Elevated rates of suicide among

Transcription

1 INJURY Suicide by occupational skill level in the Australian construction industry: data from 2001 to 2010 Allison Milner, 1 Heather Niven, 1 Anthony LaMontagne 1 Elevated rates of suicide among construction workers have been identified in several areas including Denmark, 1 Canada, 2 the United States, 3,4 the United Kingdom, 5,6 Korea 7 and Queensland, Australia. 8,9 To some extent, this may be connected to the demographic characteristics of the largely male workforce. 10 In Australia and in many other countries, the rate of suicide among males is 3-4 times that of women. 11 Past research also suggests that suicide among construction workers is connected to excessive alcohol consumption, lack of help seeking behaviour and relationship problems. 9,12 The health of workers is particularly affected by poor psychosocial working conditions, including high demand, low job control or autonomy, job insecurity and poor workplace support. 13,14 To date, there has been little research in Australia or internationally on whether there are differential vulnerabilities to suicide depending on industry and occupational skill level. Investigating this topic is particularly relevant in light of the recession, which led to increasing unemployment in Australia and many other countries. 15 Thus far, research has not examined whether there was an equivalent rise in suicide within the employed population or, more specifically, within industries at risk of suicide such as construction. For the purposes of targeting prevention initiatives, it is also important to assess whether all workers in construction are at risk or whether the burden of suicide is confined to a specific skill level. This study examined variation in suicide deaths by occupational skill level within the Abstract Objective: This study examines variation in suicide deaths by occupational skill level within the construction industry and changes in the rate of suicide over time. Methods: Suicide deaths were extracted from a national coronial database and occupations were coded. Adjusted suicide rates over the period 2001 to 2010 were calculated and incidence-rate ratios (IRRs) used to compare the overall burden of suicide in the lowest skilled group (machine operators and labourers) against skilled tradespersons in the construction industry. Results: Those employed as labourers or machine operators had an adjusted rate of 18 per 100,000 persons (95%CI 14-22) and those employed in skilled trades had an adjusted rate of 13 per 100,000 (95%CI 11-15) over the period 2001 to Compared to skilled trades, the lower skilled group had significantly elevated suicide at several time points over the period 2001 to The most observable difference in IRRs were in the years 2002 and Conclusions: Low-skilled workers in the construction industry had elevated rates of suicide compared to skilled trades workers. Implications: These workers should be targeted by prevention efforts. Key words: Construction industry, suicide, socio-economic gradient, skill level, intentional self-harm construction industry and changes in the rate of suicide over time in comparison to suicide in the general male population using data available from the National Coroners Information System (NCIS). It is the first to examine industry and occupational trends in suicide rates at a national level and in relation to the most recent global economic recession. Based on previous work demonstrating variation in suicide by socio-economic status, 16 we hypothesise that those employed in the lowest skilled jobs within the industry will have the greatest risk of suicide and there will be an observable increase in suicide during the period 2007 to 2009, coinciding with the impact of the most recent global financial crisis (GFC). Methods Study design This retrospective case-series study utilised data from the National Coroners Information System (NCIS) and focused on trends in suicide over time and by skill level. Demographic characteristics were also examined. Data source The National Coroners Information System (NCIS) is a national internet-based data storage and retrieval system for Australian coronial cases. 17,18 NCIS, which was established in 2001, is utilised by coroners, 1. McCaughey VicHealth Centre for Community Wellbeing, Melbourne School of Population and Global Health, University of Melbourne, Victoria Correspondence to: Dr Allison Milner, McCaughey VicHealth Centre for Community Wellbeing, Melbourne School of Population and Global Health, University of Melbourne, Level 5, 207 Bouverie Street, Melbourne, VIC 3010; allison.milner@unimelb.edu.au Submitted: August 2013; Revision requested: November 2013; Accepted: December 2013 The authors have stated they have no conflict of interest. Aust NZ J Public Health. 2014; 38:281-5; doi: / vol. 38 no. 3 Australian and New Zealand Journal of Public Health 281

