Is depression linked to violence? And if it is, what are the clinical implications? This Reading of the Week asks these two questions.

On the Germanwings’ Tragedy

Certainly in light of recent events, the depression-violence link is much discussed. Some have already weighed in. Consider this incredibly offensive newspaper front page.

“Madman.” “Crazed.” And if only that sort of language was restricted to tabloids. Last week, French Prime Minister Manuel Valls said: “everything is pointing towards an act that we can’t describe: criminal, crazy, suicidal.” It’s enough to make us nostalgic for the Sarkozy administration. (Who would have predicted such nostalgia three years ago?)

To be clear, though coverage has been breathless, little is known. In a thoughtful piece in The Atlantic titled “Depressed Doesn’t Mean Dangerous,” Julie Beck notes: “What evidence we have of Lubitz’s mental health or lack thereof is still scant.” Indeed, we can’t answer the most basic questions. Was the ill-fated flight’s co-pilot in active treatment for depression? Was he on medications? What was his mental state in the days leading up to the tragedy?

In The Globe and Mail, CAMH’s Dr. Jeff Daskalakis notes that:

We have been conducting a sort of psychological autopsy in the media by putting together fragments of information from his life to try to answer the question, “why did he do it?”

Noting how little we know, Dr. Daskalakis suggests we may never be able to answer that question – but he worries about inflaming stigma and stereotypes. Putting things in a larger context, Dr. Daskalakis observes:

We know that the stigma of mental illness is a major barrier that prevents people from getting care. In Ontario alone, as many as 65 percent of people with depression remain untreated.

On the Reading

But there is the unanswered question: can we link depression to violence?

As is turns out, The Lancet Psychiatry just published a major paper on this topic: “Depression and violence: a Swedish population study” by Seena Fazel et al. This excellent paper is our week’s Reading.

Drawing on Swedish databases, Fazel et al. offer a triple analysis: a look at violent crime among the depressed vs. the general population, and then a consideration of siblings and twins (three forensic psychiatry studies for the price of one). To do such work, they used several population studies from Sweden – this paper taps a huge data set. It’s a remarkable effort.

Here’s what the Oxford researchers did:

· Using 5 different registries in Sweden, they considered more than 47,000 people with depression and made comparisons. (!)

· Because admissions can be tied violence, as can other diagnoses, they excluded all people with inpatient admissions and comorbidities. (!)

· For the study, people needed to have had at least 2 episodes of depression to be included.

· They looked at convictions for violent crimes between 1972 and 2009; convictions were for homicide, assault (aggravated and common), robbery, arson, any sexual offence, sexual harassment and illegal threats.

· They then compared people with depression with people without depression (same sex, similar age) from a sample of 900,000 Swedes.

· They also compared siblings and half-siblings (15,500+ and 33,500+).

· They then compared twin pairs, looking at a sample of roughly 5,600 monozygotic twins and almost 10,000 dizygotic twins.

Here’s what they found:

· After removing the effects of low income, drug abuse, and other factors, they found an odds ratio of 3.0; in other words, people with depression were 3 times more likely to commit crime than those who didn’t have this mental illness. (!)

· The relative risk: only 3.7% of men were convicted of violent crimes with depression and 0.5% of women. (!)

· The authors note that depression combined with a history of substance, violent crime and self-harm was a potent mix, pushing up the relative risk in men to 16.3% and 1.4% in women. (!)

· In terms of siblings, the researchers found an odds ratio of 1.2 for half-siblings and 1.5 for full siblings.

· The twin data wasn’t big enough to draw meaningful conclusions.

They conclude:

In 47,158 individuals with depression, we noted three-fold increased odds of violent crime after adjustment for sociodemographic factors; this rate remained significantly elevated when we excluded patients with a previous history of substance abuse, any criminality, or self-harm.

The authors then push further:

In conclusion, we have shown an association between a diagnosis of depression and violent crime, which is independent of potential cofounders and uses two complementary designs. The magnitude of the effects certainly when compared with those for self-harm in the same population, suggest that assessment of the risk violence should be considered routinely for some individuals with clinical depression and potentially included more consistently in clinical guidelines.

Wow. Big study, big conclusions.

A few points.

1. The limitations. There are several limitations here. Start with the fact that the study only looks at Sweden and Swedish data (and I want to emphasize that some of my best friends are Swedish).

More importantly: we have almost no information on “depression” – are these people in active treatment? Are they recovered? The Mental Elf (a favourite blog of mine) lists more limitations, drawing on the work of a Cambridge professor. They note, for example, that the onset of depression was 30 – missing the adolescent population.

2. The controversial conclusion. The Lancet Psychiatry actually ran a commentary critical of the conclusion that guidelines recommend screening for violence, noting that – statistically speaking – it is impractical to screen a large population unless you have an awesomely effective tool (that is, incredibly specific and sensitive). To put that in non-technical language, if you throw a fishing-net into shallow waters, you will catch bottom feeders and mud unless you are careful.

3. The perspective. Yes, people with depression are at increased risk of violence (over the general population), but let’s remember that the majority of crime isn’t committed by people with mental illness. Or, as Dr. Fazel said about his own paper:

You can say that someone [who’s depressed] has a three- or fourfold higher risk of violence, but if that only amounts to 1 to 2% more risk in a 10-year period – that’s really a very small risk in absolute terms.

Further Reading

Dr. Daskalakis’ op ed can be found here:

The Mental Elf blog on this paper can be found here:

And for a paper putting violent crime and mental illness in a larger context, consider this earlier American Journal of Psychiatry paper by Seena Fazel:

Reading of the Week. Every week I pick a reading — often an article or a paper — from the world of Psychiatry.