[Skip to Content]
[Skip to Content Landing]
Figure 1.
Extent of Coprescription of Opioids and Benzodiazepines Before and After the Centers for Disease Control and Prevention Guideline Release
Extent of Coprescription of Opioids and Benzodiazepines Before and After the Centers for Disease Control and Prevention Guideline Release

Adjusted proportion of person-months with overlapping opioid and benzodiazepine prescriptions in the population of commercially insured individuals (A) and Medicare Advantage beneficiaries (B). Statistical significance was set at P = .05. Statistical significance for change in level is indicated next to dashed line, which denotes release of opioid guidelines. Statistical significance for change in trend (line slope) is indicated further to the right of the statistical significance of the change in level. Point estimates and 95% confidence intervals are shown in eTable 6 in the Supplement.

Figure 2.
Intensity of Coprescription of Opioids and Benzodiazepines Before and After the Centers for Disease Control and Prevention Guideline Release
Intensity of Coprescription of Opioids and Benzodiazepines Before and After the Centers for Disease Control and Prevention Guideline Release

Adjusted proportion of opioid prescription days with overlapping benzodiazepine prescriptions in the population of commercially insured individuals (A) and Medicare Advantage beneficiaries (B). Statistical significance was set at P = .05. Statistical significance for change in level is indicated next to the dashed line, which denotes the release of the opioid guidelines. Statistical significance for change in trend (line slope) is indicated further to the right of the statistical significance of the change in level. Point estimates and 95% confidence intervals are shown in eTable 6 in the Supplement.

Figure 3.
Difference in Intensity With Same vs Different Physicians Prescribing Opioids and Benzodiazepines
Difference in Intensity With Same vs Different Physicians Prescribing Opioids and Benzodiazepines

Error bars denote 95% confidence intervals, and column height indicates the percentage point increase in the rate of coprescribing (overlap days in month per opioid prescription days in month). For example, in the commercial population using long-term opioids before the guideline release, the regression-adjusted proportion of opioid prescription days with overlapped benzodiazepines was 80.5% with the same physician prescribing and 76.6% with different physicians prescribing. The absolute difference is 80.5% − 76.6% = 3.9%. All differences were statistically different from 0 when adjusted for a familywise error rate of 0.05. Point estimates and differences with 95% confidence intervals are provided in eTable 15 in the Supplement.

Table 1.  
Characteristics of Study Cohorta
Characteristics of Study Cohorta
Table 2.  
Changes in Extent and Intensity of Opioid and Benzodiazepine Coprescribing Associated With Centers for Disease Control and Prevention Guideline Release
Changes in Extent and Intensity of Opioid and Benzodiazepine Coprescribing Associated With Centers for Disease Control and Prevention Guideline Release
Supplement.

eAppendix 1. Identification of Opioids and Conversion to Morphine Milligram Equivalents

eAppendix 2. Details of the Statistical Analysis Methodology

eAppendix 3. Elixhauser Comorbidity Calculations

eAppendix 4. Cancer

eAppendix 5. Sickle Cell Disease

eAppendix 6. Hospice

eTable 1. Opioid Drugs and Combinations Included

eTable 2. Regression Results: Extent of Coprescribing for the Commercially Insured—Short-term Opioid Use Episodes

eTable 3. Regression Results: Extent of Coprescribing for the Commercially Insured—Long-term Opioid Use Episodes

eTable 4. Regression Results: Extent of Coprescribing for Medicare Advantage—Short-term Episodes

eTable 5. Regression Results: Extent of Coprescribing for Medicare Advantage—Long-term Episodes

eTable 6. Adjusted Estimates: Extent and Intensity of Coprescription

eTable 7. Regression Results: Intensity of Coprescribing for the Commercially Insured—Short-term Episodes

eTable 8. Regression Results: Intensity of Coprescribing for the Commercially Insured—Long-term Episodes

eTable 9. Regression Results: Intensity of Coprescribing for Medicare Advantage—Short-term Episodes

eTable 10. Regression Results: Intensity of Coprescribing for Medicare Advantage—Long-term Episodes

eTable 11. Regression Results: Intensity of Coprescription Among Commercial Beneficiaries With and Without Same Physician Prescribing Opioids and Benzodiazepines—Short-term Opioid Use Episodes

eTable 12. Regression Results: Intensity of Coprescription Among Commercial Beneficiaries With and Without Same Physician Prescribing Opioids and Benzodiazepines: Long-term Opioid Use Episodes

eTable 13. Regression Results: Intensity of Coprescription Among Medicare Advantage Beneficiaries With and Without Same Physician Prescribing Opioids and Benzodiazepines—Short-term Episodes

eTable 14. Regression Results: Intensity of Coprescription Among Medicare Advantage Beneficiaries With and Without Same Physician Prescribing Opioids and Benzodiazepines—Long-term Episodes

eTable 15. Intensity of Coprescribing Before and After the Guideline Release: Comparing Same Physician Prescribing to Different Physicians Prescribing Opioids and Benzodiazepines

eTable 16. Extent of Coprescription by Race/Ethnicity: Adjusted Percent of Overlapping Opioids and BZDs in Month

eTable 17. Changes in Extent of Opioid and Benzodiazepine Coprescribing Associated With CDC Guideline Release, by Race/Ethnicity

eTable 18. Regression Results Secondary Analysis: Race/Ethnicity Short-term Episodes

eTable 19. Regression Results Secondary Analysis: Race/Ethnicity Long-term Episodes

eTable 20. Extent of Coprescription by Opioid Dose: Adjusted Percent With Overlapping Opioids and Benzodiazepines in Month

eTable 21. Changes in Extent of Opioid and Benzodiazepine Coprescribing Associated With CDC Guideline Release, by Opioid Dose

eTable 22. Regression Results Sensitivity Analysis: Episodes Starting Before vs After Guideline Release

eTable 23. Adjusted Coprescribing Extent Estimates and Trends in Long-term Opioid Use Episodes Starting Before vs After the Guidelines Were Released