2 Milner, Niven and LaMontagne Article government agencies and researchers for identifying cases for death investigation, research and to monitor external causes of death in Australia. 17 Previous to the advent of this database, researchers relied on mortality data provided by the Australian Institute of Health and Welfare, which provides relatively limited information on demographic factors such as occupation. NCIS provides users with basic demographic information, including employment status and occupation at the time of death. Suicide methods are classified according to the International Classification of Disease 10th revision (ICD-10) within method specific codes X60-X84 (Chapter 10). 19 Each case on NCIS routinely includes a police text description of the circumstances and background of each extracted case, as well as coronial findings and toxicology reports. Data quality Delays in the coronial process mean that the suicide data available in NCIS is underreported from the year 2010 onwards. 20 As the system was established in 2000, there is also poor data quality for that year. 11 Aside from system issues, data on suicide may be under-reported due to definitional variation in the determination of suicide between coroners. 20,21 Regardless of these issues, NCIS offers the best available nationally representative information on suicide mortality in the country and is used to compile the official death statistics released by the Australian Bureau of Statistics. Eligibility and data extraction Figure 1 shows the process of data extraction and case eligibility classification. All cases officially recorded as intentional self-harm in NCIS were extracted in January The raw data file included 26,471 individual cases. Because of the reasons given above, only the years 2001 to 2010 were included in this study. Text descriptions of occupation and employment status at the time of death were extracted for each case. Data was also obtained on sex, method of death, age and marital status. Cases were excluded if they were not employed at the time of death. Coding of occupation Occupational information was coded by two researchers according to the Australian and New Zealand Standard Classification of Occupations (ANZSCO) (up to the six-digit level). If more than one occupation was reported, the researchers took the first listed as the primary occupation, unless the second listed occupation provided additional information that the first could not offer (e.g. construction/carpenter). Ambiguous occupational information was coded at the broadest level or was marked as uncodable. Consensus was reached via discussion. Sample definition Those in the construction industry were identified as being involved in the building of homes, dwellings, buildings, or other structures and roads, as per the Australian and New Zealand Standard Industrial Classification. 22 Construction was also defined as building work related to additions, alterations, reconstruction, maintenance and repairs. 22 After applying this definition, occupations coded as being in construction fell into two main groups and three major ANZSCO groups: Skilled trades (ANZSCO Figure 1: Case inclusion process. Cases extracted= 26,471 Cases remaining= 22,677 Cases remaining= 9,204 Cases remaining =1,998 Cases remaining = 1,947 Residential building, nonresidential building and engineering, building maintenance, and repair level 3) and machine operators and labourers (ANZSCO 7 and 8) were aggregated together due to small sample size in ANZSCO 7. Cases in higher skilled occupations such as ANZSCO 1 (Managers) and ANZSCO 2 (Professionals) were excluded as they described a generic set of non-construction related skills industry (e.g. management and human resources related job tasks) and also because there were only a small number of relevant cases. Only males were included in this study due to small number of suicides among women employed in the construction industry (n=11, as per Figure 1). Statistics Descriptive and frequency tests were used to assess suicide by age-group, method of death and relationship status by occupational skill level. We compared age to demonstrate the overall distribution of suicide by age between the groups. For each of the two occupational Excluded =3,794 from years other than 2001 to 2010 Excluded = 4,822 unemployed, 4,872 retired, 2,190 unknown employment status, 127 prisoners, 923 students, 506 home duties, 33 unspecified other Excluded (not in construction)= 7,206 ANZSCO Major grouping 1, 2, 4, 5, 6 Excluded = 11 females, 39 uncodable occupation information due to lack of detail, 1 person under 15 years 282 Australian and New Zealand Journal of Public Health 2014 vol. 38 no. 3

