TOXICOLOGY/ORIGINAL RESEARCH
Preventing Iatrogenic Overdose: A Review of In–Emergency Department Opioid-Related Adverse Drug Events and Medication Errors Francesca L. Beaudoin, MD, MS*; Roland C. Merchant, MD, MPH; Adam Janicki, MD; Donald M. McKaig, RPh; Kavita M. Babu, MD *Corresponding Author. E-mail: [email protected].
Study objective: We describe characteristics of patients with in–emergency department (ED) opioid-related adverse drug events, medication errors, and harm resulting from medication errors; identify patient-, provider-, and system-based factors associated with in-ED opioid-related medication errors and harm; and create a list of strategies to prevent future events. Methods: This retrospective study was conducted at 2 urban academic EDs. Potential iatrogenic opioid overdoses were identified by querying the ED electronic medical record for cases when naloxone was administered after an opioid was administered in the ED. Cases involving medication errors resulting in harm were reviewed qualitatively for common patient-, provider-, and systems-based factors that might have contributed to the event. Results: Of 73 ED patients with in-ED opioid-related adverse events that required reversal with naloxone, 43 had a medication error resulting in harm. Patient-, provider-, and systems-based factors that might have contributed to the events included chronic health conditions that could predispose an individual to an opioid-related adverse event, failure to adjust opioid dosing in the elderly and for hepatic or renal impairment, multiple doses and routes of administration of opioids, coadministration of opioids with other sedating medications, and systems-based problems with patient handoffs and pharmacy oversight. Conclusion: We identified patient-, provider-, and systems-based factors related to opioid-related adverse drug events and medication errors among ED patients who had received naloxone. The results from our assessment can be used to inform educational and policy initiatives aimed to prevent in-ED opioid-related adverse drug events and medication errors. [Ann Emerg Med. 2015;65:423-431.] Please see page 424 for the Editor’s Capsule Summary of this article. A podcast for this article is available at www.annemergmed.com. Continuing Medical Education exam for this article is available at http://www.acep.org/ACEPeCME/. 0196-0644/$-see front matter Copyright © 2014 by the American College of Emergency Physicians. http://dx.doi.org/10.1016/j.annemergmed.2014.11.016
INTRODUCTION Background Opioid analgesics are commonly implicated in adverse drug events.1 Although some opioid-related adverse drug events might be unavoidable, many are due to preventable medication errors. Iatrogenic inhospital opioid-related adverse drug events manifest as a range of adverse events from somnolence to potentially lifethreatening central nervous system and respiratory depression. Furthermore, opioids prescribed for acute pain in the hospital setting have been highlighted as a cause of unexpected inhospital deaths2 and are associated with longer hospital lengths of stay, higher health care costs, and greater likelihood of 30-day readmission to the hospital.3-6 The emergency department (ED) is an important environment in which to examine opioid-related adverse drug events because emergency care providers frequently administer opioids for the treatment of pain7; however, opioid-related adverse drug events and medication errors have been understudied in the ED. Volume 65, no. 4 : April 2015
Many hospitals derive estimates of adverse drug events from their voluntary incident reporting systems, which capture less than 15% of adverse drug events identified by trigger tools.8,9 Trigger tools, such as a medication order or laboratory value, can indicate the occurrence of a potential adverse drug event and prompt an external review of the patient’s care.10 The administration of naloxone, an antidote indicated for the rapid reversal of opioid-induced coma or apnea, has been described as a marker of opioid-related adverse drug events and has feasibility as a trigger tool for the review of opioidrelated adverse drug events and medication errors.11,12 Importance Although naloxone administration appears to be a feasible means of identifying potential cases of iatrogenic opioid-related adverse drug events in the ED, the identification of these adverse drug events is only the first step in addressing this problem. Patient-, provider-, and systems-level risk factors for opioid-related Annals of Emergency Medicine 423
Iatrogenic Opioid Overdose
Editor’s Capsule Summary
What is already known on this topic Opioid-related iatrogenic adverse drug events are generally predictable and avoidable. What question this study addressed What are the patient-, provider-, and systems-based factors that contribute to iatrogenic opioid adverse events in the emergency department (ED)? What this study adds to our knowledge Using the ED administration of naloxone as a marker for potential cases, 73 opioid overdose patients were identified. Iatrogenic ED events were associated with poor patient health, failure to adjust opioid dose for age or disease, repeated therapeutic dosing, coadministration with sedating medication, poor handoffs, and limited pharmacy oversight. How this is relevant to clinical practice Understanding the reasons for iatrogenic opioid adverse events can lead to improved safe medication practices.
adverse drug events also must be elucidated to develop interventions to prevent this avoidable cause of morbidity and mortality. Goals of This Investigation Our objectives were to describe characteristics of patients with in-ED opioid-related adverse drug events, medication errors, and harm resulting from medication errors, identify patient-, provider-, and system-based factors associated with in-ED opioid-related medication errors and harm, and create a list of strategies to prevent future events. In addition, we compared the number of potentially serious opioid-related adverse drug events identified with naloxone as a case finder (ie, a naloxone trigger) versus the hospitals’ voluntary reporting system.
MATERIALS AND METHODS Study Design and Setting We conducted a retrospective review and qualitative analysis of all visits during which naloxone was administered to patients from October 2009 to December 2012 at 2 large, medical school–affiliated, urban New England EDs. One ED is a Level I trauma center with more than 100,000 adult visits per year and is the tertiary care referral center for the region, whereas the other is a community hospital with more than 55,000 adult visits per year. The 2 EDs are in the same hospital system and city, employ emergency physicians who are on staff at both facilities, and are 424 Annals of Emergency Medicine
Beaudoin et al 2 of the sites for the medical school’s emergency medicine residency program. The hospitals’ institutional review board approved the study protocol. Patients who received naloxone in the ED were identified through a search of the 2 EDs’ electronic medical record system database. Naloxone medication orders were located by querying preset, free-text, and verbal orders. Patients were included in the study if they received any opioid analgesic in the ED before naloxone administration. Patients who received naloxone before the ED visit (eg, by emergency medical service providers) or who did not receive an opioid in the ED were not included. The ED electronic medical records of patients meeting study inclusion criteria were reviewed with guidelines recommended by Gilbert et al.13 A panel of 3 of the study coinvestigators (an emergency physician, an emergency physician with board certification in toxicology, and a clinical pharmacist with expertise in medication safety) independently reviewed each of the ED electronic medical records and classified the cases according to an a priori hierarchy (described below). Before conducting their review, the investigators underwent a 2-hour training session to standardize the review protocol, using sample cases. After completing their independent reviews, the investigators met, discussed their findings, and then reached a consensus on their classification of the cases, conducted rootcauses analyses of the cases to determine whether a medication error and harm resulting from that error occurred, and identified patient, provider, and system factors related to harm from opioidrelated medication errors. They then created a list of possible solutions to prevent further in-ED opioid-related adverse drug events and medication errors. The investigators agreed a priori that an outside arbitrator would mediate if consensus could not be achieved, but there were no cases for which this was required. Cases of possible opioid-related adverse drug events were classified according to the following hierarchy: (1) opioid-related adverse drug events, nonopioid-related adverse drug events, or insufficient documentation to determine whether an adverse drug event occurred or was opioid related; (2) among cases with an opioid-related adverse drug event, further classification if a medication error versus no medication error occurred; and (3) among cases with a medication error, further classification if there was harm resulting versus no harm resulting. In accordance with previous studies on this topic, opioid-related adverse drug events were defined as objective or subjective (based on ED electronic medical record documentation) respiratory distress, sedation, or hypotension that improved after naloxone administration.11 For these analyses, opioid-related nausea or vomiting, urinary retention, constipation, and dysphoria were not considered to be opioid-related adverse drug events. Cases were deemed not to be an opioid-related adverse drug event if the patient did not have a clinical response to naloxone or another cause of the adverse event was ultimately found (eg, intracranial hemorrhage, hypoglycemia). The presence of medication errors among cases with an opioid-related adverse drug event was determined through a rootcause analysis by the review panel.14 The National Coordination Council for Medication Error Reporting and Prevention Index Volume 65, no. 4 : April 2015
Beaudoin et al
Iatrogenic Opioid Overdose
Table 1. NCC-MERP index for categorizing medication errors. Category
Conceptual Definition
A B
G H
Circumstances or events that have the capacity to cause error An error occurred but the error did not reach the patient (an “error of omission” does reach the patient) An error occurred that reached the patient but did not cause patient harm An error occurred that reached the patient and required monitoring to confirm that it resulted in no harm to the patient or required intervention to preclude harm An error occurred that may have contributed to or resulted in temporary harm to the patient and required intervention An error occurred that may have contributed to or resulted in temporary harm to the patient and required initial or prolonged hospitalization An error occurred that may have contributed to or resulted in permanent patient harm An error occurred that required intervention necessary to sustain life
I
An error occurred that may have contributed to or resulted in the patient’s death
C D E F
Example of Operational Definition Not applicable* Not applicable* Not applicable* Cases not satisfying criteria E–I Hypoxia requiring supplemental oxygen Admission to the hospital directly related to opioid overdose Not applicable Bag-valve-mask ventilation or intubation Patient death
*© 2001 National Coordinating Council for Medication Error Reporting and Prevention. All rights reserved. No patients satisfied criteria for categories A through C.
