documentation does not identify prescription drug use disorder. Pain Med 2012;13:1436–43.
. Meltzer EC, Rybin D, Saitz R, Jeffrey H, Schwartz SL, Butler SF, Jane M. Identifying prescription opioid use disorder in primary care: diagnostic characteristics of the Current Opioid Misuse
Measure (COMM). Pain 2011;152:397–402.
. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Br Med J 2009;339:332–6.
. Morasco BJ, Dobscha SK. Prescription medication misuse
and substance use disorder in VA primary care patients with chronic pain. Gen Hosp Psychiatry 2008;30:93–9.
. Naliboff BD, Wu SM, Schieffer B, Bolus R, Pham Q, Baria A, Aragaki D, Van Vort W, Davis F, Shekelle P. A randomized trial of 2 prescription strategies for opioid treatment of chronic nonmalignant pain. J Pain 2011;12:288–96.
. O'Connor AB, Turk DC, Dworkin RH, Katz NP, Colucci R, Haythornthwaite J A, Klein M, O'Brien C, Posner K, Rappaport BA, Reisfield G, Adams EH, Balster RL, Bigelow GE, Burke LB, Comer SD, Cone E, Cowan P, Denisco RA, Farrar JT, Foltin RW, Haddox JD, Hertz S, Jay GW, Junor R, Kopecky EA, Leiderman DB, McDermott MP, Palmer PP, Raja SN, Rauschkolb C, Rowbotham MC, Sampaio C, Setnik B, Smith SM, Sokolowska M, Stauffer JW, Walsh SL, Zacny JP. Abuse liability measures for use in analgesic clinical trials in patients with pain: IMMPACT recommendations. Pain 2013;154:2324–34.
. Passik SD, Messina J, Golsorkhi A, Xie F. Aberrant drug-related behavior observed during clinical studies involving patients taking chronic opioid therapy for persistent pain and fentanyl buccal tablet for breakthrough pain. J Pain Symptom Manage 2011;41:116–25.
. Portenoy RK, Farrar JT, Backonja MM, Cleeland CS, Yang K, Friedman M, Colucci S V, Richards P. Long-term use of controlled-release oxycodone for noncancer pain: results of a 3-year registry study. Clin J Pain 2007;23:287–99.
. Reid MC, Engles-Horton LL, Weber MB, Kerns RD, Rogers EL, O'Connor PG. Use of opioid medications for chronic noncancer pain syndromes in primary care. J Gen Intern Med 2002;17:173–9.
. Savage SR. Assessment for addiction in pain-treatment settings. Clin J Pain 2002;18:S28–38.
. Schneider JP, Kirsh KL. Defining clinical issues around tolerance, hyperalgesia, and addiction: a quantitative and qualitative outcome study of long-term opioid dosing in a chronic pain practice. J Opioid Manag 2010;6:385–95.
. Sekhon R, Aminjavahery N, Davis CN, Roswarski MJ, Robinette C. Compliance with opioid treatment guidelines for chronic non-cancer pain (CNCP) in primary care at a Veterans Affairs Medical Center (VAMC). Pain Med 2013;14:1458–556.
. Sites BD, Beach ML, Davis MA. Increases in the use of prescription opioid analgesics and the lack of improvement in disability metrics among users. Reg Anesth Pain Med 2014;39:6–12.
. Skurtveit S, Furu K, Borchgrevink P, Handal M, Fredheim O. To what extent does a cohort of new users of weak opioids develop persistent or probable problematic opioid use? Pain 2011;152:1555–61.
. Smith SM, Dart RC, Katz NP, Paillard F, Adams EH, Comer SD, Degroot A, Edwards RR, Haddox JD, Jaffe JH, Jones CM, Kleber HD, Kopecky EA, Markman JD, Montoya ID, O'Brien C, Roland CL, Stanton M, Strain EC, Vorsanger G, Wasan AD, Weiss RD, Turk DC, Dworkin RH. Classification and definition of misuse
, abuse, and related events in clinical trials: ACTTION systematic review and recommendations. Pain 2013;154:2287–796.
. Sullivan M. Clarifying opioid misuse
and abuse. Pain 2013;154:2239–40.
. Trescot AM, Helm S, Hansen H, Benyamin R, Glaser SE, Adlaka R, Patel S, Manchikanti L. Opioids in the management of chronic non-cancer pain: an update of American Society of the Interventional Pain Physicians' (ASIPP) Guidelines. Pain Physician 2008;11:S5–S62.
. United Nations Office on Drugs and Crime. Ensuring availability of controlled medications for the relief of pain and preventing diversion and abuse: striking the right balance to achieve the optimal health outcome. 2011 Available at: www.unodc.org/docs/treatment/Pain/Ensuring_availability_of_controlled_medications_FINAL-15_March_CND_version.pdf
. Accessed 1 March 2014
. Vaglienti RM, Huber SJ, Noel KR, Johnstone RE. Misuse
of prescribed controlled substances defined by urinalysis. WV Med J 2003;99:67–70.
. Volkow ND, McLellan TA, Cotto JH, Marithanom M, Weiss RRB. Characteristics of opioid prescriptions in 2009. J Am Med Assoc 2011;305:1299–300.
. Wasan AD, Butler SF, Budman SH, Fernandez K, Weiss RD, Greenfield SF, Jamison RN. Does report of craving opioid medication predict aberrant drug behavior among chronic pain patients? Clin J Pain 2009;25:193–8.
. Webster LR, Fine PG. Approaches to improve pain relief while minimizing opioid abuse liability. J Pain 2010;11:602–11.
. Webster LR, Webster RM. Predicting aberrant behaviors opioid-treated patients: preliminary validation of the Opioid Risk Tool. Pain Med 2005;6:432–43.
. Wilsey BL, Fishman SM, Tsodikov A, Ogden C, Symreng I, Ernst A. Psychological comorbidities predicting prescription opioid abuse among patients in chronic pain presenting to the emergency department. Pain Med 2008;9:1107–17.
. Witkiewitz K, Bowen S. Depression, craving, and substance use following a randomized trial of mindfulness-based relapse prevention. J Consult Clin Psychol 2010;78:362–74.
. Witkiewitz K, Bowen S, Donovan DM. Moderating effects of a craving intervention on the relation between negative mood and heavy drinking following treatment for alcohol dependence. J Consult Clin Psychol 2011;79:54–63.
. Wu SM, Compton P, Bolus R, Schieffer B, Pham Q, Baria A, Van Vort W, Davis F, Shekelle P, Naliboff BD. The addiction behaviors checklist: validation of a new clinician-based measure of inappropriate opioid use in chronic pain. J Pain Symptom Manage 2006;32:342–51.