3 Injury Suicide in the Australian construction industry skill level groups, incidence suicide rates per 100,000 persons were calculated by dividing the total number of suicides in a year by the average number of people in that occupation using the 2001 census data. Rates were the multiplied by 100,000 persons. Following this, rates were indirectly standardised against the rate of male suicide in the general population over the period 2001 to 2010 in order to adjust for different age distributions between skilled trade occupations, machine operators and labourers, and the general male population. 23 The calculation of indirect rates required a common set of age-specific rates (the general male suicide population) which were applied to the construction populations whose suicide rates were to be standardised. 24 Indirect (rather than direct standardisation) was chosen because of the small number of suicides in the construction industry when broken down into 10-year age-groups. Incidence rate ratios with 95% confidence intervals were calculated to compare the incidence of suicide in the population of the lowest skilled (labourers and machine operators) to that of higher skilled occupational group (skilled trades workers). The incidence rate ratio (IRR) calculation was: IRR = [d1(suicides in the lowest skilled group)/ T1 (person-years in the lowest skilled group)] / d0(suicides in the highest skilled group)/ T0 (person-years in the highest skilled group)]. We used Mantel Haenszel (MH) weights (wi= weight for each rate ratio by year) to control for confounding related to year. The formula for this was IRR MH=Σ(wi* IRRi)/Σwi. Results An inspection of case reports revealed that the major method of suicide was consistent with that of the general population (e.g., hanging followed by exposure to carbon monoxide, and firearms). 11 The remaining deaths were due to overdose on drugs or other methods such as jumping from height or in front of a moving vehicle, drowning, or stabbing. Separation or divorce was noted in about 20% of cases. A slightly larger proportion of those in machine operating and labouring jobs were never married compared to skilled trades workers, who were more frequently married at the time of death. There was missing data on relationship status in a substantial proportion of male suicide cases in both skill level groups (15.8% in machine operators and labourers and 14.5% in skilled trades). Comparative changes in the burden of suicide over time Those employed as labourers or machine operators had an adjusted rate of 18 per 100,000 persons (95%CI 14-22) and those employed in skilled trades had an adjusted rate of 13 per 100,000 (95%CI 11-15) over the period 2001 to Adjusted suicide rates for skilled trades and labourers and machine operators by year can be seen in Figure 2. This shows a consistent difference in suicide among the lowest and the highest skill level groups. As a comparison, suicide among the general male population for the years 2001 to 2010 has also been plotted in Figure This shows that rates among skilled trades workers was below the general male population over the period 2001 to 2010, while those among lower skilled workers was generally above male suicide in the general population. Data in 2001 was lower than the general male population, which may reflect problems in NCIS establishing reliable data recording procedures during this time (NCIS began in 2000 and was noted to under-report suicide in its early years of operation). 20 Incidence rate ratios can be seen in Table 2. There was an overall greater and significant risk of suicide among the lowest skilled group compared to the higher skilled group (IRR 1.36, 95%CI ). On closer inspection of results by year, results revealed that the rate ratio was 1.9 times higher in 2002 (95%CI ). Labourers and machine operators also had higher suicide rates than those in skilled trades in 2007 (IRR 1.54, 95%CI ) and 2008 (IRR 1.71, 95%CI ). These differences were statistically significant at p< The Mantel Haenszel rate ratio after controlling for year was 1.35 (95%CI ). Table 1: Age-profile of suicide among those employed in the construction industry over the period 2001 to Machine operator and labourers (%) Technicians and Trades Workers (%) years years years years years and over All Ages total (n) Demographic characteristics and method of death There were 1,947 male suicides identified within the construction industry over the 10 year period of the study. About 51% were in ANZSCO group 3 (n=993, skilled trades) and 49% were labourers and machine operators (n=954, ANZSCO group 7 and 8). As can be seen in Table 1, a greater proportion of labourers and machine operators were 24 years or younger (19.9%), while a larger proportion of skilled trades workers were aged years (18.5%). Most suicide cases in both occupational groups were aged years. These age-related differences were significant in a chi-square test (χ 2 (5) = 10.6, p<0.05). Figure 2: Suicide rates among those employed in the construction industry, adjusted against the Australian population. Suicide rate per 100, Year Machine operators and labourers Skilled trades Suicide rate in all males 2014 vol. 38 no. 3 Australian and New Zealand Journal of Public Health 283

