Sage Journals: Discover world-class research

Abstract

French

Objective:

The aim of this study is to compare Canadian suicide rates with other external causes of death to examine potential poisoning misclassifications as a contributor to suicide underreporting.

Method:

The study used Statistics Canada mortality data from 2000 to 2011 to calculate sex-and age-specific ratios by external cause of injury codes.

Results:

The overall Canadian suicide rate, as well as the poisoning suicide rate, declined over the study timeframe by an average annual percentage change (AAPC) of 1.0% each year. However, unintentional and undetermined poisonings increased significantly during the timeframe. Unintentional poisoning mortality (primarily narcotics and hallucinogens, including opioids) increased in proportion to suicides for both sexes, although females were consistently higher. The undetermined death to suicide ratio was higher and increasing for females. Poisonings of undetermined intent increased over time to comprise 47% to 80% of the undetermined death category for males and females combined.

Conclusions:

Canadian poisoning suicide rates declined, in contrast to rising unintentional and undetermined poisoning mortality rates. This trend is similar to that of the United States, supporting the hypothesis that misclassification of poisoning deaths may also be an issue in Canada.

Keywords

suicide misclassification poisoning surveillance intentional injury opioids

Clinical Implications

Reliable statistics on suicide are essential to public health monitoring and suicide prevention programs in Canada.

Overall suicide rates have been decreasing in Canada; however, further exploration of the role of poisoning deaths and misclassifications to the overall suicide rate is required to understand this as a true decrease.

Limitations

This is an observational study in which underreporting of poisoning suicides due to misclassification can only be inferred.

Detailed description of type of poisoning agent involved in death is limited by ICD-10 external cause coding.

Canada has reported a modest reduction in overall suicide rates since 2000, with variation by age and sex; however, there were still almost 4000 suicides in 2011.^1
–3 The underreporting of suicide is a recognized concern in Canada and internationally.^4

–11 In particular, Rockett et al.¹² examined the countertrends of a declining suicide rate and an increasing unintentional poisoning mortality rate in the United States between 1987 and 2006 and concluded that misclassification may have contributed to changes in the suicide rate. However, there is no comparable Canadian research. Accurate surveillance of suicide is critical for effective research, prevention, and psychiatric interventions.^13,14

In Canada, all deaths that are unexpected, unexplained, or by violence are investigated by a coroner or medical examiner (C/ME). Coroners and medical examiners function under a provincial or territorial mandate. Although there are minor differences between the provinces, deaths by poisoning and deaths by suicide are investigated to some extent by all Canadian jurisdictions. The final opinions on cause and manner of death are reported to the Vital Statistics agencies of their respective jurisdictions. There is no national death investigative authority and no relevant national standards for medicolegal death investigation. Beginning in 2006, C/ME reports from most provinces have been collected and summarized within the Canadian Coroner and Medical Examiner Database (CCMED) of Statistics Canada. The CCMED still requires the signing of a number of agreements before it can qualify as a national database. The proportion of deaths investigated by a provincial coroner or medical examiner can range from 7% to 45% each year, although this varies greatly across provinces and territories.¹⁵ After an investigation, which may or may not include autopsy, the C/MEs assign a cause of death (e.g., intoxication) and a manner of death (i.e., natural, accident, undetermined, homicide, or suicide). This informs the assignment of an external cause code from the International Statistical Classification of Diseases, 10th Revision (ICD-10).

International studies suggest that suicides are underestimated due to misclassification into other ICD-10 external cause of death categories, particularly unintentional and undetermined intents and when poisoning was the method.^{16

–24} South Korea reported increasing suicide rates following improved accuracy in their manner of death classifications, suggesting an earlier masking of suicides among other categories.²⁵ A recent Canadian study by Auger et al.²⁶ reported that the inclusion of undetermined intent injuries could increase reported suicides by up to 26.5% for men and 37.7% for women; however, this study did not examine suicides misclassified as accidents. Earlier studies have estimated suicide rates to be underreported by 10% to 30%^5,11,20 due to difficulty in obtaining evidence to definitively rule on or diagnose intent for C/MEs, respectively.^27,28

