You are here

Article Text

Article menu

Download PDFPDF

Original article
Traumatic injuries and persistent opioid use in the USA: findings from a nationally representative survey
Free
  1. Suliman Alghnam1,
  2. Renan Castillo2
  1. 1King Abdullah International Medical Research Center, King Saud Bin Abdulaziz University for Health Sciences, KAIMRC, KSAU-HS, Riyadh, Saudi Arabia
  2. 2Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
  1. Correspondence to Dr Suliman Alghnam, King Abdullah International Medical Research Center, King Saud Bin Abdulaziz University for Health Sciences, KAIMRC, KSAU-HS, P.O Box 22490 Riyadh 11426, Saudi Arabia; Ghnams{at}ngha.med.sa

Abstract

Background Although opioid abuse is a rising epidemic in the USA, there are no studies to date on the incidence of persistent opioid use following injuries. Therefore, the aims of this study are: (1) to examine the incidence of persistent opioid use among a nationally representative sample of injured and non-injured populations; (2) to evaluate whether an injury is an independent predictor of persistent opioid use.

Method Data from the Medical Expenditure Panel Survey were pooled (years 2009–2012). Adults were followed for about 2 years, during which they were surveyed about injury status and opioid use every 4–5 months. To determine whether injuries are associated with persistent opioid use, weighted multiple logistic regressions were constructed.

Results While 2.3 million injured individuals received any opioid during the follow-up, 371 170 (15.6%) individuals became persistent opioid users (defined as opioid use across multiple time points). In a multiple logistic regression analysis adjusting for sociodemographic characteristics and self-reported health, those who sustained injuries were 1.4 times (95% CI 1.1 to 1.9) more likely to report persistent opioid use than those without injuries.

Conclusions We found injuries to be significantly associated with persistent opioid use in a nationally representative sample. Further investment in injury prevention may facilitate reduction of persistent opioid use and, thus, improve population health and reduce health expenditures.

https://doi.org/10.1136/injuryprev-2016-042059

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Opioid abuse is a rising epidemic in the USA.1 In 2013, there were over 16 000 opioid-related deaths, a 200% increase relative to 2000.2 The increase in opioid-related deaths was concurrent with an increase in the number of users, the number of prescriptions filled and prescription costs.3 In 2002, over 27 million people purchased one or more outpatient-prescribed opioids, while this figure increased to over 36 million in 2012.4 Due to this rising burden, deaths associated with opioid overdose are now higher than those due to MVCs in 30 states.5

    Over 39 million Americans are injured every year,6 among whom pain remains a common concern.7 Following changes in pain management recommendations in 2000, and possibly due to marketing from pharmaceutical companies, opioid utilisation increased over the past decade.5 ,8 Consequently, more injured patients may be at risk of becoming persistent opioid users due to long-term use. Persistent opioid use has been associated with numerous adverse events such as overdose,9 long hospital stays,10 myocardial infarction, traffic crashes11 and even death.12

    The percentage of the US population who become persistent opioid users following injuries is currently unknown, but it is expected to have increased over the past decades. Some estimates put the prevalence of opioid dependence between 0.2% and 3.2%.5 More comprehensive investigations of persistent opioid use may provide researchers and policy-makers with a basis to plan prevention strategies. This is instrumental for monitoring purposes and to guide future programmes aimed to reduce the burden of persistent opioid use.

    No nationally representative studies have investigated the incidence of persistent opioid use following injuries. In this study, we aim to examine the incidence of persistent opioid use among both injured and non-injured populations. Additionally, we aim to evaluate whether sustaining an injury is an independent predictor of persistent opioid use after adjusting for potential confounders. We hypothesise that traumatic injuries are a significant predictor of persistent opioid use among US non-institutionalised populations.

