Economic Insights
Employment and Social Assistance Receipt Among Overdose Fatalities in British Columbia

This article in the Economics Insights series documents the employment histories and income sources of people who died of an illicit drug overdose in British Columbia from January 1, 2007, to December 31, 2016. The data are from the British Columbia (BC) Coroners Service and from administrative data files. The majority of people who died from an illicit drug overdose in British Columbia appear to have occupied a position of economic vulnerability for a considerable portion of their lives, characterized by low and decreasing employment and increasing reliance on social assistance. There are also marked differences in employment, earnings, and social assistance receipt between women and men, younger and older age groups, and between subsets of male overdose fatalities who had sustained, marginal, or near-absent levels of labour force attachment.

Introduction

From 2011 to 2016, there were over 2,350 fatalities from illicit drug overdoses^Note  in British Columbia (Statistics Canada 2018). The British Columbia (BC) Coroners Service (2018) reported that one-fifth of overdose decedents from 2016 to 2017 were people who occasionally used illicit drugs as opposed to people who regularly used drugs. In other jurisdictions, overdoses have been disproportionately concentrated among disadvantaged groups, but have also occurred across the socioeconomic spectrum (Cairncross et al. 2018; Ho 2017). This heterogeneity suggests that different points of intervention may be needed to effectively address the increase in fatal overdoses.

In the research literature, the sociodemographic characteristics and risk behaviours of people who use illicit drugs have received the most attention, and previous studies have also focused largely on marginalized groups (Fischer et al. 2005; Kerr et al. 2007; King et al. 2014). Information on the economic characteristics of people at risk of overdose remains sparser, particularly in terms of heterogeneity within this population and longitudinal profiles. This study begins to address these gaps by documenting the employment histories and income sources of people who died from an illicit drug overdose in British Columbia from January 1, 2007, to December 31, 2016. Differences in employment, earnings, and social assistance receipt are documented among men and women and across age groups, with emphasis placed on the strength of labour force attachment by different groups and over time.

A primary objective is to provide information that may be relevant to intervention strategies. “Points of contact” between people who use drugs and formal organizations, such as businesses and public sector agencies, are one consideration in such strategies. The results below offer some perspective on the prevalence of such contacts.

Data are from the BC Coroners Service and from administrative data files containing information on employment earnings and social assistance benefits. The linkage of these files was conducted through Statistics Canada’s Social Data Linkage Environment (SDLE), which is a secure linkage environment for the creation of linked data files (Greenland 2018). The linkage was approved by senior management at Statistics Canada and complies with the agency’s Directive on Microdata Linkage. Further information on the data sources and methods is provided in the “Data and methods” section, which can be found after the “Conclusions and implications” section.

Men and women

The study sample consists of 3,128 deceased individuals. Of these 3,128 decedents, 2,399 (77%) were men and 729 (23%) were women. Attachment to paid employment, as measured by receipt of T4 earnings (earnings obtained from the T4 Statement of Remuneration [T4] file), was lower among the male and female decedents in this study compared to the average population. In the one year prior to the year of death (noted as T-1), just under one-half (48%) of men and over one-quarter (28%) of women had employment. In comparison, among British Columbian taxfilers aged 15 to 64 in 2015, 77% of males and 73% of females had employment, according to the same measure.

Among decedents, attachment to paid employment was stronger among men than women. In each year from T-5 to T-1, about one-half (48% to 51%) of men had some level of employment while this was the case for less than one-third (28% to 32%) of women. Among those employed, average annual earnings were about two times higher among men ($30,000) than women ($15,000).

From a longitudinal perspective, results over the five years prior to death show that 28% of men had employment in all five years (Table 1-1). With average annual earnings of $42,200, this group appears to have had substantial and sustained attachment to paid employment while the situation was quite different for all other decedents. Overall, 14% of women were employed in all five years, with average annual earnings of $22,000. At the other end of the distribution, one-half (51%) of women did not have employment in any of the previous five years. A further 19% of women had employment in just one or two years, with annual earnings averaging less than $5,000.

Participation in formal employment was weak or absent for some 70% of female decedents. This was the case for one-half of men: 29% had no employment and 21% had employment in one or two years out of the five-year period. Given that most female decedents did not have paid employment, workplace-based interventions would likely not have reached many of them. Among those with employment, the largest share (25%) worked in accommodation and food services. This accounted for 13% of all female decedents, reflecting their low rates of employment.

