Time Trends in Emergency Department Visits for Suicide-Related Behaviours by Girls and Boys in Alberta

Abstract

French

Objective:

In Canada, emergency departments (EDs) are a frontline setting for treating suicide-related behaviours (SRBs) among adolescents, yet description of national trends in ED SRB visits is lacking. We determined whether the SRB incidence rate and method patterns between 2002 and 2010 previously shown for Ontario adolescents were also experienced in Alberta.

Method:

A retrospective, population-based study of ED visits for SRBs (self-poisoning or self-injury, irrespective of suicidal intent) by 12- to 17-year-olds was conducted using administrative health care data from 104 EDs across Alberta, Canada. Incidence rates and 95% confidence intervals (CIs) were calculated and graphed. Rate ratios (RRs) comparing rates between time periods (2002-2005 and 2006-2010) and corresponding 95% CIs were estimated. Changes in SRB methods were also described. The time periods chosen were based on published Ontario trends.

Results:

Decreases in yearly incidence rates levelled off after 2005. Crude RRs indicated a rate decrease in 2006 to 2010 for boys (RR, 0.77; 95% CI, 0.65 to 0.90) and girls (RR, 0.80; 95% CI, 0.67 to 0.95). From 2002 to 2010, the proportion of SRB visits for self-poisoning decreased (girls, –13%; boys, –10%) while visits for self-cutting increased (girls, +13%; boys, +14%).

Conclusions:

Alberta trends were similar to those previously published for Ontario. Determining if the trends and observed changes are associated with mental health care access or availability and/or provincial suicide prevention strategies would contextualize these findings and could shape future prevention efforts. Lack of identification of suicidal intent and exclusion of fatal SRB are limitations of the current study.

Keywords

deliberate self-harm emergency department child adolescent

Clinical Implications

For both boys and girls, over time, self-poisoning made up a smaller proportion of SRB ED presentations in Alberta while self-cutting became more prominent.

Prevention strategies could be informed by determining if changes in SRB incidence rates and methods reflect changes in health care access or availability and/or prevention initiatives.

Limitations

Administrative health care data do not permit determination of suicidal intent for SRBs.

Individual-level changes such as increased frequency of repeat SRB or increased severity of injury from SRB could not be determined from population-level administrative health care data.

Introduction

Suicide-related behaviours (SRBs) are nonfatal self-poisonings or self-injuries irrespective of suicidal intent.^1,2 Emergency departments (EDs) are a frontline hospital setting for treating SRBs. Among adolescents, there are approximately 25 SRB ED visits for every documented youth suicide.^3,4 All-cause mortality for 15- to 24-year-olds presenting to hospital for SRBs is 4 times higher than for the general population, and suicide risk is roughly 10 times higher.⁵ Although SRB method may be an indicator of adolescent risk of suicide,⁶ time trends in SRB method among adolescents who receive ED care have been studied infrequently.

Description of adolescent ED SRB incidence rates and method patterns across Canada are lacking. Recent national reports of ED visits for SRB, grouped SRB, and fatal SRB (suicide) together and presentation of interprovincial trends are absent.^7
–9 A recent Ontario-based study of provincial ED SRB time trends between 2002 and 2011 identified a 30% decrease in ED SRB visit rates among 12- to 17-year-olds between 2002 and 2007 and a leveling off thereafter, although incident events increased in severity.^10,11 The Ontario data revealed that the proportion of SRB ED visits by boys for self-poisoning and self-cutting/piercing remained fairly constant over time, but the proportion of SRB by other methods decreased. For girls, for the study period of 2002 to 2007, the proportion of SRB ED visits for self-cutting initially increased but then decreased, with an inverse pattern observed for self-poisoning.¹⁰ Comparable studies from other Canadian provinces and territories are lacking.