    Methods

    This prospective observational study used data from the Medical Expenditure Panel Survey (MEPS). The survey is conducted annually by the US Agency for Healthcare Research and Quality (AHRQ) and the National Center for Health Statistics.13 The MEPS is a nationally representative sample and follows participants for about 2 years, during which participants are interviewed five times (4–5 months apart). For the present study, we used panels 14–17 that started between years 2009 and 2012. The survey comprises several data files covering areas such as expenditures, demographics, health status and medical conditions. We used three files: the population characteristics file, the prescription file and the medical conditions file. Detailed description of the survey methodology has been published previously.14

    This study included participants 18 years of age or older and who were followed for the study period. Injured individuals were selected if they reported any injury in at least one of the first three rounds of follow-up (∼first follow-up year), while the reference population included those without any reported injuries. Recidivists were those who reported an injury in more than one of the first three rounds. As this study was limited to de-identified and publicly available data, it was granted exemption by the institutional review board.

    Injury status

    In each of the five in-person interviews, medical conditions are collected as the reason for a particular episode of medical care (hospital stay, outpatient visit, emergency visit, home health episode, prescribed medication purchase or medical provider visit), the reason for one or more episodes of disability days or as a condition ‘bothering’ the person during the reference period. If a respondent reported a medical condition, he or she is then asked whether the condition was caused by an injury. Based on this response, an indicator variable was created to categorise injury status (yes vs no). In the injury group, only incidents that had occurred during follow-up were included to avoid bias due to including injured participants, who may have initiated opioids prior to the study. In addition, injured individuals, who reported opioid use prior to the injury incident, were excluded to ensure temporality between injuries and opioid use.

    Opioid use

    The prescription medication file was used to identify opioid users. The MEPS prescriptions file contains information on the medications a participant reported to have filled in a pharmacy during the study period. This study defined an opioid user as one who was prescribed opioids at least once in any round during the follow-up study. Based on the ‘Multum Lexicon classification systems’, codes that fit within two classifications of opioids were included. The first is narcotic analgesic which captured pure agonists like oxycodone. The second is narcotic analgesic combinations which captured combination drugs such as oxycodone/acetaminophen (eg, Percocet).15

    Persistent opioid use

    Use at a single time point was defined as opioid use in one round (4–5 months) of the five follow-up rounds. Persistent opioid use was defined as opioid use in more than one round (at least 8 months) following round 2 (ie, opioid use in rounds 3, 4, etc). To capture persistent opioid use, we created an indicator variable for those reporting opioid use in more than one round. This definition is consistent with several previous studies.5 ,16

    Statistical analysis

    STATA V.14 for Mac (STATA, College Station, Texas, USA) was used for all statistical analyses. Because the MEPS is a weighted sample and to account for the survey weights, strata and clustering of individuals, the survey command was used to provide nationally representative results of US non-institutionalised populations. Following AHRQ's recommendation, we divided survey weights by 2 to provide annual estimates of the injured and non-injured US populations.17 Descriptive characteristics by injury status were obtained. Bivariate analyses were explored using χ2 tests for categorical variables. A p value of <0.05 (two-sided) was considered statistically significant.

    Several variables were examined in the univariate analyses including age, gender, race/ethnicity, family income, insurance status and baseline self-reported health (excellent, very good, good, fair and poor). In addition, depression status was examined, since mental health conditions are among potential risk factors for persistent opioid use and injuries.18 ,19 Depression was explored using the Patient Health Questionnaire (PHQ-2) score. This instrument is used to screen for depressive symptoms and was found to have high sensitivity and specificity.20 The PHQ-2 score ranges between 0 and 6, and a score of 3 or higher is considered a positive screen for depressive symptoms. Because the PHQ-2 was measured twice during follow-up, we created a categorical variable: negative depressive symptoms in both years, one positive screen for depression and positive screens for depressive symptoms in both years.

    To determine whether injuries are associated with persistent opioid use, weighted simple and multiple logistic regressions were constructed adjusting for potential confounders. The following variables were significant in the univariate analysis and were included in the final model: gender (male as the reference), race/ethnicity (white as the reference, black and other), insurance status (private insurance as reference, public insurance and no insurance), baseline self-reported health and depressive symptoms. Because of its significant role in most health conditions, age (18–24 years as the reference, 25–44, 45–64 and ≥65) was included in the regression model despite its lack of significance in the univariate analysis. Interaction effects between injury status and age, gender and depression status were also investigated. The results were presented as ORs with associated 95% CIs.

