a b s t r a c t

Patients with drug-seeking behavior can be both labor and resource intensive to the emergency department (ED). Objective: To determine the effectiveness of ED care plans for individuals at high risk for drug-seeking behavior on ED visits.

Methods: A retrospective, cohort observational study.

Location: A suburban teaching hospital with an annual census of 80,000 patients. The number of ED visits was determined 1 year before and 2 subsequent years following care plan initiation.

Exclusion criteria: Unclaimed letter, incomplete data, and/or non-drug-seeking care plan.

Statistics: Two-tailed Wilcoxon signed-rank test with significance of P b .05.

Results: Sixty patients were enrolled and 7 were excluded, leaving 53 patients for analysis. Mean annual visits before care plan initiation were 7.6 (95% confidence interval [CI], 6.3-9.1). One year following imple- mentation, mean visits decreased to 2.3 (95% CI, 1.5-3.1) (P <= .0001). Two years following implementation, mean visits continued to decline to 1.5 (95% CI, 0.9-2.1) (P <= .0001). A significant reduction in visits occurred 1 and 2 years following care plan implementation.

Conclusions: Emergency department care plans are an effective method to reduce ED visits in those with drug-seeking behavior.

(C) 2015

Introduction

The national epidemic prescription-drug abuse problem affects emer- gency department (ED) physicians on a daily basis. During a 13-year period, the National Hospital Ambulatory Care Survey estimated that 44 million pain-related visits were made annually to US EDs, representing 42.6% of all visits. In 36 million of these visits (81%), the first classification code listed was pain related [1]. Between 1997 and 2007, prescriptions for opioids have increased by 700% in the United States [2]. The Drug Abuse Warning Network showed that the number of ED visits for nonmedical use of opioid analgesics increased 111% during 2004 to 2008 and in- creased 29% during 2007 to 2008 [3]. In 2010, the National Poison Data System reported more than 107,000 exposures to opioid analgesics leading to more than 27,500 hospital admissions [4]. The abuse and over- use of opioid analgesics are important national problems. Improved processes to curtail this increasing trend need to be established.

? Previous presentation: Society for Academic Emergency Medicine Scientific Assembly, Dallas 2014.

* Corresponding author at: Department of Emergency Medicine (Box 8), Morristown Medical Center, 100 Madison Ave, Morristown, NJ 07962-1965. Tel.: +1 973 971 8919;

fax: +1 973 290 7202.

E-mail address: [email protected] (F. Fiesseler).

Appropriate Opioid utilization issues have existed for many decades, and differing views have fueled contentious debates. The Harrison Nar- cotics Tax Act in 1914 regulated the distribution of opioids and forbade treatment of those with addictions. In 1973, Marks and Sachar demon- strated the severity of oligoanalgesia, suggesting that physician “mis- conceptions” led to undertreatment with narcotic analgesics and caused patients “needless suffering” [5]. Congress later passed the Pain Relief Act, intended to remove the threat of inappropriate legal liability and disciplinary action against health care professionals who prescribe pain medication to those with chronic illness. The 2012 American Col- lege of Emergency Physicians clinical policy states that ED physicians should avoid the routine prescribing of outpatient opioids for a patient with an acute exacerbation of chronic noncancer pain [6]. Analgesic underuse/overuse in those with chronic disease remains a heated debate with a rational median somewhere between the extremes.

A significant danger lies in increasing access for substance abusers, which may result in harm to themselves or others while under the influence of a controlled substance. Not prescribing opioids to those with drug- seeking behavior is nearly as important as curtailing abusers compulsive and repetitive attempts. This study looks at a subset of patients who frequent the ED seeking pain medications for chronic painful conditions.

