Article, Emergency Medicine

Effect of an emergency department opioid prescription policy on prescribing patterns

a b s t r a c t

Background: Staten Island University Hospital is located in NYC, where the opioid epidemic has resulted in signif- icant mortalities from unintentional overdoses. In 2013 as a response to the rising threat to our community, our Emergency Department (ED) administration adopted a clinical practice policy focused on decreasing the pre- scription of controlled substances. The effects of this policy on our provider prescription patterns are presented here.

Methods: A retrospective chart review of patients prescribed opioids from the ED before and after policy imple- mentation was performed. Dates chosen for analysis was November 1, 2012 through January 31, 2013 and No- vember 1, 2013 through January 31, 2014; these time periods were used to serve as a seasonally comparative group pre and post clinical practice policy implementation. Opioids written for the treatment of cough, and for children under eighteen were excluded from analysis. Patient age, sex, diagnoses, and prescription formulation, strength, and pill number was recorded for each patient receiving an Opioid prescription.

Results: There was a drop in the total prescriptions from 1756 to 1128 without a change in the average number of pills (12.78 vs 12.44) or average total dose prescribed (69.39 vs 68.98) mg of Morphine equivalent per prescrip- tion. Additionally, there were sizable reductions in Opioid prescriptions written for arthralgias/myalgias, Dental pain, soft tissue injuries, and headaches.

Conclusion: The opioid clinical policy had a clear effect in decreasing the number of patients prescribed opioids. Such policies may be the key to reducing the epidemic and saving lives from unintentional opioid overdoses.

(C) 2017


The opioid epidemic in the United States has been pervasive for many years. Our institution, located in Staten Island, New York, one of the five boroughs of New York City, faced a dismal reality when we learned that from 2000 to 2011 the rate of deaths due to unintentional overdose of opioid analgesics increased 435%, from 2.0 to 10.7 deaths per 100,000 residents [1]. Furthermore, from 2005 to 2011, disparities widened between Staten Island and the other boroughs (Bronx, Brook- lyn, Manhattan, and Queens) as the rate of opioid overdose deaths in Staten Island was 3.0-4.5 times higher than other boroughs in 2011 [2]. In addition to this population based data, recent studies have shown that opioid prescriptions from the ED are associated with pa- tients becoming long-term opioid users [3,4].

? The study was approved by the Staten Island University Hospital IRB.There was no

financial support or conflict of interest of the investigation or manuscript development.

* Corresponding author at: Department of Emergency Medicine, Staten Island University Hospital, 475 Seaview Avenue, Staten Island, NY 10305, United States.

E-mail address: [email protected] (J. Chacko).

The emergency department (ED) at Staten Island University Hospital (SIUH) is one of two on Staten Island, serving a population of approxi- mate 470,000. The ED administration made it their obligation to address the issue of opioid use/misuse for the betterment of the patients and providers of this community. The response was the creation of a clinical practice policy (Appendix) focused on decreasing the prescription of controlled substances. The concept was proposed to the ED staff in 2010 and adopted into a formal policy in February 2013.

Specifically, the policy prohibits renewing controlled substances for chronic pain conditions, prescribing of long acting substances such as methadone and Suboxone, and requires direct communication with an ED physician administrator to override. The policy does not prohibit the prescribing of controlled substance for acute painful conditions (e.g. fracture, urolithiasis). However, the provider is limited by the policy which dictates that a second prescription can only be given one time and within seven days of the initial ED visit. Of note, New York State (NYS) is one of 17 with an opioid prescribing guideline, focused on lim- iting opioid pain prescriptions [5]. Our policy supplemented the NYS mandates, placing additional limitations on the prescribing of controlled substances. What is described here are the prescribing patterns of opi- oid analgesics before and after the inception of our policy.

