Headlines raise alarm over prescription drug depression risk. But who qualifies as depressed?

Hard to tell if these headlines are using fear-mongering as clickbait, or simply riding the latest wave of stories reporting on the escalating suicide rates in the U.S.:

Millions are taking drugs, including birth control, that increase the risk of suicide (Quartz)

Study: Adults Use Medications That Can Lead to Depression and Suicide (US News & World Report)

Regardless, they’re misguided, misleading, and not supported by a study published earlier this week in JAMA . Here’s why:

• This was an observational study that cannot prove such cause-and-effect as the headlines imply.
• The study couldn’t control for other potentially significant contributors such as a pre-existing mental illness and substance abuse.
• The study can’t rule out an alternate hypothesis, which is that people taking medication are more likely to have other health problems that can lead to depression.

What the study found

This study looked at prescription drugs that list depression and suicidal thinking as side effects, to see if their use was, indeed, associated with more depressive symptoms.

It could only suggest associations, and the key findings of the study are already well known to many healthcare providers:

• There are hundreds of currently prescribed medicines that list depressive symptoms or suicidal thinking as side effects.
• The most notorious are beta blockers (for blood pressure), proton pump inhibitors (for acid reflux), female hormones, anti-seizure medications, and narcotics.
• The more of these medications a patient was taking, the more likely they were to self-report feelings of depression.

A take-home message

This latter point is key and was highlighted as a “take-home message” by the lead author in this news release.

Polypharmacy can lead to depressive symptoms … patients and health care providers need to be aware of the risk of depression that comes with all kinds of common prescription drugs, many of which are available over-the-counter … many may be surprised to learn that their medications — despite having nothing to do with mood or anxiety or any other condition normally associated with depression — can increase their risk of experiencing depressive symptoms.

That’s useful information that IS supported by the study, and I was glad to see it included in the reporting by Reuters, NPR, and even Quartz (despite their awful headline).

Who qualifies as depressed?

This is a question that’s rarely posed in news coverage of research on mood disorders. Reporting on this study was no exception. Not a single article I came across challenged the questionnaire used in this study: the Patient Health Questionnaire 9 (PHQ-9).

Here is that questionnaire:

To qualify as depressed in this study a score of 10 or more was needed. That’s right on the cusp of mild-to-moderate depression.

But there are some important things to bear in mind.

First, this is self-reported by patients. And that’s notoriously difficult to replicate. Furthermore, a recent study suggests that psychiatric studies which rely upon such questionnaires tend to OVERestimate the prevalence; most likely because the questionnaires set a very low (ie. overinclusive) bar, and patients have difficulty recalling the duration and severity of their symptoms.

Second, the tool is designed to screen for depression, NOT diagnose it. (A positive result is intended to flag people who require follow-up from a clinician to confirm the diagnosis.) And it’s certainly not designed to predict (or diagnose) the potential for suicide. So headlines like the one featured in Quartz are hyperbolic to say the least.

Finally, the questionnaire was designed in the late 90s with a grant from Pfizer which, at the time, was marketing one of the best-selling anti-depressants in history: sertraline (Zoloft®). Ask yourself: would they benefit more from setting a low or high bar for depression?

A psychiatrist’s take

All this irks Allen Frances, MD, a former chairman of psychiatry at Duke University, who also chaired the task force charged with revising the 4th edition of the Diagnostic and Statistical Manual (DSM-IV) – the standardized criteria used by many mental health care providers in making diagnoses.

“The uncritical use of these self-report questionnaires is responsible for much of the inaccurate diagnosis of mental disorders that’s so ubiquitous in clinical practice and epidemiological studies. Self-reports are inherently biased to exaggerate rates of disorder because they can’t accurately assess the severity, duration, and clinical significance of symptoms. They can only serve as screeners. Careful clinical evaluation is still needed; it’s an imperfect, but necessary check, on the over-diagnosis caused by these questionnaires.”

Advice for journalists

Many reporters made it clear that this study could not establish cause and effect. And some even included the pragmatic take-home message about polypharmacy (ie. taking multiple drugs concurrently) stressed by the lead author. That’s an important service to readers.

But most outlets did not show the same care with their headlines, which abandoned all nuance in favor of high impact framing. For example USA Today’s take — “One-third of adults in US taking drugs that may cause depression, study finds” — did not include any cautions or caveats about the study or its conclusions.

Their coverage (along with that of many other outlets) highlights an important weakness that is widespread in reporting on mental illness, and one we’d all benefit from addressing. That is, let’s show our readers how various mental illnesses are defined in the studies we’re covering. What are the criteria? How are they measured? And what are their limitations?

That so many of us are interested in stories about depression and suicide is completely understandable. That so many journalists too frequently overlook how researchers define depression is completely avoidable.