(NCC-MERP) definition of a medication error was used. In accordance with the NCC-MERP, “a medication error is any preventable event that may cause or lead to inappropriate medication use or patient harm while the medication is in the control of the health care professional, patient, or consumer. Such events may be related to professional practice, health care products, procedures, and systems, including prescribing; order communication; product labeling, packaging, and nomenclature; compounding; dispensing; distribution; administration; education; monitoring; and use.”15 Cases with an opioid-related adverse drug event and a medication error were further classified as having harm or no harm resulting, according to NCC-MERP taxonomy (Table 1).
Cases involving medication errors resulting in harm (NCCMERP categories E through H) were further reviewed for common patient, provider, and systems factors that might have contributed to the event. The analysis focused on only cases resulting in harm because the intent of the qualitative analysis was to identify factors related with and strategies to prevent both error and harm. Factors were identified through an open discussion and consensus among the 3 investigators, and the results were recorded during the deliberations. Transcripts of the discussions were not obtained or coded. A factor was included in the analysis if it was considered to be present in 2 or more cases. The investigators considered some factors a priori according to previous work by Neil et al,11 but
Figure. Classification of cases. ADE, Adverse drug event; NCC-MERP, National Coordinating Council for Medication Error Reporting and Prevention. Volume 65, no. 4 : April 2015
Annals of Emergency Medicine 425
Beaudoin et al
Iatrogenic Opioid Overdose Table 2. Patient characteristics by classification.* Opioid-Related Adverse Event
Age, y Sex, % female Medical history Diabetes Hypertension Asthma/COPD/OSA Liver disease Psychiatric comorbidity Substance abuse Cardiac disease Cancer Obesity Renal impairment Home medications Benzodiazepines Opioids Opioids administered Total opioid dose before naloxone (morphine equivalents) Route of opioid administration Intravenous Oral Intramuscular Subcutaneous Multiple routes Type of opioid Hydromorphone Morphine Fentanyl Other Multiple opioids Other sedating medication Benzodiazepines Phenothiazines Antihistamines Procedural sedation Indication for naloxone Altered mental status Bradycardia Hypotension Respiratory depression Hypoxia Other Multiple indications ED visit Disposition Admit Discharge Transfer AMA/elope ED length of stay, min Multiple doses of naloxone required Time from opioid to naloxone, min
426 Annals of Emergency Medicine
Medication Error
All (n[73)
Adverse Drug Event (n[22)
Medication Error (n[51)
Harm Resulted (n[43)
No Harm Resulted (n[8)
58 (49–76) 56.8
61 (51–82) 68.2
56 (41–70) 50.0
54 (40–70) 48.9
61 (56–71) 62.5
16.2 41.9 18.4 9.5 18.9 1.4 9.5 5.4 1.4 61.8
18.2 40.9 18.2 4.6 27.3 0 13.6 4.6 0 57.1
16.7 18.5 18.5 11.1 14.8 1.9 7.8 5.6 1.9 61.2
13.3 11.1 20.0 8.9 17.8 2.2 6.9 4.4 2.2 60.0
25.0 50.0 12.5 25.0 0 0 12.5 12.5 0 75.0
27.0 25.7
31.8 31.8
24.1 24.1
22.2 20.0
37.5 50.0
10.5 (5.8–12.1)
6.7 (4–8)
12.4 (6.7–13.7)
13.1 (6.7–21.9)
8.2 (5.4–13.4)
93.2 35.6 23.3 4.1 50.7 0 61.6 38.4 16.4 5.5 20.5
86.4 31.8 18.2 4.5 40.9
96.1 37.3 25.5 3.9 54.9
95.3 37.2 25.6 4.7 53.5
100.0 37.5 25.0 0 62.5
31.8 59.1 9.1 13.6 13.6
74.5 29.4 19.6 2.0 23.5
74.4 30.2 18.6 0 25.6
75.0 25.0 25.0 12.5 12.5
28.8 11.0 12.3 16.4
9.1 13.6 0 0
37.3 9.8 17.6 23.5
34.9 11.6 14.0 25.6
50.0 0 12.5 12.5
72.3 5.3 25.3 40.4 44.3 2.7 72.6
68.2 9.1 40.9 18.2 18.2 9.1 68.2
74.1 3.7 18.5 50.0 55.6 0 74.5
77.8 2.2 11.1 60.0 62.2 0 81.4
62.5 12.5 50.0 0 25.0 0 37.5
69.9 26.1 2.7 1.4 496 (368–621) 23.2 84 (46–152)
77.3 13.6 4.6 4.6 522 (396–650) 9 60 (38–111)
66.7 31.5 1.9 0 492 (354–585) 29.4 99.5 (57–168)
66.7 31.1 2.2 0 474 (420–511) 28 99 (50–181)
62.5 37.5 0 0 499 (354–640) 0 114 (60–140)
Volume 65, no. 4 : April 2015
Beaudoin et al
Iatrogenic Opioid Overdose
Table 2. Continued. Opioid-Related Adverse Event
Time naloxone to disposition, min Time naloxone to discharge, min
Medication Error
All (n[73)
Adverse Drug Event (n[22)
Medication Error (n[51)
Harm Resulted (n[43)
No Harm Resulted (n[8)
169 (105–267) 182 (76–245)
169 (94–363) 86 (23–568)
176.9 (110–252) 213.7 (84–238)
165 (110–286) 182 (68–238)
195 (121–221) 218 (84–251)
COPD, Chronic obstructive pulmonary disease; OSA, obstructive sleep apnea; AMA, against medical advice; cardiac disease, coronary artery disease, myocardial infarction, or congestive heart failure; liver disease, hepatitis or cirrhosis. *Categorical data are represented as a proportion (%); continuous data, as median (interquartile range).
additional factors were identified during the investigator deliberation of cases. Thematic saturation was not a component of this analysis, given the finite number of cases available for review. Next, the factors were used to create a list of strategies (actions or interventions) intended to mitigate inED opioid-related adverse drug events and medication errors identified during the qualitative analysis. Data Collection and Processing The investigators abstracted the following additional data from the ED electronic medical records for the cases: patient demographic characteristics, home medications, and medical history; medications administered in the ED; and the patient’s ED laboratory values, clinical outcomes, and disposition from the ED. Laboratory values abstracted were used to determine the presence of renal dysfunction (glomerular filtration rate 80 mg/100 mL). These data were recorded with a standardized abstraction instrument with double data entry verification. Summary statistics of patient demographic and clinical characteristics of all opioidrelated adverse drug events were generated and stratified by error and by harm classification. Because of the small sample size, formal statistical comparisons were not performed. We also queried the 2 hospitals’ voluntary error reporting system for all reported opioidrelated adverse drug events and compared the results to the number of cases identified with the naloxone case identification method outlined previously.
approximately half had opioids administered by more than 1 route (oral, intramuscular, etc). Altered mental status was cited as the most common reason for the administration of naloxone. None of the cases with an opioid-related adverse drug event identified through naloxone administration had been recorded in the hospitals’ voluntary reporting systems. In fact, during the entire study period, there were no opioid-related adverse drug events reported from the ED through the voluntary reporting system. During root-cause analysis, more than two thirds of the patients (n¼51) were determined to have had a medication error as the cause of their opioid-related adverse drug event. Of patients experiencing medication errors, most errors (n¼43) resulted in harm to the patient, according to the NCC-MERP classification system. Of the patients experiencing harm from a medication error, nearly three quarters had received hydromorphone, almost two thirds had received multiple doses of opioids, more than half had received opioids through multiple routes of administration, and half received 13 morphine milligram equivalents or more. Forty-three of the identified opioid-related adverse drug events were determined to have both a medication error and harm resulting; the results of qualitative analysis (factors and strategies) for these cases are presented in Table 3. Several factors emerged as potential targets to reduce the occurrence of iatrogenic opioid overdose and related adverse drug events, medication errors, and resultant harm. Factors related to aspects of the provider and prescribing often centered on knowledge deficits, whereas the systems-based issues revolved around resource availability, hospital culture, and policy.