In terms of social assistance, one-half of women received social assistance income in every year over the five years preceding death, and another 8% received it in four of the five years.  About one-quarter of men received social assistance in every year over the five years preceding their death, and just under 7% received it in four of the five years.

  

The results are consistent with those of Richardson et al. (2010) and Amundsen (2015), who reported low levels of employment and high levels of social assistance reliance among people who use drugs as well as people who died from an overdose. The results are also consistent with the point that women who regularly use illicit drugs tend to encounter greater barriers to formal employment than their male counterparts (Richardson et al. 2010; DeBeck et al. 2007). Finally, the socioeconomic diversity observed by Amundsen (2015) and Ho (2017) is evident, perhaps most notably by the subsets of male decedents who had sustained, marginal, or low levels of labour force attachment.

Age groups

The age at which decedents died is another dimension across which labour force ties and social income receipt varied. Chart 1 shows the shares of male and female decedents in each age group who received T4 earnings or social assistance benefits in the year prior to death. The steady decline in employment across successive age groups is clear, as is the steady increase in social assistance receipt.

Data table for Chart 1



Data table for chart 1
Table summary
This table displays the results of Data table for chart 1 T4 earnings, Social assistance, Men and Women, calculated using percent units of measure (appearing as column headers).

	T4 earnings		Social assistance
	Men	Women	Men	Women
	percent
15 to 24 years	62.2	45.8	24	36
25 to 34 years	56.8	40.1	36	69
35 to 44 years	48.5	26.3	47	79
45 to 54 years	41.9	14.9	56	83
55 years and older	27.3	Note ...: not applicable	56	80
... not applicable Note: Because of small cell counts for T4 earnings, the data for women aged 45 to 54 and for those aged 55 and older have been combined. Source: Statistics Canada, linked Longitudinal Worker File, 1989 to 2015, T5007 file, 2006 to 2015, and BC Coroners Service data, 2007 to 2016.

Further analysis is needed to determine what accounts for the decline in employment across age groups; however, the results are consistent with the argument that regular substance use gradually erodes the capacity for work. For example, Bachi et al. (2017) provide evidence that biological aging is accelerated in people who regularly use drugs, increasing their susceptibility to chronic illness and disabilities, while other studies observe that poor health is a principal barrier to obtaining employment for such people (Bauld et al. 2010; Sutton et al. 2004). In the present study sample, only male decedents under 35 years of age had employment rates above 50% in the year prior to death.

The right-hand portion of Chart 1 shows a noticeably higher rate of social assistance receipt among women and the 35-and-older age groups. About 80% of women aged 35 and older and 70% of those aged 25 to 34 received such benefits in the year prior to death. Over one-half of men aged 35 and older did so as well. This shows the prevalence of contact between these groups and provincial and other benefit-granting agencies, insofar as “contact” is meaningfully captured by receipt of income support benefits. Benefit-granting agencies were not as prevalent a point of contact for younger male decedents (those who died before age 35) as about one-third of them received social assistance income in T-1. In contrast, employers were a more prevalent point of contact. Of the young male decedents with employment, the largest share (29%) worked in the construction sector. This accounted for 24% of all young male decedents regardless of employment status.^Note Note 

Among male decedents who died at age 35 and older, less than half had employment in T-1. Again, among those who did, the largest share (24%) worked in the construction sector (Table 2). This accounted for 17% of all older male decedents regardless of employment status. By comparison, 13% of British Columbian males aged 15 to 54 received T4 earnings from construction. The construction industry in British Columbia employs over 110,000 male employees^Note  across many trades, employers and job sites.  



Shorter- and longer-term perspectives

A further issue is whether the economic characteristics of overdose decedents changed markedly in the years prior to death. For instance, were precipitous declines in employment or earnings evident that might suggest withdrawal from paid employment and increasing economic vulnerability over this period? 

Among decedents aged 25 and older, the incidence of employment declined by 6 to 7 percentage points from T-5 to T-1 among men within age groups, and by 4 to 7 percentage points among women within age groups (Chart 2). Notably, among those with employment, average annual earnings did not exhibit a clear decline from T-5 to T-1. This suggests that labour force withdrawal occurred through a transition from employment to non-employment rather than through reductions in annual hours worked.^Note  It bears noting that, overall, the magnitude of employment declines were not so large (nor consistent in the case of earnings) as to suggest widespread withdrawal from paid employment.

