Anesthesiology, Article

The utility of a statewide prescription drug-monitoring database vs the Current Opioid Misuse Measure for identifying drug-aberrant behaviors in emergency department patients already on opioids

a b s t r a c t

Background: The most recent guidelines on prescribing opioids from the United States Centers for Disease Control recommend that clinicians not prescribe opioids as first-line therapy for chronic non-cancer pain. If an opioid prescription is considered for a patient already on opioids, prescribers are encouraged to check the statewide pre- scription drug monitoring database (PDMP). Some additional guidelines recommend Screening tools such as the Current Opioid misuse Measure (COMM) which may also help identify drug-aberrant behaviors.

Objective: To compare the PDMP and the Current Opioid Misuse Measure (COMM), a commonly-recommended screening tool for patients on opioids, in detecting drug-aberrant behaviors in patients already taking opioids at the time of ED presentation.

Methods: Patients on opioids were enrolled prospectively in a mixed urban-suburban ED seeing approximately 65,000 patients per year. The sensitivity, specificity, likelihood ratios, and Diagnostic odds ratios of the PDMP and COMM were compared against objective criteria of drug-aberrant behaviors as documented in the electronic medical record (EMR) and Medical Examiner databases.

Results: Compared to the COMM, the PDMP had similar sensitivity (36% vs 45%) and similar specificity (79% vs 55%), but better positive predictive value, better negative predictive value, and better diagnostic odds ratio. The combination of the PDMP and the COMM did not improve the detection of drug-aberrant behaviors.

Conclusions: The PDMP alone is a more useful as a screening instrument than either the COMM or the combina- tion of the PDMP plus COMM in patients already taking opioids at time of ED presentation. However, the PDMP misses a majority of patients with documented drug-aberrant behaviors in the EMR, and should not be used in isolation to justify whether a particular opioid prescription is appropriate.

(C) 2019

  1. Introduction

More than 30% of Americans suffer from some form of acute or chronic pain [1]. Chronic non-cancer pain is particularly challenging to treat in the emergency department (ED), as available non-opioid op- tions either have numerous side effects or limited efficacy when used

* Corresponding author at: Department of Emergency Medicine, University of Arkansas for Medical Sciences, 4301 West Markham Street, slot #584, Little Rock, Arkansas 72205, United States of America.

E-mail address: [email protected] (M.P. Wilson).

chronically [2]. Although there is little rigorous research supporting the use of opioids in chronic noncancer pain, patients with chronic pain are the most likely to chronically ingest opioids [3]. Many of these patients do experience at least some pain relief with long-term use [4-7], but approximately 80% of patients who ingest opioids chron- ically experience at least one side effect and patients who discontinue opioids largely do so because of these adverse effects [4,8]. More trou- bling, the prevalence of addiction with chronic use of opioids varies any- where from 0%-50% [9], and the risk of overdose is increased with an associated substance use disorder [10]. For these reasons, despite the fact that existing guidelines from 2007 to 2013 largely included opioid

0735-6757/(C) 2019

treatment in their recommendations [11], many EDs have generally re- fused to prescribe opioids for chronic pain.

More recent CDC and state guidelines now recommend limiting opi- oid prescriptions for chronic noncancer pain to patients in which the “benefits for pain and function are expected to outweigh risks,” and en- courage both non-opioid and non-Pharmacologic therapy as first-line therapies [12,13]. If an opioid prescription for chronic noncancer pain is considered, the CDC states that “checking state Prescription Drug Monitoring Programs before long-term prescribing of controlled sub- stances should be a standard of care” [14], and laws in 37 states also now require physicians to check prescription drug-monitoring pro- grams (or PDMPs) before prescribing opioids in at least some circum- stances [15]. Use of the PDMP has been shown to alter physician Prescribing habits [16-18], but there is surprisingly little data about the overall ability of the PDMP to detect drug-aberrant behaviors in ED patients. Despite this, the recommendation to consult the PDMP is listed as best-practice, although this mainly relies on Type 4 evidence (i.e., evidence drawn from “clinical experience and observations, obser- vational studies with important limitations, or randomized clinical trials with several major limitations”) [12].

