When I was injecting heroin in New York in the 1980s, it was impossible not to know the risks. Even as I shrugged off the after-school special scare stories, I saw evidence of them in my own life.
One of my friends, M*, once woke up with a dead man in her bed after having given him his first dose of the drug the night before. (Frighteningly, he had snorted the heroin, and never touched a needle.) A mutual friend of ours, K, also nearly died from shooting up in M's apartment.
We joked, in an attempt to deny the horror, that M was some kind of black widow.
But while the first death was clearly an accident, K's overdose was not. As my friends carried him out of M's walk-up – both to try to get help and to avoid drawing attention to a place we often took drugs together – he weakly said they should leave him to die. That time, he survived, but less than a year later, he hung himself.
He couldn't have been older than 25.
K's hanging made his intentions quite clear: with hangings and gun suicides, it doesn't usually take an expert to figure out what happened. Except for the small minority who get sexual pleasure from asphyxia, no one uses these methods on themselves for any other reason than to die.
Unfortunately, the question of how often overdoses are really suicides is far more complex.
When someone takes a deadly drug mixture that contains an opioid – the cause of most overdose deaths – the line between the intentional and unintentional is murky at best. Many people who truly want to live (like me, even during my active addiction) sometimes take insanely high doses and mixtures. It's well known that the best advertising a dealer can get is to have someone die from their batch. To wit: All of my heroin friends wanted to try the stuff that had killed the guy in M's bed.
But in the face of an overdose epidemic that now kills more people between 25 and 64 than car accidents, determining whether people actually intend to die really matters. For one, many pain patients say they would kill themselves if they feared they were about to be cut off from their medications. If such deaths are counted as accidental overdoses, however, access to those same medications can get blamed incorrectly. And the same information that reduces unintended death risk can also be misused by people who do want to die.
"I think we have a really serious problem in the US with undercounting poisoning suicides," says Ian Rockett, a professor of epidemiology at West Virginia University and expert in overdose deaths who's worked with the Centres for Disease Control and Prevention (CDC) on a committee studying suicidal intent.
Rockett and his colleagues published a study last summer that examined over 111,000 drug overdose deaths between 2008 and 2010. It found wide differences between the 50 states in the proportion of those deaths that were attributed to suicide, divergences that seem unlikely to be linked to the actual rates of intentional and unintentional deaths. For example, in Louisiana, just under 6 percent of overdose deaths were classified as suicides, while in South Dakota, the proportion was almost 29 percent.
"I think there's reason for concern more generally that drug intoxication suicides are being seriously undercounted," he says, "But more so in some places than others."
A tendency to mischaracterise such suicides has many potential causes, like issues related to death benefits that are denied in cases of suicide, and sensitivity towards families who have already suffered a terrible loss. Also: A high level of proof is required to label a death a suicide, so in an overdose where there is no note and no one reports recent suicidal behavior, it is likely to go undetected.
Rockett's study found that deaths in the South, which has also been hit hard by opioids, were far less likely to be labeled as suicides, compared to those in the rest of the country.
And while it's possible that people who take opioids in South Dakota are more than four times as likely to be suicidal than those in Louisiana – and Kentucky opioid overdoses are half as likely to be suicides as those in New York – the sheer size of these discrepancies strongly suggests other factors are at play. For instance, suicide may be more stigmatized in Southern states, or regions with fewer fatalities may have more resources and time to do careful death investigations.
Research on calls to poison centres related to opioid overdose also suggests that suicidal intentions are commonly missed in those who actually die. A 2015 study of over 184,000 calls received by these centres (including over 1,000 that were linked to deaths) found that two thirds of the cases overall involved signs of intent to commit suicide. Among those who died, the proportion was even higher: 75 percent in people aged 20 to 59 and 86 percent in those over 60.
Further, rates of known suicides are also rising, particularly in one group that is at the highest risk for opioid overdose: whites aged 45 to 54. Between 1999 and 2013, the known suicide rate in this population rose by nearly 10 percent and the poisoning death rate increased 22 percent. If many of the "accidental" OD deaths are in fact suicides, that rate would clearly have risen even higher.
For obvious reasons, efforts to prevent suicide often require different tactics than those aimed at stopping accidental overdose. Teaching people that avoiding drug mixing dramatically reduces death risk also suggests that if you want to die, mixing is the way to go. Information about not shooting up alone and having the opioid antidote Naloxone on hand could similarly be abused.
This is not to suggest that such efforts shouldn't be made – it would be absurd to argue that we should let some people die of ignorance in order to prevent others from killing themselves. Instead, we need to look at why so many people clearly are seeking oblivion and what we can do to make their lives better.
Discovering suicidal intent is of particular concern for pain patients, who are now reporting awful experiences of not being able to get needed opioids as doctors and pharmacists become more cautious and suspicious. The rate of suicide in this group is already high – and intractable pain due to denial of medication is one cause. In our rush to prevent deaths from addiction, we need to be sure that we aren't increasing suicides in pain patients who are benefiting from medication.
Suicide has gone practically unmentioned in most discussions of the overdose problem, whether they focus on pain treatment or on addiction. And this isn't exactly surprising when the conversation is focused on or led by families who have suffered such great pain.
Rockett has made the case to the CDC that we need better ways of classifying overdose deaths, arguing that a category called "death by drug self-intoxication" would help distinguish between people who took the wrong dose and those who clearly wanted to get high, suicidal intent notwithstanding.
Rockett says officials have expressed interest in the classification. But I think the CDC should embrace this idea – and other ways of encouraging medical examiners and coroners to look a bit more closely at the possibility of suicide in overdose cases.
In order to spare other families from heartbreak, we need better data on just how bad the suicide problem is among people who overdose – and what's making so many people feel their lives aren't worth living.
*Not their real names.
If you are struggling with depression or suicide, call the National Suicide Prevention Lifeline at 1-800-273-8255.
Follow Maia Szalavitz on Twitter.