Nonetheless, social assistance receipt increased markedly over the five-year period, particularly among people who died at younger ages (Chart 3). Among decedents aged 15 to 24, social assistance receipt increased five-fold among men (from 5% to 24%) and three-fold among women (from 12% to 36%). Among decedents aged 25 to 34, the incidence of receipt increased by 11 percentage points among men and 18 percentage points among women. Increases were smaller among people who died at older ages, in part because social assistance receipt was generally prevalent by that point.

Data table for Chart 2



Data table for chart 2
Table summary
This table displays the results of Data table for chart 2. The information is grouped by Year relative to death (appearing as row headers), Men, Women, 15 to 24, 25 to 34, 35 to 44, 45 to 54, 55 and older and 45 and older, calculated using percent units of measure (appearing as column headers).

Year relative to death	Men					Women
	15 to 24	25 to 34	35 to 44	45 to 54	55 and older	15 to 24	25 to 34	35 to 44	45 and older
	percent
Year T-5	47.3	62.6	54.4	48.9	33.5	42.1	47.3	32.0	18.8
Year T-4	59.9	60.6	51.6	45.9	34.7	49.5	44.9	28.9	17.6
Year T-3	64.1	58.3	50.5	44.4	31.8	51.4	43.1	30.9	17.2
Year T-2	64.9	60.5	49.5	46.1	28.2	51.4	38.3	25.8	14.9
Year T-1	62.2	56.8	48.5	41.9	27.3	45.8	40.1	26.3	14.9
Source: Statistics Canada, linked Longitudinal Worker File, 1989 to 2015, T5007 file, 2006 to 2015, and BC Coroners Service data, 2007 to 2016.

Data table for Chart 3



Data table for chart 3
Table summary
This table displays the results of Data table for chart 3. The information is grouped by Year relative to death (appearing as row headers), Men, Women, 15 to 24, 25 to 34, 35 to 44 and 45 to 54, calculated using percent units of measure (appearing as column headers).

Year relative to death	Men				Women
	15 to 24	25 to 34	35 to 44	45 to 54	15 to 24	25 to 34	35 to 44	45 to 54
	percent
Year T-5	5	25	39	49	12	51	64	77
Year T-4	9	27	40	48	17	53	69	81
Year T-3	12	31	42	53	22	59	70	83
Year T-2	17	31	45	54	29	62	75	82
Year T-1	24	36	47	56	36	69	79	83
Source: Statistics Canada, linked Longitudinal Worker File, 1989 to 2015, T5007 file, 2006 to 2015, and BC Coroners Service data, 2007 to 2016.

Evidence over the longer term provides a more comprehensive perspective on the workforce attrition and economic marginalization of overdose decedents. To document this, male decedents were grouped by age at death into five-year categories (e.g., aged 30 to 34, 35 to 39), and the average annual unconditional T4 earnings of each age group was calculated for each of the 18 years preceding death (i.e., T-18 to T-1).^Note  Average unconditional earnings—calculated as average annual T4 earnings of all decedents regardless of whether they had employment—provide a useful summary measure reflecting both the incidence of employment and the average earnings received. Given the focus on labour force attachment, the analysis is limited to male decedents. Comparison groups of British Columbian men who were the same age as the decedents were drawn from the general population of taxfilers.

As shown in Chart 4, the difference in average unconditional earnings among male decedents aged 30 to 34 and the comparison group was less than $1,000 at age 20 to 24, but widened to over $27,000 by age 29 to 33. In slightly different terms, the average unconditional earnings of male decedents declined from 95% to 38% of those of the comparison group over the observation period. The pattern was much the same among male decedents aged 40 to 44 (Chart 5), with the difference in unconditional earnings widening from just over $4,000 in their early twenties to over $36,000 by their late thirties and early forties. In percentage terms, the average unconditional earnings of these male decedents declined from 77% to 34% of those of the comparison group. 

Data table for Chart 4



Data table for chart 4
Table summary
This table displays the results of Data table for chart 4. The information is grouped by Age groups (appearing as row headers), Comparison group and Decedents, calculated using average annual unconditional earnings units of measure (appearing as column headers).