Our study objective was to determine whether 2002 to 2010 Alberta ED SRB visit rates for boys and girls and proportions of SRB methods were similar to those previously published for Ontario.¹⁰ This reflects an interprovincial collaboration to examine ED SRB presentations among adolescents in Alberta and Ontario using comprehensive, provincial ambulatory care data sets. We used administrative health care data from 104 EDs across Alberta, Canada, from 2002 to 2010 to examine the pattern of Alberta ED SRB visit rates for boys and girls and to measure proportions of SRB methods over time.

Methods

Patient Population and Data Sources

This population-based retrospective cohort study used linked administrative databases. The underlying cohort (n = 646,975) comprised all 12- to 17-year-olds living in Alberta from April 1, 2002, to March 31, 2010, identified within the Alberta Health Care Insurance Plan cumulative population registry. Demographic (sex, residential postal code) and population data were drawn from the registry.

Similar to the Ontario-based study,¹⁰ adolescents with an ED SRB visit within the study timeframe were identified using the ED diagnostic fields of the Ambulatory Care Classification System (ACCS) database. The ACCS records ambulatory data for the province, including ED visit data.¹² Only the individual’s first ED SRB visit during the study period (index event) was retained for analysis. SRB method was identified in the ACCS database according to World Health Organization diagnostic codes¹³ (ICD-10-CA): X60 to X69 (poisoning only), X78 (cut or pierce only), X70 to X77 and X79 to X84 (other only), and multiple (more than 1 method). While SRB intent (i.e., behaviour with suicidal vs. nonsuicidal intent) is assessed by ED clinicians, this intent is not captured in the ACCS database—hence Silverman et al.’s classification of “undetermined SRB” applies to our study.¹⁴ Individuals who died at the time of the index event (n = 16) were removed from the cohort, leaving a study cohort of 4480 adolescents.

Research ethics approval was granted from the University of Alberta (Edmonton, Alberta).

Demographic Definitions

For each study subject, community population size and neighbourhood income quintile were determined using linkage to Statistics Canada Postal Code Conversion Files.¹⁵ Community population size was categorized as 1,500,000 or more, 500,000 to 1,499,999, 100,000 to 499,999, 10,000 to 99,999 (for any census metropolitan area [CMA] or census agglomeration [CA] less than 100,000), or 10,000 or less (any non-CMA/CA). Neighbourhood income quintiles were determined by ranking dissemination areas by average income (adjusted for household size) within each CMA, CA, or non-CMA/CA and then dividing the population into approximate fifths.

Statistical Analysis

Index events were summarized by SRB method(s) documented in any diagnostic field. Annual ED SRB incidence rates were calculated and graphed, by fiscal year, for boys and girls. Rates were expressed per 100,000 person years, with 95% confidence intervals (CIs). Negative binomial regression was used to compare incidence rates in 2002 to 2005 with incidence rates in 2006 to 2010 for boys and girls, after adjusting for age, community size (<100,000 vs. 100,000+), and neighbourhood income quintiles (highest income quintile vs. other). Time periods and categories were chosen to allow direct comparisons with published Ontario trends.¹⁰ Crude rates, adjusted rates, rate ratios (RRs), and the corresponding 95% CIs were calculated. Numbers and percentages of SRB methods (and corresponding 95% CIs) at the index event were calculated by fiscal year for boys and girls.

Results

SRB incidence rates are presented in Figure 1. Comparing rates from 2006 to 2010 with rates from 2002 to 2005, the crude RRs were 0.77 (95% CI, 0.65 to 0.90) and 0.80 (95% CI, 0.67 to 0.95) for boys and girls, respectively (Table 1). Results were similar after adjusting for age, community size, and neighbourhood income quintiles. From 2002 to 2010, the proportion of ED visits for self-poisoning decreased (girls, –13%; boys, –10%), while the proportion for self-cutting increased (girls, +13%; boys, +14%) (Table 2). The proportion of ED visits for other SRB methods remained fairly consistent for both boys and girls.

Figure 1.

Incidence rates of emergency department suicide-related behaviour visits by girls and boys from 2002/2003 to 2009/2010.

