INTRODUCTION
Substance use among youth in Canada is alarmingly prevalent. Canadian data indicate that 45% of students in grades 7-12 have consumed alcohol and that 33% of students have engaged in binge drinking. 1 The mean age at first alcohol use has been reported as 12.8 years. 1 Use of cannabis in the last 12 months in this same population was 21%; in addition, 5% had used ecstasy, 5% psychoactive chemicals (such as sedatives, stimulants, or pain relievers), 5% dextromethorphan, and 4% hallucinogens with the intent to “get high” within the past 12 months. 1 Prescription drug abuse is also becoming increasingly common, and youth have been identified as a high-risk group for misuse and associated harms.Reference Fischer, Rehm, Goldman and Popova 2 , 3
The morbidity and mortality associated with substance use is well described.Reference Popova, Rehm and Patra 4 Use of alcohol and illicit substances can result in serious harm, including physical injury, psychosis, loss of consciousness, and death due to cardiopulmonary collapse.Reference Black, Asbridge and Lea 5 , Reference Rehm, Gnam and Popova 6 A 2003 Canadian study of emergency department (ED) visits found that 3.1% of visits for injuries by patients ages 15–19 years were alcohol-related.Reference Black, Asbridge and Lea 5 Substance use is also associated with increased risk of sexual assault, unplanned sexual activity, behavioural problems, pregnancy, and self-harm.Reference Rehm, Gnam and Popova 6 - Reference Poulin and Graham 13 Use of more than one type of substance at the same time (termed polydrug use) is associated with a further increased risk of harm, injury, and overdose.Reference Hoare 14 Emerging evidence suggests that the overall incidence of polydrug use and associated drug-related death is increasing; however, Canadian data are limited.Reference Hingson and Zha 12 , Reference Poulin and Graham 13
The ED often serves as an entry point of care for children with acute substance use related medical or psychological issues.Reference Black, Asbridge and Lea 5 , Reference Burke, O’Sullivan and Vaughan 15 As such, the initial medical management and subsequent psychosocial needs of these at-risk youth often fall to the hands of the emergency physician. The objectives of our study were to 1) characterize the substances being used by youth who presented to a pediatric ED for complaints related to their substance use, 2) characterize the demographic descriptors of these youth, and 3) describe the acute morbidity and mortality associated with this substance use.
MATERIALS AND METHODS
Setting
This retrospective cohort study was conducted at the Stollery Children’s Hospital in Edmonton, Alberta (population 817,498). 16 This tertiary care centre provided services to children ages 0 to 16 years, with a catchment area including Northern Alberta and parts of British Columbia, Saskatchewan, the Northwest Territories, Yukon, and Nunavut. The pediatric ED with 12 pediatric beds had a yearly census of approximately 24,000 patients during the study period. It was staffed by over 20 physicians, most of whom were pediatric emergency, residency-trained.
Case identification
We included all consecutive cases of acute intoxication and unintentional poisoning in children ages 10 to 16 years who presented to the Stollery ED from January 1, 2007, to December 31, 2009. Cases were identified if their primary or secondary recorded diagnoses were coded according to the World Health Organization International Classification of Diseases, Tenth Revision (ICD-10), as any of the following: F10-19 with subsets 0.0, 0.1, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, T36.*.−50.*, T51.*−65.*, X40.*−49.*, Y10.*−19* (Table 1). We excluded cases of ingestion for intent of suicide or self-harm; accidental ingestions were also excluded, if this intention was explicitly stated in the chart. Multiple visits by the same child were recorded as separate visits. The University of Alberta Hospital Health Research Ethics Board granted ethics approval prior to implementation of the study.
Included diagnoses
Data collection
Medical records (paper-based) were reviewed between August 2010 and February 2011; this included a review of documentation by all health professionals involved in the patient’s care, as well as the results of all investigations performed during the given health care encounter. The following variables were collected from each record: demographic characteristics, past medical history of substance use and mental health disorders, guardianship status, substance(s) used, mode of arrival to the ED, Canadian Triage and Acuity Scale (CTAS) coding,Reference Murray, Bullard and Grafstein 17 presenting complaint, level of consciousness, admission status, medications and treatments provided by emergency medical services (EMS) or ED personnel, consultation with a social worker or mental health professional, and time of contact with various health professionals. The list of specific substances ingested were extracted from the chart, because they were described in the documented medical or nursing history; if available, toxicologic blood and urine tests were included as well. Data were recorded in a study-specific standardized, electronic abstraction form that was managed using Research Electronic Data Capture (REDCap), an electronic data capture tool hosted and supported by the Clinical Research Informatics Core in the Women and Children’s Health Research Institute (Edmonton, Alberta).Reference Harris, Taylor and Thielke 18
Approximately half of the medical records were abstracted by an experienced research assistant (YK), who was trained by the first author (ED); the research assistant was blinded to the study hypotheses. Due to financial constraints, the remaining records were abstracted by the first author, who was not blinded to the hypotheses. The research assistant and first author met frequently and regularly during the data abstraction phase.