Since a typical PDMP does not capture many behaviors that might be considered aberrant, other guidelines such as the Washington state “In- teragency Guideline on Prescribing Opioids for Pain” have additionally urged the use of bedside screening tools, like the Current Opioid Misuse Measure (COMM), before prescribing opioids [13]. The COMM contains 17 items that were developed by consensus of pain specialists, addiction experts, and primary care providers, and may identify drug-aberrant be- haviors in patients already taking opioids for chronic noncancer pain [19]. As the COMM assesses a number of drug-aberrant behaviors not recorded by a typical PDMP, use of this tool may potentially be a better way to detect drug-aberrant behaviors in ED patients. Somewhat sur- prisingly, however, the COMM has never been compared to a statewide PDMP for detection of drug-aberrant behaviors.

The objective of this study was to compare the performance of a

statewide PDMP, the COMM, and both measures together in detecting drug-aberrant behaviors in patients who presented for pain-related complaints or prescription refills, but who were already taking opioids by the time of ED presentation. The sensitivity and specificity of each tool were calculated in comparison to a reference standard of commonly-accepted drug-aberrant behaviors which are typically docu- mented either in the electronic medical record (EMR) or statewide medical examiner (ME) databases.

  1. Methods
    1. Participants

This is a secondary analysis of existing data from a study that col- lected patient-related data on three common opioid-risk screening tools in the ED (including the COMM), data about these same patients obtained from the EMR, and data on these same patients obtained from the PDMP [20]. As that report included all patients, whether or not they were already on opioids, the objective of the current study was to examine the COMM in patients with Opioid prescriptions at the time of ED presentation. In contrast to the previous study, this study also compares the results of the COMM to the PDMP for these same patients.


The study institution was an ED of a mixed urban-suburban hospital in Southern California, seeing approximately 65,000 patients per year. All patients presenting to the ED when a research associate (RA) was available, typically daytime Monday-Friday 8a-8p, were consecutively approached to participate in the study. As this study did not receive funding from any source, RAs were available in the ED during these

hours as chart review indicated that approximately 83% of weekday ar- rivals with pain complaints were present in the ED during this time. Par- ticipants were eligible to participate in the study if they reported pain persisting in the same body part for 6 months or longer, even if unre- lated to the current visit, or were requesting a refill request for opioids regardless of pain duration. Patients were excluded if 17 years or youn- ger, incarcerated, if they were unable to consent because of language difficulties or medical condition, or had a Cancer diagnosis. In this sec- ondary analysis of the data, patients were further excluded if a review of the statewide PDMP indicated that they did not have any prescrip- tions for schedule II or III opioids within the previous 90 days. The study was approved by the local Institutional Review Board before data collection.

Study design and procedure

After informed consent was obtained, each participant completed the COMM along with other pain measures reported separately [20]. Participants were informed that the purpose of the study was to evalu- ate opioid use in the ED and that their physician would not be told whether or not they participated in the study. This survey was filled out by the participant without the RA present, unless the participant specifically requested the RA to read the questions. After filling out the form, the participant returned the survey to the RA by placing it in a plain brown envelope.

Three months or more after the ED visit, the participant’s record in both the EMR of the ED-affiliated hospital system and the statewide PDMP were searched for all data using the triage name (which requires a driver’s license or other form of identification in the studied institu- tion) and birthdate. If the search of either the EMR or PDMP failed to show evidence that the patient was alive, the ME database was addi- tionally searched by the patient’s name and date of birth. The ME is a li- censed physician and is required by state law to determine the cause and manner of deaths that are either sudden or unexpected. The inves- tigators planned a priori to include both overdose deaths and deaths of an undetermined nature as evidence of an overdose death.

