Opioid abuse has become a major public health concern in the United States with ∼11.5 million people 12 years of age or older misusing prescription opioid pain relievers in 2016.1,2 Between 1999 and 2015, there were 183,000 prescription opioid-related fatalities in the United States. The rate of opioid overdose–related deaths increased by 200% between 2000 and 2014.3 More than 15,000 fatalities in the United States were attributed to prescription opioid overdoses in 2015.3
Methadone, oxycodone, and hydrocodone are the drugs most commonly involved in prescription opioid overdose fatalities.4 Risk of death in members who receive prescription opioid analgesics is higher for those with a history of concomitant benzodiazepine prescription.5–7 A prospective observational cohort study, conducted in a controlled substances prescription monitoring program of 2,182,374 patients in North Carolina, found that of the 629 overdose deaths involving opioids, patients codispensed benzodiazepines had a mortality rate 10 times higher than those who received an opioid analgesic alone.6
Risk of overdose death with prescription opioids has been associated with age, concomitant prescription of sedative/hypnotics, daily dosages above 40 morphine milligram equivalents (MME), and an increase in controlled substance prescriptions, pharmacies, and prescribers.7,8 In a study of 254 fatalities due to opioid overdose in Utah in 2008–2009, decedents were more likely to be middle-aged, white, less educated, unmarried, and to experience financial difficulties, unemployment, physical disability, and mental illness compared with the general population.9
In recent state-based studies, the Medicaid population has been identified as having a higher risk of death due to unintentional prescription opioid overdose.10–13 In an analysis of Medicaid prescription claims of death due to unintentional prescription opioid overdose between 2003 and 2012 in Montana, the age-adjusted mortality rate was 8 times higher than the rate for non-Medicaid adults.10 A similar 8-fold higher rate in prescription opioid overdose deaths was noted among Medicaid enrollees in Washington state during the same time frame (2004–2007).14 The Medicaid population is at higher risk of fatal outcomes because recipients have higher rates of comorbid mental health and substance abuse disorders and pain management care than the general population.10,11,13,15–17 In North Carolina in 2007, Medicaid recipients who received services for drug dependence, mental disorders, joint and back disorders, musculoskeletal system and soft tissue disorders, and respiratory and circulatory disorders were more likely to die from an unintentional overdose (opioid and nonopioid) than non-Medicaid recipients.17
Additional research is required to fully characterize individual risk factors that lead to fatal prescription opioid overdose in Medicaid members, including clinical characteristics, pharmacy and medical utilization history, and other drug exposures. The objectives of this study were to: (1) describe demographic and clinical characteristics, medical utilization, opioid use, concurrent use of benzodiazepines, risk factors, and substances involved in death for Oklahoma’s Medicaid members who died of unintentional prescription opioid poisoning; and (2) compare characteristics, medical utilization, and medication use with Medicaid members who did not have an observed death in the mortality database.
This case-control study analyzed Oklahoma state Medicaid pharmacy and medical claims provided by the Oklahoma Health Care Authority and data abstracted from medical examiner reports from the Oklahoma State Department of Health’s (OSDH) Fatal Unintentional Poisoning Surveillance System, from January 1, 2012 to June 30, 2016. Medical examiner report data consisted of member identifier, date of death, probable cause of death, and drugs reported in toxicology reports. Unintentional fatal poisoning was determined by the medical examiner through toxicological analysis.
Medical and hospital claims included outpatient physician visits, inpatient and outpatient hospital visits, diagnosis and procedure codes, dates of service, and Medicaid reimbursement amounts. Pharmacy claims consisted of quantity, days’ supply, date of service, prescriber characteristics, reimbursement amount, national drug code, and generic product identifier (which classifies drugs by group, class, subclass, name extension, dosage form and strength). Demographics such as age, sex, and self-identified race were obtained from eligibility files.
