Deconstructing the CDC Report on Suicide
The media's reaction to the CDC's report on suicide is somewhat misleading.
Amidst two high-profile celebrity deaths by suicide in the same distressing week in June, the CDC released a report finding that suicide rates in the U.S. had risen. The main findings were upsetting on their own, but in addition there was a very specific cause for concern in the way the media responded to a specific portion of the report. A particular headline, proclaiming that suicide is “more than a mental health issue” began appearing in almost every major news source shortly after the release of the report. A flurry of headlines and accompanying articles seemed to suggest that the CDC was suddenly casting a great deal of doubt on a long-held tenet in the mental health field that over 90% of suicides occur in people with mental health issues and that the two are inextricably intertwined. If this were in fact the case, it would be a big change in how the field thinks about suicide, its causes, and how to prevent it. Indeed, as one headline put it, “What leads to suicide: a new report is challenging people’s assumptions.”
But is it really true that the new CDC report is “challenging people’s assumptions?” Or are we actually dealing with something that is far more complex than these headlines make it out to be? Indeed, we are most likely dealing with the latter situation. The report does not in fact cast “widespread doubt” on the prevailing notion that mental health issues are a key component of factors contributing to suicide. In fact, for technical reasons, there would be no way for the type of study that generated the report to establish causes in this way. What we are looking at here is a case of over-simplification and misinterpretation of a statement by a CDC official that led to widespread misunderstanding of portions of this report. It is further affirmation of the fact that the intricacies of communicating complex scientific information are not always front of mind for the people we rely upon to communicate it. Nonetheless, scientists and government officials overlook these intricacies and the various ways in which this complex information can be misinterpreted at their own peril.
So why did this report suddenly seem to reverse a well-held tenet in the mental health field that mental illness is involved in the vast majority of suicides? There are a couple of technical reasons for this seeming reversal as well as communications from the CDC that could have been more careful.
On the technical side of things, the data sources the CDC used for this report were highly likely to miss many cases of mental illness when counting suicides. For this report, the CDC used three sources of information: death certificates, coroners’ reports, and police reports. This method is always going to underestimate the percentage of suicides that involve mental health diagnoses. Death certificates do not specify whether someone who died by any cause, including suicide, suffered from a psychiatric or substance use disorder at some point before dying. Coroners and medical examiners can inquire about this from family members, but they often do not do so and even when they do, their inquiries are limited. Many people with depression, substance use disorder, and other mental health conditions who die by suicide are not in treatment at the time of their deaths, for example, and therefore simply obtaining medical records will miss a lot of cases. Police do not write their reports with an eye to supplying data for CDC morbidity and mortality reports and thus have no reason to include information about mental health. Thus, the sources of data used by the CDC to determine if people who died by suicide suffered with mental illness are far from systematic. When psychological autopsies are performed following suicide, the yield of people with a mental illness is closer to 90%.
No data source is perfect and the fact that there may be under-counting of diagnoses associated with suicide for this report is not our main point here. It makes sense that the CDC would not be able to obtain psychological autopsies for a large-scale report such as this one. In general, their use of these particular data sources got the primary job done – in other words, it was an effective and practical way to measure rates of suicides across the population over an extensive time period. The problem is that the CDC did not do a good enough job of communicating the limitations of the data sources they used for the report. As a result, people took the “finding” that only 54% of people who died by suicide had a mental health diagnosis at face value, and the media ran with it as a seeming example of another medical “reversal.”
This misunderstanding was then compounded by Anne Schuchat’s statement that “Our data suggests [sic] that suicide is more than a mental health issue…We think that a comprehensive approach to suicide is what’s needed. If we only look at this as a mental health issue, we won’t make the progress that we need.” This is the statement that many media sources cited to claim that suicide is not really a mental health issue.
But that conclusion is clearly too extreme, especially given the technical limitations cited above. It’s also not really what Schuchat meant. If we look at her statement more closely, Schuchat was not necessarily claiming that there isn’t a causal link between mental health and suicide, which is what most media sources took from this statement. She’s mostly referring to how we approach suicide prevention. If we focus too much on clinical interventions that reach only people who are in treatment, we may miss a lot of people who are not in treatment, which doesn’t mean that they don’t have a mental health issue. The point is that evidence-based, comprehensive, upstream approaches to suicide prevention are the preferred route because clinical interventions on their own will always miss people who are not in treatment for a variety of reasons, even if they have a mental health issue. Most experts on suicide prevention now insist that only a public health approach can be effective. Yet in the context of a report that seemed to be proclaiming that suicide is less related to mental health than we thought, Schuchat’s statement sounds like it’s simply confirming that conclusion.
Suicide is a mysterious, terrifying behavior and the reported uptick from the CDC was certainly alarming. People were especially looking for answers that week given the high-profile suicides of Kate Spade and Anthony Bourdain. We are naturally inclined to look for causes and to misinterpret everything in our midst as a cause. That’s partially how Schuchat’s statement and some elements of the report were misinterpreted, and the media ran with it, for understandable reasons.
Yet the CDC and other public officials who deal with highly complex health and science-related topics should know about the basic tenets of how most people interpret information, like the tendency to overemphasize anything that looks like a cause. As a result, they should be exceedingly cautious and exceedingly clear when it comes to statements that could be misinterpreted, especially in a causal fashion. It is in fact in no way the case from this report that suicides have less to do with mental illness than we thought. It is also in no way clear from Schuchat’s statement that she thinks that either. In the end, we still have a lot of work to do to better understand what causes suicide and how to prevent it. In the meantime, we should be careful about too quickly discarding decades of work on the relationship between mental health and suicide and we should always ensure we are communicating in a way that does not misrepresent the data and the limitations of our scientific process.