Development of the reference standard

As no gold standard of drug abuse exists, and as the criteria from the Diagnostic and Statistical Manual (DSM-5) are difficult to assess without a structured interview which is unfeasible in typical ED prac- tice, the reference criteria for this study instead relies on objective documentation of drug-aberrant behaviors as might be found in the EMR and/or ME databases. The reference standard was constructed after a review of available literature and conversations with pain ex- perts. Criteria such as the Diagnostic and Statistical Manual typically rely largely on the patient’s self-report in order to assess addiction, but the reference standard utilized here is based solely on objective documentation. (Please see Table I for a list.) All criteria were given equal weight, and the presence of any drug-aberrant behavior was considered as sufficient to categorize the patient as having drug- aberrant behaviors.

Although the documentation of drug-aberrant behaviors in this manner does not permit a formal diagnosis of Opioid use disorder, criteria such as selling prescriptions, forging prescriptions, injecting Oral formulations, concurrent abuse of illicit drugs, and repeatedly seek- ing refills from other providers have been proposed as being especially predictive of a substance use disorder [21]. These criteria are also listed by as “more associated with medication abuse/addic- tion” [22]. Finally, as physicians are increasingly being asked to scan the EMR or PDMP for evidence of drug-aberrant behaviors before pre- scribing opioids [23], this standard is more clinically relevant to emer- gency medicine practice.

Table I

Reference criteria used in the study

Documentation of forging a prescription within the preceding 3 months and/or subsequent 3 months of enrollment ED visit.
  • Documentation of selling a prescription within the preceding 3 months and/or subsequent 3 months of enrollment ED visit.
  • Data from the electronic medical record (EMR)



    Data from the Medical Examiner (ME)

    Demographic variables such as age and gender were determined from the EMR. Each opioid prescription reported by the PDMP was con- verted into morphine milligram equivalents (MME) for each patient using a publicly-available calculator [26].

    Data processing

    Although the study enrolled patients prospectively, extraction of PDMP, ME, and EMR data post-enrollment was performed three months or more after the patient visit. These data were collected with adherence to methodological standards recommended for retrospective reviews, including the utilization of two trained RAs who were blinded to the hy-

    Overdosed or died from illegal or prescription drug X X use in the subsequent 3 months after the enrollment

    ED visit.

    Positive urine drug screen for illegal substances X (amphetamines, barbiturates, cocaine, or

    phencyclidine) within the preceding 3 or subsequent 3 months of the enrollment ED visit.

    Solicited or received a schedule II/III prescription from X 3 or more separate physicians in the preceding 3

    months or subsequent 3 months of enrollment ED visit.

    Participant had 3 or more visits to the study X institution’s ED with a chief complaint of any painful

    condition within the preceding 3 months or subsequent 3 months of enrollment ED visit.

    Solicitation or receipt of a schedule II/III prescription X refill at the ED after missing a clinic visit within 7 days

    before enrollment ED visit.

    Discharged from a medical practice due to any X drug-related aberrant behavior within the preceding 3

    months and/or subsequent 3 months of enrollment ED visit.

    Documentation of self-administration of a schedule X II/III drug not in the intended manner, including route

    or dose (such as snorting or crushing an oral medication), within the preceding 3 months and/or subsequent 3 months of enrollment ED visit.


    The COMM is a 17-item questionnaire for patients already taking opioids for chronic noncancer pain [19]. Among other questions, the COMM asks patients how often in the last 30 days they have been think- ing about opioids, how often they have had to go to someone other than their prescribing physician for pain relief, and how often they have been taking their medication differently than prescribed. The COMM is de- signed to identify whether a particular patient may be engaged in cur- rent drug-aberrant behaviors as opposed to predicting the future risk of drug-aberrant behaviors in patients being considered for long-term opioid therapy [24]. As the 90-day Test-retest reliability is high for this measure [19], we defined current drug-aberrant behaviors as occurring within the 90-days before the ED enrollment visit or within 90-days after (please see Table I). At a cutoff score of 9, the initial validation of the COMM in chronic outpatient pain management patients yielded a sensitivity of 0.77 and a specificity of 0.66.