Conditions recorded as mental health disorders included anxiety and mood disorders, attention deficit and hyperactivity disorder, oppositional defiant disorder, conduct disorder, schizophrenia, fetal alcohol syndrome, previous suicidality, and any previous or ongoing psychiatric care for any reason. A patient was considered a ward of the state if he or she were under a temporary or permanent government order, or described as being under foster care. Those who were a ward of the state were treated as not having a social work consult requested (despite the automatic involvement of such services in our ED), in order to capture the rate of “new” social work consultations. Significant head injury was defined as a skull fracture, intracranial bleed, concussion, or loss of consciousness (if health care professionals felt it to be due to head trauma rather than intoxication). Data that were missing or unclear to the abstractors were both treated as missing, and the associated denominators were adjusted; missing data were not imputed in this study.
Data analyses
Statistical analyses were performed using SAS 9.2 software by SAS Institute Inc., Cary, NC, USA. Associations between substance use and categorical variables (i.e., gender, age group) were detected using chi-square and Fisher’s exact tests. Logistic regression was used to determine the association of psychosocial factors with ward of the state, and odds ratios were calculated with 95% Wald confidence intervals. A p-value <0.05 was considered statistically significant. Because of overlap in substance use, associations for specific substances were calculated for single-drug ingestions only.
RESULTS
A total of 972 charts were potentially eligible for inclusion. This sample represented 1.3% of ED visits (972/73,166) over the study time period. Three hundred and eleven charts (32%) met the exclusion criteria, and 20 charts (2.1%) could not be located or were unreadable due to errors in microfiche processing. Thus, 641 charts were available for abstraction, including 30 charts (4.7%) that had incomplete data due to the patient leaving the ED prematurely.
Demographic data
The median age of our study population was 15 years (interquartile range [IQR] 14–16) with 63.2% (n=405/641) between the ages of 13–15 years. A total of 8% (n=53/641) of patients presented to the ED more than once; of these, 75% (n=40/53) presented twice, whereas the remaining 13 presented between three and nine times during the study period. The psychosocial determinants of health are presented in Table 2.
Demographic characteristics
Presentation to the emergency department
Over 90% of patients arrived via EMS (575/641), including 4.2% (24/575) who had been transferred from another facility. The remaining minority arrived via personal vehicle, police services vehicle, or by foot. For those who arrived via ambulance, with EMS documentation provided (407/575), 63% were activated by non-acquaintances (e.g., bystanders, police, building security, public transit operators). During the study timeframe, patients arrived to the ED accompanied by a family member (26.3%, 168/640), police (17.0%, 109/640), friend/partner (4.2%, 27/640), case worker/social worker (3.8%, 24/640), or unaccompanied aside from EMS personnel (48.8%, 312/640).
The CTAS scores in our study population were as follows: Resuscitation (1)=4.4% (28/637), Emergent (2)=55.6% (354/637), Urgent (3)=27.6% (176/637), Less Urgent (4)=12.2% (78/637), and Non-Urgent (5)=0.2% (1/637). Presenting complaints included decreased level of consciousness (47%, 300/635), erratic, aggressive or unusual behaviour (177/635), injury or assault (17%, 109/635), nausea or vomiting (15%, 94/635), and a variety of other subjective complaints, including hallucinations, palpitations, and dizziness (28%, 175/635).
Of those who had a decreased level of consciousness documented, the Glasgow Coma Score (GCS) was recorded as 3 for 3.5% (10/285), 4 to 8 for 20% (57/285), 9 to 12 for 27% (77/285), 13 or 14 for 31.2% (89/285), and 15 for 18.2% (52/285). Of those who were assaulted (60/641), 70% (42/60) were physically assaulted, and 26.7% (16/60) were sexually assaulted. Injuries sustained included minor head injuries (68%, 98/145), significant head injuries (10%, 14/145), lacerations (7%, 10/145), fractures (4%, 6/145), superficial abrasions and/or bruising (18%, 26/145), and other complaints, which included stabbing, spinal cord injury, and pneumothorax (11%, 16/145).
Substances of abuse
Specific ingestion patterns are presented in Figure 1. Of the cough/cold medications, 32% (7/22) contained dextromethorphan, and 50% (11/22) were of the brand name Coricidin. Other top cough/cold preparations ingested included Vicks Nyquil (n=2), Robitussin (n=2), and Benylin (n=1). Single-drug ingestions occurred in 66.6% of cases (426/640), whereas polydrug ingestions occurred in 33.4% (214/640). Although more ingestions of ethanol were single rather than polydrug ingestions (n=300 v. n=170, respectively), the opposite was true for all other substances. In the cases where the source of the substance(s) could be determined from the chart, disclosed sources included party or peers (n=279/331), family member or family member’s prescription (n=28/331), drug dealer (14/331), or own prescription (10/331). Table 3 outlines the specific prescription drugs reported as used by patients included in our study.