eFigure 1. Cohort Flow Chart Describing Construction of the Study Sample

eFigure 2. Race/Ethnicity: Coprescribing Extent

eFigure 3. Coprescribing Extent by Opioid Dose

eReferences

1.
Centers for Disease Control and Prevention. Prescription opioid data. https://www.cdc.gov/drugoverdose/data/prescribing.html. Published 2017. Accessed April 18, 2019.
2.
Calcaterra  S, Glanz  J, Binswanger  IA.  National trends in pharmaceutical opioid related overdose deaths compared to other substance related overdose deaths: 1999-2009.  Drug Alcohol Depend. 2013;131(3):263-270. doi:10.1016/j.drugalcdep.2012.11.018PubMedGoogle ScholarCrossref
3.
Park  TW, Saitz  R, Ganoczy  D, Ilgen  MA, Bohnert  AS.  Benzodiazepine prescribing patterns and deaths from drug overdose among US veterans receiving opioid analgesics: case-cohort study.  BMJ. 2015;350:h2698. doi:10.1136/bmj.h2698PubMedGoogle ScholarCrossref
4.
Jann  M, Kennedy  WK, Lopez  G.  Benzodiazepines: a major component in unintentional prescription drug overdoses with opioid analgesics.  J Pharm Pract. 2014;27(1):5-16. doi:10.1177/0897190013515001PubMedGoogle ScholarCrossref
5.
Jones  CM, McAninch  JK.  Emergency department visits and overdose deaths from combined use of opioids and benzodiazepines.  Am J Prev Med. 2015;49(4):493-501. doi:10.1016/j.amepre.2015.03.040PubMedGoogle ScholarCrossref
6.
Lee  SC, Klein-Schwartz  W, Doyon  S, Welsh  C.  Comparison of toxicity associated with nonmedical use of benzodiazepines with buprenorphine or methadone.  Drug Alcohol Depend. 2014;138:118-123. doi:10.1016/j.drugalcdep.2014.02.014PubMedGoogle ScholarCrossref
7.
White  JM, Irvine  RJ.  Mechanisms of fatal opioid overdose.  Addiction. 1999;94(7):961-972. doi:10.1046/j.1360-0443.1999.9479612.xPubMedGoogle ScholarCrossref
8.
Sun  EC, Dixit  A, Humphreys  K, Darnall  BD, Baker  LC, Mackey  S.  Association between concurrent use of prescription opioids and benzodiazepines and overdose: retrospective analysis.  BMJ. 2017;356:j760. doi:10.1136/bmj.j760PubMedGoogle ScholarCrossref
9.
Karaca-Mandic  P, Meara  E, Morden  NE.  The growing problem of co-treatment with opioids and benzodiazepines.  BMJ. 2017;356:j1224. doi:10.1136/bmj.j1224PubMedGoogle ScholarCrossref
10.
US Food and Drug Administration. FDA warns about serious risks and death when combining opioid pain or cough medicines with benzodiazepines; requires its strongest warning. https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-fda-warns-about-serious-risks-and-death-when-combining-opioid-pain-or. Published August 31, 2016. Accessed April 5, 2018.
11.
Dowell  D, Haegerich  TM, Chou  R.  CDC guideline for prescribing opioids for chronic pain—United States, 2016.  MMWR Recomm Rep. 2016;65(1):1-49. doi:10.15585/mmwr.rr6501e1PubMedGoogle ScholarCrossref
12.
Bauer  SR, Hitchner  L, Harrison  H, Gerstenberger  J, Steiger  S.  Predictors of higher-risk chronic opioid prescriptions in an academic primary care setting.  Subst Abus. 2016;37(1):110-117. doi:10.1080/08897077.2015.1129020PubMedGoogle ScholarCrossref
13.
Becker  WC, Starrels  JL, Heo  M, Li  X, Weiner  MG, Turner  BJ.  Racial differences in primary care opioid risk reduction strategies.  Ann Fam Med. 2011;9(3):219-225. doi:10.1370/afm.1242PubMedGoogle ScholarCrossref
14.
Burgess  DJ, Nelson  DB, Gravely  AA,  et al.  Racial differences in prescription of opioid analgesics for chronic noncancer pain in a national sample of veterans.  J Pain. 2014;15(4):447-455.PubMedGoogle Scholar
15.
Gaither  JR, Gordon  K, Crystal  S,  et al.  Racial disparities in discontinuation of long-term opioid therapy following illicit drug use among black and white patients.  Drug Alcohol Depend. 2018;192:371-376. doi:10.1016/j.drugalcdep.2018.05.033PubMedGoogle ScholarCrossref
16.
Hausmann  LR, Gao  S, Lee  ES, Kwoh  CK.  Racial disparities in the monitoring of patients on chronic opioid therapy.  Pain. 2013;154(1):46-52. doi:10.1016/j.pain.2012.07.034PubMedGoogle ScholarCrossref
17.
OptumLabs.  OptumLabs and OptumLabs Data Warehouse (OLDW) Descriptions and Citation. Cambridge, MA: OptumLabs; 2018.
18.
National Center for Injury Prevention and Control. CDC compilation of benzodiazepines, muscle relaxants, stimulants, zolpidem, and opioid analgesics with oral morphine milligram equivalent conversion factors, 2016 version. http://www.pdmpassist.org/pdf/Conversion Reference Table.xlsx. Published August 2016. Accessed May 3, 2018.
19.
Von Korff  M, Saunders  K, Thomas Ray  G,  et al.  De facto long-term opioid therapy for noncancer pain.  Clin J Pain. 2008;24(6):521-527. doi:10.1097/AJP.0b013e318169d03bPubMedGoogle ScholarCrossref
20.
Benchimol  EI, Smeeth  L, Guttmann  A,  et al; RECORD Working Committee.  The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement.  PLoS Med. 2015;12(10):e1001885. doi:10.1371/journal.pmed.1001885PubMedGoogle ScholarCrossref
21.
Quan  H, Sundararajan  V, Halfon  P,  et al.  Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.  Med Care. 2005;43(11):1130-1139. doi:10.1097/01.mlr.0000182534.19832.83PubMedGoogle ScholarCrossref
22.
Holm  S.  A simple sequentially rejective multiple test procedure.  Scand J Stat. 1979;6(2):65-70.Google Scholar
23.
StataCorp.  Stata/MP: Release 15. College Station, TX: StataCorp LLC; 2017.
24.
Bohnert  ASB, Guy  GP  Jr, Losby  JL.  Opioid prescribing in the United States before and after the Centers for Disease Control and Prevention’s 2016 opioid guideline.  Ann Intern Med. 2018;169(6):367-375. doi:10.7326/M18-1243PubMedGoogle ScholarCrossref
25.
Simpson  E.  The interpretation of interaction in contingency tables.  J R Stat Soc B. 1951;13(2):238-241.Google Scholar
26.
Hooten  WM.  Chronic pain and mental health disorders: shared neural mechanisms, epidemiology, and treatment.  Mayo Clin Proc. 2016;91(7):955-970. doi:10.1016/j.mayocp.2016.04.029PubMedGoogle ScholarCrossref
27.
Aguglia  A, Salvi  V, Maina  G, Rossetto  I, Aguglia  E.  Fibromyalgia syndrome and depressive symptoms: comorbidity and clinical correlates.  J Affect Disord. 2011;128(3):262-266. doi:10.1016/j.jad.2010.07.004PubMedGoogle ScholarCrossref
28.
Raphael  KG, Janal  MN, Nayak  S, Schwartz  JE, Gallagher  RM.  Psychiatric comorbidities in a community sample of women with fibromyalgia.  Pain. 2006;124(1-2):117-125. doi:10.1016/j.pain.2006.04.004PubMedGoogle ScholarCrossref
29.
Thieme  K, Turk  DC, Flor  H.  Comorbid depression and anxiety in fibromyalgia syndrome: relationship to somatic and psychosocial variables.  Psychosom Med. 2004;66(6):837-844. doi:10.1097/01.psy.0000146329.63158.40PubMedGoogle ScholarCrossref
30.
Giannakopoulos  NN, Keller  L, Rammelsberg  P, Kronmüller  KT, Schmitter  M.  Anxiety and depression in patients with chronic temporomandibular pain and in controls.  J Dent. 2010;38(5):369-376. doi:10.1016/j.jdent.2010.01.003PubMedGoogle ScholarCrossref
31.
Manfredini  D, Borella  L, Favero  L, Ferronato  G, Guarda-Nardini  L.  Chronic pain severity and depression/somatization levels in TMD patients.  Int J Prosthodont. 2010;23(6):529-534.PubMedGoogle Scholar
32.
Demyttenaere  K, Bruffaerts  R, Lee  S,  et al.  Mental disorders among persons with chronic back or neck pain: results from the World Mental Health Surveys.  Pain. 2007;129(3):332-342. doi:10.1016/j.pain.2007.01.022PubMedGoogle ScholarCrossref
33.
Dersh  J, Gatchel  RJ, Mayer  T, Polatin  P, Temple  OR.  Prevalence of psychiatric disorders in patients with chronic disabling occupational spinal disorders.  Spine (Phila Pa 1976). 2006;31(10):1156-1162. doi:10.1097/01.brs.0000216441.83135.6fPubMedGoogle ScholarCrossref
34.
Von Korff  M, Crane  P, Lane  M,  et al.  Chronic spinal pain and physical-mental comorbidity in the United States: results from the national comorbidity survey replication.  Pain. 2005;113(3):331-339. doi:10.1016/j.pain.2004.11.010PubMedGoogle ScholarCrossref
35.
Balliet  WE, Edwards-Hampton  S, Borckardt  JJ,  et al.  Depressive symptoms, pain, and quality of life among patients with nonalcohol-related chronic pancreatitis.  Pain Res Treat. 2012;2012:978646.PubMedGoogle Scholar
36.
Van Oudenhove  L, Törnblom  H, Störsrud  S, Tack  J, Simrén  M.  Depression and somatization are associated with increased postprandial symptoms in patients with irritable bowel syndrome.  Gastroenterology. 2016;150(4):866-874. doi:10.1053/j.gastro.2015.11.010PubMedGoogle ScholarCrossref
37.
Burris  JL, Cyders  MA, de Leeuw  R, Smith  GT, Carlson  CR.  Posttraumatic stress disorder symptoms and chronic orofacial pain: an empirical examination of the mutual maintenance model.  J Orofac Pain. 2009;23(3):243-252.PubMedGoogle Scholar
38.
Monteiro  DR, Zuim  PR, Pesqueira  AA, Ribeiro  PdoP, Garcia  AR.  Relationship between anxiety and chronic orofacial pain of temporomandibular disorder in a group of university students.  J Prosthodont Res. 2011;55(3):154-158. doi:10.1016/j.jpor.2010.11.001PubMedGoogle ScholarCrossref
39.
Arnold  LM, Hudson  JI, Keck  PE, Auchenbach  MB, Javaras  KN, Hess  EV.  Comorbidity of fibromyalgia and psychiatric disorders.  J Clin Psychiatry. 2006;67(8):1219-1225. doi:10.4088/JCP.v67n0807PubMedGoogle ScholarCrossref
40.
Cunningham  JL, Craner  JR, Evans  MM, Hooten  WM.  Benzodiazepine use in patients with chronic pain in an interdisciplinary pain rehabilitation program.  J Pain Res. 2017;10:311-317. doi:10.2147/JPR.S123487PubMedGoogle ScholarCrossref
41.
Nielsen  S, Lintzeris  N, Bruno  R,  et al.  Benzodiazepine use among chronic pain patients prescribed opioids: associations with pain, physical and mental health, and health service utilization.  Pain Med. 2015;16(2):356-366. doi:10.1111/pme.12594PubMedGoogle ScholarCrossref
42.
Lin  LA, Bohnert  ASB, Kerns  RD, Clay  MA, Ganoczy  D, Ilgen  MA.  Impact of the Opioid Safety Initiative on opioid-related prescribing in veterans.  Pain. 2017;158(5):833-839. doi:10.1097/j.pain.0000000000000837PubMedGoogle ScholarCrossref
43.
Cabana  MD, Rand  CS, Powe  NR,  et al.  Why don’t physicians follow clinical practice guidelines? a framework for improvement.  JAMA. 1999;282(15):1458-1465. doi:10.1001/jama.282.15.1458PubMedGoogle ScholarCrossref
44.
Pathman  DE, Konrad  TR, Freed  GL, Freeman  VA, Koch  GG.  The awareness-to-adherence model of the steps to clinical guideline compliance: the case of pediatric vaccine recommendations.  Med Care. 1996;34(9):873-889. doi:10.1097/00005650-199609000-00002PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 1,270
    Original Investigation
    Public Health
    August 2, 2019

