(This is the third post in this series; my first post discusses the most popular current theory of suicide as well as some statistics on it, while my second post talks about what suicide prevention means in practice.)
In the previous posts about World Suicide Prevention Day, I looked at what research currently says about suicide, and at what ways organisations and societies try to prevent suicide. In this post, I’m going to look more closely at what these methods assume about suicide prevention, and if those assumptions make sense.
Assumption 1) There is a good reason to exclude workplaces and jobs from this conversation.
When first searching generally online, links to suicide “post-vention” appear more readily than for prevention. However, there are some useful resources online, mostly created by Australian mental health organisations.
Many workplaces do have well-being plans, and some have specific suicide prevention programs. But their effectiveness is unclear, because many programs cannot be easily compared or evaluated. In a 2014 systematic review (a study which critiques and compares multiple studies on the same subject), the researchers could only include 13 workplace-specific studies. Only five programs had published evaluations online. However, those five suggested that the initiatives helped; employees reported a variety of improvements including an increased likelihood of asking for help, an increased knowledge of risk factors, and (in the two longest-running studies) a decrease in suicide rates.
Workplace initiatives seem like they could help, based on the few which have been evaluated. But there’s another aspect to consider too. The study authors criticised these initiatives, because they started from the point of helping people who were feeling low or at risk of suicide, but did not take the proactive step of analysing whether parts of the workplaces themselves may be risk factors or contributors. That’s an oversight which needs fixing; if workplaces could make small changes which reduced potential risk factors in the first place, then fewer employees would become at-risk. If workplaces could connect to other aspects of suicide prevention, rather than being their own island, even more so.
Based on the evidence so far, it seems like workplaces shouldn’t be left out of conversations. They can do some good already, and have the potential to do even more, especially if they are studied and tested in the same way as other types of initiative.
Assumption 2) We can measure whether specific suicide prevention initiatives help.
In the previous post I discussed some types of initiatives, which I’ll summarise here:
Examples of specific initiatives include: screening tests which assess people’s current level of suicidal feelings or their risk of suicide; environmental changes which reduce people’s access to lethal environments and means; and social reduction. Social reduction includes community programs; education about risk factors; resilience programs for at-risk people; in-school wellness programs; reducing the maximum purchasable amount of over-the-counter medications; reducing substance misuse and domestic violence; and influencing how the news reports on suicide.
Means reduction is the most widely studied initiative so far. A research summary from the Action Alliance for Suicide Prevention discussed some examples of successful means reduction efforts. In the 1960’s, the most common suicide method in the UK was inhaling then-toxic domestic gas: a country-wide switch to a non-toxic gas dramatically reduced suicide rates. Similarly, Sri Lanka experienced fewer suicides after legislation prevented the sale of some commonly-used toxic pesticides. Other forms of means reduction such as barrier fences are thought to be helpful, but these are harder to get data on because they represent a much smaller proportion of suicides.
Community programs are often overlooked by research, while widespread changes such as reducing substance misuse and domestic violence or influencing news reports are much harder to test in a controlled way. Finally, although screening tests have been studied to an extent, they are also limited. For ethical reasons, people who are severely ill or very vulnerable cannot usually be studied, so the most at-risk people cannot be included in research. In a summary of USA-based research on various suicide prevention initiatives, screening tests could identify people who were at-risk but at the cost of picking up thousands of times more false positives. For example, if a test could identify ten people who were at high risk of attempting suicide, it might also wrongly say categorise 3000 other people as high-risk. These large results can hide who actually is at risk.
Also, none of the included studies could test the larger question of whether screening in primary care (at a GP/family doctor) reduced actual death rates. So the biggest piece of that puzzle is still unclear.
Assumption 3) We can accurately understand who is likely to be suicidal, to better target services
Unfortunately, predicting who may become suicidal isn’t feasible yet.
Researchers in the USA carried out a meta-analysis of 365 different studies on suicide risk factors, which aimed to combine and understand 50 years of research. Over the last half-century, studies have found consistent risk factors such as previous depression, social isolation, and a family history of suicide. However, these studies could not predict people’s behaviour accurately. The studies could only predict people’s behaviour slightly better than chance alone could, and even with more knowledge the studies didn’t make better predictions over time.
One explanation might be that studying risk factors isn’t the right approach. Risk factors aren’t strange, rare traits – they’re everywhere. From family history, to previous medical conditions, to past and current life events, every aspect of life contains risk factors. Everyone has multiple factors in their life which increase their risk of ever completing suicide, as well as multiple factors which reduce their risk. However, each individual risk factor has a tiny effect on overall danger. The likelihood of a specific person completing suicide in a particular year is so small that even if a risk factor could double or triple their chance, it would still be almost undetectable. According to the meta-analysis authors, a risk factor would have to increase someone’s likelihood of attempting suicide in a given year by 30x or their likelihood of completing suicide in a given year by 750x to be useful.
One potential solution could be to switch from studying solitary risk factors to developing risk algorithms, which would try to make predictions based on combinations of many different factors. While early research on risk algorithms suggests they will eventually be more effective, they are currently too undeveloped to be helpful.
There is a fourth assumption here too:
Assumption 4) All suicides are the result of mental health issues/illness/life events, and every suicide is preventable.
The question of whether anyone can be rationally suicidal, of whether they can decide upon it with a clear mind, is often given to philosophy rather than psychology. It’s not a question I’m going to wade into here, as I don’t know enough to give a nuanced answer, but I will link to a paper which discusses multiple viewpoints.
Suicide is commonly presented being as the result of mental health issues, including the widespread claim that 90% of people who complete suicide are experiencing current depression. This claim seems to originate from an analysis of 154 studies which used the “psychological autopsy” approach, where researchers use interviews with family members of the deceased person alongside medical data to try and establish what circumstances they were facing. In this analysis, an average of 90% of participants per study seemed to be experiencing mental health issues.
However, studies from other cultures suggest lower rates than this. In an (unfortunately paywalled) study based in China, only 45% of the young people studied met the criteria of a mental illness at the time of their death, while another China-based study found that 65% of participants met that criteria. While two studies provides a much smaller base of evidence than 154 studies, these findings still suggest that the connections between mental illness and suicide might depend on culture rather than being universal.