Age groups	Comparison group	Decedents
	average annual unconditional earnings
20 to 24	15,400	14,600
21 to 25	18,200	16,600
22 to 26	21,200	17,200
23 to 27	24,100	16,700
24 to 28	27,000	17,100
25 to 29	30,100	18,100
26 to 30	33,300	18,400
27 to 31	36,700	18,200
28 to 32	40,200	17,700
29 to 33	43,600	16,400
Source: Statistics Canada, linked Longitudinal Worker File, 1989 to 2015, T5007 file, 2006 to 2015, and BC Coroners Service data, 2007 to 2016.

Data table for Chart 5



Data table for chart 5
Table summary
This table displays the results of Data table for chart 5. The information is grouped by Age groups (appearing as row headers), Comparison group and Decedents, calculated using average annual unconditional earnings units of measure (appearing as column headers).

Age groups	Comparison group	Decedents
	average annual unconditional earnings
22 to 26	18,600	14,300
23 to 27	20,600	14,400
24 to 28	22,700	14,500
25 to 29	24,900	17,100
26 to 30	27,100	17,600
27 to 31	29,800	17,400
28 to 32	31,100	18,400
29 to 33	33,900	18,800
30 to 34	36,100	19,500
31 to 35	37,700	20,900
32 to 36	40,100	21,000
33 to 37	42,200	19,700
34 to 38	45,500	18,400
35 to 39	46,900	18,800
36 to 40	48,300	21,600
37 to 41	50,600	18,900
38 to 42	52,700	18,400
39 to 43	54,600	18,500
Source: Statistics Canada, linked Longitudinal Worker File, 1989 to 2015, T5007 file, 2006 to 2015, and BC Coroners Service data, 2007 to 2016.

Conclusions and implications

The majority of decedents in this study appear to have occupied a position of economic vulnerability for a considerable portion of their lives, characterized by no or little formal employment and increasing reliance on social assistance over time. Observed incomes were generally modest or low.

However, within the context of economic marginalization, heterogeneity is evident. Considerable differences in the labour force attachment and social assistance receipt are observed between men and women, and across age groups. In addition, while most decedents were economically vulnerable, a subset of the male population had sustained employment and substantial earnings over the five years prior to death. This latter group appears to be distinct from all other overdose decedents.

This heterogeneity may be important with respect to intervention strategies. Employers and workplaces, particularly those in the construction sector, are points of intervention that warrant consideration for reaching younger men. The study observed that one in four male overdose decedents who died before age 35 worked in this sector in the years prior to their deaths. Given low rates of employment among female overdose decedents and for most male overdose decedents in older age groups, interventions geared toward employers and workplaces may offer less promise. Contact between these groups and social service agencies were more prevalent.

The inclusion of additional administrative data files in subsequent research could be helpful in providing a more complete picture of income sources and other potential points of intervention. For example, overdose decedents may be over-represented among recipients of Canada Pension Plan Disability benefits given their age (i.e., non-seniors), low rate of employment, and possible health problems. Similarly, more detailed information on social assistance benefits is available from provincial files than from T5007 Statements of Benefits, such as month of last benefit receipt relative to month of death. This information could be helpful in piecing together key transitions among at-risk populations. Beyond income alone, a more comprehensive assessment of points of contact could be provided by looking at interactions with social service agencies, the criminal justice system, and health care providers.

Data and methods

Five administrative data files were linked together for this analysis. First, the universe of people who died of a confirmed illicit drug overdose in British Columbia from January 1, 2007, to December 31, 2016, was drawn from the British Columbia (BC) Coroners Service Database. Second, information on the paid employment of decedents was drawn from the T4 Statement of Remuneration (T4) file. Employers are required to complete and submit T4 slips for employees paid over $500 in a tax year or for whom Canada Pension Plan or Quebec Pension Plan contributions, employment insurance benefits, or income taxes were deducted. Third, information on social assistance benefits received was drawn from the T5007 Statements of Benefits (T5007) file. Provincial and other benefit-granting agencies are required to submit T5007 slips for payments over $500 made to beneficiaries in a tax year. Social assistance refers to payments to people for basic needs (food, clothing, and shelter) and is based on a means, needs, or income test. Fourth, information on industry of employment of T4 recipients was drawn from the Longitudinal Employment Analysis Program (LEAP) maintained by Statistics Canada. Finally, the T1 Family File was used to produce earnings trajectories for a comparison group of taxfilers in the general population of British Columbia.

Both the T4 and T5007 files are submitted directly to the Canada Revenue Agency (CRA) by the payers. No action on the part of employees or beneficiaries is required to be included in these files. Hence, these sources provide information that is both comprehensive and robust. Information on other government transfers, such as employment insurance and Canada Pension Plan Disability benefits, and other sources of market income, such as self-employment, are not available in the linked analytical file.