    Rates of Physician Coprescribing of Opioids and Benzodiazepines After the Release of the Centers for Disease Control and Prevention Guidelines in 2016

    Author Affiliations
    • 1Division of Health Care Policy Research, Mayo Clinic, Rochester, Minnesota
    • 2Department of Anesthesiology and Perioperative Medicine, Mayo Clinic, Rochester, Minnesota
    • 3Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts
    • 4Department of Medicine, Massachusetts General Hospital, Boston
    • 5National Bureau of Economic Research, Cambridge, Massachusetts
    • 6Section of General Internal Medicine, Department of Medicine, Yale University School of Medicine, New Haven, Connecticut
    • 7Department of Health Policy and Management, Yale University School of Public Health, New Haven, Connecticut
    • 8Center for Outcomes Research and Evaluation, Yale-New Haven Hospital, New Haven, Connecticut
    • 9Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, Mayo Clinic, Rochester, Minnesota
    • 10OptumLabs, Cambridge, Massachusetts
    • 11Department of Finance, Carlson School of Management, University of Minnesota, Minneapolis
    JAMA Netw Open. 2019;2(8):e198325. doi:10.1001/jamanetworkopen.2019.8325
    Key Points español 中文 (chinese)

    Question  Was the release of the Centers for Disease Control and Prevention’s 2016 opioid prescribing guidelines associated with changes in the rate of coprescription of opioids and benzodiazepines?

    Findings  This cohort study of administrative data for 4 897 464 patients found small, statistically significant decreases in the rate of coprescription in the 2 years after the guideline release. These decreases were seen only in the targeted guideline population—people using opioids long term—not in those receiving short courses of opioids.

    Meaning  Coprescribing rates of opioids and benzodiazepines may have decreased after the Centers for Disease Control and Prevention guideline release, but associations with patient outcomes are not yet known.

    Abstract

    Importance  The Centers for Disease Control and Prevention guidelines in 2016 recommended avoiding concurrent use of opioids and benzodiazepines.

    Objective  To determine whether the release of the guidelines was associated with changes in coprescription of opioids and benzodiazepines.

    Design, Setting, and Participants  This retrospective cohort study used claims data obtained from a US national database of medical and pharmacy claims for 3 598 322 adult commercially insured patients and 1 299 142 Medicare Advantage (MA) beneficiaries with no recent history of cancer, sickle cell disease, or hospice care who ever used prescribed opioids during the study period, January 1, 2014, through March 31, 2018.

    Exposures  Overlapping opioid and benzodiazepine prescriptions filled.

    Main Outcomes and Measures  The extent (proportion of person-months with any overlapping days of prescription of opioids and benzodiazepines) and intensity (proportion of days with opioids prescribed where benzodiazepines were also available) of coprescription.

    Results  Of 4 897 464 patients (with 13.4 million person-months of opioid use), the total number of unique commercially insured individuals was 3 598 322 (1 974 731 women [54.9%]), and the total number of unique MA beneficiaries was 1 299 142 (770 256 women [59.3%]). Among 128 576 participants experiencing chronic pain episodes, more than one-half of person-months of long-term opioid use occurred in women (52.7% of person-months among those with commercial insurance and 62.4% of person-months among MA beneficiaries). The median (interquartile range) age of the participants was 51 (41-58) years for patients in the commercial insurance group and 70 (61-77) years for those in the MA group. The mean (SE) extent of coprescription was 23.0% (0.18%) for the commercial insurance group and 25.7% (0.18%) for the MA group. The extent of coprescription decreased in the targeted guideline population—individuals with long-term opioid use—after the guideline release (postguideline slope, −0.95 percentages point per year [95% CI, −1.44 to −0.46 percentage points per year] for the commercial insurance group and −1.06 percentage points per year [95% CI, −1.49 to −0.63 percentage points per year] for the MA group). Nontargeted short-term episodes of opioid use were associated with no change or small declines in trend (for the MA group, postguideline slope of 0.47 percentage point per year [95% CI, 0.35-0.59 percentage point per year]; for the commercial insurance group, postguideline slope of −0.05 percentage point per year [95% CI, −0.12 to 0.02 percentage point per year]). High coprescribing intensity was seen, with 79.3% (95% CI, 78.9%-79.6%) of opioid prescription days in the commercial insurance group and 83.9% (95% CI, 83.7%-84.2%) in the MA group overlapping with benzodiazepines. There was no change in the intensity of coprescribing. Intensity of coprescription was higher when the same clinician prescribed opioids and benzodiazepines.