    Results

    Of the 36 824 participants included in the study, 4713 (12.7%) met the injury inclusion criteria (table 1). After taking sampling weights, clusters and strata into account, the injured population was representative of over 25 million injured individuals, of which 2.3 million (9.1%) used opioids at least once during follow-up. The data suggest that 371 170 individuals (15.6%) became persistent opioid users during the 2-year period (figure 1). Injured individuals also represented about 10.6% of the approximately 3.5 million individuals who reported persistent opioid use in the USA.

    View this table:
    Table 1

    Description of the MEPS study population by injury status

    Figure 1
    Figure 1

    Weighted frequencies of Medical Expenditure Panel Survey (MEPS) participants by injury status and across opioid use categories. *Injuries sustained prior to the follow-up study. **Injuries sustained in either round 3 or 4.

    In the univariate analyses, injured individuals were more likely to be males, white and privately insured than those without injuries (p<0.01). While there were no differences in income and education, baseline self-reported health and depression status were significantly worse among injured individuals than those without injuries (table 1). In addition, injured individuals reported higher incidence of persistent opioid use than the injury-free population (2.8% vs 1.6%, p<0.01).

    Regression analyses indicate that a traumatic injury is a significant predictor of persistent opioid use (table 2). Those who sustained an injury were 73% more likely to report persistent opioid use (OR=1.7, 95% CI 1.3 to 2.2). Increased odds of persistent opioid use remained significant even after adjusting for other covariates (OR=1.4, 95% CI 1.1 to 1.9). Interaction effects between injury status and age, gender or depression status were not significant.

    View this table:
    Table 2

    Logistic regression analysis of predictors of long-term opioid use in the USA (N=36 824)

    Other health conditions were also significant predictors of persistent opioid use (table 2). Individuals who had two positive screens for depressive symptoms were three times more likely to report persistent opioid use than those without depressive symptoms, despite adjusting for other covariates (OR=3.0, 95% CI 2.2 to 4.2). Even higher odds of persistent opioid use were reported among patients with lower perceived health. Those who reported ‘poor’ baseline health were over 11 times more likely to report persistent opioid use than those reporting ‘excellent’ health adjusting for other variables (OR=11.8, 95% CI 7.2 to 18.4).

    Discussion

    This nationally representative study found traumatic injuries to be a significant predictor of persistent opioid use. A significant outcome of the study is that among injured individuals who used opioids at least once, over 15% became persistent opioid users. This is higher than the estimated 11.4%, who become chronic users after initiating opioid treatment in Canada.21 Our results also indicate that injured individuals, who became persistent opioid users, represent about 10.6% of the persistent opioid users in the USA. Public health programmes should target individuals with new injuries to reduce the burden of opioid dependence, abuse and associated medical costs.

    The evidence presented here, that injuries are associated with increased likelihood of persistent opioid use, lends further credence to previous studies.5 ,22 Our estimate of 2.8% is fairly similar to a study by Fishbain et al23 who conducted a systematic review of 67 studies and found that opioid abuse rates varied between 0.2% and 3.2%. However, other studies found higher estimates of opioid use among patients with trauma.18 ,22 Helmerhorst et al18 examined persistent opioid use among patients with trauma, and found 28% reported using opioids 1–2 months after musculoskeletal injuries. A study by Cannon et al22 found 20% patients were on prescription opioids at the time of the admission. Nevertheless, this study examined opioid incidence at the time of admission rather than long-term use and was limited to a single hospital. Hence, their estimate is not directly comparable with ours, and several confounders may have affected their findings. Another potential explanation for the discrepancy between our findings and prior literature has to do with inclusion criteria. The present study examined persistent opioid use in a subset of the population which was not injured, was not on opioids, then became injured prior to opioid use. Therefore, it is not unlikely that rates of persistent opioid use may differ from other populations.