The proposed causes of suicide misclassification are complex and include stigma, frequency, and extent of autopsies; disagreement amongst death investigators on burden of proof; lack of national death reporting standards; variation in local resources; changing attitudes of officials (e.g., police, coroners) and families involved; age and sex or subgroup of the decedent^6,8,10,21; rise in opioid poisonings^24,29,30; and the method or injury mechanism of death.⁷ Violent forms of suicide, including firearms, asphyxia (includes hanging), and jumping, are more likely to be classified as suicide compared with other methods such as poisoning, which may provide more time to intervene, if discovered.^27,31

The purpose of this study is to present Canadian suicide rates between 2000 and 2011 by method or injury mechanism (e.g., poisoning), age, and sex in comparison to other external causes of death (unintentional, undetermined) to examine potential suicide misclassification.

Methods

Mortality data from 2000 to 2011, for ages 15 years and older, by sex, and selected external causes of injury (hanging, strangulation, and suffocation [henceforth referred to as hanging]; firearms; or poisoning, which are the leading causes of suicide) were retrieved from the Public Health Agency of Canada holdings of the Canadian Vital Statistics: Death Database (CVS: D) and the Canadian Socio-Economic Information Management System (CANSIM; accessed 2014 Dec). The external cause of injury or poisoning death was classified according to the ICD-10, adopted in Canada in 2000.

Statistical Analysis

Mortality rates for time trends were age-standardized using the direct method and the 1991 Canadian census population.³² The average annual percentage change (AAPC), which is a weighted average of the annual percent changes (APCs), was calculated using Joinpoint software, providing the magnitude of trends for time periods identified by Joinpoints.³³

Based on the methodology described by Rockett et al.,¹² sex- and age-specific ratios were calculated by external cause of injury codes in each year (2000-2011):

Ratio A (number of poisoning suicides/number of suicides) provides a measure of the relative use of poisonings as a means of suicide and allows observation of the change in this proportion over time.

Ratio B (number of undetermined deaths [all nature of injuries]/number of suicides): by providing a ratio of unclassified intent deaths (undetermined) to those definitively classified as suicides, as well as assuming that some proportion of undetermined cases are in fact unidentified suicides, a crude measure of suicide underenumeration can be obtained. Since it is unlikely that all deaths of unclassified intent are actually misreported suicides, this ratio represents the maximum possible proportion of suicides misclassified as undetermined.

Ratio C (number of unintentional poisonings/number of suicides) provides a second potential measure of suicide underenumeration, assuming that some proportion of unintentional poisonings are misclassified suicides, and again provides the maximal value possible.

Ratio D (number of undetermined poisoning deaths/number of undetermined deaths [all nature of injuries]) provides a measure of what proportion of undetermined deaths are due to poisoning.

Results

The overall age-standardized suicide rate among Canadians 15 years and older has decreased slightly between 2000 and 2011, from 14.4/100,000 to 12.6/100,000 (AAPC, –1.0%; 95% CI, –1.6 to –0.4). The suicide rate decrease is due primarily to the decline in the male-standardized rates from 22.6 to 19.2/100,000 (AAPC, –2.0%; 95% CI, –3.0 to –0.9) while female suicide rates held constant at around 6.0/100,000 during this time period. Among males, the overall decrease is explained by declining rates of suicide by firearms (AAPC, –3.7%; 95% CI, –4.8 to –2.6) and poisoning (AAPC, –1.9%; 95% CI, –3.0 to –0.8). The female suicide rate overall remained flat, as the leading methods of suicide, poisoning and hanging, demonstrated no significant change.

A Joinpoint model was applied to poisoning rates by intent, as illustrated in Figure 1. Unintentional poisonings increased significantly by an APC of 7.4% between 2004 and 2011, while the undetermined poisoning rate jumped between 2000 and 2003 by an APC of 41.8% and then flattened. Suicide poisoning rates decreased by an AAPC of 1.2%, a steady but slight decline over the timeframe.

Figure 1.

Poisoning mortality rates by intent, males and females combined, 15 years and older, Canada, 2000 to 2011 (age-standardized Canada, 1991).