4 Milner, Niven and LaMontagne Article Discussion The first main result of this study was that the suicide rate among labourers and machine operators was significantly greater than the suicide rate of skilled trades workers within the construction industry over the period 2001 to Second, there was an increased divergence among those employed in lower versus higher skilled jobs during the most recent economic recession ( ) that was largely driven by increasing rates among the lowest-skilled occupations. The methods of death in those employed in the construction industry were predominately by hanging and generally by younger males, consistent with the trend in the general male population. This difference between those employed in the lowest versus higher occupations corresponds to a recent meta-analysis across 34 studies. 25 This study found that the risk of suicide among unskilled labourers was close to twice that in the general employed population, while the rates of suicide among skilled trades workers were only marginally above those in the general employed population. A study in the UK 6 also showed an increasing divergence in the suicide rates of manual workers compared to higher skilled jobs over the period from 1979 to Overall, the suicide rates among labourers rose markedly over the 26-year period of the study, while rates in the highest skilled occupations did not. There was an increase in risk of suicide among lower skilled construction jobs during the most recent economic recession. The transiency of this increase and its decline is consistent with the hypothesis that lower skilled workers in the construction industry were particularly affected by the GFC. From a peak in 2007, suicide rates in the lowest skilled level occupations reduced, following the overall trend of suicide rates in the general population, which have been declining following a peak in the late 1990s. 11 There have been several other reports of increased suicide in response to the most recent recession in England, 26 Greece, 27 the United States 28 and several other European countries. 29 Job loss and declining income are thought to be among the main drivers of rising suicide rates in these countries. Potentially, the declining general population rate in Australia disguised an elevated risk among low-skilled workers most affected by fluctuations in the global economic market Table 2: Incident rate-ratios of those employed as machine operators and labourers compared to skilled trades workers and the general male population, 2001 to Exposed Non-exposed Mach. labour (Standardised rate) Skilled trades (Standardised rate) during the most recent recession. However, we also note that there was an increase in suicide rates in 2004, which was before the economic recession. It is also relevant to consider that the recession may have led to elevated rates in other vulnerable groups not included in this paper, such as the unemployed population. There are a number of possible explanations for differences in the burden of suicide by occupational skill level within the industry. First, it is possible that the higher burden reflects characteristics of younger male workers employed in labouring professions, who also tend to have higher suicide rates than other age groups. 11 We note that the age as a potential confounder was controlled through the age-standardisation of rates. Elevated rates may also be connected to life-related stressors, as well as higher levels of impulsivity and greater use of alcohol and drugs. 30 This aligns with research on suicides in the construction industry, which found that alcohol and drugs, untreated mental health conditions and relationship issues precipitated a large proportion of suicide deaths. 9 Failure to seek or accept help for mental health issues has also been identified as a key risk factor for young males in the construction industry. 9 At a population level, the higher burden of suicide in labourers and machine operators may reflect risk associated with lower socio-economic status and wider social disadvantage. 