Medical examiner data for unintentional prescription overdose deaths were abstracted onto Excel spreadsheets by OSDH staff and transferred to researchers. They were linked with pharmacy, medical, and hospital claims data through the member identifier and subsequently deidentified before analysis. The research was approved by the pharmacy director of the state Medicaid program and the University of Oklahoma Health Sciences Center Institutional Review Board after ethics review.
Cases were required to be Medicaid eligible during year of death, had an opioid involved in probable cause of death, and had ≥1 opioid prescription claim between January 1, 2011 and June 30, 2016. Case index date was defined as date of death recorded by the medical examiner. Some of the cases may not have been eligible at the time of death. Decedents without any opioid prescription claims during the time period were also described.
Nonevent controls were living Medicaid members with ≥1 opioid prescription claim during the study period. The nonevent controls were matched 3:1 to cases through propensity score matching using age, sex, race, Deyo-Charlson Comorbidity Index (CCI) score, and number of opioid prescription claims. Members dually eligible for Medicare and Medicaid were excluded (because of incomplete records). Sample selection is presented in Supplementary Figure 1 (Supplemental Digital Content 1, http://links.lww.com/MLR/B598). Median time-to-event from the first opioid claim to date of death in cases was added to the first opioid claim for each control to create dummy index dates (ie, replicating when the event could have occurred relative to matched controls). Demographics, comorbidities, pharmacy, and medical utilization were examined 12 months before index date. Prevalence of substances attributed to death was gathered from the cause of death field in medical examiner data.
Member demographics consisted of age, sex, self-identified race (excluding ethnicity), residence type [metropolitan (≥50,000 population), micropolitan (10,000–49,999 population), and rural (other areas)], and Oklahoma regions grouped into 6 regions (Northwest, Southwest, Northeast, Southeast, Oklahoma City, and Tulsa). Comorbidities were identified from diagnosis fields in medical claims with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and Tenth Revision (ICD-10) codes for CCI comorbidity conditions related to substance use disorders, and chronic pain (identified by ≥2 claims at least 90 d apart from the same pain code set).
Medical utilization was defined as the number of all-cause hospitalizations and the number of emergency room (ER) visits in medical claims. Opioid and benzodiazepine utilization was determined from prescription claims data, including the number of claims, days’ supply, and type of medication. Generic Product Identifier codes used to define opioid and benzodiazepine prescriptions are listed in Supplementary Table 1 (Supplemental Digital Content 2, http://links.lww.com/MLR/B599). Opioid amounts were converted to daily MME using the Centers for Disease Control and Prevention methodology.18 Daily MME doses were categorized into 1–29, 30–59, 60–89, 90–179, and ≥180 mg. Concomitant opioid and benzodiazepine use was defined as calendar days during which drugs from both classes were available to the member based on the dates of service and days’ supply. Days of overlap were categorized into 0, 1–6, 7–29, 30–89, 90–179, and ≥180 days.
Substances per member were abstracted from the cause of death column on the medical examiner spreadsheet. Substances were reported by frequencies and percentages involved for each death and by combination for dealths with multiple substances listed.
Demographics, clinical characteristics, and medical and pharmacy utilization in the 12 months before death were reported as means and SDs for continuous measures, and frequencies and percentages for categorical measures. The prevalence of substances attributed to death were reported as frequencies and percentages. Differences between cases and controls were evaluated with Wilcoxon-Mann-Whitney tests for nonparametric continuous variables, χ2 goodness of fit for categorical variables, and the Fisher exact test where ≥1 or cells had an expected frequency of ≤5.
A logistic regression model was used to assess factors associated with death. Model selection included collinearity assessment and evaluation of predictors through fit statistics. Final model covariates included age, sex, race, residence type, modified CCI (excluding liver conditions), nonopioid substance abuse, prior opioid dependence and poisonings, hepatitis, anxiety, major depressive disorder, mild depression, bipolar disorder, schizophrenia, number of chronic pain diagnoses, number of inpatient hospitalizations and ER visits, average daily MME dosage, days of opioid/benzodiazepine overlap, and year. All odds ratios presented are adjusted. Analyses were conducted with SAS, version 9.4 (SAS Institute, Cary, NC) and STATA, version 14.2 (StataCorp, College Station, TX).