Table 1 provides analysis of poisoning deaths by intent and substance. Drugs account for the majority of all poisoning deaths, regardless of intent (i.e., 83% in 2011, up from 72% in 2000). Both unintentional and undetermined poisonings were consistently and overwhelmingly comprised of drug deaths but experienced no important shift in the proportion of cases with drug involvement or the predominance of male cases (around two-thirds). However, the proportion of alcohol-related deaths did increase to surpass gases/vapours in both these categories. These patterns held when comparing males and females over time. Among intentional poisonings, drug-related deaths for both sexes combined rose from 57% in 2000 to 77% in 2011 and, in contrast to other intents, were equally distributed by sex; however, the proportion of female poisoning suicides involving drugs reached 89% in 2011 compared with 67% among males. This increase was accompanied by a decrease in gas/vapour suicides among both sexes, yet still accounting for 32% of male poisoning suicides in 2011.

Table 1.

Poisoning Deaths in Canada, 2011, by ICD-10 External Cause Code, Males and Females Combined, 15 Years and Older.

Year	Drugs^a		Alcohol		Gases/Vapours
Year	X40-X44	%	X45	%	X46-X49	%	Total
Unintentional poisonings
2000	791	83	70	7	93	10	954
2001	797	86	87	8	66	6	950
2002	865	85	68	7	79	8	1012
2003	751	85	57	6	71	8	879
2004	797	84	75	8	78	8	950
2005	895	88	58	6	63	6	1016
2006	973	90	59	5	48	4	1080
2007	1105	82	151	11	87	6	1343
2008	1149	80	222	15	63	4	1434
2009	1241	82	202	13	63	4	1506
2010	1314	84	188	12	60	4	1562
2011	1378	84	185	11	68	4	1631
	Drugs^a		Alcohol		Gases/Vapours
	X60-X64	%	X65	%	X66-X69	%	Total
Intentional poisonings
2000	499	57	14	2	359	41	872
2001	553	57	15	2	399	41	967
2002	534	60	9	1	342	39	885
2003	613	66	14	2	296	32	923
2004	635	67	7	1	301	32	943
2005	619	68	7	1	284	31	910
2006	566	67	12	1	264	31	842
2007	687	72	14	1	257	27	958
2008	679	73	10	1	244	26	933
2009	717	73	15	2	251	26	983
2010	730	75	10	1	232	24	972
2011	738	77	11	1	212	22	961
	Drugs^a		Alcohol		Gases/Vapours
	Y10-Y14	%	Y15	%	Y16-Y19	%	Total
Undetermined poisonings
2000	161	86	7	4	19	10	187
2001	183	89	10	5	12	6	205
2002	228	88	8	3	23	9	259
2003	425	89	35	7	18	4	478
2004	421	92	22	5	16	3	459
2005	482	91	29	5	19	4	530
2006	407	89	26	6	25	5	458
2007	447	87	45	9	21	4	513
2008	479	90	35	7	19	4	533
2009	488	90	42	8	12	2	542
2010	531	90	38	6	20	3	589
2011	443	89	42	8	10	2	495

ICD-10, International Statistical Classification of Diseases, 10th Revision.

^aIncludes opioids, nonopioids, and hallucinogens.

Figure 2 illustrates unintentional poisonings associated with certain drug categories or alcohol (ICD-10 codes X40-X44, X45), where there was a rise in mortality associated with narcotics and hallucinogens (X42) (which includes opioids), as well as other and unspecified drugs (X44) and alcohol (X45) in 2006, all contributing to the rising unintentional poisoning mortality rates. Although there were around twice as many unintentional poisoning cases among males as females, similar patterns were observed among both sexes.

Figure 2.

Rates of unintentional poisonings associated with drugs and alcohol, ICD-10 codes X40 to X42, X44 to X45, males and females combined, 2000 to 2011, age 15 years and older, Canada (age-standardized Canada, 1991). Drugs: [X40], Nonopioid analgesics, antipyretics and antirheumatics; [X41], Antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs; [X42], Narcotics and psychodysleptics (hallucinogens); [X44], Other/unspecified drugs, medicaments and biological substances; Alcohol: [X45], Alcohol.