    The PDMP is a statewide database which records all controlled sub- stance prescriptions filled within the state, and has an additional reporting mechanism for forged or stolen prescriptions. Use of this data- base is more reliable for detecting drug-aberrant behavior than physi- cian impression [25]. As the PDMP has no standard cut-off, patients were considered to have evidence of drug-aberrant behaviors if they re- ceived opioid prescriptions from >=3 physicians in the preceding 3 months of the enrollment ED visit or if they had forged or stolen a pre- scription. Although the duration over which to measure doctor- shopping behavior often varies between studies, this definition is similar to previous studies of this type [23]. Although participation by physi- cians in the PDMP was voluntary at the time of this study, all pharmacies in the state were required by law to participate.

    pothesis of the study, who were blinded to the outcome measures of the study, and who used a prespecified data abstraction tool [27]. All ex- tracted data were then compared between RAs, and disagreements re- solved by consensus.

    Data analyses

    The primary measures of interest were the sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratios, and diagnostic odds ratio of the PDMP (using a cutoff score of >=3 pre- scriptions from >=3 providers or sale/forgery of prescriptions within the preceding 3 months) as compared to the COMM (using a cutoff score of >=9) for detecting current drug-aberrant behaviors. These instruments were then compared alone and in combination against drug-aberrant behaviors as documented in the EMR + ME databases (see Table I). Age is reported as mean +- standard deviation. 95% confidence intervals for the odds ratios and other statistical analyses were performed with RStudio 1.1.463 running R version 3.4.1 and Microsoft Excel 2016. Sig- nificance testing of the diagnostic odds ratio, as an overall measure of the effectiveness of the test, was performed using confidence intervals. Confidence intervals N1 were assumed to represent a more useful test.

    1. Result

    After meeting all Inclusion/exclusion criteria, 154 participants were enrolled in the study. Eighty-two of these patients (40 females, average age 49 +- 13 years) were subsequently determined to be on opioids at time of their ED presentation, and so are included in the analyses below. No participants were determined to have died during review of the ME database, and so all further analyses below report COMM and PDMP data that WERE compared to the EMR only.

    Current opioid-using participants had prescriptions for an average of 6240 MME, or an average of 69 MME per person per day for the preced- ing 90 days. Forty-four of these participants (54%) were subsequently determined to have drug-aberrant behaviors after a review of the criteria in Table I. When compared against the reference standard, the PDMP had a similar sensitivity (0.36 vs 0.45), a similar specificity (0.79 vs 0.55), but better likelihood ratios than the COMM (see Table II). The combination of the COMM and the PDMP was calculated in the following manner. A positive test result on either measure was considered as a positive test, while a negative test result required both the COMM and the PDMP to be negative. The sensitivity of the combina- tion was improved over the PDMP alone (0.61 vs 0.36) but the specific- ity was decreased (0.39 vs 0.79) as would be expected from tests administered in parallel that are combined in this manner [28].

    The diagnostic odds ratio, defined as the odds that a test is positive if the subject has the disease divided by the odds of the test being negative if the patient does not have a disease, was only N1 for the PDMP. This in- dicates that, overall, the PDMP was a more useful test than either the COMM or PDMP plus COMM combined in detecting drug-aberrant be- haviors. However, the PDMP alone failed to identify a majority of pa- tients with documented drug-aberrant behaviors in the EMR.

    Table II

    Performance of the COMM vs the PDMP for patients already on opioids before the enrollment visit.

    Average score

    Median score







    Diagnostic Odds Ratio

    COMM compared to EMR for 1 or more aberrant behaviors









    1.03 [0.16, 1.90]


    Evidence of drug-aberrant behaviors (COMM>=9)

    PDMP compared to EMR for 1 or more aberrant behaviors









    2.14 [1.15, 3.14]


    Evidence of drug-aberrant behaviors (>=3 rx in PDMP during 3 months prior,

    sold/forged rx)

    COMM or PDMP compared to EMR for 1 or more aberrant behaviors









    1.04 [.15, 1.93]


    Positive test (COMM or PDMP+), negative test (COMM and PDMP-)

    1. Discussion

    This study confirms findings from studies of outpatients, namely that at present there are few diagnostic tools which reliably predict drug-aberrant behaviors [6,29]. A recent study by Hawk and colleagues previously noted that a majority of patients with a known opioid use disorder are not captured by the PDMP [30]. This study extends that finding to patients with drug-aberrant behaviors who are already taking opioids by time of ED presentation. Nonetheless, use of the PDMP may still be a better screen for ED patients, especially when compared to available instruments such as the COMM that have only been validated in the outpatient setting [20].