Table 1.

Incidence Rate of Suicide-Related Behaviour in Time 1 (2006-2007 to 2009-2010) Compared to Time 2 (2002-2003 to 2005-2006), in Boys and Girls in Alberta.

	Girls			Boys
	Time 2 (95% CI)^a	Time 1 (95% CI)^a	Time 2 vs. Time 1 RR (95% CI)^a	Time 2: (95% CI)^a	Time 1: (95% CI)^a	Time 2 vs. Time 1 RR (95% CI)^a
Model 1, crude, n = 8^b	257.01 (226.79–291.26)	323.12 (285.75–365.41)	0.80 (0.67–0.95)	93.97 (83.61–105.63)	122.21 (109.46–136.45)	0.77 (0.65–0.90)
Model 2, age-adjusted, n = 48^c	213.43 (196.93–231.35)	272.14 (252.18–293.72)	0.78 (0.71–0.87)	68.60 (61.84–76.11)	90.06 (81.83–99.12)	0.76 (0.68–0.85)
Model 2, community size-adjusted, n = 16^d	263.26 (236.80–292.69)	328.27 (296.02–364.06)	0.80 (0.69–0.93)	96.09 (85.15–108.44)	123.87 (110.47–138.90)	0.78 (0.66–0.92)
Model 2, income-adjusted, n = 16^e	230.26 (206.57–256.68)	301.16 (271.49–334.03)	0.76 (0.66–0.89)	83.74 (74.70–93.89)	109.15 (98.18–121.34)	0.77 (0.67–0.88)

Abbreviations: CI, confidence interval; RR, rate ratio.

^arate per 100,000 person years.

^b8 years.

^c8 years by 6 age groups.

^d8 years by 2 community size groups.

^e8 years by 2 income groups.

Table 2.

Summary of Suicide-Related Behaviour Method^a at Index Visit by Year, Percentage (95% Confidence Interval).

	No. of Emergency Department Visits	Self-Cutting	Self-Poisoning	Other
Total	4480
Girls
2002-2003	565	17 (15 to 21)	77 (74 to 80)	5 (3 to 7)
2003-2004	431	22 (20 to 28)	72 (67 to 76)	4 (2 to 6)
2004-2005	424	29 (26 to 35)	64 (60 to 69)	4 (3 to 7)
2005-2006	363	34 (31 to 40)	57 (52 to 62)	6 (5 to 10)
2006-2007	362	37 (34 to 44)	56 (51 to 61)	4 (2 to 7)
2007-2008	335	29 (26 to 35)	64 (59 to 70)	5 (3 to 7)
2008-2009	393	31 (28 to 37)	61 (56 to 66)	6 (4 to 9)
2009-2010	341	30 (25 to 35)	64 (59 to 69)	6 (3 to 8)
Boys
2002-2003	218	19 (15 to 26)	60 (53 to 66)	20 (14 to 25)
2003-2004	184	22 (17 to 29)	57 (50 to 64)	18 (14 to 30)
2004-2005	160	26 (21 to 35)	49 (42 to 57)	22 (16 to 30)
2005-2006	149	34 (27 to 42)	44 (36 to 52)	21 (15 to 28)
2006-2007	135	33 (25 to 41)	48 (40 to 57)	19 (12 to 25)
2007-2008	133	33 (25 to 41)	47 (38 to 55)	20 (13 to 27)
2008-2009	146	23 (16 to 30)	64 (56 to 71)	12 (8 to 18)
2009-2010	141	33 (25 to 40)	50 (42 to 59)	16 (11 to 23)

^aMultiple methods were reported in 61 visits for girls and 18 visits for boys. We do not report these visits by year due to small cell sizes. Reporting these small numbers could potentially identify a child/event when you take into account the other data presented in the article.

