What predicts differences in suicide judgements between Coroners and a Suicide Register?

Why Do We Agree to Disagree? Agreement and Reasons for Disagreement in Judgements of Intentional Self-Harm from Coroners and a Suicide Register in Queensland, Australia, from 2001 to 2015

By Stuart Leske, Bridget Weir, Ghazala Adam and Kairi Kõlves

Published 30 December 2023

What's the issue?

Having accurate suicide data is important to prevent suicide, measure the impact of suicide and decide where to focus suicide prevention efforts to try to reduce suicide rates.

Sometimes it can be difficult to know if a person died by suicide. Without clear intent, a death may be determined to be unintentional or accidental.

In Australia, the National Coronial Information System (NCIS) provides information about suicide deaths reported to coroners. Previous research suggests coronial systems underestimate suicides in Australia^1,2 as practices are not standardised between coroners and jurisdictions.

The authors of this study sought to assess the direction, extent and any predictors of differences between the reporting of suicide in Queensland and Queensland NCIS data.

What was done?

The authors looked at all suspected suicides reported to the Queensland Suicide Register from 2001 to 2015, comparing it to Queensland data in the NCIS. They used binary logistic regression to explore factors that might predict a discrepancy between the two systems.

What was found?

There were 9,520 cases of suspected suicides between 2001 to 2015 available for analysis. Overall, the Queensland Suicide Register reported 510 more suicides than the NCIS over the 15 years.

Many factors were found to predict a discrepancy in judgements between the Queensland Suicide Registry and the NCIS. For example, the likelihood of discrepant judgements was:

• 1.9 times higher when the person was female, compared to male
• 2.5 times higher when the person was unemployed, compared to full-time, part-time or casual employment
• 4.1 times higher for people in major cities, compared to remote and very remote areas
• 4.2 times higher for a homeless person, compared to living with a spouse
• 3.7 times higher when the person had dementia
• 1.9 times higher when the person had bipolar
• 1.8 times higher when the person had a substance abuse condition
• 1.7 times higher when the person was non-Indigenous, compared to Indigenous
• 1.4 times higher for people not experiencing relationship conflict or separation, compared to those experiencing conflict
• 2.2 times higher when the deceased had no known history of sexual abuse
• 7.2 times higher when no suicide note was left.

Why are findings important?

The findings suggest multiple factors may predict discrepancies in judgements between the two systems. Groups that die by suicide at lower rates (such as females) or in more complex circumstances (such as people with dementia) seemed to be more likely to have a discrepancy in judgements.

Common risk factors appeared to reduce the likelihood of discrepant judgements (such as relationship issues). However, some increased the likelihood, such as a substance use condition or bipolar disorder.

There are no objective criteria to decide if a person died by suicide. Whilst the independent judgements made by the Queensland Suicide Register are not necessarily the gold standard, the system is not limited by legal processes.

Overall, the authors argue that the number of discrepancies (N = 510) over a 15-year period is too high. This study adds to the evidence that there may be some underreporting of suicide in the coronial system. The authors conclude that underreporting is a well-known limitation and is acknowledged by those analysing suicide data held by the NCIS. Predictors of judgement discrepancy may provide others with useful information about the characteristics of potential deaths by suicide according to other frameworks.