Substances Abused by Patients Presenting to the Pediatric Emergency Department (n=640).
*Included crack cocaine, GHB, heroin, mushrooms, and ketamine.
Prescription medications of abuse (n=37 patients; multidrug ingestions were possible and noted)
* Includes naproxen, ranitidine, quinine, oral contraceptives, and diabetes medication.
Emergency management
Specific interventions provided by EMS or in the ED included IV fluids 67.9% (423/623), medications 41.6% (259/623), intubation 3.5% (22/623), and chest compressions 0.2% (1/623). The provincial poison control centre was contacted in 3.6% (22/623) of all cases. Nine percent of patients (57/623) were admitted; of these, one third were admitted to the pediatric intensive care unit (PICU). In the PICU, another 2 were intubated, and 1 was placed on extracorporeal membrane oxygenation. Of all admissions, 14.6% (31/213) were for polydrug ingestions, whereas 6.4% (26/408) were for single drug ingestions. The admission rate per substance is presented in Figure 1.
Patient disposition
The median length of stay in the ED was 4.58 hours (IQR 4.45 to 4.93). The median length of stay for admitted patients was 2 days (range: 1–3). Thirty-two percent of non-ward of the state patients (161/509) received a new consultation from the ED social worker, and 12.6% (78/619) had a mental health consultation performed in the ED or requested as an outpatient. Other follow-up appointments (not including mental health) were booked for 18% (111/619) of patients. Two deaths were captured during our study period, and both occurred in the PICU, after the patient had left the ED; both patients had ingested methylenedioxymethamphetamine (MDMA).
Demographic trends and correlations
Age: Younger age (<16 years old) was associated with use of cannabis (72.7%, 16/22), prescription medications (69.2%, 9/13), and cough/cold medications (85.7%, 12/14), whereas older age (≥13 years old) was associated with the use of alcohol (96%, 288/300), MDMA (95.1%, 39/41), and other illicit substances (100%, 15/15). This overall association between age and specific single-drug ingestion (n=405) was statistically significant (p=0.048).
Sex: Female patients tended to use more MDMA (68.3%, 28/41) and alcohol (57.3%, 172/300), whereas male patients used more of all other substances (e.g., cannabis) (63.6%, 14/22), other illicit substances (60%, 9/15), prescription medications (61.5%, 8/13), and cough and cold medications (57.1%, 8/14). However, none of the associations between sex and specific single-drug ingestions (n=405) were statistically significant (p=0.067).
GCS: Lower presenting GCS (≤12) was seen with use of ethanol (57.1%, 100/175), whereas higher presenting GCS (13–15) was seen with use of cannabis (83.3%, 5/6), MDMA (100%, 1/1), other illicit substances (66.7%, 2/3), and cough/cold medications (66.7%, 2/3). However, the overall association between GCS and single-drug ingestion (n=196) was not statistically significant (p=0.227).
Admission to hospital: Overall, hospital admission (n=22/405) was associated with specific single-drug ingestions (p<0.001). Hospital admission was also associated with polydrug use, with 14.6% (n=31/214) of those with polydrug ingestion being admitted versus 6.1% (n=26/426) of those with single-drug ingestion (p=0.001).
Assault and injury: Overall, sustaining a physical injury (10.7%, 43/403) was associated with the ingestion of single-specific substances (p<0.001). Ethanol ingestion was associated with more physical injury, whereas all other substances were associated with less injury.
Psychosocial factors: In our study population, those who admitted to a previous substance abuse problem (31%, 193/622) (odds ratio [OR] =1.8, CI 1.2–2.7, p=0.004) and those who reported homelessness (6.4%, 41/639) (OR =2.9, CI 1.5–5.6, p=0.002) were more likely to be wards of the state (20.9%, 132/632). In addition, those who disclosed a pre-existing mental health problem (16.1%, 102/619) were more likely to claim homelessness (6.4%, 41/639) (OR =2.8, CI 1.4–5.6, p=0.005). There was no significant increase in risk of polydrug ingestion with age, previous substance use history, mental health condition, being a ward of the state, or homeless.