Discussion

The 2 key findings in this Alberta-wide study—a decline to 2005 and then a leveling of yearly SRB incidence rates and changing trends in self-cutting and self-poisoning during the study period—are trends also reported for Ontario.¹⁰ These studies provide the first direct comparisons of ED SRB incidence rate and method patterns among young people with ED visits across Canada. Below we consider temporal associations for the Alberta findings and make recommendation for further study.

In this study, the leveling of ED SRB rates coincided with the Alberta Suicide Prevention Strategy.¹⁶ While the effects of the strategy on SRB rates are unknown, it is possible that, introduction of the strategy mitigated against a SRB rate increase. Such a mitigating phenomenon has been suggested with the introduction of Local Health Integration Networks in Ontario.¹⁰ The Alberta suicide prevention strategy was introduced across the province in 2006 with specific goals to improve intervention and treatment for those at risk of suicide and to reduce access to lethal means of suicide. Work by others has established the validity of ED visit rates as a proxy for effectiveness of suicide prevention and reported decreased SRB ED visit rates following suicide prevention strategy implementation.¹⁷ Examining the relationship of provincial policy to ED visits for SRBs was outside the scope of this project. The SRB ED visit rate trends found in the current study could likewise reflect changes in the availability or accessibility of local health services or school-based programs. Future investigations could generate specific recommendations to guide suicide prevention policy and interventions.

The meaning and implications of the observed trends in ED SRB methods are worth further investigation. Among UK youth, ED SRB visits for self-cutting have been associated with increased risk of subsequent suicide compared to those with visits for self-poisonings.⁶ These findings did not rely on identification of suicidal intention; similar to our study, suicidal intent was not identified. Further study of ED visits for self-cutting with and without known suicidal intent could inform ED clinical care (including disposition planning) and practice recommendations for post-ED follow-up care. In Alberta, closer examination of SRB methods trends and specific Alberta suicide prevention strategies to reduce access to lethal means and self-poisoning could clarify this study’s findings and guide future prevention interventions. When considering trends in SRB methods, it is important to consider the impact of antidepressant prescribing patterns. For instance, 1 UK-based study has noted a reduction in both prescribing and self-poisoning with antidepressants in 12- to 19-year-olds following reduced medication availability subsequent to a UK-issued regulatory warning.¹⁷ The impact of antidepressant prescribing on SRB rates, however, is complex since both prescribing (availability of medication for self-poisoning, antidepressant-associated suicidal ideation) and not prescribing (untreated depression) can potentially lead to increased rates of SRB. While these population-level temporal associations have been reported by several authors,^18,19 it remains an ecological phenomenon without individual-level verification.

Limitations

This study has several limitations. First, it was impossible for us to distinguish between intent (suicidal vs. nonsuicidal behaviours) using International Classification of Diseases diagnostic codes. For our study, ED SRBs were not assessed according to suicidal intent and medical severity, and we are unclear on how the implications of study findings will differ according to these characteristics. While ED SRBs have been associated with an increase in all-cause mortality and suicide risk,⁵ whether this outcome/risk is different among subpopulations of young people who engage in SRB is important to study. Second, we did not have data on use of other health services (i.e., primary care, mental health services) and are not able to interpret ED SRB visit trends in the context of other service use. Third, we studied ED visits for behaviours; therefore, our results do not generalize to visits where ideation was present in the absence of enacted behaviour (i.e., ED visit by a youth with anxiety disorder who also had thoughts of self-harm) and may not reflect SRBs in the general adolescent population.

Conclusions

Alberta ED SRB visit rates for 2002 to 2010 for boys and girls and proportions of SRB methods were similar to those previously published for Ontario. Determining if the trends and observed changes are associated with mental health care access or availability and/or provincial suicide prevention strategies would clarify findings from these studies and could shape future prevention efforts. Study of SRB with suicidal intent and fatal SRB is also necessary to understand whether trends, outcomes, and risks are different among subpopulations of young people who engage in SRB.

Footnotes

Role of the Sponsor

The sponsors had no role in the specific conduct of the review; in the collection, management, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.

