Author
Li Zhuoyang; et al.
Attributable risk of psychiatric and socio-economic factors for suicide from individual-level, population-based studies: a systematic review.
Journal citation/publication details
Social Science and Medicine, 72(4), February 2011, pp.608-616.
Summary
Fourteen case-control and cohort studies were included in this systematic review and meta-analysis. Although relative risk for suicide was much higher for psychiatric disorders compared to socioeconomic factors, the population attributable risk for some socioeconomic factors and some psychiatric disorders were found to be of the same order of magnitude. The suggestion that public health policy on suicide prevention should focus on lower risk, but higher prevalence, socioeconomic factors is discussed.
Context
Suicide has been associated with a range of risk factors both distal and proximal. The two most commonly studied risk factors are mental disorders and socioeconomic status, SES. The aim of this study was to systematically review ‘the risk of suicide with high prevalence [psychiatric] disorders and key SES measures in population-based studies of suicide’ and ‘estimate the [associated] population attributable risk of suicide’.
Methods
What sources were searched?
The online databases Medline and EMBASE were searched from 1966 and 1980, respectively, to December 2009. The reference lists of relevant articles and previous reviews were hand searched for additional studies.
What search terms/strategies were used?
Searches were carried out using the keywords ‘suicide’ and (‘affective disorder’ or ‘anxiety disorder’ or ‘psychosis’ or ‘personality disorder’ or ‘substance abuse’) and (’socio-economic factors’) and (‘case-control studies’ or ‘cohort studies’).
What criteria were used to decide on which studies to include?
Studies had to be case control or cohort studies with suicide as the main outcome and in which relative risk of suicide was reported, or could be estimated, for either mental disorder or an individual level SES measure. Population-based controls in case-control studies were allowed. Separate gender estimates were required and cohort studies had to include a minimum follow-up period of one year. Psychiatric disorders, based on an ICD or DSM definition, could come from a self-reported or clinical diagnosis.
Who decided on their relevance and quality?
Study selection was based firstly on titles only, and then on abstracts. The selection process is outlined in Figure 1 and includes reasons for exclusion. There is no indication of who was responsible for study selection. Study quality was not formally assessed.
How many studies were included and where were they from?
The database searches yielded 2,142 references, 691 of which were deemed potentially relevant. A total of 478 papers were rejected based on their abstracts, leaving 213 studies for further evaluation. Fourteen articles were finally included in the systematic review and meta-analysis. Study settings included North America, Scandinavia, and North Asia.
How were the study findings combined?
Data was extracted and coded. Studies were grouped according to five psychiatric disorder categories and four socio-economic factor categories. Studies were also grouped according to whether the risk estimates were reported for males, females, or as a total for those where analyses were not stratified by gender. Meta-analysis was carried out using Stata Version 10.1. Relative risk, RR, and population attributable risk, PAR, associated with suicide were estimated for psychiatric disorders and socio-economic status. RR estimates were stratified by age group, geographic region and study design. Further details of statistical methods are available in the text.
Findings of the review
The magnitude of risk for suicide was four to five times higher for psychiatric disorders compared to socioeconomic factors. The relative risk of suicide associated with any mental disorder was 7.49 (95% CI 6.21 to 9.05) in males and 11.94 (95% CI 9.87 to 14.44) in females. For socioeconomic factors the overall relative risk of suicide for those in the lowest socioeconomic group compared to the highest group was 2.06 (95% CI 1.51 to 2.82) for males and 1.51 (95% CI 1.22 to 1.88) for females.
Population attributable risk was of a similar magnitude for some psychiatric disorders and socioeconomic factors. In males median PAR for low educational achievement was 40.9% (range 18.5% to 46.5%) and low occupational status was 33.1% (20.5% to 43.7%) while PAR for affective disorder was 26.3% (range 6.7% to 45.4%) and substance use disorders was 9.0% (range 5.0% to 23.9%). In females the PAR for low educational achievement was 20.3% (range 18.7% to 21.8%) compared to a similar magnitude for affective disorders (31.6%, range 18.6% to 66.8%), substance use disorders (25.4% range 4.8% to 31.9%) and anxiety disorders (11.7%, range 5.5% to 21.6%).
A number of gender differences were reported: there was a higher relative risk in males than females for personality disorders and the socioeconomic factors of education and occupation, and a higher relative risk for females than males for substance use disorders and anxiety disorders.
Authors' conclusions
Both psychiatric factors and socio-economic factors are associated with increased risk of suicide in populations. The population attributable risk estimates for suicide were similar in magnitude for socio-economic factors and some psychiatric factors. ‘Interventions relating to prevention of suicide should focus not only on individual psychiatric factors (high risk, low prevalence) but also the potential effects of interventions relating to broader socio-economic factors (low risk, high prevalence).’
Implications for policy or practice
None are discussed.