0735-6757/(C) 2017

1328 J. Chacko et al. / American Journal of Emergency Medicine 35 (2017) 13271329


This is a retrospective, descriptive study of the opioid prescribing patterns from the SIUH ED before and after the implementation of the clinical practice policy. Prior approval was granted by the Institutional Review Board for Northwell Health. Data was obtained from the Emer- gency Department Information Systems (EDIS, Allscripts EDIS V.7.2.1, Raleigh, North Carolina USA) which is used as the electronic health re- cord system for the SIUH ED. Patient age, sex, primary and secondary di- agnoses, as well as prescription drug name, dose, and quantity of pills were extracted from the EDIS from November 1, 2012 through January 31, 2013 and November 1, 2013 through January 31, 2014. These time periods were used to serve as a seasonally comparative group. Prescrip- tions written for patients under 18 years of age and for upper respirato- ry tract symptoms were excluded. These formulations included but were not limited to Tylenol #3, Hycodan (hydrocodone/homatropin), Tussionex (hydrocodone/chlorpheniramine), Phernergan Elixir with Codeine, and Promethazine/Codeine syrup. Using the opioid conversion table as described in Goldfrank’s Toxicological Emergencies [6], all pre- scriptions were converted to equivalent milligrams of morphine. Analy- sis was conducted to evaluate the change in number of opioid prescriptions written, total pills ordered, dosage of medication written, type of opioid formulation, and indications as per discharge diagnoses. Statistical calculations were performed using Microsoft Excel Profes- sional 2010 (Microsoft Corporation, Redmond, WA).


1804 prescriptions were written in the 2012-2013 time period and 1155 prescriptions were written in the 2013-2014 time period. A total of 59 prescriptions were excluded from the pre-policy group and 28 prescriptions from the post-policy group. 23 of those prescriptions from the pre-policy group and 8 prescriptions from the post-policy group were excluded due to prescriptions written for patients under 18 years of age. The other exclusions were due to cough related opioid prescriptions. Ultimately, 1745 prescriptions from the 2012-2013 time period and 1127 prescriptions from the 2013-2014 time period were included in the final analysis. No significant difference in was found in the sex of the patients between the two study populations (46.0% male vs 46.4% male). The average age in the pre- and post-clini- cal policy was 45.9 vs 47.3 years, respectively (Table 1).

As noted above, during the three months prior to the implementa- tion of the clinical policy 1745 prescriptions were written for opioids in our study group. In the same three month period of the following cal- endar year 1127 prescriptions were written. There was a similar decline in the number of pills and adjusted milligrams of morphine prescribed (Fig. 1). Prescribing patterns per prescription were also evaluated. Quantity of pills given, milligrams of morphine prescribed and milli- grams of morphine per prescription were compared in both groups. Av- erage pills per prescription (12.78 vs 12.44, p = 0.095) and average

Table 1

Demographics of pre-policy and post-policy populations.



Prescriptions (patients)




1130 (64.8%)

961 (85.3%)


529 (30.3%)

84 (7.5%)


78 (4.5%)

78 (6.9%)


5 (0.3%)

4 (0.4%)


3 (0.2%)


Males (p = 0.838)






Age (p = 0.023)

45.9 +- 17.0

47.3 +- 16.3




Milligrams of morphine



Fig. 1. Opioid prescriptions in number of pills and adjusted milligrams of morphine.

milligrams of morphine per prescription (69.39 vs 68.98, p = 0.801) were not statistically different pre- and post-policy.

The opioid formulations were analyzed and the greatest decline in prescriptions was found for hydrocodone, with a reduction of 84%. Diag- noses were also analyzed with considerable decreases in prescriptions for arthralgias/myalgias, dental pain, soft tissue injuries (contusions, cellulitis), and headaches. Some smaller diagnosis groups also saw a re- duction in post-policy prescriptions, including abdominal pain, chest pain, corneal abrasions, and medication refills. Final diagnoses that did not point toward any painful pathology, such as allergic reaction or con- junctivitis, were lumped together in the “Other” category which saw a 47.8% decline post-policy implementation (Fig. 2).


We saw a significant change in the number of pills and total adjusted milligrams of morphine prescribed after the implementation of the opi- oid clinical policy. When corrected for the number of prescriptions writ- ten, there was no significant change in number of pills (12.8 +- 4.9 vs 12.4 +- 5.5, p = 0.095) or milligrams or morphine (69.4 +- 42.7 vs 69.0 +- 41.7, p = 0.801) prescribed. Therefore, the difference is solely due to the number of prescriptions written. When looking at the distri- bution of opioid formulations, there was a major drop in hydrocodone prescriptions. One contributing factor was the rescheduling of hydrocodone from C-III to C-II.