    Conclusion and Relevance  This study observed a reduction in the extent but not intensity of coprescribing of benzodiazepines for patients with long-term opioid use.

    Introduction

    Rates of opioid prescribing and deaths attributed to prescription opioid overdose have quadrupled in the United States over the past 20 years.1 One-quarter2 to one-half3 of opioid-related deaths also involve a benzodiazepine, which is consistent with the known risks of adverse events when these 2 drug classes are used concomitantly.4-7 In particular, the risk of overdose increases when people take opioids and benzodiazepines at the same time.3,4,8 Coprescribing of opioids and benzodiazepines in the United States has previously been shown to be common.8 One study9 found that, in 2013, 17% of prescription opioid users also used a benzodiazepine. Recognizing the risk of coprescribing, in 2016 the US Food and Drug Administration10 issued a warning to patients and clinicians about the potential risks of combined opioid and benzodiazepine use.

    In March 2016, the US Centers for Disease Control and Prevention (CDC) released guidelines11 intended to help primary care clinicians treat patients with chronic non–cancer-related pain. These guidelines specifically included a category A recommendation to avoid coprescription of opioids and benzodiazepines “whenever possible.”11 The panel writing the guidelines noted that there may be clinically appropriate situations where the drugs may be used together, giving the example of a person taking a stable low dose of a benzodiazepine who experiences acute pain. However, the overall recommendation in the context of long-term opioid use was to avoid benzodiazepines if possible, given patient needs.

    To better understand the association of the CDC opioid guidelines with coprescribing, changes in extent of coprescribing of benzodiazepines among patients newly initiating long-term opioid therapy were compared before and after release of the guidelines in a large population of commercially insured patients and Medicare Advantage (MA) beneficiaries. The primary outcomes were the extent and intensity of coprescription of opioids and benzodiazepines. Because of the well-recognized racial differences in opioid prescribing practices,12-16 changes in coprescribing by race and ethnicity were also analyzed.

    Methods
    Study Population

    Episodes of opioid use between January 1, 2014, and March 31, 2018, were identified in the OptumLabs Data Warehouse,17 which includes deidentified claims data for privately insured participants and MA enrollees in a large, private, US health plan. Participants were stratified into 2 insurance groups, commercially insured and MA beneficiaries. Prescriptions for all opioids, including tramadol, filled through the health plan during the study period were captured (eAppendix 1 and eTable 1 in the Supplement).18

    An episode of opioid use was defined as a period when the patient went no more than 30 days without using opioids. Opioids were considered to be available to a beneficiary from the prescription fill date to the time the prescription would be expected to be completed if the medication was consumed according to prescriber instructions. Episodes of opioid use that spanned at least 90 calendar days and included at least 120 days’ supply or 10 fills (ie, how many times a participant received medication from the pharmacy) were classified as long term.19 Episodes of opioid use not meeting these criteria were classified as short term and were used as a comparison group to estimate trends unrelated to the release of the CDC guidelines.

    Inclusion criteria for incident episodes of opioid use were (1) the first opioid prescription filled in an episode was preceded by 90 days of insurance enrollment with no opioids available, and (2) the participant continued insurance enrollment for at least 120 days after the first prescription fill in the episode. The latter criterion was necessary to ascertain whether an episode represented a long-term episode of opioid use. Exclusion criteria included (1) episodes of opioid use with 1 or more prescriptions written for more than 1000 morphine milligram equivalents (MMEs) per day, (2) individuals younger than 18 years, (3) individuals with cancer, (4) individuals with sickle cell disease, (5) individuals receiving hospice services, and (6) episodes of opioid use that included buprenorphine, methadone, or naloxone, potentially suggesting treatment for opioid use disorder. A flowchart is provided in eFigure 1 in the Supplement.

    This study was deemed exempt from review by the University of Minnesota and Mayo Clinic institutional review boards because data were existing and deidentified. This report follows the Reporting of Studies Conducted Using Observational Routinely Collected Health Data (RECORD) reporting guidelines for observational studies using routinely collected health data.20

    Outcomes

    The primary outcomes were the extent and intensity of coprescription of opioids and benzodiazepines. The extent of coprescription was measured with a flag to identify each calendar month during which a patient had overlapping opioid and benzodiazepine prescriptions filled. The intensity of coprescription was measured as the proportion of opioid prescription days where benzodiazepines were also available. In a sensitivity analysis focusing on long-term episodes starting before and after the guideline release, the outcome variable was an indicator of whether there was any benzodiazepine prescription filled in the first 90 days of the opioid use episode.

    Covariates

    Beneficiary covariates included age, sex, race/ethnicity, state of residence, and indicator variables for each of the Elixhauser comorbidities21 measured on a rolling 6-month basis. Coprescribing by the same physician (as opposed to 1 physician prescribing the benzodiazepine and another prescribing the opioid) was assessed for person-months with 1 or more days of overlap. The categorical variable could take the following values: (1) at least 1 day of overlap with opioid and benzodiazepine prescriptions written by different physicians, (2) at least 1 day of overlap with prescriptions written by the same physician, or (3) at least 1 day of overlap, but it was not possible to assess whether the same physician prescribed the drugs because of unknown prescriber identification numbers on 1 or both of the prescriptions.

    Statistical Analysis

    Changes in coprescribing were analyzed both as a 1-time change in level at the time of CDC guideline release and as a change in trend after the guideline release. To implement this approach, piecewise regression with a single knot spline was used, with the knot placed at the release of the guidelines (March 2016).

    The unit of observation was a person-month during which the beneficiary had opioids available for at least 1 day. Generalized linear models were specified with clustered SEs to account for repeated observations on the same patients.

    The Holm22 sequentially rejective multiple test procedure (2-sided) was used to control the familywise error rate at 0.05. The tables, text, and figures denote whether comparisons are statistically significant after this adjustment. Summary statistics are presented as percentage (SE) for categorical variables and as quartiles (median, 25th percentile, and 75th percentile) for continuous variables.

    Details on model specifications and secondary analyses are provided in eAppendixes 2, 3, 4, 5, and 6 in the Supplement. Figures present adjusted rates of overlapping opioid and benzodiazepine prescriptions filled and adjusted proportions of opioid prescription days where benzodiazepines were also available, with 95% confidence intervals shown in the tables. All analyses were conducted with Stata/MP statistical software version 15.0 (StataCorp).23

    Results

    The study sample included 13.4 million person-months of opioid use contributed by 4 897 464 beneficiaries. In both the commercial and MA populations, non-Hispanic white patients accounted for a larger proportion of long-term compared with short-term opioid use episode person-months (Table 1). The total number of unique commercial beneficiaries was 3 598 322 (1 974 731 women [54.9%]), and the total number of unique MA beneficiaries was 1 299 142 (770 256 women [59.3%]). Among 128 576 beneficiaries experiencing chronic pain episodes, more than one-half of person-months of long-term opioid use occurred in women (52.7% of person-months among those with commercial insurance and 62.4% of person-months among MA beneficiaries). The median (interquartile range) age was 51 (41-58) years for those with commercial insurance and 70 (61-77) years for MA beneficiaries. The median age was older for patients with long-term compared with short-term opioid use episodes in the commercial population; the opposite was true in the MA population.