It is our hypothesis that an effective ED-based care plan developed by the ED in conjunction with the patient’s primary care physician

http://dx.doi.org/10.1016/j.ajem.2015.08.038

0735-6757/(C) 2015

1800 F. Fiesseler et al. / American Journal of Emergency Medicine 33 (2015) 1799–1801

(PCP) would reduce visits to the ED. Decreasing use and misuse of ED services for chronic pain or drug-seeking behavior has the potential to improve patient care and help control societal costs.

Methods

This study was a retrospective, cohort observational study in the ED of a suburban teaching hospital with an annual census of 80,000 patients. A care plan was initiated when a patient was identified as concerning for drug-seeking behavior by an ED staff member. Any treating staff member could initiate this process. Although specific criteria were not utilized, physicians generally consider multiple previ- ous ED visits where patients would be administered opioids with a neg- ative workup, requesting opioids specifically, allergies to NSAIDs, or stating that NSAIDs did not work, as markers. Before referral, medical records were reviewed by a nurse administrator in conjunction with the ED director for possible referral to the PCP. If concerns existed, an ED administrator then would contact the patient’s PCP. If the PCP con- curred, an ED care plan was formulated. All plans encouraged the pa- tient to visit the ED for any new or concerning symptoms. The patient was informed that upon ED arrival, he or she would be screened for any new disease condition. If no new aliment existed, the care plan would specifically limit or forbid opioid administration. All patients re- ceived a certified letter regarding their care plan and were encouraged to contact the ED with any concerns. Copies of the certified letter and care plan were centrally located in the ED, and medical records were

marked at triage that a care plan was applicable to this patient.

Sixty patients were randomly selected for review. All personal iden- tifying data were blinded to study personnel before data collection. The number of ED visits was determined 1 year before and 2 subsequent years following implementation of each enrolled patient. The initiation date was based on the date of the patient’s certified letter. Care plans were subdivided into those that allowed limited opioid usage vs those that recommend against use. Data were extracted using an electronic ED patient charting system and imported onto a Microsoft Excel data sheet for analysis. Exclusion criteria included an unclaimed registered let- ter, incomplete data, or non-drug-seeking care plan. Statistical analysis used two-tailed Wilcoxon signed-rank test with significance of P b .05. Sta- tistical analysis was performed using Analyze-it version 2.24 Excel 12+ (Leeds, UK) and Statgraphics Centurion XVI version 16.1.11 (Warrenton, VA). This study was approved by our institutional review board.

Results

Sixty patients were enrolled. Excluded subjects included the following: incomplete data (n = 2) (1 preceded our charting system, 1 care plan was initiated b 2 years ago), did not claim their registered letter (n = 2), and non-drug-seeking care plan (n = 3). The remaining 53 patients’ medical records were analyzed. Mean age was 38 years (interquartile range, 27-46 years). Male sex comprised 54% of participants (Table 1).

Mean annual ED visits before care plan initiation for each patient was 7.6 (95% confidence interval [CI], 6.3-9.1). Women averaged 9.5

(95% CI, 7.4-11.7) ED visits; and men, 5.9 (95% CI, 4.6-7.2) (P = .009)

visits. One year following implementation, there was an overall decline to 2.3 (95% CI, 1.5-3.1; P b .0001) ED patient visits per year. Women averaged 3.0 (95% CI, 1.6-4.4) and men 2.1 (95% CI, 1.5-3.9; P = .21)

visits 1 year after care plan initiation. Two years following care plan

Table 1

Sex difference in ED visits

	Male	Female	P value
Mean age at initiation, y	37	41
Mean visits 1 y before care plan	5.9	9.5	.009
Mean visits 1 y following care plan	2.1	3	.21
Mean visits 2 y following care plan	1.2	1.9	.13

implementation, the mean annual ED visit rate was 1.5 (95% CI, 0.9-2.1; P b .0001) overall. Again, women had a higher ED visit rate at 2 years post care plan initiation: 1.9 (95% CI, 1.0-2.7) vs 1.2 (95% CI, 0.4-2.1; P = .13) for men. The reduction in visits that occurred beyond the first year following care plan implementation, 2.3 annual ED visits compared with the second year rate of 1.5 (P = .02), was statistically significant.