Prescribing patterns for painful pathologies classically treated with opioids such as fractures (-8.9%), Neuropathic pain (-4.3%), and uro- lithiasis (-18.8%) were minimally reduced. In contrast, there were con- siderable reductions in opioid prescriptions for more vague diagnoses such as abdominal pain (- 34.7%), arthalgia/myalgia (- 46.8%), and chest pain (-60.9%). There were also reductions in painful pathologies that can be managed without opioids, such as burn (-56.7%), corneal abrasion (- 75.0%), and dental pain (- 51.1%). Other diagnoses had smaller reductions such as abscess (- 26.2%), back pain (- 24.7%), and UTI (-23.1%).

One subtle finding is the variance in average age; while this differ- ence is statistically significant, we do not believe it is clinically signifi- cant. Although we do not believe that overall patient age played a role in provider prescribing patterns, there was a drop in the pediatric exclu- sions from 23 to 8, which may reflect some judicious thinking in the younger, more vulnerable population.


This study has a several limitations. First, as a retrospective study we are able only to report data documented in the ED chart. Next, there was a short time period used for analysis. However, the sole use of electronic medical records in opioid prescribing guaranteed a complete capture of all prescriptions written in the study periods. While there was not a di- rect analysis of the total ED census during the study periods, it is unlike- ly but possible less patients presented with painful complaints to the ED

J. Chacko et al. / American Journal of Emergency Medicine 35 (2017) 13271329 1329

Fig. 2. Effects of policy on diagnoses with special attention to top ten pre-policy diagnoses.

in the post-implementation period. Lastly, the external application of this study may be limited as social media, local new agencies, and news- papers at the time of the study period provided increased public aware- ness of the opioid epidemic possibly influencing providers practicing patterns and patients’ willingness to be prescribed opioid medication.


The opioid clinical policy was correlated with a decrease in the num- ber of patients prescribed opioids, with the opioid formulation hydrocodone having the greatest reduction in prescriptions prescribed. However, no difference in the number of pills and adjusted milligrams of morphine prescribed per prescription pre- and post-policy imple- mentation was found. Additionally, there was a considerable reduction in opioid prescriptions for many diagnoses which were previously com- monly prescribed opioids.

Appendix A. Staten Island University hospital emergency depart- ment clinical policy — controlled substance ordering and prescrib- ing in the ED

    • Methadone and Suboxone (Buprenorphine/naloxone) cannot be or- dered or prescribed by an ED physician or a PA.
    • Benzodiazepine prescriptions cannot be written by an ED physician or a PA.
    • Controlled substance Prescriptions cannot be renewed by an ED phy- sician. Specifically, if a patient’s chronic pain is managed by a Primary Care Provider (PCP) or a pain management physician, and a patient “runs out of his medication” and his physician is “on vacation,” an ED physician is not permitted to renew his Medication prescription.
    • Controlled substance pain prescriptions may be “extended” if a pa- tient was seen in the ED for an acute injury and has exhausted the sup- ply of medication originally prescribed to him. This may only be performed if the patient returns to the ED within 7 days and may be performed one time only.
    • This guideline does not prohibit an ED physician from ordering one time doses of pain medication or benzodiazepines while the patient is still in the ED.
    • This guideline applies to controlled stimulants such as those used for ADHD.
    • If an ED attending physician or a PA has a valid medical reason to over- ride this guideline, the ED physician administrator on-call must be contacted for an approval.


  1. Phone D, Tuazon E, Kattan J, et al. Decrease in rate of opioid analgesic overdose deaths

    — Staten Island, New York City, 2011-2013. MMWR Morb Mortal Wkly Rep 2015;64: 491-4.

    Unintentional Opioid Analgesic Poisoning (Overdose) Deaths in New York City, 2011. York City Department of Health and Mental Hygiene; May 2013Web. 20 Oct. 2016 databrief27.pdfN.

  2. Barnett ML, Olenski AR, Jena AB. Opioid-prescribing patterns of emergency physicians and risk of long-term use. NEJM 2017;376:663-73.
  3. Hoppe JA, Kim H, Heard K. Association of emergency department opioid initiation with recurrent opioid use. Ann Emerg Med 2015;65(5):493-9.
  4. Broida RI, Gronowski T, Kalnow AF, et al. State emergency department opioid guide- lines current status. West J Emerg Med 2017;18(3):340-4.
  5. Nelson Lewis S, Olsen Dean. “Opioids” Goldfrank’s Toxicologic Emergencies. 9th ed. New York: McGraw-Hill; 2015 568Print.

Leave a Reply

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