Demographics and Clinical Characteristics
Overall, 639 members had fatal unintentional prescription opioid overdoses, of whom 321 had ≥1 opioid prescription claim and 318 had no opioid prescription claims in the year before death. Members with claims were compared with 963 controls.
Demographic and clinical characteristics of cases and controls are presented in Table 1. The mean±SD age for the cases with claims was 44.5±11.2 years; most (64.2%) were female, self-identified white (80.7%), and had mean CCI score of 1.5±2.0. No difference was seen with the controls for the aforementioned characteristics. For the 318 decedents without opioid prescription claims, the mean±SD age was 47.0±12.9 years, approximately half (51.6%) were female, and 84.3% self-identified as white. Controls had an average of 9.9±3.9 months of eligibility before their index dates, whereas cases averaged 10.4±3.3 months of eligibility before their index dates, a statistically significant difference (P<0.05).
Compared with controls, cases had greater proportions of nonopioid substance use disorders (67.3% vs. 44.9%; P<0.001), opioid dependence/abuse (23.7% vs. 7.3%; P<0.001), hepatitis (14.6% vs. 4.9%; P<0.001), gastrointestinal bleeding (24.6% vs. 18.7%; P=0.022), trauma not involving motor vehicle accidents (53.9% vs. 43.6%; P=0.001), and nonopioid poisonings (20.3% vs. 3.7%; P<0.001) (Table 1). A significantly higher proportion of cases also had anxiety (61.7% vs. 32.7%; P<0.001), mild depression (38.0% vs. 23.7%; P<0.001), major depressive disorder (30.8% vs. 17.0%; P<0.001), bipolar disorder (25.2% vs. 9.2%; P<0.001), and schizophrenia (9.4% vs. 2.9%; P<0.001). Cases were reported to have higher common chronic pain types than controls: neck or joint pain (32.4% vs. 26.4%; P=0.037) and low back pain (43.6% vs. 26.6%; P<0.001). The mean number of unique pain diagnoses was also higher for cases (1.5±1.5 vs. 1.0±1.3 diagnoses; P<0.001).
The preindex utilization for both groups is presented in Table 2. The cases presented to the ER more often than controls (2.6±2.9 vs. 1.5±2.3 visits per patient; P<0.001). Cases had significantly greater exposure to opioids in terms of daily MME doses (67.2±74.4 vs. 47.2±50.9 mg; P<0.001) compared with controls. Cases had more recorded use of long-acting opioids (28.7% vs. 15.9%; P<0.001) and oxycodone (49.2% vs. 39.5%; P=0.002); however, the use of short-acting opioids was comparable between cases and controls. Nearly twice as many cases than controls had opioid/benzodiazepine overlap (70.4% vs. 35.9%; P<0.001), with a longer mean duration (119.1±128.2 vs. 48.3±94.7 d; P<0.001) of overlap.
Substances Involved in Death
The most common opioids which were reported to be involved in case fatalities included oxycodone (36.5%), hydrocodone (35.5%), morphine (16.5%), methadone (12.5%), and fentanyl (11.2%) (Table 3). The most common benzodiazepine was alprazolam (22.1%). The top-specific single and multiple substances involved in individual deaths were single opioids [oxycodone (9.4%), hydrocodone (5.3%), fentanyl (4.4%), methadone (3.7%), and morphine (3.4%)], opioid with opioid [ie, hydrocodone and oxycodone (2.5%)], and opioids with benzodiazepines [ie, alprazolam and hydrocodone (3.1%), and alprazolam and oxycodone (2.2%)]. Benzodiazepines were reported to be involved in 29.3% of deaths during the study. Decedents without prescription opioid claims reported similar proportions to cases, although morphine (20.1%) and methadone (15.7%) were slightly higher. In total, 32 deaths (10.0%) involved methamphetamine, 6 (1.9%) involved cocaine, and 2 (<1%) involved heroin among cases. For those deaths without claims, 33 (10.4%) involved methamphetamine, 9 (2.8%) involved cocaine, and 2 (<1%) involved heroin.