    Extent of Coprescription
    Unadjusted Rates

    Among patients using long-term opioids, approximately one-quarter of person-months involved overlapping opioid and benzodiazepine prescriptions (mean [SE], commercial group, 23.0% [0.18%]; MA group, 25.7% [0.18%]). Opioid and benzodiazepine coprescription was less common during short-term opioid use episodes compared with long-term opioid episodes (Table 1).

    Adjusted Rates

    Before the CDC guidelines were released, the extent of coprescription was increasing in both the commercial and MA long-term opioid use populations (preguideline slope for commercial long-term use population, 0.52 percentage point per year [95% CI, 0.02-1.03 percentage points per year]; preguideline slope for MA long-term use population, 1.09 percentage points per year [95% CI, 0.56-1.62 percentage points per year]) (Figure 1 and Table 2). The release of the CDC guidelines was associated with a decrease in trend for both of these populations, resulting in a negative slope after release of approximately 1 percentage point per year (−0.95 percentage point per year [95% CI, −1.44 to −0.46 percentage points per year] for the commercial group and −1.06 percentage points per year [95% CI, −1.49 to −0.63 percentage points per year] for the MA group; the postguideline trend minus preguideline trend was −0.12% per month for the commercial population and −0.18% per month for the MA population). These changes were statistically significant after controlling for a familywise error rate of 0.05 across all comparisons in extent and intensity of coprescription. The guidelines were associated with a 0.87–percentage point increase in the level of coprescribing in the commercial population, resulting in similar coprescribing rates at the beginning and end of the study period. There was no statistically significant change in level for the MA population using opioids long term. Complete regression results are available in eTables 2, 3, 4, and 5 in the Supplement; adjusted rates with confidence intervals are available in eTable 6 in the Supplement.

    The comparison group of short-term opioid use episodes in the MA and commercial populations showed smaller changes or no change. In the MA short-term use population, the extent of coprescribing was increasing at approximately the same rate before and after the guideline release (slope, 0.44 percentage point [95% CI, 0.32-0.56 percentage point] before and 0.47 percentage point [95% CI, 0.35-0.59 percentage point] after; P = .70); a statistically significant decrease of −1.02 percentage points (95% CI, −1.21 to −0.83; P < .001 after controlling for familywise error rate of 0.05) attenuated the overall trend toward increasing coprescribing in this population. In the commercially insured population with short-term opioid use, the trend before guideline release was a slight increase (slope, 0.19 percentage point per year [95% CI, 0.13-0.24 percentage point per year]). After guideline release, the slope was no longer increasing (not statistically different from 0); this change was statistically significant (postguideline slope, −0.05 percentage point per year [95% CI, −0.12 to 0.02 percentage point per year]; P < .001 after controlling for familywise error rate of 0.05). There was a small, statistically significant change of −0.25 percentage point (95% CI, −0.35 to −0.15; P < .001 after controlling for familywise error rate of 0.05) associated with guideline release in the commercial short-term use group.

    Intensity of Coprescription
    Unadjusted Rates

    Intensity of coprescription was measured for participants with long-term opioid use episodes who had at least 1 day of overlapping benzodiazepine and opioid prescriptions in the months evaluated. Coprescription intensity was 79.3% (95% CI, 78.9%-79.6%) in the commercial group and 83.9% (95% CI, 83.7%-84.2%) in the MA population (Table 1).

    Adjusted Rates

    There were no statistically significant changes in either level or trend of intensity of coprescription of opioids and benzodiazepines in any beneficiary or opioid use group (Figure 2 and Table 2). Slopes were positive for all beneficiary and opioid use groups and periods except for long-term opioid use episodes in the commercial group in the postguideline period, where the slope was not statistically significantly different from 0. However, these slopes were exceptionally small; for example, among commercial insurance beneficiaries with long-term opioid use episodes, the intensity of coprescription was increasing by 0.57% per year (95% CI, 0.04%-1.11% per year) before the guidelines were released, on a baseline intensity rate of 78.36% (95% CI, 77.53%-79.185%), for a relative change of 0.73% per year. Complete regression results are available in eTables 7, 8, 9, and 10 in the Supplement.

    Same Physician Prescribing
    Unadjusted Proportions

    Overlapping prescriptions for opioids and benzodiazepines were frequently written by the same physician. For commercial participants with long-term use episodes, a mean (SE) of 59.1% (0.12%) of overlapping prescriptions were from the same physician, and for MA beneficiaries, a mean (SE) of 60.3% (0.10%) of overlapping prescriptions were from the same physician (Table 1).

    Adjusted Proportions

    Figure 3 shows the difference in rates of overlap for patients with the same prescribing physician compared with those with different physicians prescribing opioids and benzodiazepines. In months with any coprescription during long-term opioid use episodes, having the same physician prescribe both opioids and benzodiazepines was associated with a small but consistent increase in the proportion of overlap days in both patient populations. This association was present both before and after the release of the CDC guidelines.

    For beneficiaries with short-term opioid use episodes, conditional on having any overlap in the month, MA beneficiaries experienced greater coprescription intensity when the same physician prescribed both drugs, compared with having different prescribers for the 2 drugs, whereas commercial insurance beneficiaries actually had reduced coprescription intensity when the same physician prescribed both drugs. Complete regression results are available in eTables 11, 12, 13, and 14 in the Supplement; adjusted rates with confidence intervals are available in eTable 15 in the Supplement.

    Secondary Analyses
    Race/Ethnicity-Stratified Analysis

    White and Hispanic beneficiaries had higher rates of coprescription than did black beneficiaries over the entire study period (eFigure 2 and eTables 16, 17, 18, and 19 in the Supplement). There were no statistically significant changes in trend or level for black and Hispanic beneficiaries with long-term opioid use episodes (mean [SD] monthly sample size, 4522 [1684] for black and 2565 [839] for Hispanic beneficiaries).

    Opioid Dose–Stratified Analysis

    The risk of overdose in people taking both opioids and benzodiazepines increases with opioid dose.3 We compared coprescribing extent by opioid dose: less than 20 MMEs, 20 to less than 50 MMEs, 50 to less than 90 MMEs, 90 to less than 120 MMEs, and 120 or more MMEs of opioids per opioid use day. We found that the extent of coprescribing was highest in the 120 or more MME group, with adjusted proportions at the end of the study of 35.9% (95% CI, 33.8%-38.0%) of person-months in the long-term opioid use population and 19.6% (95% CI, 18.7%-20.4%) of person-months in the short-term opioid use population (eFigure 3, eTable 20, and eTable 21 in the Supplement). Statistically significant decreases in slope were seen in the 20 to less than 50 MME and 50 to less than 90 MME long-term opioid use dose groups. Statistically significant increases in slope were found in the short-term opioid use group at those same dose levels (20 to < 50 MME and 50 to < 90 MME).