Ratio A: Poisoning Suicides/Suicides

Overall, poisoning suicides accounted for about 20% of all male and almost 50% of all female suicides aged 25 years and older; poisoning was a less frequent method of suicide among 15- to 24-year-old males and females (around 10% and 20%, respectively). Over the study period, these proportions manifested no clear trend among either sex (data not shown), with the exception of 25- to 44-year-old males, who registered a slight decrease, and among females 45 years and older, for whom poisoning became a more prevalent method of suicide (50%-60%).

Ratio B: Undetermined Deaths/Suicides

Figure 3 illustrates the ratio of all undetermined intent death classifications (regardless of nature of injury; i.e., drowning, poisoning, firearm) to all suicide death classifications. The male ratio held steady after 2003, whereas the female ratio was higher than that of males and increasing.

Figure 3.

Ratio B: all undetermined intent deaths to all suicides, age 15 years and older, by sex, Canada, 2000 to 2011.

Ratio C: Unintentional Poisoning Deaths/Suicides

Unintentional poisoning deaths rose in relation to suicides beginning in 2003 (Figure 4) for both sexes, although females were consistently higher.

Figure 4.

Ratio C: unintentional poisoning deaths to all suicides, 15 years and older, by sex, Canada, 2000 to 2011.

Ratio D: Undetermined Poisoning Deaths/Undetermined Deaths

The proportion of undetermined deaths, attributed to poisoning, increased from 47% in 2000 to 80% in 2011 (both sexes combined), with females demonstrating a consistently higher ratio over the time period. This is a particular burden in the 24- to 64-year age group, among both males (80%) and females (90%) (data not shown).

Discussion

During the 12-year observation period, the overall poisoning mortality rate increased, due in part to steadily increasing unintentional poisoning rates beginning in 2004, as well as to the undetermined poisoning rates, which doubled initially until 2003 and then flattened. Although suicide poisoning mortality rates decreased slightly overall, this held as a primary method among females but was less prevalent among males. Undetermined poisoning deaths have risen to comprise the large majority of undetermined deaths among both sexes, but again, with a higher burden among females overall, and in particular for males and females between 25 and 64 years of age.

Across all poisoning intent categories, drugs and medications (overdoses) were the primary cause of death, but it was solely suicide poisonings where an increasing proportion of drug involvement was observed. Over the same period, the proportion of alcohol-related deaths increased within the unintentional and undetermined poisoning categories. Alcohol and drug abuse are known risk factors for suicide, oftentimes consumed in combination, leading potentially to interactions, disinhibition, and impulsivity.³⁴ The difficulty in disentangling mixed drug and alcohol cases to identify the primary cause of death is complex^35

–39 and may contribute in some part to this trend.

Canadian research has addressed overall changes during this timeframe in youth suicide rates.¹ The Canadian regulatory warnings on antidepressant prescribing initially in youth (but later expanded to all ages) and the serious implications for youth suicide rates, notably in girls^40
–42 and suicide-related morbidity, were explored.^43,44 However, less attention has been given to changes in poisoning suicide rates among adults and, even more fundamentally, the validity of poisoning rates overall.

C/MEs experience difficulties collecting sufficient evidence to ascertain true intent, particularly in poisoning cases. This is of particular concern in cases where the agents involved (e.g., opioids) may produce a secondary gain (e.g., euphoric effects), whereas the intent of a poisoning death involving carbon monoxide is more readily ascertained. We surmise that this issue primarily affects the validity of suicide and unintentional poisoning deaths and rates, since poisoning homicides are exceedingly rare in Canada. These opposing trends are similar to patterns reported in the United States and England and supports the hypothesis that misclassification of poisoning deaths may also be problematic for Canada. Further research is required to determine the extent to which unintentional poisonings (drug overdoses) are truly increasing and what proportion, if any, are misclassified suicides to improve data reliability. At the extreme limit of possibility, the estimates in this study suggest the proportion of suicides potentially misclassified as undetermined deaths could range from 15% to 25% and from 40% to 60% due to unintentional poisonings in 2011.