    Sole use of the PDMP, however, is likely to miss significant numbers of patients with drug-aberrant behaviors. Such low sensitivity may be acceptable if the PDMP is used as a rapid screen for deciding whether or not to refer patients to treatment. However, there may be potential harm in using both the PDMP and the COMM if providers are either falsely reassured [31] or use the results as justification to prescribe more opioids than originally intended [32].

    The reasons for the low sensitivity of both the COMM and PDMP are unclear, but there are several possible explanations. First, some studies have documented that patients are moving from Prescription opioids to illicit drugs such as heroin or fentanyl analogs which are not captured by the PDMP [33,34]. Second, many patients may be diverting opioids to family or friends, which would also not be tracked by the PDMP [1]. Fi- nally, some percentage of patients are engaged in active deception. Al- though the number of these patients is likely b10% [6], both the COMM and the PDMP would be expected to miss many of these pa- tients. Such active deception might only be noted in the EMR, as was the case in this study.


    There are several limitations to this study. First, this is a relatively small study limited to a single-center in which the reference stan- dard is at least partially dependent on documentation in the EMR. However, this does not alter the main findings of the study, namely that both the PDMP and the COMM failed to identify a majority of drug-aberrant behaviors that were already documented in the EMR (i.e., each measure had a number of false negatives). Second, this study was not designed or powered to examine prevalence of drug- aberrant behaviors. Although the prevalence of drug-aberrant be- haviors in this sample is actually quite high (54%) and is similar to previous studies on this topic [9], the true prevalence may be not be representative of different settings.

    Third, there are potential difficulties with the reference standard used in this study, particularly the definition for doctor-shopping, defined as >=3 prescriptions from >=3 or more physicians in the 3 months before or after the enrollment visit. This definition is con- troversial, and is much less conservative than a study of opioid

    shopping by Cepeda and colleagues which utilized the definition of

    >=2 prescriptions filled at >=3 pharmacies over 18 months [35]. Al- though the definition used in this study could conceivably capture pain patients who are managing their conditions by frequent trips to the ED, this population is nonetheless at higher risk of overdose. As evidence on this point, an investigation of over 600,000 Medicare patients found that patients who received 3 prescriptions from dif- ferent providers in 6 months had an absolute risk of overdose of 4.8 per 1000 beneficiary years, which is approximately 2.5 times the risk of a patient with only one prescriber [23]. Consequently, al- though less conservative than the definition used by Cepeda et al. [35], the definition of doctor-shopping in this study is both scientif- ically reasonable and clinically relevant, as it allowed for chronic non-cancer patients to receive at least 1 additional prescription from another physician during this timeframe for presumably innoc- uous reasons.

    1. Conclusions

    Existing tools like the COMM and the statewide PDMP poorly predict drug-aberrant behaviors in ED patients already taking opioids, and should not be used in isolation in order to decide whether or not a par- ticular opioid prescription is appropriate. Although the PDMP is rela- tively insensitive to drug-aberrant behaviors, it has a higher positive predictive value, a higher negative predictive value, and a better diag- nostic odds ratio than a screening instrument such as the COMM. The performance of the PDMP is not improved with the addition of the COMM. The PDMP is therefore more likely a useful test overall to iden- tify patients with drug-aberrant behaviors, and may be particularly use- ful as a rapid screen if the goal is to identify ED patients who may benefit from substance use treatment.