    Coprescribing Extent in Episodes Starting Before vs After Guideline Release

    A sensitivity analysis was specified post hoc to look only at long-term episodes starting before vs after the guideline release. This analysis found no change in the extent of coprescribing level or trend in the first 90 days of the episode for commercial beneficiaries and a statistically significant decrease in trend for MA beneficiaries (MA preguideline slope, 0.22 percentage point per year [95% CI, −0.54 to 0.99 percentage point per year]; MA postguideline slope, −1.23 percentage points per year [95% CI, −2.26 to −0.32 percentage points per year]) (eTable 22 and eTable 23 in the Supplement).

    Discussion

    After the release of the CDC guidelines addressing long-term opioid prescribing, previously increasing trends in the extent of coprescribing of opioids and benzodiazepines during long-term opioid use episodes were observed to reverse, with decreasing trends over time seen after the CDC guideline release. The coprescribing extent in short-term opioid use episodes saw smaller decreases or no change. There were no differences in the intensity of coprescribing after the release of the CDC guidelines.

    Bohnert et al24 reported that, according to IQVIA retail pharmacy data for the years 2015 to 2017, coprescribing across all patient populations was declining both before and after the CDC guidelines. We did not find this in our observed populations, who demonstrated increasing or stable rates of coprescribing before the CDC guidelines. This difference could be due to a difference in the populations studied, because the article by Bohnert et al24 included people who paid in cash or used Medicaid, who are not present in our analysis. This difference may also reflect differences in study designs. Unlike the study by Bohnert et al,24 we separately report trends in long-term vs short-term opioid use episodes and for MA vs commercial insurance beneficiaries. The differences between the 2 studies could be an example of the Simpson paradox,25 where trends in stratified analyses are reversed in the unstratified analysis. As a concrete example, given the large difference in overall coprescription rates between long-term and short-term opioid episodes, if the relative proportion of short-term episodes increased (or if the length of long-term episodes increased), the overall rate of coprescription would decrease, even while the coprescription rate in short-term episodes increased.

    Our study adds to the marketwide findings by disaggregating the population into people using opioids long term, who are the target of the guidelines and at greatest risk for adverse events, vs people using opioids short-term. We also considered both how common coprescription is (ie, its extent) as well as how much overlap patients experience when they have coprescribed drugs (ie, the intensity of coprescription). We found that the release of the guidelines was associated with a reduction in the extent of coprescribing in long-term opioid use episodes, but not with a change in intensity of coprescribing.

    We also characterized coprescribing by the same physician vs the case when different physicians prescribe the opioids and benzodiazepines. We found that most of the time, the same physician was prescribing both opioids and benzodiazepines and that when the same physician prescribed both opioids and benzodiazepines, the intensity of coprescription increased. In other words, coprescription seemed not to be driven by lack of information or coordination among treating physicians.

    One explanation for this finding could be that physicians coprescribe opioids and benzodiazepines to treat co-occurring pain and mental health conditions. Chronic pain and mental health disorders are common in the general population, and high rates of co-occurrence have been reported.26 The estimated current or 12-month frequency of depressive symptoms exceeds 50% among adults with fibromyalgia,27-29 temporomandibular joint disorder,30,31 chronic spinal pain,32-34 and chronic abdominal pain.35,36 Similarly, the 12-month frequency of anxiety exceeds 50% in individuals with temporomandibular joint disorder,37,38 fibromyalgia,29,39 and chronic abdominal pain.35,36 Although benzodiazepines are generally indicated for the treatment of anxiety disorders, their use in adults with chronic pain has been associated with greater levels of depression, greater pain scores, and use of other psychotropic medications.40,41

    Physicians treating people with co-occurring pain and anxiety may be choosing to treat with opioids and benzodiazepines, perhaps after other treatments have failed. They may also have patients with an established regimen of either opioids or benzodiazepines who later develop anxiety (in the case of opioids) or pain (in the case of benzodiazepines). Coprescription may result because the patient does not want to or is unable to discontinue one of the drugs.

    Having seen relatively minor changes in coprescribing, we specified a post hoc analysis to explore whether physicians were less likely to initiate coprescription after the CDC guidelines were released. We reasoned that for beneficiaries who did not have an established history of using 1 or both drugs, physicians may have faced fewer barriers to avoiding this practice altogether. This assumption was tested with a comparison of rates of coprescription in the first 90 days of new long-term episodes of opioid use that started before the guidelines were released vs those that started after the guidelines were released. In the MA population, we observed a decrease in trend for new coprescribing after the guidelines, although no change was seen in the commercial population. The contradictory findings may suggest that physicians consider MA beneficiaries to face higher risks from coprescription, but our study was not designed to test this hypothesis.

    Guidelines are a passive intervention to change physician behavior and might not be expected to yield substantial impact. We can compare our results with the 2013 implementation of the Opioid Safety Initiative, a Veterans Health Administration initiative to reduce coprescription of benzodiazepines and high-dose opioids.42 At the beginning of the study period, rates of coprescribing were already decreasing. Segmented regression found the Opioid Safety Initiative was associated with a change in trend of −0.11% per month. This is similar to the association seen in long-term opioid use episodes in our study (postguideline trend minus preguideline trend was −0.12% per month for the commercial population and −0.18% per month for the MA population).

    Studies of the uptake of clinical guidelines have explored the reasons physicians do not adhere to guideline recommendations.43 The framework by Pathman et al44 suggests that, before adhering to a guideline, physicians must become aware of the guideline, agree with the content, and adopt the recommendations. The increase in overlap intensity when the same physician prescribes both drugs may suggest that some physicians do not agree with the CDC’s recommendation to avoid coprescribing opioids and benzodiazepines. Alternatively, some physicians may have adopted the recommendation but made the clinical decision not to adhere to the recommendation for select patients either because some patients were resistant to making the recommended pharmacological changes or the physician believed the patient was benefiting from the coprescription of both drugs. Observing little change in coprescription intensity could suggest the guidelines had little impact or, alternatively, that physicians who continued to coprescribe were coprescribing for patients for whom the balance of risks and benefits was favorable. However, the high intensity of coprescribing, particularly among those with the highest opioid doses, suggests that some patients remain at risk because of inappropriate prescribing. Future studies should explore changes in coprescribing at the physician and patient levels to further understand changes in physician behavior and its impact on patients.

    Limitations

    As a retrospective study using administrative data, this study is subject to limitations. The study results represent an association rather than a causal effect of the guideline release on coprescription of opioids and benzodiazepines. The data also may be missing prescriptions filled by beneficiaries who paid cash at the pharmacy, which could affect the ascertainment of the outcome measures. Our data provide trends for 2.25 years before and 2 years after the guideline release, and trends may change if they are observed for shorter or longer periods. Also, this administrative data set does not include details necessary to determine the appropriateness of coprescribing.

    Conclusions

    Coprescription of opioids and benzodiazepines was common among this sample of commercially insured and MA beneficiaries, both before and after the release of the CDC guidelines. After guideline release, the extent of coprescribing decreased by a modest amount in people using opioids long term, but not those using opioids short term. There was no change associated with the guidelines in the intensity of coprescribing for any population. Future studies focused on identifying patterns of physician response to these guidelines may provide insight into who was affected by the guideline release on coprescription of opioids and benzodiazepines.

    Back to top
    Article Information

    Accepted for Publication: June 11, 2019.

    Published: August 2, 2019. doi:10.1001/jamanetworkopen.2019.8325

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Jeffery MM et al. JAMA Network Open.

    Corresponding Author: Molly M. Jeffery, PhD, Division of Health Care Policy Research, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (jeffery.molly@mayo.edu).

    Author Contributions: Dr Jeffery had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Jeffery, Hooten, Jena, Shah, Karaca-Mandic.

    Acquisition, analysis, or interpretation of data: Jeffery, Ross, Karaca-Mandic.