RESULTS
LIMITATIONS
The classification of cases for this study is depicted in the Figure, and the demographic and clinical characteristics of the patients are summarized in Table 2. There were 87 cases during which naloxone was administered after an opioid was administered in the ED, of which 73 were determined to be opioid-related adverse drug events. Among the 73 patients with an opioid-related adverse drug event (Table 2), most were women, most had renal impairment, and approximately one fourth had reported benzodiazepines or opioids among their home medications. Approximately half of the patients received 10 morphine milligram equivalents or more in the ED before receiving naloxone (usually intravenously). Most patients received hydromorphone, and
We identified cases of in-ED opioid-related adverse drug events and medication errors through a search of in-ED naloxone administration. Cases of opioid toxicity that resulted in prolonged observation, oxygen administration, noninvasive positive airway pressure, intubation, admission to a higher level of care, or death would not have been detected by this methodology if naloxone had not been administered. Any cases in which naloxone was not ordered or recorded in the electronic medical record also would have been missed, although such occurrences would be unlikely. Also, there was no comparison group analyzed in this study. As such, it is impossible to determine risk for an opioid-related adverse drug event by this study’s methodology. To determine
Volume 65, no. 4 : April 2015
Annals of Emergency Medicine 427
Beaudoin et al
Iatrogenic Opioid Overdose
Table 3. Results of qualitative analysis to identify patient-, provider-, and system-based factors, and potential strategies to reduce in-ED opioid-related medication errors and harm. Patient Factors
Advanced age Alcohol intoxication Altered mental status (eg, head injury, dementia) COPD Dementia Hepatic impairment Obesity Renal impairment Sleep apnea
Possible interventions
Exercise caution with the comorbidities outlined above under “Patient factors” and consider dose adjustment
Provider
Systems Based
Coadministration of opioids with other sedating medications (eg, benzodiazepines) Failure to obtain an accurate medication history Inappropriate use of opioids (eg, altered mental status, hypercarbic at baseline) No dose adjustment (hepatic or renal failure, geriatric) Provider inexperience (level of training) Route (IM or SC) Repeated opioid dosing, short interval between doses Lack of adherence to procedural sedation guidelines Lack of familiarity with naloxone dosing or duration Delayed recognition of the opioid overdose Lack of adequate discharge instructions or disclosure to the patient Inadequate documentation in the patient record Develop education for providers around: Appropriate use of opioids Dose adjustment (hepatic or renal failure, geriatric) Pharmacokinetics (onset and duration of action between different opioids and routes of administration) Recognition of signs of opioid overdose Naloxone dosing and duration of action Improve measures for accurate medication histories and medication reconciliation Ensure adequate supervision of inexperienced providers (residents and midlevels) Create guidelines for the use of opioids and procedural sedation Document, disclose, discuss: Document the event in the patient record Disclose the event to the patient Discuss with inpatient team or PCP
Inadequate monitoring (oximetry, capnography) Resuscitative equipment not readily available Transitions of care, change of shift Communication failures Verbal ordering Multiple providers involved No reporting through voluntary system Inadequate observation period and monitoring for recurrent symptoms No formal review by medication safety committee or peer review committee Lack of formalized “trigger tool” and review process Ensure adequate monitoring, particularly in high-risk patients Use standardized handoffs (SBAR) for transitions of care Use electronic ordering Limit use of verbal orders Judicious use of opioids as part of nursing protocol orders Ensure naloxone and resuscitative equipment are readily available Monitor for recurrent symptoms Consider admission or higher level of care Debrief with the care team Report in the voluntary reporting system Formal review by medication safety committee or peer review committee Formalize use of a “trigger tool” and review process Perform mock or simulated opioid-overdose scenarios with care team
SBAR, Situation background assessment recommendation; IM, intramuscular; SC, subcutaneous; PCP, primary care physician.
risk, one would have to assess all patients who received opioids in the ED for the outcome of an opioid-related adverse drug event. Because the study authors performed the qualitative analysis, they were not blind to the aims of the study, potentially introducing bias. Additionally, interrater agreement was not assessed because the analysis involved a consensus open discussion. Furthermore, the qualitative analysis was not exhaustive and thematic saturation was not a possible goal, given the finite number of cases involved. As such, the qualitative analysis likely missed factors related to medication errors and opioid-related adverse drug events. For instance, prescribing naloxone at discharge to “highrisk” patients who are discharged home with opioids did not emerge as a factor but could be a relevant area of future study. This study also is limited by its retrospective nature and reliance on the electronic medical record for information surrounding the opioid-related adverse drug event. Incomplete or vague descriptions of events in the electronic medical records could have led to misclassification. In particular, there were 7 cases with inadequate 428 Annals of Emergency Medicine
documentation to determine the contribution of an opioid analgesic to the event. A prospective study using naloxone trigger methodology in an electronic medical record that requires clinicians to provide additional information when naloxone is ordered should be considered for future studies. This study also was conducted within 1 hospital system. Although 2 study sites were used, the clinician provider pool between the 2 EDs is largely the same. Practice patterns and opioid prescribing might vary among providers and institutions and could affect generalizability of the study findings. Additionally, we did not assess for outlying analgesic ordering practices or patterns of behavior among individual providers.
DISCUSSION In-ED opioid-related adverse drug events represent an infrequent but important source of morbidity and mortality. For many of the patients in this investigation, their acute medical problem was not life threatening (eg, ankle injury, migraine, Volume 65, no. 4 : April 2015
Beaudoin et al abdominal pain), but the administration of opioid analgesics led to a near-fatal event. Opioid-related adverse drug events were not preventable in all cases (n¼22), which is a reminder that even when prescribed appropriately, opioids can lead to harm. However, more than two thirds of our cases with opioid-related adverse drug events were due to medication errors. More important, of the 51 patients who experienced an opioidrelated medication error, 17 required a lifesaving intervention (NCC-MERP category H), and opioids were believed to have contributed to 1 patient’s death (NCC-MERP category I). These instances reasonably can be classified as “never events,” preventable health-care-related injuries that should never occur.16 We identified patient-, provider-, and systems-level factors that likely contributed to opioid-related adverse drug events in these EDs, and we generated a list of strategies aimed to reduce medication errors and harm according to these factors. This information can inform future educational or policy interventions concerning opioid prescribing to prevent opioidrelated medication errors and patient harm. A large proportion of the patients in this investigation had chronic health conditions: renal dysfunction (>60%), liver disease, or cardiopulmonary disease. Our findings corroborate previous research findings from the inpatient setting. In a study of hospitalized patients who developed life-threatening respiratory problems while receiving analgesia, the following comorbid characteristics were present more frequently than in a comparison population: congestive heart failure, postoperative acute renal failure, obstructive sleep apnea, dysrhythmia, diabetes mellitus, coronary artery disease, and hypertension.17 As we note in our list of recommended strategies, providers should assess for the presence of these comorbid conditions when prescribing opioids. Hepatic impairment can lead to increased bioavailability of opioids because of decreased first-pass metabolism and reduced oxidation by liver enzymes; distribution may be altered by a reduction in the production of drug-binding proteins.18 Many opioids and their metabolites are eliminated by the kidneys; therefore, caution and dose adjustment are warranted in cases of renal insufficiency.19 The high prevalence of liver dysfunction and renal insufficiency in our patient series might indicate that ED clinicians did not adequately adjust the dose of opioids for patients with these conditions. Lack of a renal- or hepatic-adjusted dose (based on the upper end of manufacturers’ recommended starting doses) was identified as a factor during the qualitative analysis. We recommend that emergency medicine providers be familiar with dose adjustments of opioids. Hydromorphone was the most commonly administered opioid among cases analyzed in this study, followed by morphine and fentanyl. In an unpublished review of quality assurance data from these 2 EDs, the ratio of morphine to hydromorphone prescribing frequency was greater than 2:1. Other studies have found morphine to be involved in more opioid-related adverse drug events than hydromorphone.11 The number of opioidrelated adverse drug events associated with hydromorphone may reflect the fact that hydromorphone is a high-potency opioid administered in small milligram quantities compared with Volume 65, no. 4 : April 2015