    Compared with our result, a higher percentage of persistent opioid use was found among patients with trauma in other countries.24 ,25 A study by Zwisler et al,24 among Danish population, found that 7.0% of patients with trauma were persistent opioid users 6 months after injuries. Another study examined persistent opioid use among Australian workers and found 6.3% of those injured to be on opioid beyond the first year of injury.25 It is possible that the discrepancy with our findings is due to differences in the study design or, more likely, to prescribing practices between the USA and other countries.

    Among the study population, our estimate of persistent opioid use was lower than expected. Previous reports estimated that approximately 3.0% to 4.0% of the adult US population are persistent opioid users.26 Population-based estimates from other countries are much lower than that of the USA. A study by Mellbye et al27 examined persistent opioid use among the Norwegian population and found 0.6% to be persistent users. To the best of our knowledge, the only nationally representative US study was conducted by Sites et al15 to examine trends in the use of prescription opioid analgesics in the USA. The authors found that 11.8% of the population were opioid users, but their study examined any opioids rather than persistent opioid use. In 2010, it was estimated that over 27 million were opioid users, and according to our finding, 3.5 million of those are persistent opioid users.28 This is a major concern from a public health point of view and merits further investigation.

    We found other health conditions to be associated with persistent opioid use, one of which is depression. Depressed individuals were three times more likely to be persistent opioid users than those without depression symptoms. This is very similar to the threefold to fourfold increase in the likelihood of persistent opioid use reported by prior findings.5 Interestingly, a previous study using MEPS found depression to be associated with injuries and with injury recidivism.19 There is a need for further studies to investigate whether addressing depressive symptom is associated with reduction of persistent opioid use.

    Our study has important implications for public health. First, the estimated incidence of persistent opioid use provides a basis for initiatives aiming to reduce opioid dependence and measure the success of implemented programmes over time. For example, future studies of the incidence of persistent opioid use following the recent CDC opioid prescribing guidelines for chronic pain may provide valuable findings regarding its effectiveness.29 A study by Franklin et al30 in Washington State examined opioid use before and after implementation of a specific opioid dosing guideline and found a decline in the number of opioid-related deaths among injured workers. Second, we found injuries to be independently associated with persistent opioid use. Therefore, investing in preventing injuries may facilitate a sizeable decline of persistent opioid use. Third, our findings suggest depression is a strong predictor of persistent opioid use; thus, public health advocates could use these findings to call for initiatives in order to improve mental health and reduce persistent opioid use.

    The study has several limitations. First, the data are self-reported and may not necessarily reflect true prescription or persistent opioid use. However, a previous study, using the MEPS dataset, validated self-reported prescriptions and found that respondents accurately reported their medications.31 Second, even if patients filled those prescriptions, actual adherence cannot be ascertained. The possibility remains that patients stopped using medications and were not persistent opioid users. Third, important variables such as alcohol abuse are not collected by the MEPS and, therefore, we were not able to assess their association with persistent opioid use. Fourth, injury mechanism is not collected in the MEPS, which limits the ability to design more specific injury prevention programmes. Finally, our study followed individuals for about 2 years. Therefore, it is unknown whether opioid use will continue in the upcoming years. Future longer prospective studies should take this into account to provide more comprehensive evidence.

    Our findings may underestimate the true prevalence of persistent opioid use for a number of reasons. First, we excluded hospitalised patients with more severe injuries as well as patients who died due to drug overdose or misuse, which may have occurred following persistent opioid use. Second, those who used opioids prior to injuries, including for other conditions, were also excluded and thus not represented in our sample. Third, individuals who were prescribed opioids during the study and then transitioned to be persistent users through illegal means may not appear in our data. Additionally, substance abuse may also lead to persistent opioid use as suggested by previous studies.32 Prescription opioids represent a major contributor to illegal recreational drugs in the USA, second after marijuana.33 As a result, the burden may have been underestimated since the sample was limited to those followed for a 2-year period. Finally, it is possible that for some participants injury may have occurred toward the end of the 2-year study period. Those who became persistent opioid users will not be captured due to our limited follow-up.