    If the purpose of using the PDMP, on the other hand, is to prevent in- appropriate opioid prescriptions to patients with drug-aberrant behav- iors, the PDMP should not be the sole criterion used as it is likely to miss significant numbers of these patients. There is a potential harm in using the PDMP if providers are either falsely reassured [31] or use results of the PDMP as a justification to prescribe more opioids than originally intended [32].


    The Arkansas Department of Health does not guarantee the accuracy of the information, and the views expressed in this paper are not neces- sarily those of the Arkansas Department of Health.


    The authors would like to specially thank the UAMS Clinician Scien- tist Program, which generously supported a portion of Dr. Wilson’s time during the analysis and drafting of the manuscript.


    1. Volkow ND, McLellan AT. Opioid abuse in chronic pain–misconceptions and mitiga- tion strategies. N Engl J Med 2016;374(13):1253-63.
    2. Hansen GR. Management of chronic pain in the acute care setting. Emerg Med Clin North Am 2005;23(2):307-38.
    3. Shah A, Hayes CJ, Martin BC. Factors influencing long-term opioid use among opioid naive patients: an examination of initial prescription characteristics and pain etiolo- gies. J Pain 2017;18(11):1374-83.
    4. Kalso E, Edwards JE, Moore RA, McQuay HJ. Opioids in chronic non-cancer pain: sys- tematic review of efficacy and safety. Pain 2004;112(3):372-80.
    5. Ballantyne JC, LaForge KS. Opioid dependence and addiction during opioid treatment

    of chronic pain. Pain 2007;129(3):235-55.

    1. Volkow N, Benveniste H, McLellan AT. Use and misuse of opioids in chronic pain. Annu Rev Med 2018;69:451-65.
    2. Brennan F. The US congressional “decade on pain control and research” 2001-2011: a review. J Pain Palliat Care Pharmacother 2015;29(3):212-27.
    3. Noble M, Tregear SJ, Treadwell JR, Schoelles K. Long-term opioid therapy for chronic noncancer pain: a systematic review and meta-analysis of efficacy and safety. J Pain Symptom Manage 2008;35(2):214-28.
    4. Hojsted J, Sjogren P. Addiction to opioids in chronic pain patients: a literature re- view. Eur J Pain 2007;11(5):490-518.
    5. Webster LR. Risk factors for opioid-use disorder and overdose. Anesth Analg 2017; 125(5):1741-8.
    6. Nuckols TK, Anderson L, Popescu I, Diamant AL, Doyle B, Di Capua P, et al. Opioid prescribing: a systematic review and critical appraisal of guidelines for chronic pain. Ann Intern Med 2014;160(1):38-47.
    7. Dowell D, Haegerich TM, Chou R. CDC guideline for prescribing opioids for chronic pain–United States, 2016. JAMA 2016;315(15):1624-45.
    8. Washington State Agency Medical Directors’ Group. Interagency guideline on pre- scribing opioids for pain 2015 [cited 2018 July 14]. Available from: http://www.
    9. United States Centers for Disease Control. Vital signs: overdoses of prescription opi- oid pain relievers and other drugs among women — United States, 1999-2010 [cited 2018 May 9]. 62(26):537-542:[Available from: preview/mmwrhtml/mm6226a3.htm.
    10. . National Alliance for Model State Drug laws. Mandated use of PMPs – 37 states with specified circumstances requiring prescriber access 2018 [cited 2018 Nov 27]. Available from: 20PMPs%20-%20State%20Map%20(1-2-18)/.
    11. Bao Y, Pan Y, Taylor A, Radakrishnan S, Luo F, Pincus HA, et al. Prescription drug monitoring programs are associated with sustained reductions in opioid prescribing by physicians. Health Aff (Millwood) 2016;35(6):1045-51.
    12. Gugelmann HM, Perrone J. Can prescription drug monitoring programs help limit opioid abuse? JAMA 2011;306(20):2258-9.
    13. Manchikanti L, Kaye AM, Kaye AD. Current state of opioid therapy and abuse. Curr Pain Headache Rep 2016;20(5):34.
    14. Butler SF, Budman SH, Fernandez KC, Houle B, Benoit C, Katz N, et al. Development and validation of the current opioid misuse measure. Pain 2007;130(1-2):144-56.
    15. Chalmers CE, Mullinax S, Brennan J, Vilke GM, Oliveto A, Wilson MP. Screening tools validated in the outpatient pain management setting poorly predict opioid misuse in

    the emergency department: a pilot study. J Emerg Med 2019. 1016/j.jemermed.2019.03.018 [epub ahead of print].