The rising use, misuse, and abuse of opioids in North America are well documented,^45

–48 as is the associated suicide risk.^36,38 Distinguishing between an intentional and unintentional overdose can be problematic for C/MEs, particularly in opioid overdose cases.^39,45 Canada was the second largest per capita consumer of prescription opioids in 2013, after the United States.⁴⁹ Canadian researchers examined opioid prescribing rates among those receiving social assistance in Ontario, reporting an increase of 16.2% in the annual prescribing rate between 2003 and 2008. They found a direct association between opioid dose and opioid-related mortality.⁵⁰

In the United States, opioid misuse has been rising since the 1990s, along with an increase in associated mortality,^51,52 and opioid users are reported to have increased rates of substance use disorders in comparison to nonopioid drug users.⁵³ Opioid prescription overdose deaths increased by 415% among females and 265% among males between 1999 and 2010.⁵⁴ Opioid misuse is a growing issue among women,^55,56 as women are more likely to be prescribed opioids than men, and in higher doses, with a more severe clinical profile.^56,57 Opioid overdose is reported as being involved in 1 out of 10 suicides among women in the United States,⁵⁴ but we hypothesize that the true proportion is higher.

It is beyond the scope of this study to claim that the rise in unintentional and undetermined poisonings is due entirely to misclassified suicides or, alternatively, that poisoning suicides are decreasing due to prevention efforts, while accidental overdoses are escalating due to greater access to opioids; the truth, most likely, lies between these 2 extremes. Acceptance that the increase in the unintentional poisoning mortality rate reflects a true rise in accidental drug overdoses would point to an urgent need for prevention that addresses related prescription for pain and possible mood addiction issues, particularly with regards to women. However, if instead the increase is partially a function of suicide misclassification, as found in prior research,^16

–21 then suicide specific prevention strategies, including restricting access to lethal means,^58,59 can be added to the ameliorative mix. Overall, the explanation for the rise may include changes in a complex pathway beginning with substance use, including increased access to drugs, developing into abuse and associated physiological and social consequences, and ending, in some as yet indeterminate number of suicides.^37,39

A high burden of proof is necessary for C/MEs to categorize a death as suicide.^60,61 The subjective and complex nature of investigating and determining intent, particularly in poisoning cases, is made more complicated by the nature of the premorbid emotional/physical state of the individuals in question, who themselves may not be completely aware of their own intent prior to the event^16,30,39,60 or who may experience ambivalence towards death.^24,61,62 When a suicide cannot be definitively declared, the only remaining injury manner-of-death options are unintentional or undetermined. This uncertainty induces underreporting of suicide and overreporting of the other intents.³⁰ The proportion of deaths subject to autopsy (e.g., may include toxicological and microscopic analysis) in Canada declined from 9.9% in 2000 to 4.8% in 2011 (SC Table 102-0510), further subjecting suicides to misclassification.^6
–8 It is difficult to ascertain if there have been changes over time in the C/ME assessments, as procedures and standards differ by province. No standard protocol is followed.

An international multidisciplinary research team has proposed a new working definition, death from drug self-intoxication (DDSI), to inform users of data from the US Centers for Disease Control and Prevention’s (CDC) National Violent Death Reporting System (NVDRS). Pertinent to poisoning mortality cases, estimated DDSIs comprise all fatal drug self-intoxications whose manner of death was suicide and the great preponderance classified as undetermined or accident.⁶¹ An enhanced category of self-injury mortality, comprising suicides by any method and nonsuicide DDSIs, would represent the 8^th leading cause of death in the US versus 10^th in the case of suicide alone.⁶² Drug use/abuse is a high-risk behaviour that is widely understood to increase the probability of death, leading to difficulties in classifying these cases as accidents, when the complex intent of these deaths may fall along a continuum between purely intentional and unintentional.⁶³ In Canada, we recommend the inclusion of standardized information by the C/MEs, which would then further enrich the CCMED and inform prevention.⁶⁴ We recommend further study and consensus to determine the best approach for Canada.

This study has a number of limitations. As it is a retrospective observational study, it can only describe Canadian trends and suggest potential explanations. Currently, diagnostic codes that detail the specific nature of injury and associated agent or agents for mortality data are unavailable for analysis in Canada and therefore prevent identification of specific drug involvement (i.e., opioid drugs cannot be separated from other types within the X42 category). The administrative data employed provide limited variables and prevent more complex analysis involving potential comorbidities. Unlike the United States, Canada does not yet have a national violent death reporting system that combines multiple data sources, including state and local medical examiner, coroner, law enforcement, crime lab, and vital statistics records, into one surveillance tool. However, the CCMED¹⁵ may be used in the future to help fill this void.