    In conclusion, we have shown that injures are significantly associated with persistent opioid use in a nationally representative sample. Public health interventions may use these findings as a basis for nationwide initiatives to reduce the burden of persistent opioid use among both injured and non-injured populations. Further investment in injury prevention may facilitate reduction of opioid abuse and, thus, improve population health and reduce health expenditures.

    What is already known on the subject?

    • Persistent opioid use is associated with overdose death and other negative health consequences.

    • Previous observational studies provided little evidence on the association between injuries and persistent opioid use, because they only examined short-term opioid use after injuries or their generalisability was not established because of limitations in study populations.

    What this study adds?

    • This research presents the first nationally representative estimates of persistent opioid use among both injured and non-injured US populations.

    • Among injured individuals who used opioids at least once, over 15% became persistent opioid users.

    • Injured individuals who became persistent opioid users represent about 10.6% of persistent opioid users in the USA.

    Keeping cyclists from getting ‘Doored’

    Some otherwise desirable bicycle lanes place cyclists in a position where they are squeezed between car lanes and parked cars. Thus, when drivers open their door without looking, the door may hit a passing cyclist. One solution worth adopting while waiting for better urban designs is the ‘Dutch Reach’. This entails teaching drivers to reach over with their right arm to open their door. This causes drivers to look back and see oncoming bicyclists. (In the case of right-side drivers, the left hand should be used.)

    OFFENDER restraints can be fatal

    In the UK, an internal risk assessment in young offender institutions shows that certain restraint procedures used against non-compliant children carry a 40%–60% chance of causing injuries. Many impair breathing or circulation. Comment: I wonder if these injuries would be classified as intentional or unintentional. Although one inquest judged the event to be accidental, the restraint technique used was withdrawn 2 months later.