Iatrogenic Opioid Overdose morphine. In other words, providers in these EDs might have prescribed 1 to 2 mg of hydromorphone without realizing its greater potency because the total amount of milligrams prescribed seemed small. Additionally, emergency providers might not know the equianalgesic conversion between morphine and hydromorphone or the onset and duration of hydromorphone action. In an opioid-naive patient, assuming no cross-tolerance, 1 mg of hydromorphone is generally considered equivalent to 6.7 mg of morphine. Other factors related to medication prescribing and administration identified in our patient population included the administration of more than 1 opioid, multiple routes of administration, and coadministration of other sedating medications to the same patient. These factors are consistent with those described among a Canadian inpatient sample.11 Providers should be aware that there is a delayed onset of action in intramuscular or subcutaneous preparations. Repeated dosing after an initial intramuscular or subcutaneous dose could lead to dose stacking and delayed adverse effects. In addition, opioid-induced respiratory depression can last longer than the analgesic effect of opioids, especially for fentanyl.20 Therefore, patients who receive more than 1 dose of opioids for pain could be more susceptible to respiratory depression. Automated systems can be put in place to warn providers when a second dose of opioids is being ordered or when there is a short interval between dosing. Other factors relevant to the prevention of opioid-related adverse drug events that emerged were related more to systemsbased practice rather than patient or prescriber factors. Problems identified included multiple providers giving orders for opioids at approximately the same time, change of shift (suggesting communication failures), and use of verbal orders. One potential issue that was not examined was provider clustering: whether particular providers account for a large proportion of cases. This aspect was not reviewed because of provider confidentiality concerns but might be of interest for subsequent study or quality improvement projects. System factors that also were affected by opioid-related adverse drug events were the need for continued monitoring, higher levels of care, and admission. In fact, a large proportion of patients in this study were admitted (z70%); this observation could reflect the acuity of the population receiving opioid analgesics, or it might indicate sequelae of in-ED opioid overdose that led to hospitalization (eg, hypoxia). Of particular importance, none of the opioid-related adverse drug events in this investigation were identified through the hospital’s voluntary reporting system. Although this absence might reflect institutional culture, this deficit has been corroborated by previous studies that have demonstrated the superiority of a naloxone trigger methodology over voluntary reporting in identifying opioid-related adverse drug events.10,21,22 In a previous study of an automated naloxone trigger tool program for pediatric inpatients, there were 34 adverse drug events identified among 59 triggered cases, yielding a positive predictive value of 57.6%.10 Another study reported a positive predictive value of 68.3% in the postoperative period.23 However, the identification of Annals of Emergency Medicine 429
Beaudoin et al
Iatrogenic Opioid Overdose opioid-related adverse drug events is only the first step in addressing iatrogenic opioid toxicity. The qualitative analysis in this study highlights areas that can be addressed by education or process improvements to decrease the overall incidence of opioidrelated adverse drug events, particularly those resulting from medication errors. Although not a part of this study analysis, the findings from this investigation led to structured provider education, an emphasis on standardized handoffs, and plans to implement naloxone trigger methodology at the study institutions. The implementation and prospective study of naloxone trigger methodology in the ED is a logical next research step. In addition, naloxone trigger methodology can also be used to prompt real-time action. For example, after several cardiac arrests were attributed to opioid administration in an inpatient setting at another institution, rapid response teams were triggered by naloxone administration, which led to a decrease in opioidrelated cardiac arrest.12 In summary, although the provision of analgesia for acute pain is an essential element of emergency care, excessive doses of opioids and inappropriate prescribing practices likely lead to lifethreatening opioid-related adverse drug events. We particularly urge caution when prescribing opioids for patients with hepatic or renal impairment or multiple comorbidities and recommend following the manufacturer-recommended dose adjustments for patients with these conditions. In addition, providers should familiarize themselves with the pharmacokinetics of opioids, especially for intramuscular and subcutaneous parenteral preparations, and exercise caution when coadministering other sedating medications with opioids. Last, future research should evaluate the effectiveness of naloxone trigger methodology in the ED setting. The authors would like to acknowledge the contributions of Drs. Traci Green and Jason Hack to the proposed work.
Supervising editor: Lewis S. Nelson, MD Author affiliations: From the Department of Emergency Medicine, Rhode Island Hospital, the Alpert Medical School of Brown University, Providence, RI (Beaudoin, Merchant, Janicki); the Department of Medicine, Division of Clinical Pharmacology, Rhode Island Hospital, Providence, RI (McKaig); and the Division of Medical Toxicology, Department of Emergency Medicine, University of Massachusetts Medical School, Worcester, MA (Babu). Author contributions: FLB and KMB conceived the study and designed the methodology. KMB obtained research funding and supervised the study and data collection. FLB, DMM, and KMB performed data collection. AJ managed the data. RCM provided statistical advice on study design. FLB and RCM analyzed the data. FLB, AJ, and RCM drafted the article, and all authors contributed substantially to its revision. FLB takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships
430 Annals of Emergency Medicine
in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist and provided the following details: Funded by the Lifespan Risk Management Grant Award. Publication dates: Received for publication July 16, 2014. Revisions received September 2, 2014, and November 6, 2014. Accepted for publication November 19, 2014. Available online December 18, 2014. Presented at the American College of Medical Toxicology Annual Scientific Meeting, Phoenix, AZ, March 2014.
REFERENCES 1. Davies EC, Green CF, Taylor S, et al. Adverse drug reactions in hospital in-patients: a prospective analysis of 3695 patient-episodes. PLoS One. 2009;4:e4439. 2. Lynn LA, Curry JP. Patterns of unexpected in-hospital deaths: a root cause analysis. Patient Saf Surg. 2011;5:3. 3. Oderda GM, Evans RS, Lloyd J, et al. Cost of opioid-related adverse drug events in surgical patients. J Pain Symptom Manage. 2003;25:276-283. 4. Oderda GM, Said Q, Evans RS, et al. Opioid-related adverse drug events in surgical hospitalizations: impact on costs and length of stay. Ann Pharmacother. 2007;41:400-406. 5. Barletta JF. Clinical and economic burden of opioid use for postsurgical pain: focus on ventilatory impairment and ileus. Pharmacotherapy. 2012;32(9 suppl):12S-18S. 6. Pizzi LT, Toner R, Foley K, et al. Relationship between potential opioidrelated adverse effects and hospital length of stay in patients receiving opioids after orthopedic surgery. Pharmacotherapy. 2012;32:502-514. 7. Volkow ND, McLellan TA, Cotto JH, et al. Characteristics of opioid prescriptions in 2009. JAMA. 2011;305:1299-1301. 8. Rozich JD, Haraden CR, Resar RK. Adverse drug event trigger tool: a practical methodology for measuring medication related harm. Qual Saf Health Care. 2003;12:194-200. 9. Classen DC, Pestotnik SL, Evans RS, et al. Computerized surveillance of adverse drug events in hospital patients. JAMA. 1991;266: 2847-2851. 10. Muething SE, Conway PH, Kloppenborg E, et al. Identifying causes of adverse events detected by an automated trigger tool through in-depth analysis. Qual Saf Health Care. 2010;19:435-439. 11. Neil K, Marcil A, Kosar L, et al. Retrospective analysis of opioid medication incidents requiring administration of naloxone. Can J Hosp Pharm. 2013;66:280-288. 12. Guirgis FW, Gerdik C, Wears RL, et al. Naloxone triggering the RRT: a human antidote? J Patient Saf. 2014 Mar 10. [Epub ahead of print]. 13. Gilbert EH, Lowenstein SR, Koziol-McLain J, et al. Chart reviews in emergency medicine research: where are the methods? Ann Emerg Med. 1996;27:305-308. 14. Webster LR, Cochella S, Dasgupta N, et al. An analysis of the root causes for opioid-related overdose deaths in the United States. Pain Med. 2011;(12 suppl 2):S26-S35. 15. NCC-MERP. About medication errors. Available at: http://www.nccmerp. org/aboutMedErrors.html. Accessed July 8, 2014. 16. de Wet C, O’Donnell C, Bowie P. Developing a preliminary “never event” list for general practice using consensus-building methods. Br J Gen Pract. 2014;64:e159-e167. 17. Ramachandran SK, Haider N, Saran KA, et al. Life-threatening critical respiratory events: a retrospective study of postoperative patients found unresponsive during analgesic therapy. J Clin Anesth. 2011;23:207-213. 18. Bosilkovska M, Walder B, Besson M, et al. Analgesics in patients with hepatic impairment: pharmacology and clinical implications. Drugs. 2012;72:1645-1669. 19. Dean M. Opioids in renal failure and dialysis patients. J Pain Symptom Manage. 2004;28:497-504.
Volume 65, no. 4 : April 2015
Beaudoin et al 20. Bailey PL, Streisand JB, East KA, et al. Differences in magnitude and duration of opioid-induced respiratory depression and analgesia with fentanyl and sufentanil. Anesth Analg. 1990;70:8-15. 21. Takata GS, Mason W, Taketomo C, et al. Development, testing, and findings of a pediatric-focused trigger tool to identify medication-related harm in US children’s hospitals. Pediatrics. 2008;121:e927-e935.
Iatrogenic Opioid Overdose 22. Takata GS, Taketomo CK, Waite S. Characteristics of medication errors and adverse drug events in hospitals participating in the California Pediatric Patient Safety Initiative. Am J Health Syst Pharm. 2008;65:2036-2044. 23. Eckstrand JA, Habib AS, Williamson A, et al. Computerized surveillance of opioid-related adverse drug events in perioperative care: a crosssectional study. Patient Saf Surg. 2009;3:18.
Meet the Editorial Board See photos and read the biographical sketches of the Annals Editorial Board. Information on the entire editorial board is available at www.annemergmed.com.
