Pattern of Opioid Overdoses and Interventions at the Emergency Department: Impact of COVID-19

Background: Opioid-related overdoses cause substantial numbers of preventable deaths. Naloxone is an opioid antagonist available in take-home naloxone (THN) kits as a lifesaving measure for opioid overdose. As the emergency department (ED) is a primary point of contact for patients with high-risk opioid use, evidence-based recommendations from the Society of Hospital Pharmacists of Australia THN practice guidelines include the provision of THN, accompanied by psychosocial interventions. However, implementation of these guidelines in practice is unknown. This study investigated ED opioid-related overdose presentations, concordance of post-overdose interventions with the THN practice guidelines, and the impact, if any, of the SARS-CoV-2 (COVID-19) pandemic on case presentations.MethodsA single-centre retrospective audit was conducted at a major tertiary hospital of patients presenting with overdoses involving opioids and non-opioids between March to August 2019 and March to August 2020. Patient presentations and interventions delivered by the paramedics, ED and upon discharge from the ED were collated from medical records and analysed using descriptive statistics, chi square and independent T-tests.ResultsThe majority (66.2%) of patients presented to hospital with mixed drug overdoses involving opioids and non-opioids. Pharmaceutical opioids were implicated in a greater proportion (72.1%) of overdoses than illicit opioids. Fewer patients presented in March to August 2020 as compared with 2019 (26 vs. 42), and mixed drug overdoses were more frequent in 2020 than 2019 (80.8% vs. 57.1%). Referral to outpatient psychology (22.0%) and drug and alcohol services (20.3%) were amongst the most common post-discharge interventions. Naloxone was provided to 28 patients (41.2%) by the paramedics and/or ED. No patients received THN upon discharge.ConclusionsThis study highlights opportunities to improve ED provision of THN and other interventions post-opioid overdose. Large-scale multi-centre studies are required to ascertain the capacity of EDs to provide THN and the impact of COVID-19 on opioid overdose presentations.

conjunction with counselling about its use [25]. THN should be provided alongside psychosocial interventions, such as psychiatric assessment, peer-recovery coach programs and social services referral, to prevent recurrent overdose [16,26]. Since this Society of Hospital Pharmacists of Australia guideline was released in November 2020, there is currently no information about its implementation in practice.
Studies investigating the capacity to which these potentially life-saving interventions are delivered post-overdose are limited. Information regarding opioid overdose presentation numbers, and the delivery of post-discharge interventions, is required to provide an indication of the capacity of Australian EDs to meet the WHO recommendations and Society of Hospital Pharmacists of Australia guidelines. As the ED is a primary, and perhaps only, point of contact with the healthcare system for patients living with SUDs, further understanding of this topic is critical to address the upward shift in opioid-induced mortality. Furthermore, to date, changes in opioid overdose presentations with the onset of the COVID-19 pandemic have not been evaluated in Australia.

Aim
This study aimed to assess ED overdose presentations involving opioids and how they align with current Society of Hospital Pharmacists of Australia practice guidelines for THN in Australian hospitals. Speci cally, the study aimed to determine: (1) the frequency of overdose presentations that involved pharmaceutical and non-pharmaceutical opioids, (2) the characteristics and comorbidities of opioid overdose patients and their ED presentations, (3) the interventions and post-discharge management strategies provided after opioid overdose presentations and (4) differences in opioid overdose presentations before and during the COVID-19 pandemic.

Study Design
A single-centre, retrospective observational audit was conducted at a tertiary hospital in Australia, using data obtained from medical records. Governance Evidence Knowledge Outcome (GEKO) (Quality Activity number 39844; approved 26 February 2021) and reciprocal Curtin human research ethics (HRE2021-0095) approval were obtained prior to data collection.

Participants and Sampling
Patients 18 years and older admitted to the ED with opioid-related overdose from March to August 2019 (pre-COVID- 19) and March to August 2020 (COVID-19) were included. Opioid-related overdoses were de ned as overdoses where opioids were the sole contributor or contributed in combination with other drugs (mixed drug overdoses). Patients presenting with overdose are admitted under the toxicology team for immediate review, investigations and treatment.
A list of eligible cases was obtained from the Medical Records Department using the Australian ICD-10-AM clinical codes: EKB00 (drug and alcohol), X42 (accidental poisoning by and exposure to narcotics and hallucinogens) and X62 (intentional self-poisoning by and exposure to narcotics and hallucinogens). Patient numbers for each ICD-10-AM code were tabulated (Supplementary Table 1). Following this, ICD-10-AM codes indicative of opioid-related overdose were identi ed: T40.1 (heroin), T40.6 (other and unspeci ed narcotics), T40.2 (other opioids), X42 and X62. These codes were used to shortlist appropriate patients and data were collated from corresponding les where the inclusion criteria were met.