    References

    1. ↵
      1. Nelson LS,
      2. Juurlink DN,
      3. Perrone J
      . Addressing the opioid epidemic. JAMA 2015;314:1453–4. doi:10.1001/jama.2015.12397
      OpenUrlCrossRefPubMed
    2. ↵
      Increases in Drug and Opioid Overdose Deaths—United States, 2000–2014. http://www.cdc.gov/mmwr/preview/mmwrhtml/mm64e1218a1.htm (accessed 23 Dec 2015).
    3. ↵
      1. Mosher HJ,
      2. Krebs EE,
      3. Carrel M, et al
      . Trends in prevalent and incident opioid receipt: an observational study in veterans health administration 2004–2012. J Gen Intern Med 2015;30:597–604. doi:10.1007/s11606-014-3143-z
      OpenUrlCrossRefPubMed
    4. ↵
      1. Stagnitti M
      . Trends in prescribed outpatient opioid use and expenses in the U.S. civilian noninstitutionalized population, 2002–2012. Rockville, 2015. http://meps.ahrq.gov/mepsweb/data_files/publications/st478/stat478.shtml
    5. ↵
      1. Sullivan MD,
      2. Howe CQ
      . Opioid therapy for chronic pain in the United States: promises and perils. Pain 2013;154 Suppl 1(Suppl):S94–100. doi:10.1016/j.pain.2013.09.009
      OpenUrl
    6. ↵
      1. McCaig LF,
      2. Burt CW
      . National Hospital Ambulatory Medical Care Survey: 2002 emergency department summary. Adv Data 2004:340:1–34.
      OpenUrlPubMed
    7. ↵
      1. Rivara FP,
      2. Mackenzie EJ,
      3. Jurkovich GJ, et al
      . Prevalence of pain in patients 1 year after major trauma. Arch Surg 2008;143:282–7; discussion 288. doi:10.1001/archsurg.2007.61
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Weisberg D,
      2. Stannard C
      . Lost in translation? Learning from the opioid epidemic in the USA. Anaesthesia 2013;68:1215–19. doi:10.1111/anae.12503
      OpenUrlCrossRefPubMed
    9. ↵
      1. Miller M,
      2. Barber CW,
      3. Leatherman S, et al
      . Prescription opioid duration of action and the risk of unintentional overdose among patients receiving opioid therapy. JAMA Intern Med 2015;175:608–15. doi:10.1001/jamainternmed.2014.8071
      OpenUrlCrossRefPubMed
    10. ↵
      1. Pandya U,
      2. O'Mara MS,
      3. Wilson W, et al
      . Impact of preexisting opioid use on injury mechanism, type, and outcome. J Surg Res 2015;198:7–12. doi:10.1016/j.jss.2015.05.033
      OpenUrlCrossRefPubMed
    11. ↵
      1. Gomes T,
      2. Redelmeier DA,
      3. Juurlink DN, et al
      . Opioid dose and risk of road trauma in Canada: a population-based study. JAMA Intern Med 2013;173:196–201. doi:10.1001/2013.jamainternmed.733
      OpenUrlCrossRefPubMed
    12. ↵
      1. Chou R,
      2. Turner JA,
      3. Devine EB, et al
      . The effectiveness and risks of long-term opioid therapy for chronic pain: a systematic review for a national institutes of health pathways to prevention workshop. Ann Intern Med 2015;162:276. doi:10.7326/M14-2559
      OpenUrlCrossRefPubMed
    13. ↵
      1. Alghnam S,
      2. Palta M,
      3. Remington P, et al
      . The association between motor vehicle injuries and health-related quality of life: a longitudinal study of a population-based sample in the United States. Qual Life Res 2014;23:119–27. doi:10.1007/s11136-013-0444-3
      OpenUrlCrossRefPubMed
    14. ↵
      1. Cohen JW,
      2. Monheit AC,
      3. Beauregard KM, et al
      . The Medical Expenditure Panel Survey: a national health information resource. Inquiry 1996;33:373–89.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Sites BD,
      2. Beach ML,
      3. 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. doi:10.1097/AAP.0000000000000022
      OpenUrlCrossRefPubMed
    16. ↵
      1. Heins SE,
      2. Feldman DR,
      3. Bodycombe D, et al
      . Early opioid prescription and risk of long-term opioid use among US workers with back and shoulder injuries: a retrospective cohort study. Inj Prev 2016;22:211–15. doi:10.1136/injuryprev-2015-041630
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Alghnam S,
      2. Vanness DJ,
      3. Gaskin DJ, et al
      . Estimating annual medical and out-of-pocket expenditures associated with traumatic injuries in the United States. J Trauma Acute Care Surg 2016;80:258–64. doi:10.1097/TA.0000000000000910
      OpenUrlCrossRefPubMed
    18. ↵
      1. Helmerhorst GT,
      2. Vranceanu AM,
      3. Vrahas M, et al
      . Risk factors for continued opioid use one to two months after surgery for musculoskeletal trauma. J Bone Joint Surg Am 2014;96:495–9. doi:10.2106/JBJS.L.01406
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Alghnam S,
      2. Tinkoff GH,
      3. Castillo R
