Over the decades that there has been a focus on suicide prevention, there have been many lists of warning signs for suicide. These lists of signs persist today. However, the lists are particularly problematic, because they tend to induce guilt and shame in those who have lost someone by suicide rather than forming an effective prevention strategy.

To make the conversation concrete, let’s use the now disproved mnemonic IS PATH WARM from the American Association of Suicidology. It is:

• Ideation
• Substance Use
• Purposelessness
• Anxiety
• Trapped
• Hopelessness
• Withdraw
• Anger
• Recklessness
• Mood Changes

It’s not that this list isn’t correlated with suicide attempts – it is. The problem is the degree to which it’s correlated and how useful the signal is. Coming back to mood changes for a moment, there is no precision about degree. Do they seem happier than they normally are? Maybe they smile once more per hour or seem to have an extra spring in their step?

Most people in the public who read “mood changes” believe that it’s only when they become less happy or joyful. But as the research attests, the riskiest time for those with depression is as their depression starts to lift. Mood change is hard to tell prospectively but is often the source of retrospective guilt.

To return to the original point about the relationship between the indicator and the probability of an attempt, we’ll look at ideation – short for “suicidal ideation” – as a warning sign for suicide. We know that ideation impacts at least one-third of the population, and that suicide fatality rates are in the range of 14 per 100,000. That means that roughly 1 person in 2,400 who has ideation will die by suicide. If we estimate that there are 20 attempts for every suicide death, that means 1 in 120 people who have suicidal ideation will make an attempt.

Clearly, even ideation, which the ideation-to-action framework for suicide prevention would imply is a very strong indicator, isn’t very discriminatory. That is, it doesn’t discriminate between those who will attempt and those who won’t.

Even if we were to find a highly discriminatory warning sign, we have the problem that some people who consider suicide are actively trying to hide this fact from others. They’d work very hard to ensure that the warning sign wasn’t detectable by others for fear that they’d respond – and some people don’t want others to stop an attempt.

The final issue with warning signs is their detectability. Even if a highly discriminatory warning sign existed and the person didn’t attempt to suppress it, could we detect it in the noise of our daily life? The answer is, unfortunately, probably not. The work on biases shows that we’re able to detect and correct for these biases only when we’re actively thinking about them. Daniel Kahneman, who wrote Thinking, Fast and Slow, which is about biases, said that even writing the book didn’t help prevent him from avoiding them at all times.

Collectively, this is a strong argument that not everyone sends warning signs, and they couldn’t help us discriminate between those who want to die by suicide and those who don’t. Even when they don’t want to suppress the warning signs, we’re unlikely to detect them if they did.

That being said, three indicators are sufficiently specific that, if present, should be a strong indication that a person needs help. They are:

1. They’re directly stating that they want to die or are considering suicide.
2. They’re giving away their most prized possessions.
3. They’re indicating they will not be around.

It’s not that paying attention to the warning signs is bad, it’s about realizing they have limits.