Disclaimer

This study is based in part on data provided by Alberta Health. The interpretation and conclusions contained herein are those of the researchers and do not necessarily represent the views of the Government of Alberta. Neither the Government of Alberta nor Alberta Health expresses any opinion in relation to this study.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this project was provided by the Canadian Institutes of Health Research (CIHR) (SEC 117128). Dr. Newton holds a CIHR New Investigator Award. Dr. Rosychuk is a Health Scholar with Alberta Innovates–Health Solutions.

References

National Collaborating Centre for Mental Health. Self-harm: the short-term physical and psychological management and secondary prevention of self-harm in primary and secondary care. Leicester, UK: British Psychological Society & The Royal College of Psychiatrists; 2004.

Silverman

Berman

Sanddal

. Rebuilding the Tower of Babel: a revised nomenclature for the study of suicide and suicidal behaviors. Part 2: suicide-related ideations, communications and behaviors. Suicide Life Threat Behav. 2007;37(3):264–277.

Bethell

Bondy

Lou

WYW

. Emergency department presentations for self-harm among Ontario youth. Can J Public Health. 2013;104(2):e124–e130.

Statistics Canada. Suicides and suicide rate, by sex and age group [Internet] [cited 2015 Feb 20]. Available from: www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/hlth66d-eng.htm.

Hawton

Harriss

. Deliberate self-harm in young people: characteristics and subsequent mortality in a 20-year cohort of patients presenting to hospital. J Clin Psychiatry. 2007;68(10):1574–1583.

Bergen

Hawton

Waters

. How do methods of non-fatal self-harm relate to eventual suicide? J Affect Disord. 2012;136:526–533.

Parachute. The cost of injury in Canada. Toronto (ON): Parachute; 2015.

Canadian Institute for Health Information. Self-harm and assault: a closer look at children and youth. Ottawa (ON): CIHI; 2014.

Canadian Institute for Health Information. Health indicators 2011. Ottawa (ON): CIHI; 2011.

10.

Rhodes

Bethell

Carlisle

. Time trends in suicide-related behaviours in boys and girls. Can J Psychiatry. 2014;59(3):152–159.

11.

Rhodes

Skinner

. Time trends in medically serious suicide-related behaviours in boys and girls. Can J Psychiatry. 2014;59(10):556–560.

12.

Alberta Health and Wellness. Ambulatory care in Alberta using ambulatory care classification system data. Edmonton (AB): Alberta Health and Wellness; 2004.

13.

World Health Organization. International statistical classification of diseases and related health problems [Internet] [cited 2011 May 6]. Available from: www.who.int/classifications/icd/en/.

14.

Silverman

Berman

Sanddal

O’carroll

Joiner

. Rebuilding the tower of Babel: a revised nomenclature for the study of suicide and suicidal behaviors: Part 2: suicide-related ideations, communications, and behaviors. Suicide Life Threat Behav. 2007;37:264–277.

15.

Wilkins

. PCCF+ version 5E user’s guide (geocodes/PCCF) automated geographic coding based on Statistics Canada postal code conversion files. Ottawa (ON): Statistics Canada; 2009.

16.

Alberta Mental Health Board. A call to action: the Alberta Suicide Prevention Strategy [Internet] [cited 2015 Mar 31]. Available from: http://www.albertahealthservices.ca/MentalHealthWellness/hi-mhw-sps-main-2006-03.pdf.

17.

Bergen

Hawton

Waters

. Epidemiology and trends in non-fatal self-harm in three centres in England: 2000-2007. Br J Psychiatry. 2010;197:493–498.

18.

Gibbons

Hur

Bhaumik

Mann

. The relationship between antidepressant prescription rates and rate of early adolescent suicide. Am J Psychiatry. 2006;163:1898–1904.

19.

Gibbons

Brown

Hur

. Early evidence on the effects of regulators’ suicidality warnings on SSRI prescriptions and suicide in children and adolescents. Am J Psychiatry. 2007;164:1356–1363.