      . Longitudinal assessment of injury recidivism among adults in the United States: findings from a population-based sample. Inj Epidemiol 2016;3:5. doi:10.1186/s40621-016-0071-x
      OpenUrlCrossRefPubMed
    20. ↵
      1. Godha D,
      2. Shi L,
      3. Mavronicolas H
      . Association between tendency towards depression and severity of rheumatoid arthritis from a national representative sample: the Medical Expenditure Panel Survey. Curr Med Res Opin 2010;26:1685–90. doi:10.1185/03007991003795808
      OpenUrlCrossRefPubMed
    21. ↵
      1. Kaplovitch E,
      2. Gomes T,
      3. Camacho X, et al
      . Sex differences in dose escalation and overdose death during chronic opioid therapy: a population-based cohort study. PLoS ONE 2015;10:e0134550. doi:10.1371/journal.pone.0134550
      OpenUrlCrossRefPubMed
    22. ↵
      1. Cannon R,
      2. Bozeman M,
      3. Miller KR, et al
      . The prevalence and impact of prescription controlled substance use among injured patients at a {Level I} trauma center. J Trauma Acute Care Surg 2014;76:172–5. doi:10.1097/TA.0b013e3182ab10de
      OpenUrlCrossRefPubMed
    23. ↵
      1. Fishbain DA,
      2. Cole B,
      3. Lewis J, et al
      . What percentage of chronic nonmalignant pain patients exposed to chronic opioid analgesic therapy develop abuse/addiction and/or aberrant drug-related behaviors? a structured evidence-based review. Pain Med 2008;9:444–59. doi:10.1111/j.1526-4637.2007.00370.x
      OpenUrlCrossRefPubMed
    24. ↵
      1. Zwisler ST,
      2. Hallas J,
      3. Larsen MS, et al
      . Opioid prescriptions before and after high-energy trauma. J Opioid Manag 2015;11:313–18. doi:10.5055/jom.2015.0280
      OpenUrlCrossRefPubMed
    25. ↵
      1. Berecki-Gisolf J,
      2. Collie A,
      3. McClure RJ
      . Prescription opioids for occupational injury: results from workers’ compensation claims records. Pain Med 2014;15:1549–57. doi:10.1111/pme.12421
      OpenUrlCrossRefPubMed
    26. ↵
      1. Boudreau D,
      2. Von Korff M,
      3. Rutter CM, et al
      . Trends in long-term opioid therapy for chronic non-cancer pain. Pharmacoepidemiol Drug Saf 2009;18:1166–75. doi:10.1002/pds.1833
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Mellbye A,
      2. Karlstad O,
      3. Skurtveit S, et al
      . Co-morbidity in persistent opioid users with chronic non-malignant pain in Norway. Eur J Pain 2014;18: 1083–93. doi:10.1002/j.1532-2149.2014.00449.x
      OpenUrlCrossRefPubMed
    28. ↵
      1. Zatzick D,
      2. Donovan DM,
      3. Jurkovich G, et al
      . Disseminating alcohol screening and brief intervention at trauma centers: a policy-relevant cluster randomized effectiveness trial. Addiction 2014;109:754–65. doi:10.1111/add.12492
      OpenUrlCrossRefPubMed
    29. ↵
      1. Dowell D,
      2. Haegerich TM,
      3. Chou R
      . CDC guideline for prescribing opioids for chronic pain—United States, 2016. JAMA 2016;315:1624–45. doi:10.1001/jama.2016.1464
      OpenUrlCrossRefPubMed
    30. ↵
      1. Franklin GM,
      2. Mai J,
      3. Turner J, et al
      . Bending the prescription opioid dosing and mortality curves: Impact of the Washington State opioid dosing guideline. Am J Ind Med 2012;55:325–31. doi:10.1002/ajim.21998
      OpenUrlCrossRefPubMed
    31. ↵
      1. Hill SC,
      2. Zuvekas SH,
      3. Zodet MW
      . Implications of the accuracy of MEPS prescription drug data for health services research. Inquiry 2011;48:242–59.
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Weisner CM,
      2. Campbell CI,
      3. Ray GT, et al
      . Trends in prescribed opioid therapy for non-cancer pain for individuals with prior substance use disorders. Pain 2009;145:287–93. doi:10.1016/j.pain.2009.05.006
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      Results from the 2008 National Survey on Drug Use and Health: national findings. Rockville, 2011. http://archive.samhsa.gov/data/NSDUH/2k8nsduh/2k8results.pdf (accessed 9 Mar 2016).

    Footnotes

    • Twitter Follow Suliman Alghnam at @alghnams

    • Contributors SA conceived, developed and refined the study design. SA and RC came up with the research idea, analysed the data and prepared the manuscript. SA wrote a major part of the manuscript and RC contributed to the method and discussion sections. Both authors critically reviewed the manuscript and approved the final draft.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; internally peer reviewed.

    • Ethical approval Because it only used de-identified and publically available information, this study was granted an exemption status by the Institutional Review Board at Johns Hopkins University.