Data Collection
Two auditors extracted data from patients' medical records. To ensure accuracy and robustness in the data collected, the auditors attended a one-week ED visit with the toxicology team to familiarise themselves with procedures and documentation in relation to overdose presentations prior to data collection. To further improve data accuracy, the rst 10% of patients were reviewed by both auditors and cross-checked. Disparities were evaluated and resolved prior to continuation of the audit.
Data transcribed from medical records comprised of demographic information, current medications and comorbidities, overdose factors (e.g. type of opioids used and their Standard for the Uniform Scheduling of Medicines and Poisons (SUSMP) Schedule), ED presentation characteristics (e.g. date of admission, length of stay and arrival mode) and interventions delivered after medical assistance arrived (pre-ED), in the ED and upon discharge from the ED. Pharmaceutical opioids were de ned as those classi ed under the SUSMP as Schedule 2, 3, 4 and 8, while non-pharmaceutical opioids were classi ed as Schedule 9 [27].

Data Analysis
Data were analysed using SPSS version 27. Continuous and categorical variables were analysed using descriptive statistics. Inferential statistics (chi square tests and independent T-tests) were used to explore associations between patient characteristics, their presentations and the interventions provided, and to detect differences in overdose presentations between the audit periods in 2019 and 2020.

Description of Sample
A total of 108 medical records of patients presenting during the audit period with a diagnosis indicative of opioid overdose were identi ed. Of these, 42 were either not overdoses (e.g. adverse drug reaction) or were not opioid-related overdoses (e.g. cannabis overdose) and, therefore, were excluded. Data were extracted from 66 patients who presented with opioid-related overdose. One patient presented 3 times during the audit period and, therefore, 68 cases were included in the primary analysis. There was 32 male and 36 female cases included and the mean age of patients at presentation was 40.6 ± 16.8 years (Table 1).

Emergency Department Presentation
Of the 68 cases reviewed, 42 (61.8%) presented in 2019 and 26 (38.2%) presented during the 2020 6-month audit period of March to August ( Table 1). The majority of patients arrived by ambulance (N= 55, 80.9%), while the remaining arrived via private transport (N= 10, 14.7%), hospital transfer (N= 2, 2.9%) or were brought in by police (N= 1, 1.5%). There was a non-signi cant trend of a reduction in arrival by ambulance during the COVID-period of 2020 (88.1% vs. 69.2%). The median length of stay was 7 hours (IQR: 9). Eighteen patients (26.5%) were admitted as an inpatient and, of these, the majority (N= 11, 61.1%) were admitted to the intensive care unit. Half of the inpatients returned to the ED prior to discharge. Aspiration pneumonia was reported for 7 of the 18 patients (38.9%) as the primary reason for hospital admission. One patient died in hospital as a consequence of their overdose. Men were more likely to report drug-related comorbidities than females (71.9% vs. 41.7%; p<0.05) ( Table 2). Excluding SUDs, the most common comorbidity was mental, behavioural or neurodevelopmental disorders (ICD-11 code 06) (    Interventions Delivered Of the 55 patients that arrived by ambulance, 9 (16.4%) received naloxone administered by paramedics (Table 5). One patient received naloxone delivered intranasally by a bystander prior to arrival of the paramedics. Within the ED, naloxone was delivered to 33.8% (N= 23) of patients. The median total dose of naloxone provided within the ambulance and ED was 550 micrograms, though there was signi cant variation from 100 micrograms to 14,212 micrograms. Assessments by the drug and alcohol service were provided to 20.6% (N= 14) of patients. A further 4 patients declined, 14 were admitted as inpatients, 1 patient was assessed on a recent admission and 1 was discharged against medical advice. Of the patients who were discharged from the ED, no patients received THN. A recommendation for patients to follow up with their general practitioner was most frequently documented within the post-discharge management plan (N= 23, 39.0%), followed by psychiatry (N= 13, 22.0%) and indirect referral or continued engagement with drug and alcohol services (N= 12, 20.3%). For patients with SUDs, drug and alcohol service engagement was a more common post-discharge intervention in 2020 as compared with 2019 (56.3% vs. 26.3%).