Volume 65, no. 4 : April 2015
Annals of Emergency Medicine 431
Preventing Iatrogenic Overdose: A Review of In–Emergency Department Opioid-Related Adverse Drug Events and Medication Errors Francesca L. Beaudoin, MD, MS*; Roland C. Merchant, MD, MPH; Adam Janicki, MD; Donald M. McKaig, RPh; Kavita M. Babu, MD *Corresponding Author. E-mail: [email protected].
Study objective: We describe characteristics of patients with in–emergency department (ED) opioid-related adverse drug events, medication errors, and harm resulting from medication errors; identify patient-, provider-, and system-based factors associated with in-ED opioid-related medication errors and harm; and create a list of strategies to prevent future events. Methods: This retrospective study was conducted at 2 urban academic EDs. Potential iatrogenic opioid overdoses were identified by querying the ED electronic medical record for cases when naloxone was administered after an opioid was administered in the ED. Cases involving medication errors resulting in harm were reviewed qualitatively for common patient-, provider-, and systems-based factors that might have contributed to the event. Results: Of 73 ED patients with in-ED opioid-related adverse events that required reversal with naloxone, 43 had a medication error resulting in harm. Patient-, provider-, and systems-based factors that might have contributed to the events included chronic health conditions that could predispose an individual to an opioid-related adverse event, failure to adjust opioid dosing in the elderly and for hepatic or renal impairment, multiple doses and routes of administration of opioids, coadministration of opioids with other sedating medications, and systems-based problems with patient handoffs and pharmacy oversight. Conclusion: We identified patient-, provider-, and systems-based factors related to opioid-related adverse drug events and medication errors among ED patients who had received naloxone. The results from our assessment can be used to inform educational and policy initiatives aimed to prevent in-ED opioid-related adverse drug events and medication errors. [Ann Emerg Med. 2015;65:423-431.] Please see page 424 for the Editor’s Capsule Summary of this article. A podcast for this article is available at www.annemergmed.com. Continuing Medical Education exam for this article is available at http://www.acep.org/ACEPeCME/. 0196-0644/$-see front matter Copyright © 2014 by the American College of Emergency Physicians. http://dx.doi.org/10.1016/j.annemergmed.2014.11.016
INTRODUCTION Background Opioid analgesics are commonly implicated in adverse drug events.1 Although some opioid-related adverse drug events might be unavoidable, many are due to preventable medication errors. Iatrogenic inhospital opioid-related adverse drug events manifest as a range of adverse events from somnolence to potentially lifethreatening central nervous system and respiratory depression. Furthermore, opioids prescribed for acute pain in the hospital setting have been highlighted as a cause of unexpected inhospital deaths2 and are associated with longer hospital lengths of stay, higher health care costs, and greater likelihood of 30-day readmission to the hospital.3-6 The emergency department (ED) is an important environment in which to examine opioid-related adverse drug events because emergency care providers frequently administer opioids for the treatment of pain7; however, opioid-related adverse drug events and medication errors have been understudied in the ED. Volume 65, no. 4 : April 2015
Many hospitals derive estimates of adverse drug events from their voluntary incident reporting systems, which capture less than 15% of adverse drug events identified by trigger tools.8,9 Trigger tools, such as a medication order or laboratory value, can indicate the occurrence of a potential adverse drug event and prompt an external review of the patient’s care.10 The administration of naloxone, an antidote indicated for the rapid reversal of opioid-induced coma or apnea, has been described as a marker of opioid-related adverse drug events and has feasibility as a trigger tool for the review of opioidrelated adverse drug events and medication errors.11,12 Importance Although naloxone administration appears to be a feasible means of identifying potential cases of iatrogenic opioid-related adverse drug events in the ED, the identification of these adverse drug events is only the first step in addressing this problem. Patient-, provider-, and systems-level risk factors for opioid-related Annals of Emergency Medicine 423
Iatrogenic Opioid Overdose
Editor’s Capsule Summary
What is already known on this topic Opioid-related iatrogenic adverse drug events are generally predictable and avoidable. What question this study addressed What are the patient-, provider-, and systems-based factors that contribute to iatrogenic opioid adverse events in the emergency department (ED)? What this study adds to our knowledge Using the ED administration of naloxone as a marker for potential cases, 73 opioid overdose patients were identified. Iatrogenic ED events were associated with poor patient health, failure to adjust opioid dose for age or disease, repeated therapeutic dosing, coadministration with sedating medication, poor handoffs, and limited pharmacy oversight. How this is relevant to clinical practice Understanding the reasons for iatrogenic opioid adverse events can lead to improved safe medication practices.
adverse drug events also must be elucidated to develop interventions to prevent this avoidable cause of morbidity and mortality. Goals of This Investigation Our objectives were to describe characteristics of patients with in-ED opioid-related adverse drug events, medication errors, and harm resulting from medication errors, identify patient-, provider-, and system-based factors associated with in-ED opioid-related medication errors and harm, and create a list of strategies to prevent future events. In addition, we compared the number of potentially serious opioid-related adverse drug events identified with naloxone as a case finder (ie, a naloxone trigger) versus the hospitals’ voluntary reporting system.
MATERIALS AND METHODS Study Design and Setting We conducted a retrospective review and qualitative analysis of all visits during which naloxone was administered to patients from October 2009 to December 2012 at 2 large, medical school–affiliated, urban New England EDs. One ED is a Level I trauma center with more than 100,000 adult visits per year and is the tertiary care referral center for the region, whereas the other is a community hospital with more than 55,000 adult visits per year. The 2 EDs are in the same hospital system and city, employ emergency physicians who are on staff at both facilities, and are 424 Annals of Emergency Medicine
Beaudoin et al 2 of the sites for the medical school’s emergency medicine residency program. The hospitals’ institutional review board approved the study protocol. Patients who received naloxone in the ED were identified through a search of the 2 EDs’ electronic medical record system database. Naloxone medication orders were located by querying preset, free-text, and verbal orders. Patients were included in the study if they received any opioid analgesic in the ED before naloxone administration. Patients who received naloxone before the ED visit (eg, by emergency medical service providers) or who did not receive an opioid in the ED were not included. The ED electronic medical records of patients meeting study inclusion criteria were reviewed with guidelines recommended by Gilbert et al.13 A panel of 3 of the study coinvestigators (an emergency physician, an emergency physician with board certification in toxicology, and a clinical pharmacist with expertise in medication safety) independently reviewed each of the ED electronic medical records and classified the cases according to an a priori hierarchy (described below). Before conducting their review, the investigators underwent a 2-hour training session to standardize the review protocol, using sample cases. After completing their independent reviews, the investigators met, discussed their findings, and then reached a consensus on their classification of the cases, conducted rootcauses analyses of the cases to determine whether a medication error and harm resulting from that error occurred, and identified patient, provider, and system factors related to harm from opioidrelated medication errors. They then created a list of possible solutions to prevent further in-ED opioid-related adverse drug events and medication errors. The investigators agreed a priori that an outside arbitrator would mediate if consensus could not be achieved, but there were no cases for which this was required. Cases of possible opioid-related adverse drug events were classified according to the following hierarchy: (1) opioid-related adverse drug events, nonopioid-related adverse drug events, or insufficient documentation to determine whether an adverse drug event occurred or was opioid related; (2) among cases with an opioid-related adverse drug event, further classification if a medication error versus no medication error occurred; and (3) among cases with a medication error, further classification if there was harm resulting versus no harm resulting. In accordance with previous studies on this topic, opioid-related adverse drug events were defined as objective or subjective (based on ED electronic medical record documentation) respiratory distress, sedation, or hypotension that improved after naloxone administration.11 For these analyses, opioid-related nausea or vomiting, urinary retention, constipation, and dysphoria were not considered to be opioid-related adverse drug events. Cases were deemed not to be an opioid-related adverse drug event if the patient did not have a clinical response to naloxone or another cause of the adverse event was ultimately found (eg, intracranial hemorrhage, hypoglycemia). The presence of medication errors among cases with an opioid-related adverse drug event was determined through a rootcause analysis by the review panel.14 The National Coordination Council for Medication Error Reporting and Prevention Index Volume 65, no. 4 : April 2015
Beaudoin et al
Iatrogenic Opioid Overdose
Table 1. NCC-MERP index for categorizing medication errors. Category
Conceptual Definition
A B
G H
Circumstances or events that have the capacity to cause error An error occurred but the error did not reach the patient (an “error of omission” does reach the patient) An error occurred that reached the patient but did not cause patient harm An error occurred that reached the patient and required monitoring to confirm that it resulted in no harm to the patient or required intervention to preclude harm An error occurred that may have contributed to or resulted in temporary harm to the patient and required intervention An error occurred that may have contributed to or resulted in temporary harm to the patient and required initial or prolonged hospitalization An error occurred that may have contributed to or resulted in permanent patient harm An error occurred that required intervention necessary to sustain life
I
An error occurred that may have contributed to or resulted in the patient’s death
C D E F
Example of Operational Definition Not applicable* Not applicable* Not applicable* Cases not satisfying criteria E–I Hypoxia requiring supplemental oxygen Admission to the hospital directly related to opioid overdose Not applicable Bag-valve-mask ventilation or intubation Patient death
*© 2001 National Coordinating Council for Medication Error Reporting and Prevention. All rights reserved. No patients satisfied criteria for categories A through C.