Conclusion

Reliable and valid mortality statistics are essential to public health and clinical medicine. Without this foundation, suicide and other injury surveillance, etiologic research, treatment, rehabilitation, policy, prevention, and evaluation are severely compromised. The findings of this study recommend further research to quantify any potential misclassification and remediate, if necessary, to prevent future error that limits comparability both within Canada, as well as internationally. Consideration should be given in Canada to the need for continual standardized monitoring and analysis of poisoning deaths associated with drugs, across the continuum of decedent intentionality, to enhance prevention and achieve a more accurate accounting of suicide and other self-injury mortality.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

References

Skinner

McFaull

. Suicide among children and adolescents in Canada: trends and sex differences, 1980-2008. CMAJ. 2012;184(9):1029–1034.

Navaneelan

Suicide rates: an overview

[Internet]. Ottawa (ON): Statistics Canada; 2012. [cited 2014 Sep 20]. Available from: http://www.statcan.gc.ca/pub/82-624-x/2012001/article/11696-eng.pdf.

Statistics Canada. Table 102-0551—Deaths and mortality rate, by selected causes, by sex and age group, Canada, CANSIM (database) [Internet]. [cited 2015 Sep 4]. Available from: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/hlth66a-eng.htm.

O’Carroll

. A consideration of the validity and reliability of suicide mortality data. Suicide Life Threat Behav. 1989;19(1):1–16.

Speechley

Stavraky

. The adequacy of suicide statistics for use in epidemiology and public health. Can J Public Health. 1991;82(1):38–42.

Kapusta

Tran

Rockett

. Declining autopsy rates and suicide misclassification: a cross-national analysis of 35 countries. Arch Gen Psychiatry. 2011;68(10):1050–1057.

Rockett

IRH

. Counting suicides and making suicide count as a public health problem. Crisis. 2010;31(5):227–230.

Rockett

IRH

Kapusta

Bhandari

. Suicide misclassification in an international context: revisitation and update. Suicidology Online. 2011;2:48–61.

Tollefsen

Hem

Ekeberg

. The reliability of suicide statistics: a systematic review. BMC Psychiatry. 2012;12:9.

10.

De Leo

Dudley

Aebersold

. Achieving standardised reporting of suicide in Australia: rationale and program for change. MFA. 2010;192(8):452–456.

11.

Donaldson

Larsen

Fullerton-Gleason

. Classifying undetermined poisoning deaths. Inj Prev. 2006;12:338–343.

12.

Rockett

IRH

Hobbs

De Leo

. Suicide and unintentional poisoning mortality trends in the United States, 1987-2006: two unrelated phenomena? BMC Public Health. 2010;10:705.

13.

Scott

Swartz

Warburton

. The psychological autopsy: solving the mysteries of death. Psychiatr Clin North Am. 2006;29(3):805–822.

14.

Potter

Powell

Kachur

. Suicide prevention from a public health perspective. Suicide Life Threat Behav. 1995;25(1):82–91.

15.

Canadian coroner and medical examiner database: annual report, 2006-2008 [Internet]. Ottawa (ON): Statistics Canada and Public Health Agency of Canada. [cited 2014 Sep 25]. Available from: www.statcan.gc.ca/pub/82-214-x/82-214-x2012001-eng.pdf.

16.

Breiding

Wiersema

. Variability of undetermined manner of death classification in the US. Inj Prev. 2006;12(Suppl 2):ii49–ii54.

17.

Linsley

Schapira

Kelly

. Open verdict v. suicide: importance to research. Br J Psychiatry. 2001;178:465–468.

18.

Mohler

Earls

. Trends in adolescent suicide: misclassification bias? Am J Public Health. 2001;91:150–153.

19.

O’Donnell

Farmer

. The limitations of official suicide statistics. Br J Psychiatry. 1995;166:458–461.

20.

Ohberg

Lonnqvist

. Suicides hidden among undetermined deaths. Acta Psychiatr Scand. 1998;98:214–218.

21.

Sorenson

Shen

Kraus

. Undetermined manner of death: a comparison with unintentional injury, suicide, and homicide death. Eval Rev. 1997;21:43–57.