Discussion
This study observed a greater proportion of opioid-related overdoses involving pharmaceutical opioids as compared to heroin. Mixed drug overdoses involving non-opioids were more frequent than opioid only overdoses. This nding was more pronounced during March to August 2020 as compared to the same period in 2019 (pre-COVID- 19), and for those presenting with intentional overdoses. A reduction in the number of patients presenting with opioid-related overdoses between 2019 and 2020 was evident, which may be attributed to the COVID-19 pandemic.
Mental health comorbidities were prominent amongst the patient population, especially depression and anxiety. As part of the post-discharge plan for patients presenting with opioid-related overdose, several strategies were documented including psychiatry follow-up and drug and alcohol services referral. However, no patients received THN. Barriers to the provision of THN in hospital EDs need to be addressed to complement and support wider efforts to promote THN access and prevent opioid-induced mortality.
To our knowledge, this is the rst study examining the provision of THN and other interventions post-opioid overdose in Australian EDs.
Providing and educating people on THN is an effective strategy to reverse opioid overdose and prevent mortality [16]. The 2020 Society of Hospital Pharmacists of Australia guidelines state that THN should be provided to patients presenting with opioid toxicity, those who inject opioids, use opioid substitution therapy or are prescribed opioids for chronic pain [25]. While THN is available free-of-charge from 225 community drug and alcohol services and pharmacies in the Australian state of Western Australia under the THN pilot, only 1 hospital is registered with this program and 2 other hospitals are supplying THN to ED patients according to anecdotal reports [28]. Equity of access and care needs to be addressed, as no hospitals outside of the inner city are currently providing THN. Internationally, the underutilisation of THN programs is attributed to multiple barriers, including lack of time, training and institutional support [21,29]. Furthermore, some healthcare workers hold stigmatising and inaccurate assumptions about naloxone distribution [30,31]. There is a need to improve understanding and shift attitudes regarding THN to ensure survivors of opioid overdose, and those at risk of overdose from pharmaceutical and non-pharmaceutical drug use, have access to this life-saving intervention from the ED [25].
An expected nding of this study was the signi cant proportion of patients with comorbid mental health conditions. This was evident for those who intentionally overdosed (75%), as well as those who unintentionally overdosed (51.4%), which re ects a well-established relationship between SUDs and mental health conditions [32][33][34]. Those who intentionally overdosed were less likely to be given naloxone than those who unintentionally overdosed (p<0.001). This may indicate intentional overdoses were less severe or opioids contributed to a lesser extent. Consistent with the latter, those who intentionally overdosed were signi cantly more likely to present with mixed drug overdoses (p<0.05).
Psychiatric team follow-up was more commonly documented in the post-discharge plan of those who intentionally overdosed than those who unintentionally overdosed (p<0.05). However, referral rates could be improved. Addressing the circumstances surrounding patients' intentional overdoses and implementing treatment strategies is crucial to preventing self-harm [35].
A signi cant number of patients were referred to outpatient drug and alcohol services. However, outpatient referrals often indicated the provision of contact numbers and relied on patients to organise appointments. This is inconsistent with an ideal continuity of care approach where, for example, a clinical handover is conducted between healthcare providers with patient involvement or peer recovery coaches conduct motivational interviewing and facilitate referrals [15,36]. A complicating factor with regards to organising referrals may be patients' lack of motivation to change their drug use behaviours [37]. Requesting discharge and declining services was a common occurrence amongst the patient population, which may demotivate healthcare professionals to encourage patient change. The drug and alcohol team (at this hospital) extended their hours of operation considerably from 2019 to 2020, which was accompanied by a 30% increase in referral rates. Positive changes such as this are required to address the disconnect between EDs and community drug and alcohol services and optimise patient engagement in outpatient services following ED overdose presentations.
Pharmaceutical opioids were implicated in the majority of opioid-related overdoses included in this audit (72.1%). Comparatively, the Australian Institute of Health and Welfare reported the annual hospitalisation rate for overdoses involving pharmaceutical opioids was more than double that of overdoses involving heroin or opium (9.1 vs. 3.4 per 100,000 population) [38]. This is consistent with US trends, where in 2018 9.9 million people reported past-year use of prescription opioids for non-medical purposes and 800,000 reported heroin use [1]. In the current audit, pharmaceutical opioids prescribed for the patient were implicated in a substantial number of cases (N= 24). In recognition of the harm attributed to pharmaceutical opioids, the Therapeutic Goods Administration of Australia has established the Opioid Regulatory Advisory Group to alter opioid prescribing and dispensing practices [39]. For example, smaller pack sizes for immediate-release prescription opioid medications have become available to allow dispensing of the quantity required for the patient. This prevents surplus opioids circulating within the community that may cause inadvertent or deliberate harm [39]. Continual review of opioid use is required to lessen the burden of opioid-related harm.
A signi cant reduction in the number of ED presentations for opioid-related overdose was seen between 2019 and 2020 (N= 42 vs. 26, respectively). This nding correlates with anecdotal reports from healthcare workers at the hospital of fewer overdose presentations with the COVID-19 pandemic onset. In contrast, US studies have observed an increase in opioid overdose presentations from March 2020 [4,7]. There are a number of possible explanations for the decline observed in this audit including (1) changes in drug access with border closures, (2) comparatively fewer well-meaning bystanders available to phone ambulances during isolation periods imposed in Australia, (3) increased fatalities (4) community expansion of the Australian THN pilot [20], and (4) a small sample size. To support the second of these hypotheses, patients were more likely to present by means other than ambulance transport during 2020 than 2019 (30.8% vs. 11.9%). While there was no signi cant difference in the proportion of overdoses involving pharmaceutical and non-pharmaceutical opioids, mixed drug overdoses involving non-opioids were more frequent in 2020 than 2019 (80.8% vs. 57.1%), as well as overdoses involving between 4 and 7 drugs (42.3% vs. 19.0%). These trends provide preliminary evidence to suggest a shift in drug use behaviours during 2020 as compared with 2019.