DISCUSSION
The ED represents a key point of health care system access for youth who use substances.Reference Burke, O’Sullivan and Vaughan 15 Our study suggests that this group is medically vulnerable, with high rates of morbidity and medical co-morbidity, as well as socially vulnerable, as evidenced by the number of homeless youth and those considered as wards of the state. Many youth presented to the ED without any social supports (only 30% were accompanied by a family member or friend), and, in the majority of cases, EMS was activated by a non-acquaintance. Approximately 17% of youth sustained an injury or an assault prior to their ED presentation, and the majority (60%) required emergent physician assessment (CTAS 1 or 2). This, coupled with the fact that youth substance use is known to predict problematic substance use and poor psychosocial functioning as an adult,Reference Sussman, Skara and Ames 19 - Reference Rohde, Lewinsohn and Seeley 21 suggests that early identification and treatment, which is beyond the acute medical management for youth presenting to the ED with a substance use disorder, are urgently needed.
Although screening/brief intervention/referral-to-treatment has been advocated for routine use in the ED setting,Reference Burke, O’Sullivan and Vaughan 15 , Reference Newton, Gokiert and Mabood 22 such an approach has not been widely implemented. This may be due to the fact that some ED providers feel that asking about alcohol consumption may be seen as offensive, and many feel pressured by a perceived lack of time and resources to deal with these issues.Reference Mabood, Zhou and Dong 23 Despite this, alcohol- and drug-using adolescents have clearly been identified as a population that ED care providers need to be able to expertly manage and link to appropriate resources.Reference Newton, Gokiert and Mabood 22 , Reference Hollman and Alderman 24 Despite these recommendations, the majority of patients were not admitted, and only a minority of patients in our study was referred to mental health or substance use services at the time of ED or hospital discharge.
Emergency physicians need to be alert to the changing patterns of drug use in their community. Prescription drug misuse is a growing problem in Canada.Reference Fischer, Rehm, Goldman and Popova 2 , 3 Canada is the world’s second largest per capita consumer of opioids, one of the most dangerous classes of prescription medications that is used recreationally. 3 Youth have been identified at particularly high risk of harm when using these substances, and existing research has shown that youth view prescription drugs as “safer” than illegal drugs.Reference Twombly and Holtz 25 Of concern in our study is that, although only 6% of presentations were related to prescription drug use, this group had the highest admission rate (38% v. an average admission rate of 6.1% for single-drug ingestions).
Despite awareness of both the immediate and delayed burden of medical and social illness faced by these patients, there remains room for improvement in their care in the ED. Similarly, community-based interventions (e.g., drug screening by primary care providers or school personnel) could be tailored, based on the observed trends in this study. Implementing a systematic way to identify high-risk substance using adolescents and creating connections that ensure a seamless transition from acute care to appropriate community-based treatment options are urgently needed to prevent future morbidity and mortality. While this is an extremely complex problem that cannot be easily solved, the presentation of these youth to the ED must, at the very least, be used as an opportunity to connect them with resources aimed at prevention, including mental health and addictions counseling or a social worker.
LIMITATIONS
Limitations for this study review include many that are inherent to medical record review methodology. These include missing information due to incomplete charts, lack of standardization of the records, and records missed due to incorrect coding. Response bias may have affected this retrospective study due to the sensitive nature of the patients’ medical presentations and their potential reluctance to disclose all details to the treating team. Further, the potential inability for patients to accurately describe what substances that they ingested might have introduced ascertainment bias. Our data were not re-abstracted to check for inter-rater reliability, and one abstractor was not blinded to study hypothesis, due to financial constraint of the study. The repeat visits for 8% (53/641) of our included patients were treated as unique events, because per-patient-based statistical calculations were beyond the scope of this project. Given that the data are from 2007–2009, it is possible that drug abuse trends have changed in the interim. Finally, this study reflects one Canadian pediatric centre and, as such, is not necessarily generalizable to all other centres.
CONCLUSION
Youth who presented to the ED for substance use represented a socially vulnerable population whose use of recreational substances resulted in high medical acuity and significant morbidity. Such youth required acute medical resources and spent significant time in the hospital and ED. Health professionals in the ED may not be meeting their follow-up needs, given how few are documented as having been referred to social workers and mental health services. As such, it behooves us to think about referral systems and how we treat patients with acute intoxication in the ED. Further study of the economic impact of these presentations should be considered, as should the development and study of intervention programs for these high-risk youth.
Acknowledgements
We would like to thank Yvonne Klatt for her assistance in data abstraction, Hitesh Bhatt for his statistical support, and Melissa Gutland for her administrative support for this project. We would also like to acknowledge the Women and Children’s Health Research Institute (WCHRI) for their financial support of this research through the WCHRI Trainee Grant Competition.
Competing interests: This project was funded through the Women and Children’s Health Research Institute Trainee Grant Program (2010). Dr. Newton receives New Investigator funding from the CIHR (Canadian Institutes of Health Research) and Dr. Rosychuk receives funding from AHFMR (Alberta Heritage Foundation for Medical Research).