Logistic Regression Analyses
Members diagnosed with opioid dependence had a 1.67 [95% confidence interval (CI), 1.08–2.57] times higher adjusted odds of death than members without a diagnosis (Table 4). Members also showed higher odds of mortality than controls with a diagnosis of any hepatitis A, B, or C [adjusted odds ratio (aOR), 2.52; 95% CI, 1.47–4.33] and nonopioid poisonings (aOR, 3.18; 95% CI, 1.87–5.40). Bipolar disorder was associated with an aOR of 1.78 (95% CI, 1.20–2.65) and schizophrenia with an aOR of 1.15 (95% CI, 1.04–1.27).
Analyses of unintentional opioid prescription overdose fatality by total average daily MME doses and the days of opioid/benzodiazepine overlap are presented in Figure 1. For total average daily MMEs, odds ratios were statistically significant at 90–179 MMEs/day (aOR, 3.07; 95% CI, 1.75–5.38) and at ≥180 MMEs/day (aOR, 3.30; 95% CI, 1.60–6.81) compared with 1–29 MMEs/day. Relative to no days of opioid/benzodiazepine overlap, higher odds of mortality were associated with exposure to 30–89 days of overlap (aOR, 3.00; 95% CI, 1.84–4.89), 90–179 days (aOR, 2.66; 95% CI, 1.51–4.68), and ≥180 days (aOR, 3.55; 95% CI, 2.34–5.39).
This case-controlled study described opioid poisoning and substances involved in the deaths of the Oklahoma Medicaid members who died of unintentional prescription opioid overdose and identified risk factors associated with fatal poisoning. As a whole, the population was middle-aged, majority female, predominantly white, with a low CCI index. At baseline, cases had a higher incidence of nonopioid substance use disorders, opioid dependence and abuse, hepatitis, gastrointestinal bleeding, trauma, and nonopioid poisonings, and more chronic pain diagnoses than those of controls. Cases also experienced more hospitalizations and ER visits than controls. Multivariable analyses found that opioid dependence, hepatitis, nonopioid poisonings, bipolar disorder, schizophrenia, greater exposure in average total daily MMEs (range, 90–>180 MMEs/d), and greater length of opioid/benzodiazepine overlap (range, 30–>180 d) were associated with higher odds of unintentional opioid prescription overdose deaths in the cases compared with the controls.
The results of this study, showing that higher exposure to prescription opioids, other substance use disorders, and higher opioid/benzodiazepine overlap were associated with opioid deaths, are consistent with other studies that examined the characteristics of unintentional opioid prescription deaths.6,9,19,20 Previous research has also showed the high rate of substance abuse among prescription drug decedents,13,17,21 and potential unrecognized substance use disorders among patients treated with prescription opioids.22 In the current study, a high proportion of deaths consisted of polydrug poisoning, and the most frequent (nearly 30%) combination involved opioids and benzodiazepines. A greater proportion of cases had opioid/benzodiazepine overlap and a longer mean duration of overlap than controls. Benzodiazepines have been reported frequently in deaths involving opioid analgesics.5,6,12,23,24 In a prospective, observational cohort study of North Carolina residents in 2010, benzodiazepines were prescribed in the previous year to 80% of patients who received opioid analgesics, and were involved in 61% of opioid overdose deaths. The rate of overdose death was 10 times higher in patients who received both benzodiazepines and opioids compared with opioids alone.6 Similarly, Fulton-Kehoe and colleagues reported that among Medicaid members in Washington state between 2006 and 2010, 44.5% of methadone poisoning decedents and 48.1% of other opioid poisoning decedents had sedative (eg, benzodiazepine) prescriptions during the month before death.23 A case-cohort study among United States military veterans revealed that 49% of opioid overdose decedents had concurrent opioid analgesic and benzodiazepine prescriptions.5 A significant increase in opioid mortality risk was noted with recent concurrent use of benzodiazepines in a Washington Medicaid population in 2010.12
Cases had significantly greater exposure to opioids in daily MMEs (90–>180 MMEs/d compared with the lower doses). This is consistent with other studies showing a dose-response relationship between higher daily opioid doses and risk of death.7,12 Dunn et al25 reported an 8.9-fold increase in overdose risk in patients receiving ≥100 MME/day. Dasgupta et al6 used a prescription monitoring program with name-linked mortality data to reveal that death rates increased gradually across the range of average daily MME. Results from the present study support recommendations for increased clinical vigilance with high-dose prescriptions.26 However, caution should be used when considering these average daily doses in the absence of other contexts, as a recent literature review determined that, though doses ranging from 40 to 200 MME/day were associated with increases in opioid overdose and mortality,7 no dose threshold for opioid overdose death could be established.