22.

Gunnell

Bennewith

Simkin

. Time trends in coroners’ use of different verdicts for possible suicides and their impact on officially reported incidence of suicide in England: 1990-2005. Psychol Med. 2013;43:1415–1417.

23.

Whitt

. Where did the bodies go? The social construction of suicide data, New York City, 1976-1992. Sociol Inq. 2006;76(2):166–187.

24.

Bohnert

McCarthy

Ignacio

. Misclassification of suicide deaths: examining the psychiatric history of overdose decedents. Inj Prev. 2013;19(5):326–330.

25.

Chan

Caine

Chang

. The impact of improving suicide death classification in South Korea: a comparison with Japan and Hong Kong. PLoS One. 2015;10(5):e01257 30.

26.

Auger

Burrows

Gamache

. Suicide in Canada: impact of injuries with undetermined intent on regional rankings. Inj Prev. 2016;22:76–78.

27.

Rockett

IRH

Kapusta

Coben

. Beyond suicide: action needed to improve self-injury mortality accounting. JAMA Psychiatry. 2014;71:3.

28.

Maybank

Bowes

Barber

. Drug overdose deaths in Halifax, Nova Scotia from 1996 to 2010: a descriptive and comparative study [Internet]. [cited 2015 July 24]. Available from: http://library2.smu.ca/handle/01/24891#.VNETD2dp2oo.

29.

Drug Abuse Warning Network. National estimates of drug-related emergency department visits [Internet]. Rockville (MD): US Department of Health and Human Services; 2011. [cited 2015 Jul 7]. Available from: http://www.samhsa.gov/data/sites/default/files/DAWN2k11ED/DAWN2k11ED/DAWN2k11ED.pdf.

30.

Suicide among adults aged 35-64—United States, 1999-2010. MMWR. 2013;62(17)321–325.

31.

Parai

Kreiger

Tomlinson

. The validity of the certification of manner of death by Ontario coroners. Ann Epidemiol. 2006;16(11):805–811.

32.

Statistics Canada. Table 051-0001—estimates of population, by age group and sex for July 1, Canada, provinces and territories, annual (persons unless otherwise noted), CANSIM (database) [Internet]. [cited 2014 Nov 26]. Available from: http://www5.statcan.gc.ca/cansim/a26?lang=eng&id=510001.

33.

Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute. Joinpoint regression program, version 4.1.1—August 2014. Bethesda, MD: National Cancer Institute.

34.

Division of Violence Prevention, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. Suicides due to alcohol and/or drug overdose [Internet]. [cited 2015 Aug 9]. Available from: http://www.cdc.gov/ViolencePrevention/pdf/NVDRS_Data_Brief-a.pdf.

35.

Cavanagh

Carson

Sharpe

. Psychological autopsy studies of suicide: a systematic review [published correction appears in Psychol Med. 2003;33:947]. Psychol Med. 2003;33:395–405.

36.

Wilcox

Conner

Caine

. Association of alcohol and drug use disorders and completed suicide: an empirical review of cohort studies. Drug Alcohol Depend. 2004;76(Suppl):S11–S19.

37.

Miller

Mahler

Gold

. Suicide risk associated with drug and alcohol dependence. J Addict Dis. 1991;10(3):49–61.

38.

Roy

. Characteristics of drug addicts who attempt suicide. Psychiatry Res. 2003;121:99–103.

39.

Bohnert

ASB

Roeder

Ilgen

. Unintentional overdose and suicide among substance users: a review of overlap and risk factors. Drug Alcohol Depend. 2010;110:183–192.

40.

Katz

Kozyrskyj

Prior

. Effect of regulatory warnings on antidepressant prescription rates, use of health services and outcomes among children, adolescents and young adults. CMAJ. 2008;178(8):1005–1111.

41.

Rhodes

Skinner

McFaull

. Canada-wide effect of regulatory warnings on antidepressant prescribing and suicide rates in boys and girls. Can J Psychiatry. 2013;58(11):640–645.

42.

Isacsson

Rich

. Antidepressant drugs and the risk of suicide in children and adolescents. Paediatr Drugs. 2014;16(2):115–122.