The study sample consists of 3,128 people, 2,399 males and 729 females. Age refers to a person’s age group at death. The year of death is identified as year T, while years prior to the year of death are identified as T-1, T-2, and so on. Incidence of employment is defined as the share of T4 earners of over $500 (from all jobs) among all decedents in a given year. Total observed income is defined as the sum of T4 earnings and social assistance income. All earnings and income estimates are calculated on an annual basis and expressed in 2016 constant dollars.

References

Amundsen, E.J. 2015. “Drug-Related Causes of Death: Socioeconomic and Demographic Characteristics of the Deceased.” Scandinavian Journal of Public Health 43 (6): 571–579.

Bachi, K., S. Sierra, N.D. Volkow, R.Z. Goldstein, and N. Alia-Klein. 2017. “Is Biological Aging Accelerated in Drug Addiction?” Current Opinion in Behavioral Science 13: 34–39.

Bauld, L., G. Hay, J. McKell, and C. Carroll. 2010. Problem Drug Users’ Experiences of Employment and the Benefit System. Research Report No. 640. Norwich, United Kingdom: Department for Work and Pensions.

British Columbia (BC) Coroners Service. 2018. Illicit Drug Overdose Deaths in BC: Findings of Coroners’ Investigations. Victoria, BC: Ministry of Public Safety and Solicitor General.

Cairncross, Z.F., J. Herring, T. van Ingen, B.T. Smith, P. Leece, B. Schwartz, and K. Hohenadel. 2018. “Relation between Opioid-Related Harms and Socioeconomic Inequalities in Ontario: A Population-Based Descriptive Study.” CMAJ Open 6 (4): E478–E485.

DeBeck, K., K. Shannon, E. Wood, K. Li, J. Montaner, and T. Kerr. 2007. “Income Generating Activities of People Who Inject Drugs.” Drug and Alcohol Dependence 91 (1): 50–56.

Fischer, B., J. Rehm, S. Brissette, and S. Brochu. 2005. “Illicit Opioid Use in Canada: Comparing Social, Health, and Drug Use Characteristics of Untreated Users in Five Cities (OPICAN Study).” Journal of Urban Health: Bulletin of the New York Academy of Medicine 82 (2): 250–266.

Galarneau, D., R. Morissette, and J. Usalcas. 2013. “What Has Changed for Young People in Canada?” Insights on Canadian Society. Statistics Canada Catalogue no. 75-006-X. Ottawa, Ontario: Statistics Canada.

Greenland, J. 2018. Social Data Linkage Environment (SDLE) Methodology Report: Linkage between British Columbia Coroner’s File (2007-2017) and the SDLE Derived Record Depository (version 16). Ottawa, Ontario: Statistics Canada.

Ho, J.Y. 2017. “The Contribution of Drug Overdose to Educational Gradients in Life Expectancy in the United States, 1992-2011.” Demography 54 (3): 1175–1202.

Kerr, T., N. Fairbairn, M. Tyndall, D. Marsh, K. Li, J. Montaner, and E. Wood. 2007. “Predictors of Non-Fatal Overdose Among a Cohort of Polysubstance-Using Injection Drug Users.” Drug and Alcohol Dependence 87 (1): 39–45.

King, N.B., V. Fraser, C. Boikos, R. Richardson, and S. Harper. 2014. “Determinants of Increased Opioid-Related Mortality in the United States and Canada, 1990-2013: A Systemic Review.” American Journal of Public Health 104 (4): e32–e42.

Richardson, L., E. Wood, K. Li, and T. Kerr. 2010. “Factors Associated with Employment among a Cohort of Injection Drug Users.” Drug and Alcohol Review 29 (3): 292–300.

Statistics Canada. 2018. “Drug Overdose Crisis: Socioeconomic Characteristics of those Dying of Illicit Drug Overdoses in British Columbia, 2011 to 2016.” The Daily, November 13. Statistics Canada Catalogue no. 11-001-X. Ottawa, Ontario: Statistics Canada.

Sutton, L., A. Cebulla, C. Heaver, and N. Smith. 2004. Drug and Alcohol Use as Barriers to Employment: A Review of the Literature. CRSP Research Report no. 499S. Loughborough, United Kingdom: Loughborough University.