In the present study, cases had higher common chronic pain types, mainly neck or joint pain and low back pain. Death rates from opioid analgesic poisoning were reported as increased in members with chronic pain compared with those with no chronic pain. A high rate of chronic pain has previously been reported among prescription opioid decedents.27 However, the balance of long-term risks and benefits of use of opioid analgesic in pain is poorly understood, and further research is required to evaluate the risk factors for opioid overdose mortality rates in patients with chronic pain.25
The study revealed that approximately half of current or former Oklahoma Medicaid members with a fatal unintentional prescription opioid overdose did not have a covered prescription claim in the year before death. This finding may represent the reality that some people were not eligible near the time of death, implying that they obtained their prescriptions through other insurance coverage, other forms of payment, or diversion. Prescription utilization for this group was likely underestimated. This finding reveals the necessity of investigating out-of-pocket payments, such as through prescription drug monitoring program databases, to help state Medicaid programs form a more comprehensive picture of opioid utilization.28 Such data may assist in identifying beneficiaries with opioid abuse.
Although these results were consistent with previous studies, limitations of the present study should be considered. The exact cause of deaths in cases with multiple or mixed opioid poisoning is difficult to determine because it is unknown whether the opioids were the primary cause of death or were secondary to another substance. In addition, deaths are assumed to be unintentional unless there is evidence that proves otherwise. The study did not analyze utilization patterns beyond Medicaid coverage and included the claims data for only 1 state, which may limit generalizability. The sources for substances associated with opioid deaths were not reported by the medical examiner, therefore, the study was unable to determine whether opioid substances involved in death were obtained illicitly or by prescription. Some cases may not have been eligible for Medicaid at the time of death. Finally, each death occurred in different contexts with varying levels of evidence. Therefore, a uniform standard for determining cause of death was not possible and some overdose deaths could be classified as having another cause.
Decedents had greater proportions of mental health disorders, pain, and comorbidities associated with addiction compared with the controls. The findings also emphasize potentially problematic opioid exposure including higher daily MMEs and longer durations of opioid/benzodiazepine overlap. A large proportion of deaths involved benzodiazepines and illicit substances use. The risks of concurrent opioid and benzodiazepine use should be carefully considered, even for a short course of therapy, especially since labeling information for benzodiazepines have a black box warning regarding concomitant use of benzodiazepines and opioids.29 Prescribers and policymakers should continue to monitor overlaps in benzodiazepine and opioid prescriptions, paying particular attention to state guidelines and regulations regarding use of prescription monitoring programs. State health care agencies should provide education on safety regarding concurrent use of these 2 class of drugs. Certain patient populations, such as those who receive Medicaid services, have high-risk factors (eg, comorbid mental health and substance use disorders) for unintentional opioid overdose death, and alternatives to benzodiazepines, such as nonbenzodiazepine anxiolytics, may be required for these patients if the risk of overdose is considered too great.9–11,13,15–17 Polysubstance use or abuse risk is compounded by patients having multiple prescriptions of different medications from various providers.30 For example, a pain management specialist may prescribe an opioid to a patient, whereas a psychiatrist prescribes a benzodiazepine, with neither the specialist nor the psychiatrist being aware of other prescriptions. Care coordination models are required to bring disparate specialties together, raise awareness of possible drug-drug interactions, and decrease the risk of adverse events.