30 Workplace and industry-related factors may also influence suicide within construction. IRR Low CI Upp CI Overall M-H Notes: IRR =Incident rate ratio; Low CI = Lower confidence interval; Upp CI =Upper confidence interval; MH=Mantel-Haenszel estimate of the rate ratio. Findings from job stress research indicates adverse psychosocial working conditions are associated with worse mental health while positive elements of a job, such as high control over the organisation and structure of work and the ability to use a variety skills, are associated with better mental health. 31,32 This may explain the lower rate of suicide in skilled trades such as builders and electricians, who have greater latitude and control over their work than those in lowerskilled jobs, who have the worst working conditions. 16 Although, we note that the construction industry in general is believed to be exposed to a number of negative working experiences. 33,34 In saying this, there are few studies specifically investigating the influence of adverse working environments on suicide This emphasises the need for more research on this topic in the future, particularly as to whether a worsening in working conditions led to an increase in suicide among lower-skilled workers during the GFC. The limitations of this study included issues in suicide data available from NCIS in its initial years of operation. 20,39 Other potential problems affecting this research include the miscoding of occupation, despite the fact that independent coding was conducted by two researchers and the use of a structured approach to the classification of occupation. We also acknowledge that there may be a number of suicides coded as other causes of death and intent categories 20,21 (i.e. undetermined intent) not considered in this 284 Australian and New Zealand Journal of Public Health 2014 vol. 38 no. 3

5 Injury Suicide in the Australian construction industry paper and may further contribute to under reporting. The strengths of this study are that it is based on the best available data and provides information on occupationally coded suicide at a national level. The use of this data contributes substantially to current research on suicide in the construction industry of Australia, which has previously been confined to specific sample areas and states. 8,9,38 Conclusions and implications The main contribution of this paper is that it demonstrates differences in the risk of suicide by skill level within the construction industry. 25 From a prevention perspective, this suggests that low-skilled workers in construction should be prioritised in prevention and intervention initiatives. In 2012, the Australian government invested a considerable amount in a suicide prevention program within the construction industry. This program, Mates in Construction, seeks to raise general awareness of suicide and train gatekeepers to recognise and refer those at risk. Preliminary evidence suggests that this training has received positive feedback within industry. 12 However, long-term efforts are needed to assess the efficacy of this intervention in reducing suicidality. Intervention efforts also need to address possible work-related risk factors for suicide. From a wider research perspective, further work is needed to assess the risk of suicide among other occupational and skill level groups over time, and to identify industryspecific risk and protective factors that can be used in workplace suicide prevention initiatives. Acknowledgements This work was supported by a National Health and Medical Research Council Capacity Building Grant in Population Health (grant number ) and McCaughey Centre funding from Victorian Health Promotion Foundation. References 1. Agerbo E, Gunnell D, Bonde JP, Mortensen PB, Nordentoft M. Suicide and occupation: the impact of socio-economic, demographic and psychiatric differences. Psychol Med. 2007;37: Mustard CA, Bielecky A, Etches J, Wilkins R, Tjepkema M, Amick BC, et al. Suicide mortality by occupation in Canada, Can J Psychiatry. 2010;55: van Wijngaarden E. An exploratory investigation of suicide and occupational exposure. J Occup Environ Med. 2003;45: Kposowa AJ. Suicide mortality in the United States: differentials by industrial and occupational groups. Am J Ind Med. 1999;36: Meltzer H, Griffiths C, Brock A, Rooney C, Jenkins R. Patterns of suicide by occupation in England and Wales: Br J Psychiatry. 2008;193: Roberts SE, Jaremin B, Lloyd K. High-risk occupations for suicide. Psychol Med. 2013:43(6): Kim MD, Hong SC, Lee SY, Kwak YS, Lee CI, Hwang SW, et al. Suicide risk in relation to social class: a national register-based study of adult suicides in Korea, Int J Soc Psychiatry. 2006;52: Andersen K, Hawgood J, Klieve H, Kolves K, De Leo D. Suicide in selected occupations in Queensland: evidence from the State suicide register. Aust N Z J Psychiatry. 