Ninety-one percent (n = 48) of care plans recommended no opioid ad-

ministration, whereas the remainder allowed a specific allotment. Compar- ing these 2 types of plans, the annual ED visit rates at 2 years decreased to

1.5 (95% CI, 0-5.4) and 2.4 (95% CI, 0-29), respectively (P = .52). No patients required subsequent admission on succeeding visits.

Discussion

This study demonstrates that ED-initiated care plans aimed at chronic pain patients suspected of drug-seeking behavior reduce recurrent visits. Patients with drug-seeking behavior can be time consuming, disruptive, and burdensome to a busy ED. A number of previous studies have attempted to help address this problem with varied results. Our tech- nique is inexpensive, is reproducible, and requires little technical involve- ment. Few prior studies have been able to demonstrate significant declines in ED revisit rates. Careful planning in association with a detailed medical plan in association witha patient’s primary care doctor, as in our study, may help reduce health care cost and improve patient care.

Care plan utilization in the ED is not an uncommon practice [7]. Pope et al [8] described a Case management program for frequent visitors to an inner-city ED. The 24 enrolled patients had a 26.5 median annual ED visit rate. After the implementation of the ED care plan, “super-uti- lizers” decreased to a median of 6.5 visits per year. Although this study supports the use of ED care plans to manage patients that are very frequent users of the ED, their study did not focus on patients iden- tified as “drug-seeking.” Only 5 (21%) in their study population were identified as such. In addition, this study used a team of hospital/social workers to devise the care plan without discernable input from PCPs [8]. Pillow et al [9] also attempted to reduce visits in “Frequent ED users” using an electronic database care plan system. They looked at the top 50 ED patient visitors. Ultimately, they could only demonstrate a trend towards decreasing monthly ED visits, without an effect on the rate of admissions. Their study population consisted of 22% with a history of substance abuse [9]. Similar to the study of Pope et al, this study did

not focus on the “drug-seeking” ED population [8].

Another method used to decrease drug-seeking behavior is one that most states have implemented: a central monitoring system for pre- scriptions of controlled substance. It is not clear if these state-run pre- scription monitoring systems have led to decreased prescription drug use or decreased ED utilization. However, these programs appear to be excellent resources to assist in the acute treatment of a suspected drug seeker and may serve as a drug-seeking behavior Risk assessment tool for ED patients. Weiner et al [10] demonstrated that the prescrip- tion drug monitoring program in Massachusetts affected prescribing in only 9.5% of patients. They also concluded that emergency providers had only fair agreement with objective criteria from the Prescription Drug Monitoring Program in suspecting drug-seeking behavior. Baehren et al [11] demonstrated that prescription monitoring programs changed clinical management in 41% (n = 74) of cases. The majority (61%; n = 45) resulted in fewer or no opioid medications prescriptions, whereas 39% (n = 29) resulted in more opioid medication than previously planned. Prescription monitoring programs hold promise in assisting ED physicians in managing and identifying potential drug-seeking pa- tients. However, they do not appear to be 100% effective and likely will not be the only resource needed for this patient population. Incor- porated into a departmental plan, this is one tool that will help to iden- tify and document those patients who exhibit drug-seeking behavior.

In another before-and-after study, Pradel et al [12] monitored pre-

scribing trends for buprenorphine, comparing prescribing pattern in more than 2600 patients using an electronic prescription monitoring

F. Fiesseler et al. / American Journal of Emergency Medicine 33 (2015) 1799–1801 1801

database. After implementation, doctor shopping decreased rapidly, suggesting an immediate benefit of this program. The prescribed quan- tity did not change after implementation. Despite prescription drug monitoring programs providing an intuitive perception of benefit for the medical community, there are only limited data to indicate any be- nefit of these programs for improving patient outcomes or reducing the misuse of prescription drugs [13].