    1. Michna E, Ross EL, Hynes WL, Nedeljkovic SS, Soumekh S, Janfaza D, et al. Predicting aberrant drug behavior in patients treated for chronic pain: importance of abuse his- tory. J Pain Symptom Manage 2004;28(3):250-8.
    2. Aberrant drug taking behaviors. Available from: default/files/files/AberrantDrugTakingBehaviors.pdf.
    3. Carey CM, Jena AB, Barnett ML. Patterns of potential opioid misuse and subsequent adverse outcomes in medicare, 2008 to 2012. Ann Intern Med 2018;168(12): 837-45.
    4. Current Opioid Misuse Measure (COMM). [cited 2018 Nov 27]. Available from 20Measure-ENG.pdf.
    5. Weiner SG, Griggs CA, Mitchell PM, Langlois BK, Friedman FD, Moore RL, et al. Clini- cian impression versus prescription drug monitoring program criteria in the assess- ment of drug-seeking behavior in the emergency department. Ann Emerg Med 2013;62(4):281-9.
    6. Narcotic equivalence converter. [cited 2018 May 9]. Available from http://www.
    7. Auten J, Ishimine P. How to design a study that everyone will believe: retrospective reviews. In: Wilson M, Guluma K, Hayden S, editors. Doing research in emergency and acute care: making order out of chaos. Oxford, England: Wiley-Blackwell; 2015. p. 85-92.
    8. Weinstein S, Obuchowski NA, Lieber ML. Clinical evaluation of Diagnostic tests. AJR Am J Roentgenol 2005;184(1):14-9.
    9. Chou R, Fanciullo GJ, Fine PG, Miaskowski C, Passik SD, Portenoy RK. Opioids for chronic noncancer pain: prediction and identification of aberrant drug-related be- haviors: a review of the evidence for an American Pain Society and American Acad- emy of pain medicine clinical practice guideline. J Pain 2009;10(2):131-46.
    10. Hawk K, D’Onofrio G, Fiellin DA, Chawarski MC, O’Connor PG, Owens PH, et al. Past- year prescription drug monitoring program opioid prescriptions and self-reported opioid use in an emergency department population with opioid use disorder. Acad Emerg Med 2018;25(5):508-16.
    11. McAllister MW, Aaronson P, Spillane J, Schreiber M, Baroso G, Kraemer D, et al. Im- pact of prescription drug-monitoring program on controlled substance Prescribing in the ED. Am J Emerg Med 2015;33(6):781-5.
    12. Katz N, Panas L, Kim M, Audet AD, Bilansky A, Eadie J, et al. Usefulness of prescription monitoring programs for surveillance–analysis of Schedule II opioid prescription data in Massachusetts, 1996-2006. Pharmacoepidemiol Drug Saf 2010;19(2): 115-23.
    13. Faryar KA, Freeman CL, Persaud AK, Furmanek SP, Guinn BE, Mattingly WA, et al. The effects of Kentucky’s comprehensive opioid legislation on patients presenting with prescription opioid or heroin abuse to one urban emergency department. J Emerg Med 2017;53(6):805-14.
    14. Rudd RA, Seth P, David F, Scholl L. Increases in drug and opioid-involved overdose deaths – United States, 2010-2015. MMWR Morb Mortal Wkly Rep 2016;65 (50-51):1445-52.
    15. Cepeda MS, Fife D, Chow W, Mastrogiovanni G, Henderson SC. Opioid shopping be- havior: how often, how soon, which drugs, and what payment method. J Clin Pharmacol 2013;53(1):112-7.

    Leave a Reply

    Your email address will not be published. Required fields are marked *