(NCC-MERP) definition of a medication error was used. In accordance with the NCC-MERP, “a medication error is any preventable event that may cause or lead to inappropriate medication use or patient harm while the medication is in the control of the health care professional, patient, or consumer. Such events may be related to professional practice, health care products, procedures, and systems, including prescribing; order communication; product labeling, packaging, and nomenclature; compounding; dispensing; distribution; administration; education; monitoring; and use.”15 Cases with an opioid-related adverse drug event and a medication error were further classified as having harm or no harm resulting, according to NCC-MERP taxonomy (Table 1).
Cases involving medication errors resulting in harm (NCCMERP categories E through H) were further reviewed for common patient, provider, and systems factors that might have contributed to the event. The analysis focused on only cases resulting in harm because the intent of the qualitative analysis was to identify factors related with and strategies to prevent both error and harm. Factors were identified through an open discussion and consensus among the 3 investigators, and the results were recorded during the deliberations. Transcripts of the discussions were not obtained or coded. A factor was included in the analysis if it was considered to be present in 2 or more cases. The investigators considered some factors a priori according to previous work by Neil et al,11 but
Figure. Classification of cases. ADE, Adverse drug event; NCC-MERP, National Coordinating Council for Medication Error Reporting and Prevention. Volume 65, no. 4 : April 2015
Annals of Emergency Medicine 425
Beaudoin et al
Iatrogenic Opioid Overdose Table 2. Patient characteristics by classification.* Opioid-Related Adverse Event
Age, y Sex, % female Medical history Diabetes Hypertension Asthma/COPD/OSA Liver disease Psychiatric comorbidity Substance abuse Cardiac disease Cancer Obesity Renal impairment Home medications Benzodiazepines Opioids Opioids administered Total opioid dose before naloxone (morphine equivalents) Route of opioid administration Intravenous Oral Intramuscular Subcutaneous Multiple routes Type of opioid Hydromorphone Morphine Fentanyl Other Multiple opioids Other sedating medication Benzodiazepines Phenothiazines Antihistamines Procedural sedation Indication for naloxone Altered mental status Bradycardia Hypotension Respiratory depression Hypoxia Other Multiple indications ED visit Disposition Admit Discharge Transfer AMA/elope ED length of stay, min Multiple doses of naloxone required Time from opioid to naloxone, min
426 Annals of Emergency Medicine
Medication Error
All (n[73)
Adverse Drug Event (n[22)
Medication Error (n[51)
Harm Resulted (n[43)
No Harm Resulted (n[8)
58 (49–76) 56.8
61 (51–82) 68.2
56 (41–70) 50.0
54 (40–70) 48.9
61 (56–71) 62.5
16.2 41.9 18.4 9.5 18.9 1.4 9.5 5.4 1.4 61.8
18.2 40.9 18.2 4.6 27.3 0 13.6 4.6 0 57.1
16.7 18.5 18.5 11.1 14.8 1.9 7.8 5.6 1.9 61.2
13.3 11.1 20.0 8.9 17.8 2.2 6.9 4.4 2.2 60.0
25.0 50.0 12.5 25.0 0 0 12.5 12.5 0 75.0
27.0 25.7
31.8 31.8
24.1 24.1
22.2 20.0
37.5 50.0
10.5 (5.8–12.1)
6.7 (4–8)
12.4 (6.7–13.7)
13.1 (6.7–21.9)
8.2 (5.4–13.4)
93.2 35.6 23.3 4.1 50.7 0 61.6 38.4 16.4 5.5 20.5
86.4 31.8 18.2 4.5 40.9
96.1 37.3 25.5 3.9 54.9
95.3 37.2 25.6 4.7 53.5
100.0 37.5 25.0 0 62.5
31.8 59.1 9.1 13.6 13.6
74.5 29.4 19.6 2.0 23.5
74.4 30.2 18.6 0 25.6
75.0 25.0 25.0 12.5 12.5
28.8 11.0 12.3 16.4
9.1 13.6 0 0
37.3 9.8 17.6 23.5
34.9 11.6 14.0 25.6
50.0 0 12.5 12.5
72.3 5.3 25.3 40.4 44.3 2.7 72.6
68.2 9.1 40.9 18.2 18.2 9.1 68.2
74.1 3.7 18.5 50.0 55.6 0 74.5
77.8 2.2 11.1 60.0 62.2 0 81.4
62.5 12.5 50.0 0 25.0 0 37.5
69.9 26.1 2.7 1.4 496 (368–621) 23.2 84 (46–152)
77.3 13.6 4.6 4.6 522 (396–650) 9 60 (38–111)
66.7 31.5 1.9 0 492 (354–585) 29.4 99.5 (57–168)
66.7 31.1 2.2 0 474 (420–511) 28 99 (50–181)
62.5 37.5 0 0 499 (354–640) 0 114 (60–140)
Volume 65, no. 4 : April 2015
Beaudoin et al
Iatrogenic Opioid Overdose
Table 2. Continued. Opioid-Related Adverse Event
Time naloxone to disposition, min Time naloxone to discharge, min
Medication Error
All (n[73)
Adverse Drug Event (n[22)
Medication Error (n[51)
Harm Resulted (n[43)
No Harm Resulted (n[8)
169 (105–267) 182 (76–245)
169 (94–363) 86 (23–568)
176.9 (110–252) 213.7 (84–238)
165 (110–286) 182 (68–238)
195 (121–221) 218 (84–251)
COPD, Chronic obstructive pulmonary disease; OSA, obstructive sleep apnea; AMA, against medical advice; cardiac disease, coronary artery disease, myocardial infarction, or congestive heart failure; liver disease, hepatitis or cirrhosis. *Categorical data are represented as a proportion (%); continuous data, as median (interquartile range).
additional factors were identified during the investigator deliberation of cases. Thematic saturation was not a component of this analysis, given the finite number of cases available for review. Next, the factors were used to create a list of strategies (actions or interventions) intended to mitigate inED opioid-related adverse drug events and medication errors identified during the qualitative analysis. Data Collection and Processing The investigators abstracted the following additional data from the ED electronic medical records for the cases: patient demographic characteristics, home medications, and medical history; medications administered in the ED; and the patient’s ED laboratory values, clinical outcomes, and disposition from the ED. Laboratory values abstracted were used to determine the presence of renal dysfunction (glomerular filtration rate 80 mg/100 mL). These data were recorded with a standardized abstraction instrument with double data entry verification. Summary statistics of patient demographic and clinical characteristics of all opioidrelated adverse drug events were generated and stratified by error and by harm classification. Because of the small sample size, formal statistical comparisons were not performed. We also queried the 2 hospitals’ voluntary error reporting system for all reported opioidrelated adverse drug events and compared the results to the number of cases identified with the naloxone case identification method outlined previously.
approximately half had opioids administered by more than 1 route (oral, intramuscular, etc). Altered mental status was cited as the most common reason for the administration of naloxone. None of the cases with an opioid-related adverse drug event identified through naloxone administration had been recorded in the hospitals’ voluntary reporting systems. In fact, during the entire study period, there were no opioid-related adverse drug events reported from the ED through the voluntary reporting system. During root-cause analysis, more than two thirds of the patients (n¼51) were determined to have had a medication error as the cause of their opioid-related adverse drug event. Of patients experiencing medication errors, most errors (n¼43) resulted in harm to the patient, according to the NCC-MERP classification system. Of the patients experiencing harm from a medication error, nearly three quarters had received hydromorphone, almost two thirds had received multiple doses of opioids, more than half had received opioids through multiple routes of administration, and half received 13 morphine milligram equivalents or more. Forty-three of the identified opioid-related adverse drug events were determined to have both a medication error and harm resulting; the results of qualitative analysis (factors and strategies) for these cases are presented in Table 3. Several factors emerged as potential targets to reduce the occurrence of iatrogenic opioid overdose and related adverse drug events, medication errors, and resultant harm. Factors related to aspects of the provider and prescribing often centered on knowledge deficits, whereas the systems-based issues revolved around resource availability, hospital culture, and policy.