Washington state recently implemented a 7-step program for Me- dicaid recipients which includes ED information exchange, patient education, disseminating lists of frequent users, care plans, enrollment in a state Prescription Monitoring Program, strict guidelines for pre- scribing narcotics, and regular feedback reports. This collaborative effort between legislature and EDs is reducing Medicaid spending [14]. Initial results are promising in this multifactorial approach, although we do not know which elements are most contributory.

The American College of Emergency Physicians recently published a clinical policy for treating noncancer pain in those who are at risk for Opioid abuse. They reported that state prescription monitoring pro- grams may help (level C) identify patients who are at high risk for pre- scription opioid diversion or doctor shopping. The policy recommends that for patients with an acute exacerbation of noncancer chronic pain, (1) physicians should avoid the routine prescribing of outpatient opioids; (2) if opioids are prescribed on discharge, the prescription should be for the lowest practical dose for a limited duration; (3) the prescriber should consider the patient’s risk for Opioid misuse, abuse, or diversion; and (4) the clinician should, if practicable, honor existing patient-physician pain contracts/treatment agreements [6].

In 2003, Geiderman [15] discussed the ethical, legal, and regulatory considerations surrounding the use of “habitual patient files.” The article acknowledged common and informal use of such files and pro- moted the development of formal standards for their use. We believe that our study highlights the potential use of an ED-based care plan that is transparent and involves the patient and PCP.

The majority of physicians report marked ambivalence concerning controlled drug prescribing: their desire to relieve pain and distress vs the fear of creating addiction and being investigated by law enforce- ment [16]. Some emergency physicians procure pay increases based on patient satisfaction and timely pain control [17]. The complexity and competing interests of clinical practice often do not lead to straight- forward answers. Our care plans demonstrate an economical way for ED physicians to curtail visits in those with chronic pain syndromes. Relieving pain and suffering is a fundamental responsibility of emergency medicine, and we have a concurrent duty to limit the personal and societal harm that can result from prescription drug abuse.

Limitations and future questions

The first limitation of our study was our small enrollment size. It was determined with post hoc analysis that our data were robust enough and would not benefit from additional enrollees. Secondly, collaboration be- tween the ED and PCPs may be difficult, often with many patients not having primary care available. Although this study design limits enrollees to those with a PCP, other types of physician follow-up may be used. The main advantage to using a PCP is the established relationship with the patient and the likelihood for higher rates of post-ED care follow-up.

Another concern regarding our success in decreasing ED visits is that patients are merely shifted to another ED. Reviewing state databases could help resolve this contention that these patients are simply “relocated.” Although New Jersey has such a database and has been ex- tremely helpful in our ED, we were unable to obtain approval for use of this database for research purposes. One prior study suggests that patients who were told that they “would receive no further narcotics” at a given fa- cility subsequently received controlled substances from another hospital in 93% of cases and from the same facility in 71% of cases [18]. This issue along with possibility of shared ED care plans needs to be addressed in fu- ture research.

Defining who is at high risk for drug-seeking behavior also appears to be difficult. risk assessment tools for opioid misuse exist but do not ap- pear to work with high levels of sensitivity and specificity. One particular limitation of these tools is that they have never been compared with the prescription drug-monitoring program in regard to their prognostic accu- racy [19,20]. In fact, 22 of 49 states with prescription drug monitoring pro- grams mandatED prescribers to query the system before writing for controlled substances with recognized potential for abuse in an attempt to eliminate preconceptions [21]. To address these problems appropriate- ly, physicians need adequate education in substance abuse. A survey by the National Center of Addiction and Substance Abuse reveals that physi- cians do not feel that they are well versed to identify signs of substance abuse–a skill that should be taught in medical school and residency pro- grams [22]. Obviously, our patients were only triggered as being high risk not by a formal assessment tool but by the gestalt of an ED staff member. Further education and research on this topic are warranted.