    Drafting of the manuscript: Jeffery, Hooten, Karaca-Mandic.

    Critical revision of the manuscript for important intellectual content: Jeffery, Jena, Ross, Shah, Karaca-Mandic.

    Statistical analysis: Jeffery, Jena, Karaca-Mandic.

    Obtained funding: Jeffery, Karaca-Mandic.

    Administrative, technical, or material support: Hooten, Shah.

    Supervision: Jena, Karaca-Mandic.

    Conflict of Interest Disclosures: Dr Jeffery reported grants from the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (NIH) during the conduct of the study, and her spouse owns stock in Vireo Health. Dr Jena reported personal fees from Pfizer, Bristol-Myers Squibb, Novartis, Amgen, Eli Lilly, Vertex, AstraZeneca, Celgene, Tesaro, Sanofi-Aventis, Biogen, Precision Health Economics, and Analysis Group outside the submitted work. Dr Ross reported grants from the NHLBI and Agency for Healthcare Research and Quality (AHRQ) during the conduct of the study and grants from the US Food and Drug Administration (FDA), Center for Medicare & Medicaid Services, Johnson & Johnson, Medtronic, Inc, Blue Cross Blue Shield Association, and Laura and John Arnold Foundation outside the submitted work. Dr Shah reported grants from AHRQ during the conduct of the study and grants from the FDA outside the submitted work and, in the past 36 months, has received research support through Mayo Clinic from the FDA to establish the Yale-Mayo Clinic Center for Excellence in Regulatory Science and Innovation program, the CMS Innovation Center under the Transforming Clinical Practice Initiative, AHRQ, NHLBI, the National Science Foundation, and the Patient Centered Outcomes Research Institute to develop a Clinical Data Research Network (LHSNet). Dr Karaca-Mandic reported grants from AHRQ and NIH during the conduct of the study, personal fees from Tactile Medical and Precision Health Economics, and grants from American Cancer Society outside the submitted work. No other disclosures were reported.

    Funding/Support: This project was supported by grant R56HL130496 from NHLBI and grant R01HS025164 from AHRQ.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or AHRQ.

    Additional Information: OptumLabs data are available for research through a virtual data warehouse. We are unable to share analytic data files.