43.

Rhodes

Skinner

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

44.

Rhodes

Bethell

Carlisle

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

45.

Cantor

McTaggart De Leo

. Misclassification of suicide-the contribution of opiates. Psychpathology. 2001;34:140–146.

46.

King

Fraser

Richardson

Harper

. Determinants of increased opioid-related mortality in the United States and Canada, 1990-2013: a systematic review. Am J Public Health. 2014;104(8):e32–e42.

47.

Centers for Disease Control and Prevention (CDC). Vital signs: overdoses of prescription opioid pain relievers—United States, 1999-2008. MMWR Morb Mortal Wkly Rep. 2011;60:1487–1492.

48.

Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, Division of Unintentional Injury Prevention. Policy impact: prescription painkiller overdoses [Internet]. [cited 2014 Dec 5]. Available from: www.cdc.gov/homeandrecreationalsafety/rxbrief/.

49.

Report of the International Narcotics Control Board for 2013 [Internet]. Vienna (Austria): International Narcotics Control Board; 2013. Available from: www.incb.org/documents/Publications/AnnualReports/AR2013/English/AR_2013_E.pdf.

50.

Gomes

Juurlink

Dhalla

. Trends in opioid use and dosing among socio-economically disadvantage patients. Open Med. 2011;5(1):e13–e22.

51.

Warner

Chen

Makuc

. Increase in fatal poisonings involving opioid analgesics in the United States, 1999-2006 [Internet]. NCHS Data Brief 22:1-8. [cited 2015 Jul 2]. Available from: http://www.cdc.gov/nchs/data/databriefs/db22.pdf.

52.

Denisco

Chandler

Compton

. Addressing the intersecting problems of opioid misuse and chronic pain treatment. Exp Clin Psychopharmacol. 2008;16(5):417–428.

53.

Woody

Yang

. How do prescription opioid users differ from users of heroin or other drugs in psychopathology: results from the National Epidemiologic Survey on Alcohol and Related Conditions. J Addict Med. 2011;5(1):28–35.

54.

MMWR, vital signs: overdoses of prescription opioid pain relievers and other drugs among women—United States, 1999-2010 [Internet]. Atlanta (GA): Centers for Disease Control and Prevention; 2013 [cited 5 June 2015]. Available from: http://www.cdc.gov/mmwr/preview/mmwrhtml/mm6226a3.htm.

55.

Wastila

Ritter

Strickler

. Gender and other factors associated with the nonmedical use of abusable prescription drugs. Simoni Subst Use Misuse. 2004;39(1):1–23.

56.

Greenfield

Back

Lawson

Brady

. Substance abuse in women. Psychiatr Clin North Am. 2010;33(2):339–355.

57.

Williams

Sampson

Kalilani

. Epidemiology of opioid pharmacy claims in the United States. J Opioid Manage. 2008;4(3):145–152.

58.

Sarchiapone

Mandelli

Iosue

. Controlling access to suicide means. Int J Environ Res Public Health. 2011;8(12):4550–4562.

59.

Hawton

Bergen

Waters

. Impact of withdrawal of the analgesic Co-proxamol on nonfatal self-poisoning in the U.K. Crisis. 2011;32(2):81–87.

60.

Silverman

Berman

Sanddal

. Rebuilding the Tower of Babel: a revised nomenclature of the study of suicide and suicidal behaviors part 1: background, rationale and methodology. Suicide Life Threat Behav. 2007;37(3):264–277.

61.

Rockett

IRH

Smith

Caine

. Confronting death from drug self-intoxication (DSSI): prevention through a better definition. Am J Public Health. 2014;104(12):e49–e55.

62.

Rockett

IRH

Caine

. Self-injury is the eighth leading cause of death in the United States: it is time to pay attention. JAMA Psychiatry. 2015;72(11):1069–1070.

63.

Rockett

IRH

Hobbs

. Variable classification of drug-intoxication suicides across US states: a partial artifact of forensics? PLoS ONE. 2015;10(8):e0135296.

64.

Campbell

Jackson

Basset

Bowes

Donahue

Cartwright

. Can we use medical examiners’ records for suicide surveillance and prevention research in Nova Scotia? CDIC. 2011;31(4):165–171.