2010;44: Heller TS, Hawgood JL, Leo DD. Correlates of suicide in building industry workers. Arch Suicide Res. 2007;11: Australian Bureau of Statistics Gender Indicators, Australia, Jan Canberra (AUST) ABS; Australian Bureau of Statistics Causes of Death Data, Canberra (AUST): ABS; Gullestrup J, Lequertier B, Martin G. MATES in construction: Impact of a multimodal, communitybased program for suicide prevention in the construction industry. Int J Environ Res Public Health. 2011;8: LaMontagne AD, Keegel T, Vallance D, Ostry A, Wolfe R. Job strain- attributable depression in a sample of working Australians: assessing the contribution to health inequalities. BMC Public Health. 2008;8: Vanroelen C, Levecque K, Louckx F. Differential exposure and differential vulnerability as counteracting forces linking the psychosocial work environment to socioeconomic health differences. J Epidemiol Community Health. 2010;64: Australian Bureau of Statistics Australian Labour Market Statistics. Canberra (AUST): ABS; Milner A, Spittal MJ, Pirkis J, LaMontagne AD. Suicide by occupation: A systematic review and meta-analysis. Br J Psychiatry. 2013;203(6): Bugeja L, Clapperton AJ, Killian JJ, Stephan KL, Ozanne- Smith J. Reliability of ICD-10 external cause of death codes in the National Coroners Information System. HIM J. 2010;39: Daking L, Dodds L. ICD-10 mortality coding and the NCIS: A comparative study. HIM J. 2007;36:11-23; discussion World Health Organisation. ICD-10 Classifications of Mental and Behavioural Disorder: Clinical Descriptions and Diagnostic Guidelines. Geneva (CHE) WHO; De Leo D, Dudley MJ, Aebersold CJ, Mendoza JA, Barnes MA, Harrison JE, et al. Achieving standardised reporting of suicide in Australia: rationale and program for change. Med J Aust. 2010;192: Simon-Davies J. Suicide in Australia [Internet]. Canberra (AUST): Parliament of Australia; 2011 [cited 2013 Jun 10]. Available from: About_Parliament/Parliamentary_Departments/ Parliamentary_Library/pubs/BN/ /Suicide 22. Australian Bureau of Statistics. Australian and New Zealand Standard Industrial Classification: Division E- Construction. Canberra (AUST): ABS; Berry J, Harrison J. A Guide to Statistical Methods for Injury Surveillance. Canberra (AUST): Australian Institute of Health and Welfare; Naing NN. Easy way to learn standardization: Direct and indirect methods. Malays J Med Sci. 2000;7: Milner A, Spittal MJ, Pirkis J, LaMontagne AD. Suicide by occupation: A systematic review and meta-analysis. Br J Psychiatry. 2013;203: Barr B, Taylor-Robinson D, Scott-Samuel A, McKee M, Stuckler D. Suicides associated with the economic recession in England: time trend analysis. BMJ. 2012;345:e5142. doi: /bmj.e Economou M, Madianos M, Peppou LE, Theleritis C, Patelakis A, Stefanis C. Suicidal ideation and reported suicide attempts in Greece during the economic crisis. World Psychiatry. 2013;12: Reeves A, Stuckler D, McKee M, Gunnell D, Chang S-S, Basu S. Increase in state suicide rates in the USA during economic recession. Lancet. 2012;380: Stuckler D, Basu S, Suhrcke M, Coutts A, McKee M. Effects of the 2008 recession on health: a first look at European data. Lancet. 2011;378: Pitman A, Krysinska K, Osborn D, King M. Suicide in young men. Lancet 2012;379: Stansfeld S, Candy B. Psychosocial work environment and mental health-a meta-analytic review. Scand J Work Environ Health. 2006;32: Bonde JPE. Psychosocial factors at work and risk of depression: A systematic review of the epidemiological evidence. J Occup Environ Med. 2008;65: Lingard H, Francis V. The work-life experiences of office and site-based employees in the Australian construction industry. Constr Mang Econ. 2004;22: Lingard H, Francis V. Does work family conflict mediate the relationship between job schedule demands and burnout in male construction professionals and managers? Constr Manag Econ. 2005;23: Schneider B, Grebner K, Schnabel A, Hampel H, Georgi K, Seidler A. Impact of employment status and workrelated factors on risk of completed suicide. A casecontrol psychological autopsy study. Psychiatry Res. 2011;190: Tsutsumi A, Kayaba K, Ojima T, Ishikawa S, Kawakami N, Jichi Med School Cohort Study Grp. Low control at work and the risk of suicide in Japanese men: A prospective cohort study. Psychother Psychosom. 2007;76: Takada M, Suzuki A, Shima S, Inoue K, Kazukawa S, Hojoh M. Associations between Lifestyle Factors, Working Environment, Depressive Symptoms and Suicidal Ideation: A Large-scale Study in Japan. Industrial Health. 2009;47: Bottomley J, Neith M. Suicide and Work: The Need for Improved Data Collection on Work Factors in Suicide as a Contribution to Suicide Prevention. St Kilda (AUST): Creative Ministries Network; Australian Bureau of Statistics nformation Paper: External Causes of Death, Data Quality, Canberra (AUST): ABS; vol. 38 no. 3 Australian and New Zealand Journal of Public Health 285