RESULTS
LIMITATIONS
The classification of cases for this study is depicted in the Figure, and the demographic and clinical characteristics of the patients are summarized in Table 2. There were 87 cases during which naloxone was administered after an opioid was administered in the ED, of which 73 were determined to be opioid-related adverse drug events. Among the 73 patients with an opioid-related adverse drug event (Table 2), most were women, most had renal impairment, and approximately one fourth had reported benzodiazepines or opioids among their home medications. Approximately half of the patients received 10 morphine milligram equivalents or more in the ED before receiving naloxone (usually intravenously). Most patients received hydromorphone, and
We identified cases of in-ED opioid-related adverse drug events and medication errors through a search of in-ED naloxone administration. Cases of opioid toxicity that resulted in prolonged observation, oxygen administration, noninvasive positive airway pressure, intubation, admission to a higher level of care, or death would not have been detected by this methodology if naloxone had not been administered. Any cases in which naloxone was not ordered or recorded in the electronic medical record also would have been missed, although such occurrences would be unlikely. Also, there was no comparison group analyzed in this study. As such, it is impossible to determine risk for an opioid-related adverse drug event by this study’s methodology. To determine
Volume 65, no. 4 : April 2015
Annals of Emergency Medicine 427
Beaudoin et al
Iatrogenic Opioid Overdose
Table 3. Results of qualitative analysis to identify patient-, provider-, and system-based factors, and potential strategies to reduce in-ED opioid-related medication errors and harm. Patient Factors
Advanced age Alcohol intoxication Altered mental status (eg, head injury, dementia) COPD Dementia Hepatic impairment Obesity Renal impairment Sleep apnea
Possible interventions
Exercise caution with the comorbidities outlined above under “Patient factors” and consider dose adjustment
Provider
Systems Based
Coadministration of opioids with other sedating medications (eg, benzodiazepines) Failure to obtain an accurate medication history Inappropriate use of opioids (eg, altered mental status, hypercarbic at baseline) No dose adjustment (hepatic or renal failure, geriatric) Provider inexperience (level of training) Route (IM or SC) Repeated opioid dosing, short interval between doses Lack of adherence to procedural sedation guidelines Lack of familiarity with naloxone dosing or duration Delayed recognition of the opioid overdose Lack of adequate discharge instructions or disclosure to the patient Inadequate documentation in the patient record Develop education for providers around: Appropriate use of opioids Dose adjustment (hepatic or renal failure, geriatric) Pharmacokinetics (onset and duration of action between different opioids and routes of administration) Recognition of signs of opioid overdose Naloxone dosing and duration of action Improve measures for accurate medication histories and medication reconciliation Ensure adequate supervision of inexperienced providers (residents and midlevels) Create guidelines for the use of opioids and procedural sedation Document, disclose, discuss: Document the event in the patient record Disclose the event to the patient Discuss with inpatient team or PCP
Inadequate monitoring (oximetry, capnography) Resuscitative equipment not readily available Transitions of care, change of shift Communication failures Verbal ordering Multiple providers involved No reporting through voluntary system Inadequate observation period and monitoring for recurrent symptoms No formal review by medication safety committee or peer review committee Lack of formalized “trigger tool” and review process Ensure adequate monitoring, particularly in high-risk patients Use standardized handoffs (SBAR) for transitions of care Use electronic ordering Limit use of verbal orders Judicious use of opioids as part of nursing protocol orders Ensure naloxone and resuscitative equipment are readily available Monitor for recurrent symptoms Consider admission or higher level of care Debrief with the care team Report in the voluntary reporting system Formal review by medication safety committee or peer review committee Formalize use of a “trigger tool” and review process Perform mock or simulated opioid-overdose scenarios with care team
SBAR, Situation background assessment recommendation; IM, intramuscular; SC, subcutaneous; PCP, primary care physician.
risk, one would have to assess all patients who received opioids in the ED for the outcome of an opioid-related adverse drug event. Because the study authors performed the qualitative analysis, they were not blind to the aims of the study, potentially introducing bias. Additionally, interrater agreement was not assessed because the analysis involved a consensus open discussion. Furthermore, the qualitative analysis was not exhaustive and thematic saturation was not a possible goal, given the finite number of cases involved. As such, the qualitative analysis likely missed factors related to medication errors and opioid-related adverse drug events. For instance, prescribing naloxone at discharge to “highrisk” patients who are discharged home with opioids did not emerge as a factor but could be a relevant area of future study. This study also is limited by its retrospective nature and reliance on the electronic medical record for information surrounding the opioid-related adverse drug event. Incomplete or vague descriptions of events in the electronic medical records could have led to misclassification. In particular, there were 7 cases with inadequate 428 Annals of Emergency Medicine
documentation to determine the contribution of an opioid analgesic to the event. A prospective study using naloxone trigger methodology in an electronic medical record that requires clinicians to provide additional information when naloxone is ordered should be considered for future studies. This study also was conducted within 1 hospital system. Although 2 study sites were used, the clinician provider pool between the 2 EDs is largely the same. Practice patterns and opioid prescribing might vary among providers and institutions and could affect generalizability of the study findings. Additionally, we did not assess for outlying analgesic ordering practices or patterns of behavior among individual providers.
DISCUSSION In-ED opioid-related adverse drug events represent an infrequent but important source of morbidity and mortality. For many of the patients in this investigation, their acute medical problem was not life threatening (eg, ankle injury, migraine, Volume 65, no. 4 : April 2015
Beaudoin et al abdominal pain), but the administration of opioid analgesics led to a near-fatal event. Opioid-related adverse drug events were not preventable in all cases (n¼22), which is a reminder that even when prescribed appropriately, opioids can lead to harm. However, more than two thirds of our cases with opioid-related adverse drug events were due to medication errors. More important, of the 51 patients who experienced an opioidrelated medication error, 17 required a lifesaving intervention (NCC-MERP category H), and opioids were believed to have contributed to 1 patient’s death (NCC-MERP category I). These instances reasonably can be classified as “never events,” preventable health-care-related injuries that should never occur.16 We identified patient-, provider-, and systems-level factors that likely contributed to opioid-related adverse drug events in these EDs, and we generated a list of strategies aimed to reduce medication errors and harm according to these factors. This information can inform future educational or policy interventions concerning opioid prescribing to prevent opioidrelated medication errors and patient harm. A large proportion of the patients in this investigation had chronic health conditions: renal dysfunction (>60%), liver disease, or cardiopulmonary disease. Our findings corroborate previous research findings from the inpatient setting. In a study of hospitalized patients who developed life-threatening respiratory problems while receiving analgesia, the following comorbid characteristics were present more frequently than in a comparison population: congestive heart failure, postoperative acute renal failure, obstructive sleep apnea, dysrhythmia, diabetes mellitus, coronary artery disease, and hypertension.17 As we note in our list of recommended strategies, providers should assess for the presence of these comorbid conditions when prescribing opioids. Hepatic impairment can lead to increased bioavailability of opioids because of decreased first-pass metabolism and reduced oxidation by liver enzymes; distribution may be altered by a reduction in the production of drug-binding proteins.18 Many opioids and their metabolites are eliminated by the kidneys; therefore, caution and dose adjustment are warranted in cases of renal insufficiency.19 The high prevalence of liver dysfunction and renal insufficiency in our patient series might indicate that ED clinicians did not adequately adjust the dose of opioids for patients with these conditions. Lack of a renal- or hepatic-adjusted dose (based on the upper end of manufacturers’ recommended starting doses) was identified as a factor during the qualitative analysis. We recommend that emergency medicine providers be familiar with dose adjustments of opioids. Hydromorphone was the most commonly administered opioid among cases analyzed in this study, followed by morphine and fentanyl. In an unpublished review of quality assurance data from these 2 EDs, the ratio of morphine to hydromorphone prescribing frequency was greater than 2:1. Other studies have found morphine to be involved in more opioid-related adverse drug events than hydromorphone.11 The number of opioidrelated adverse drug events associated with hydromorphone may reflect the fact that hydromorphone is a high-potency opioid administered in small milligram quantities compared with Volume 65, no. 4 : April 2015