Addressing the opioid epidemic requires an informed policy. State Medicaid agencies already implement a variety of strategies.31 Worldwide, policies to address prescription opioid misuse and associated harms have varied. Some Eastern European countries set hard restrictions, such as specifying a maximum daily dose or authorizing only a few physicians prescribing power.32 The Canadian province of Ontario removed high-dose opioids from formulary coverage.33 Guidelines in Australia, Canada, and Germany differ in daily dose thresholds requiring caution, from 50 to 200 mg.34 The findings of this study may contribute to recommendations for establishing rational opioid dose thresholds and use of benzodiazepines in clinical guidelines and government policies. It also adds to the existing body of literature by providing a greater level of detail on the characteristics and utilization of Medicaid members who experienced a fatal unintentional prescription drug overdose. Specifically, it directly examined of the scale of opioid/benzodiazepine overlap and the drugs involved in death among this population.
The authors thank Kim Poinsett-Holmes, PharmD, and Sharad Upadhyay, M Pharm, of Evidera (Bethesda, MD) for providing medical writing support, which was funded by Purdue Pharma L.P., Stamford, CT. The authors also thank Nancy J. Nesser, PharmD, JD, of the Oklahoma Health Care Authority, for allowing access to the Oklahoma Medicaid medical and pharmacy claims databases and Claire B. Nguyen, MS, of the Oklahoma State Department of Health for providing prescription drug overdose death data.
1. Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results From the 2016 National Survey on Drug Use and Health [HHS Publication No (SMA) 17 5044, NSDUH Series H 52]. Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration; 2017. Available at: www.samhsa.gov/data/
. Accessed November 1, 2017.
4. Rudd RA, Aleshire N, Zibbell JE, et al. Increases in drug and opioid
overdose deaths—United States, 2000-2014. MMWR. 2016;64:1378–1382.
5. Park TW, Saitz R, Ganoczy D, et al. Benzodiazepine prescribing patterns and deaths from drug overdose
among US veterans receiving opioid analgesics
: case-cohort study. BMJ. 2015;350:h2698.
6. Dasgupta N, Funk MJ, Proescholdbell S, et al. Cohort study of the impact of high-dose opioid analgesics
on overdoes mortality. Pain Med. 2016;17:85–98.
7. King NB, Fraser V, Boikos C, et al. Determinants of increased opioid
-related mortality in the United States and Canada, 1990-2013: a systematic review. Am J Public Health. 2014;104:e32–e42.
8. Paulozzi LJ, Kilbourne EM, Shah NG, et al. A history of being prescribed controlled substances and risk of drug overdose death
. Pain Med. 2012;13:87–95.
9. Johnson EM, Lanier WA, Merrill RM, et al. Unintentional prescription opioid
-related overdose deaths: description of decedents by next of kin or best contact, Utah, 2008-2009. J Gen Intern Med. 2013;28:522–529.
10. Fernandes JC, Campana D, Harwell TS, et al. High mortality rate of unintentional poisoning due to prescription opioids in adults enrolled in Medicaid compared to those not enrolled in Medicaid in Montana. Drug Alcohol Depend. 2015;153:346–349.
11. Mack KA, Zhang K, Paulozzi L, et al. Prescription practices involving opioid analgesics
among Americans with Medicaid, 2010. J Health Care Poor Underserved. 2015;26:182–198.
12. Garg RK, Fulton-Kehoe D, Franklin GM. Patterns of opioid
use and risk of opioid
among Medicaid patients. Med Care. 2017;55:661–668.
13. Paulozzi L, Baldwin G, Franklin G, et al. CDC grand rounds: prescription drug overdoses–a US epidemic. MMWR. 2012;61:10–13.
14. Coolen P, Best S, Lima A, et al. Overdose deaths involving prescription opioids among Medicaid enrollees–Washington, 2004-2007. MMWR. 2009;58:1171–1175.