Conclusions

Emergency department care plans are an effective way to reduce ED visits in those identified as having drug-seeking behavior. These effects appear to be sustained, lasting up to 2 years.

References

1. Pletcher MJ, Kertesz SG, Kohn MA, Gonzales R. Trends in opioid prescribing by race/ ethnicity for patients seeking care in US emergency departments. JAMA 2008; 299(1):70-8.
2. Department of Justice Drug Enforcement Administration. Automation of reports and con- solidation orders system (ARCOS). http://www.deadversion.usdoj.gov/arcos/index.html.
3. Emergency department visits involving nonmedical use of selected prescription drugs–United States 2004-2008. Morb Mortal Wkly Rep 2010;59(23):705-9.
4. American Association of Poison control centers. National Poison Data System 2010 annual report. http://www.aapcc.org.
5. Marks RM, Sachar EJ. Undertreatment of medical inpatients with narcotic analgesics. Ann Intern Med 1973;78(2):173-81.
6. Clinical policy: critical issues in the prescribing of opioids for adult patients in the emergency department. Ann Emerg Med 2012;60(4):499-525.
7. Graber MA, Gjerde C, Bergus G, Ely J. The unofficial “problem patient” files and inter- institutional information transfer in emergency medicine in Iowa. Am J Emerg Med 1995;13:509-11.
8. Pope D, Fernandes CM, Bouthillette F, Etherington J. Frequent users of the emergen- cy department: a program to improve care and reduce visits. Can Med Assoc J 2000; 162(7):1017-20.
9. Pillow MT, Doctor S, Brown S, Carter K, Mulliken R. An emergency department- initiated, web-based, multidisciplinary approach to decreasing emergency depart- ment visits by the top frequent visitors using patient care plans. J Emerg Med 2013;44(4):853-60.
10. 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.
11. Baehren DF, Marco CA, Droz DE, Sinha S, Callan EM, Akpunonu P. A statewide pre- scription monitoring program affects emergency department prescribing behaviors. Ann Emerg Med 2010;56(1):19-23.e1-3.
12. Pradel V, Frauger E, Thirion X, Ronfle E, Lapierre V, Masut A, et al. Impact of a pre- scription monitoring program on doctor-shopping for high dose buprenorphine. Pharmacoepidemiol Drug Saf 2009;18(1):36-43.
13. Paulozzi LJ, Kilbourne EM, Desai HA. Prescription drug monitoring programs and Death rates from drug overdose. Pain Med 2011;12(5):747-54.
14. Collaborative effort in Washington state slashes non-essential use of the ED by Medicaid patients, delivering millions in projected savings. ED Manag 2013;25(4):41-4.
15. Geiderman JM. Keeping lists and naming names: habitual patient files for suspected nontherapeutic drug-seeking patients. Ann Emerg Med 2003;41(6):873-81.
16. Parran Jr T. Prescription drug abuse. A question of balance. Med Clin N Am 1997; 81(4):967-78.
17. Poon SJ, Greenwood-Erickson MB. The Opioid prescription epidemic and the role of emergency medicine. Ann Emerg Med 2014;64(5):490-5.
18. Zechnich AD, Hedges JR. Community-wide emergency department visits by patients suspected of drug-seeking behavior. Acad Emerg Med 1996;3(4):312-7.
19. Hoppe J, Perrone J, Nelson LS. Being judge and jury: a new skill for emergency physicians. Ann Emerg Med 2013;62(4):290-2.
20. McNabb C, Foot C, Ting J, Breeze K, Stickley M. Diagnosing drug-seeking behaviour in adult emergency department. Emerg Med Australas 2006;18(2):138-42.
21. Haffajee R, Jena A, Weiner S. Mandatory use of prescription drug monitoring programs. J Am Med Assoc 2015;313(9):891-2.
22. Friedman RA. The changing face of teenage drug abuse–the trend toward prescription drugs. N Engl J Med 2006;354(14):1448-50.