Iatrogenic Opioid Overdose morphine. In other words, providers in these EDs might have prescribed 1 to 2 mg of hydromorphone without realizing its greater potency because the total amount of milligrams prescribed seemed small. Additionally, emergency providers might not know the equianalgesic conversion between morphine and hydromorphone or the onset and duration of hydromorphone action. In an opioid-naive patient, assuming no cross-tolerance, 1 mg of hydromorphone is generally considered equivalent to 6.7 mg of morphine. Other factors related to medication prescribing and administration identified in our patient population included the administration of more than 1 opioid, multiple routes of administration, and coadministration of other sedating medications to the same patient. These factors are consistent with those described among a Canadian inpatient sample.11 Providers should be aware that there is a delayed onset of action in intramuscular or subcutaneous preparations. Repeated dosing after an initial intramuscular or subcutaneous dose could lead to dose stacking and delayed adverse effects. In addition, opioid-induced respiratory depression can last longer than the analgesic effect of opioids, especially for fentanyl.20 Therefore, patients who receive more than 1 dose of opioids for pain could be more susceptible to respiratory depression. Automated systems can be put in place to warn providers when a second dose of opioids is being ordered or when there is a short interval between dosing. Other factors relevant to the prevention of opioid-related adverse drug events that emerged were related more to systemsbased practice rather than patient or prescriber factors. Problems identified included multiple providers giving orders for opioids at approximately the same time, change of shift (suggesting communication failures), and use of verbal orders. One potential issue that was not examined was provider clustering: whether particular providers account for a large proportion of cases. This aspect was not reviewed because of provider confidentiality concerns but might be of interest for subsequent study or quality improvement projects. System factors that also were affected by opioid-related adverse drug events were the need for continued monitoring, higher levels of care, and admission. In fact, a large proportion of patients in this study were admitted (z70%); this observation could reflect the acuity of the population receiving opioid analgesics, or it might indicate sequelae of in-ED opioid overdose that led to hospitalization (eg, hypoxia). Of particular importance, none of the opioid-related adverse drug events in this investigation were identified through the hospital’s voluntary reporting system. Although this absence might reflect institutional culture, this deficit has been corroborated by previous studies that have demonstrated the superiority of a naloxone trigger methodology over voluntary reporting in identifying opioid-related adverse drug events.10,21,22 In a previous study of an automated naloxone trigger tool program for pediatric inpatients, there were 34 adverse drug events identified among 59 triggered cases, yielding a positive predictive value of 57.6%.10 Another study reported a positive predictive value of 68.3% in the postoperative period.23 However, the identification of Annals of Emergency Medicine 429
Beaudoin et al
Iatrogenic Opioid Overdose opioid-related adverse drug events is only the first step in addressing iatrogenic opioid toxicity. The qualitative analysis in this study highlights areas that can be addressed by education or process improvements to decrease the overall incidence of opioidrelated adverse drug events, particularly those resulting from medication errors. Although not a part of this study analysis, the findings from this investigation led to structured provider education, an emphasis on standardized handoffs, and plans to implement naloxone trigger methodology at the study institutions. The implementation and prospective study of naloxone trigger methodology in the ED is a logical next research step. In addition, naloxone trigger methodology can also be used to prompt real-time action. For example, after several cardiac arrests were attributed to opioid administration in an inpatient setting at another institution, rapid response teams were triggered by naloxone administration, which led to a decrease in opioidrelated cardiac arrest.12 In summary, although the provision of analgesia for acute pain is an essential element of emergency care, excessive doses of opioids and inappropriate prescribing practices likely lead to lifethreatening opioid-related adverse drug events. We particularly urge caution when prescribing opioids for patients with hepatic or renal impairment or multiple comorbidities and recommend following the manufacturer-recommended dose adjustments for patients with these conditions. In addition, providers should familiarize themselves with the pharmacokinetics of opioids, especially for intramuscular and subcutaneous parenteral preparations, and exercise caution when coadministering other sedating medications with opioids. Last, future research should evaluate the effectiveness of naloxone trigger methodology in the ED setting. The authors would like to acknowledge the contributions of Drs. Traci Green and Jason Hack to the proposed work.
Supervising editor: Lewis S. Nelson, MD Author affiliations: From the Department of Emergency Medicine, Rhode Island Hospital, the Alpert Medical School of Brown University, Providence, RI (Beaudoin, Merchant, Janicki); the Department of Medicine, Division of Clinical Pharmacology, Rhode Island Hospital, Providence, RI (McKaig); and the Division of Medical Toxicology, Department of Emergency Medicine, University of Massachusetts Medical School, Worcester, MA (Babu). Author contributions: FLB and KMB conceived the study and designed the methodology. KMB obtained research funding and supervised the study and data collection. FLB, DMM, and KMB performed data collection. AJ managed the data. RCM provided statistical advice on study design. FLB and RCM analyzed the data. FLB, AJ, and RCM drafted the article, and all authors contributed substantially to its revision. FLB takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships
430 Annals of Emergency Medicine
in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist and provided the following details: Funded by the Lifespan Risk Management Grant Award. Publication dates: Received for publication July 16, 2014. Revisions received September 2, 2014, and November 6, 2014. Accepted for publication November 19, 2014. Available online December 18, 2014. Presented at the American College of Medical Toxicology Annual Scientific Meeting, Phoenix, AZ, March 2014.
REFERENCES 1. Davies EC, Green CF, Taylor S, et al. Adverse drug reactions in hospital in-patients: a prospective analysis of 3695 patient-episodes. PLoS One. 2009;4:e4439. 2. Lynn LA, Curry JP. Patterns of unexpected in-hospital deaths: a root cause analysis. Patient Saf Surg. 2011;5:3. 3. Oderda GM, Evans RS, Lloyd J, et al. Cost of opioid-related adverse drug events in surgical patients. J Pain Symptom Manage. 2003;25:276-283. 4. Oderda GM, Said Q, Evans RS, et al. Opioid-related adverse drug events in surgical hospitalizations: impact on costs and length of stay. Ann Pharmacother. 2007;41:400-406. 5. Barletta JF. Clinical and economic burden of opioid use for postsurgical pain: focus on ventilatory impairment and ileus. Pharmacotherapy. 2012;32(9 suppl):12S-18S. 6. Pizzi LT, Toner R, Foley K, et al. Relationship between potential opioidrelated adverse effects and hospital length of stay in patients receiving opioids after orthopedic surgery. Pharmacotherapy. 2012;32:502-514. 7. Volkow ND, McLellan TA, Cotto JH, et al. Characteristics of opioid prescriptions in 2009. JAMA. 2011;305:1299-1301. 8. Rozich JD, Haraden CR, Resar RK. Adverse drug event trigger tool: a practical methodology for measuring medication related harm. Qual Saf Health Care. 2003;12:194-200. 9. Classen DC, Pestotnik SL, Evans RS, et al. Computerized surveillance of adverse drug events in hospital patients. JAMA. 1991;266: 2847-2851. 10. Muething SE, Conway PH, Kloppenborg E, et al. Identifying causes of adverse events detected by an automated trigger tool through in-depth analysis. Qual Saf Health Care. 2010;19:435-439. 11. Neil K, Marcil A, Kosar L, et al. Retrospective analysis of opioid medication incidents requiring administration of naloxone. Can J Hosp Pharm. 2013;66:280-288. 12. Guirgis FW, Gerdik C, Wears RL, et al. Naloxone triggering the RRT: a human antidote? J Patient Saf. 2014 Mar 10. [Epub ahead of print]. 13. Gilbert EH, Lowenstein SR, Koziol-McLain J, et al. Chart reviews in emergency medicine research: where are the methods? Ann Emerg Med. 1996;27:305-308. 14. Webster LR, Cochella S, Dasgupta N, et al. An analysis of the root causes for opioid-related overdose deaths in the United States. Pain Med. 2011;(12 suppl 2):S26-S35. 15. NCC-MERP. About medication errors. Available at: http://www.nccmerp. org/aboutMedErrors.html. Accessed July 8, 2014. 16. de Wet C, O’Donnell C, Bowie P. Developing a preliminary “never event” list for general practice using consensus-building methods. Br J Gen Pract. 2014;64:e159-e167. 17. Ramachandran SK, Haider N, Saran KA, et al. Life-threatening critical respiratory events: a retrospective study of postoperative patients found unresponsive during analgesic therapy. J Clin Anesth. 2011;23:207-213. 18. Bosilkovska M, Walder B, Besson M, et al. Analgesics in patients with hepatic impairment: pharmacology and clinical implications. Drugs. 2012;72:1645-1669. 19. Dean M. Opioids in renal failure and dialysis patients. J Pain Symptom Manage. 2004;28:497-504.
Volume 65, no. 4 : April 2015
Beaudoin et al 20. Bailey PL, Streisand JB, East KA, et al. Differences in magnitude and duration of opioid-induced respiratory depression and analgesia with fentanyl and sufentanil. Anesth Analg. 1990;70:8-15. 21. Takata GS, Mason W, Taketomo C, et al. Development, testing, and findings of a pediatric-focused trigger tool to identify medication-related harm in US children’s hospitals. Pediatrics. 2008;121:e927-e935.
Iatrogenic Opioid Overdose 22. Takata GS, Taketomo CK, Waite S. Characteristics of medication errors and adverse drug events in hospitals participating in the California Pediatric Patient Safety Initiative. Am J Health Syst Pharm. 2008;65:2036-2044. 23. Eckstrand JA, Habib AS, Williamson A, et al. Computerized surveillance of opioid-related adverse drug events in perioperative care: a crosssectional study. Patient Saf Surg. 2009;3:18.
Meet the Editorial Board See photos and read the biographical sketches of the Annals Editorial Board. Information on the entire editorial board is available at www.annemergmed.com.
Volume 65, no. 4 : April 2015
Annals of Emergency Medicine 431