15. McAdam-Marx C, Roland CL, Cleveland J, et al. Costs of opioid
abuse and misuse determined from a Medicaid database. J Pain Palliat Care Pharmacother. 2010;24:5–18.
16. Adelmann PK. Mental and substance use disorders among Medicaid recipients: prevalence estimates from two national surveys. Adm Policy Ment Health. 2003;31:111–129.
17. Whitmire JT, Adams GW. Unintentional overdose deaths in the North Carolina Medicaid population: prevalence, prescription drug use, and medical care services. SCHS Studies. 2010;162:1–12.
18. Prescription Drug Monitoring Program Training and Technical Assistance Center. Centers for Disease Control and Prevention compilation of benzodiazepines
, muscle relaxants, stimulants, zolpidem, and opioid analgesics
with oral morphine milligram equivalent conversion factors. 2016. Available at: www.pdmpassist.org/pdf/Conversion%20Reference%20Table.xlsx
. Accessed October 30, 2016.
19. Cerdá M, Ransome Y, Keyes KM, et al. Prescription opioid
mortality trends in New York City, 1990–2006: examining the emergence of an epidemic. Drug Alcohol Depend. 2013;132:53–62.
20. Chen LH, Hedegaard H, Warner M. Drug-poisoning Deaths Involving Opioid Analgesics
: United States, 1999–2011 NCHS Data Brief, No 166. Hyattsville, MD: National Center for Health Statistics; 2014. Available at: www.cdc.gov/nchs/products/databriefs/db166.htm
. Accessed November 11, 2017.
21. Hall AJ, Logan JE, Toblin RL, et al. Patterns of abuse among unintentional pharmaceutical overdose fatalities. JAMA. 2008;300:2613–2620.
22. Braden JB, Russo J, Fan MY, et al. Emergency department visits among recipients of chronic opioid
therapy. Arch Intern Med. 2010;170:1425–1432.
23. Fulton-Kehoe D, Sullivan MD, Turner JA, et al. Opioid
poisonings in Washington State Medicaid: trends, dosing, and guidelines. Med Care. 2015;53:679–685.
24. Jones CM, Mack KA, Paulozzi LJ. Pharmaceutical overdose deaths, United States, 2010. JAMA. 2013;309:657–659.
25. Dunn KM, Saunders KW, Rutter CM, et al. Opioid
prescriptions for chronic pain and overdose: a cohort study. Ann Intern Med. 2010;152:85–92.
26. Dowell D, Haegerich TM, Chou R. CDC guideline for prescribing opioids for chronic pain—United States, 2016. MMWR Recomm Rep. 2016;65:1–49.
27. Toblin RL, Paulozzi LJ, Logan JE, et al. Mental illness and psychotropic drug use among prescription drug overdose
deaths: a medical examiner chart review. J Clin Psychiatry. 2010;71:491–496.
28. Hartung DM, Ahmed SM, Middleton L, et al. Using prescription monitoring program data to characterize out-of-pocket payments for opioid
prescriptions in a state Medicaid program. Pharmacoepidemiol Drug Saf. 2017;26:1053–1060.
29. Pharmacia and Upjohn Company LLC. Xanax (Package Insert). New York, NY: Pfizer Inc; 2017.
30. O’Brien PL, Karnell LH, Gokhale M, et al. Prescribing of benzodiazepines
and opioids to individuals with substance abuse disorders. Drug Alcohol Depend. 2017;178:223–230.
32. Cherny NI, Baselga J, de Conno F, et al. Formulary availability and regulatory barriers to accessibility of opioids for cancer pain in Europe: a report from the ESMO/EAPC Opioid
Policy Initiative. Ann Oncol. 2010;21:615–626.
33. Dyer O. Ontario plans to stop funding high dose opioids. BMJ. 2016;354:i4300.
34. Hauser W, Schug S, Furlan AD. The opioid
epidemic and national guidelines for opioid
therapy for chronic noncancer pain: a perspective from different continents. Pain Rep. 2017;2:e599.