    References
    1.
    Centers for Disease Control and Prevention. Prescription opioid data. https://www.cdc.gov/drugoverdose/data/prescribing.html. Published 2017. Accessed April 18, 2019.
    2.
    Calcaterra  S, Glanz  J, Binswanger  IA.  National trends in pharmaceutical opioid related overdose deaths compared to other substance related overdose deaths: 1999-2009.  Drug Alcohol Depend. 2013;131(3):263-270. doi:10.1016/j.drugalcdep.2012.11.018PubMedGoogle ScholarCrossref
    3.
    Park  TW, Saitz  R, Ganoczy  D, Ilgen  MA, Bohnert  AS.  Benzodiazepine prescribing patterns and deaths from drug overdose among US veterans receiving opioid analgesics: case-cohort study.  BMJ. 2015;350:h2698. doi:10.1136/bmj.h2698PubMedGoogle ScholarCrossref
    4.
    Jann  M, Kennedy  WK, Lopez  G.  Benzodiazepines: a major component in unintentional prescription drug overdoses with opioid analgesics.  J Pharm Pract. 2014;27(1):5-16. doi:10.1177/0897190013515001PubMedGoogle ScholarCrossref
    5.
    Jones  CM, McAninch  JK.  Emergency department visits and overdose deaths from combined use of opioids and benzodiazepines.  Am J Prev Med. 2015;49(4):493-501. doi:10.1016/j.amepre.2015.03.040PubMedGoogle ScholarCrossref
    6.
    Lee  SC, Klein-Schwartz  W, Doyon  S, Welsh  C.  Comparison of toxicity associated with nonmedical use of benzodiazepines with buprenorphine or methadone.  Drug Alcohol Depend. 2014;138:118-123. doi:10.1016/j.drugalcdep.2014.02.014PubMedGoogle ScholarCrossref
    7.
    White  JM, Irvine  RJ.  Mechanisms of fatal opioid overdose.  Addiction. 1999;94(7):961-972. doi:10.1046/j.1360-0443.1999.9479612.xPubMedGoogle ScholarCrossref
    8.
    Sun  EC, Dixit  A, Humphreys  K, Darnall  BD, Baker  LC, Mackey  S.  Association between concurrent use of prescription opioids and benzodiazepines and overdose: retrospective analysis.  BMJ. 2017;356:j760. doi:10.1136/bmj.j760PubMedGoogle ScholarCrossref
    9.
    Karaca-Mandic  P, Meara  E, Morden  NE.  The growing problem of co-treatment with opioids and benzodiazepines.  BMJ. 2017;356:j1224. doi:10.1136/bmj.j1224PubMedGoogle ScholarCrossref
    10.
    US Food and Drug Administration. FDA warns about serious risks and death when combining opioid pain or cough medicines with benzodiazepines; requires its strongest warning. https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-fda-warns-about-serious-risks-and-death-when-combining-opioid-pain-or. Published August 31, 2016. Accessed April 5, 2018.
    11.
    Dowell  D, Haegerich  TM, Chou  R.  CDC guideline for prescribing opioids for chronic pain—United States, 2016.  MMWR Recomm Rep. 2016;65(1):1-49. doi:10.15585/mmwr.rr6501e1PubMedGoogle ScholarCrossref
    12.
    Bauer  SR, Hitchner  L, Harrison  H, Gerstenberger  J, Steiger  S.  Predictors of higher-risk chronic opioid prescriptions in an academic primary care setting.  Subst Abus. 2016;37(1):110-117. doi:10.1080/08897077.2015.1129020PubMedGoogle ScholarCrossref
    13.
    Becker  WC, Starrels  JL, Heo  M, Li  X, Weiner  MG, Turner  BJ.  Racial differences in primary care opioid risk reduction strategies.  Ann Fam Med. 2011;9(3):219-225. doi:10.1370/afm.1242PubMedGoogle ScholarCrossref
    14.
    Burgess  DJ, Nelson  DB, Gravely  AA,  et al.  Racial differences in prescription of opioid analgesics for chronic noncancer pain in a national sample of veterans.  J Pain. 2014;15(4):447-455.PubMedGoogle Scholar
    15.
    Gaither  JR, Gordon  K, Crystal  S,  et al.  Racial disparities in discontinuation of long-term opioid therapy following illicit drug use among black and white patients.  Drug Alcohol Depend. 2018;192:371-376. doi:10.1016/j.drugalcdep.2018.05.033PubMedGoogle ScholarCrossref
    16.
    Hausmann  LR, Gao  S, Lee  ES, Kwoh  CK.  Racial disparities in the monitoring of patients on chronic opioid therapy.  Pain. 2013;154(1):46-52. doi:10.1016/j.pain.2012.07.034PubMedGoogle ScholarCrossref
    17.
    OptumLabs.  OptumLabs and OptumLabs Data Warehouse (OLDW) Descriptions and Citation. Cambridge, MA: OptumLabs; 2018.
    18.
    National Center for Injury Prevention and Control. CDC compilation of benzodiazepines, muscle relaxants, stimulants, zolpidem, and opioid analgesics with oral morphine milligram equivalent conversion factors, 2016 version. http://www.pdmpassist.org/pdf/Conversion Reference Table.xlsx. Published August 2016. Accessed May 3, 2018.
    19.
    Von Korff  M, Saunders  K, Thomas Ray  G,  et al.  De facto long-term opioid therapy for noncancer pain.  Clin J Pain. 2008;24(6):521-527. doi:10.1097/AJP.0b013e318169d03bPubMedGoogle ScholarCrossref
    20.
    Benchimol  EI, Smeeth  L, Guttmann  A,  et al; RECORD Working Committee.  The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement.  PLoS Med. 2015;12(10):e1001885. doi:10.1371/journal.pmed.1001885PubMedGoogle ScholarCrossref
    21.
    Quan  H, Sundararajan  V, Halfon  P,  et al.  Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.  Med Care. 2005;43(11):1130-1139. doi:10.1097/01.mlr.0000182534.19832.83PubMedGoogle ScholarCrossref
    22.
    Holm  S.  A simple sequentially rejective multiple test procedure.  Scand J Stat. 1979;6(2):65-70.Google Scholar
    23.
    StataCorp.  Stata/MP: Release 15. College Station, TX: StataCorp LLC; 2017.
    24.
    Bohnert  ASB, Guy  GP  Jr, Losby  JL.  Opioid prescribing in the United States before and after the Centers for Disease Control and Prevention’s 2016 opioid guideline.  Ann Intern Med. 2018;169(6):367-375. doi:10.7326/M18-1243PubMedGoogle ScholarCrossref
    25.
    Simpson  E.  The interpretation of interaction in contingency tables.  J R Stat Soc B. 1951;13(2):238-241.Google Scholar
    26.
    Hooten  WM.  Chronic pain and mental health disorders: shared neural mechanisms, epidemiology, and treatment.  Mayo Clin Proc. 2016;91(7):955-970. doi:10.1016/j.mayocp.2016.04.029PubMedGoogle ScholarCrossref
    27.
    Aguglia  A, Salvi  V, Maina  G, Rossetto  I, Aguglia  E.  Fibromyalgia syndrome and depressive symptoms: comorbidity and clinical correlates.  J Affect Disord. 2011;128(3):262-266. doi:10.1016/j.jad.2010.07.004PubMedGoogle ScholarCrossref
    28.
    Raphael  KG, Janal  MN, Nayak  S, Schwartz  JE, Gallagher  RM.  Psychiatric comorbidities in a community sample of women with fibromyalgia.  Pain. 2006;124(1-2):117-125. doi:10.1016/j.pain.2006.04.004PubMedGoogle ScholarCrossref
    29.
    Thieme  K, Turk  DC, Flor  H.  Comorbid depression and anxiety in fibromyalgia syndrome: relationship to somatic and psychosocial variables.  Psychosom Med. 2004;66(6):837-844. doi:10.1097/01.psy.0000146329.63158.40PubMedGoogle ScholarCrossref
    30.
    Giannakopoulos  NN, Keller  L, Rammelsberg  P, Kronmüller  KT, Schmitter  M.  Anxiety and depression in patients with chronic temporomandibular pain and in controls.  J Dent. 2010;38(5):369-376. doi:10.1016/j.jdent.2010.01.003PubMedGoogle ScholarCrossref
    31.
    Manfredini  D, Borella  L, Favero  L, Ferronato  G, Guarda-Nardini  L.  Chronic pain severity and depression/somatization levels in TMD patients.  Int J Prosthodont. 2010;23(6):529-534.PubMedGoogle Scholar
    32.
    Demyttenaere  K, Bruffaerts  R, Lee  S,  et al.  Mental disorders among persons with chronic back or neck pain: results from the World Mental Health Surveys.  Pain. 2007;129(3):332-342. doi:10.1016/j.pain.2007.01.022PubMedGoogle ScholarCrossref
    33.
    Dersh  J, Gatchel  RJ, Mayer  T, Polatin  P, Temple  OR.  Prevalence of psychiatric disorders in patients with chronic disabling occupational spinal disorders.  Spine (Phila Pa 1976). 2006;31(10):1156-1162. doi:10.1097/01.brs.0000216441.83135.6fPubMedGoogle ScholarCrossref
    34.
    Von Korff  M, Crane  P, Lane  M,  et al.  Chronic spinal pain and physical-mental comorbidity in the United States: results from the national comorbidity survey replication.  Pain. 2005;113(3):331-339. doi:10.1016/j.pain.2004.11.010PubMedGoogle ScholarCrossref
    35.
    Balliet  WE, Edwards-Hampton  S, Borckardt  JJ,  et al.  Depressive symptoms, pain, and quality of life among patients with nonalcohol-related chronic pancreatitis.  Pain Res Treat. 2012;2012:978646.PubMedGoogle Scholar
    36.
    Van Oudenhove  L, Törnblom  H, Störsrud  S, Tack  J, Simrén  M.  Depression and somatization are associated with increased postprandial symptoms in patients with irritable bowel syndrome.  Gastroenterology. 2016;150(4):866-874. doi:10.1053/j.gastro.2015.11.010PubMedGoogle ScholarCrossref
    37.
    Burris  JL, Cyders  MA, de Leeuw  R, Smith  GT, Carlson  CR.  Posttraumatic stress disorder symptoms and chronic orofacial pain: an empirical examination of the mutual maintenance model.  J Orofac Pain. 2009;23(3):243-252.PubMedGoogle Scholar
    38.
    Monteiro  DR, Zuim  PR, Pesqueira  AA, Ribeiro  PdoP, Garcia  AR.  Relationship between anxiety and chronic orofacial pain of temporomandibular disorder in a group of university students.  J Prosthodont Res. 2011;55(3):154-158. doi:10.1016/j.jpor.2010.11.001PubMedGoogle ScholarCrossref
    39.
    Arnold  LM, Hudson  JI, Keck  PE, Auchenbach  MB, Javaras  KN, Hess  EV.  Comorbidity of fibromyalgia and psychiatric disorders.  J Clin Psychiatry. 2006;67(8):1219-1225. doi:10.4088/JCP.v67n0807PubMedGoogle ScholarCrossref
    40.
    Cunningham  JL, Craner  JR, Evans  MM, Hooten  WM.  Benzodiazepine use in patients with chronic pain in an interdisciplinary pain rehabilitation program.  J Pain Res. 2017;10:311-317. doi:10.2147/JPR.S123487PubMedGoogle ScholarCrossref
    41.
    Nielsen  S, Lintzeris  N, Bruno  R,  et al.  Benzodiazepine use among chronic pain patients prescribed opioids: associations with pain, physical and mental health, and health service utilization.  Pain Med. 2015;16(2):356-366. doi:10.1111/pme.12594PubMedGoogle ScholarCrossref
    42.
    Lin  LA, Bohnert  ASB, Kerns  RD, Clay  MA, Ganoczy  D, Ilgen  MA.  Impact of the Opioid Safety Initiative on opioid-related prescribing in veterans.  Pain. 2017;158(5):833-839. doi:10.1097/j.pain.0000000000000837PubMedGoogle ScholarCrossref
    43.
    Cabana  MD, Rand  CS, Powe  NR,  et al.  Why don’t physicians follow clinical practice guidelines? a framework for improvement.  JAMA. 1999;282(15):1458-1465. doi:10.1001/jama.282.15.1458PubMedGoogle ScholarCrossref
    44.
    Pathman  DE, Konrad  TR, Freed  GL, Freeman  VA, Koch  GG.  The awareness-to-adherence model of the steps to clinical guideline compliance: the case of pediatric vaccine recommendations.  Med Care. 1996;34(9):873-889. doi:10.1097/00005650-199609000-00002PubMedGoogle ScholarCrossref
    ×