The latest and greatest miracles, break-throughs, and fads are trumpeted in the news, and the newsroom's dwindling coffers combined with a "have to beat the other guys to it" 24-hour news cycle means consumers are often shortchanged.
Think about O.J. Simpson. When the story first broke, speculation ran rampant. In our country it's now perceived guilty until proven innocent, and even then people might think you're still guilty. Most of us weren't in the courtroom, and even fewer were jurors privy to the evidence. We instead had to rely on reporters to convey the news. Most people made judgments on whether O.J. was innocent before a jury of his peers rendered a decision, and then moved on with their daily lives.
Now think about a story about, say, a new finding that says eating chocolate cupcakes every day may help prevent cancer. (I'm still waiting for this research to be conducted.) Nightly news anchors would interview various experts, news wire services would send out reports to be picked up by local dailies, Google News would feature it as a top story. Here's the problem - what happens if people make partially informed, and possibly harmful, decisions - like stocking up on chocolate cupcakes - based on early reports of research before more detailed analyses can be disseminated?
Let's switch to a real-life story, this one about antidepressants. One of the big players in the pharmaceutical world, based on sales, is the antidepressant class (which account for $10 billion in sales in 2008 in the U.S. alone1). Of the top 15 products in the U.S. by sales, three are antidepressants - Abilify, Effexor XR, and Lexapro. (Of the top selling generics last year, two were antipsychotics and one was an antidepressant - budeprion, better known as Wellbutrin.) And of prescriptions dispensed last year, antidepressants ranked third at 164.2 million dispensed. In other words, antidepressants are big business, and a lot of people use them (about 27 million people, according to 2005 data).
News on antidepressants, then, will be of interest to many patients and - as any story that impacts decision-making - should be careful to present all the facts and context so consumers can make the best choices.
A flood of vagueness
An article in the January 6, 2010, issue of JAMA has entered deluge-status, flooding media outlets. Here are a few randomly selected headlines:
Pills seen as less helpful for mild depression2
Study: Drugs work best for severe depression3
Antidepressants Work For Severe, Not Mild, Depression4
Study: Antidepressant lift may be all in your head5
Placebos: Pretty Good for Depression6
Limits to antidepressants' effectiveness: study7
Popular Drugs May Help Only Severe Depression8
Milder depression may not respond to antidepressants9
Antidepressants: Best for Severe Depression?10
Antidepressants aren't the only remedy for milder forms of depression: study11
Study finds medication of little help to patients with mild, moderate depression12
Scanning these headlines, what do you think the JAMA study was about? That antidepressants in general don't really work unless you're standing on a ledge? The actual JAMA article title doesn't do much in the way of shedding light: "Antidepressant Drug Effects and Depression Severity: A Patient-Level Meta-analysis."
Looking beyond the headlines, some of the articles do at least a silver, and possibly gold, medal job or reporting on the study: the Columbus Dispatch, the New York Times, LA Times, CNN, and WebMD. The bronze goes to Reuters, USA Today, and possibly the Wall Street Journal and the NY Daily News. Fierce Pharma isn't even in the running.
Why such a discrepancy among news organizations? Surely the Wall Street Journal has more star power than the Columbus Dispatch. And shouldn't the basic facts stay constant throughout? Sadly, no.
What does the JAMA article actually say?
As Fournier et al. write in their abstract, "antidepressant medications represent the best established treatment for major depressive disorder (MDD), but there is little evidence that they have a specific pharmacological effect relative to pill placebo for patients with less severe depression." True - many therapies approved for depression do show effects in severely depressed patients (we'll get to definitions in a bit), and there is a paucity of studies on the true effect on mildly to moderately depressed patients. Fournier et al. conducted a meta-analysis of six clinical trials "to estimate the relative benefit of medication versus placebo across a wide range of initial symptom severity in patients diagnosed with depression."
Given that about 30% of people who seek treatment for depression would be considered very severely depressed, you might be tempted to conclude that placebo is about as effective as an antidepressant in 70% of treated patients. This isn't entirely correct, as we'll soon see.
Fournier et al. examined six studies that, unlike previous analyses, included a range of Hamilton Depression Rating Scale (HDRS) scores. The HDRS is one of the most commonly used scales for rating depression13. Scores range from 8-13 (mild depression), 14-18 (moderate), 19-22 (severe) and greater than or equal to 23 (very severe). Most clinical studies employ two methods to show a greater effect - a placebo washout period, and HDRS scores greater than 20 or so. The placebo washout period allows exclusion of patients that improve upon receiving placebo, and the high HDRS level ensures that the greatest change from baseline to end is achieved. (This isn't necessarily nefarious; it makes statistical sense.) Fournier's analysis included scores ranging from the low teens to the 30s; a treatment/placebo comparison of at least 6 weeks' duration, and scores available from both intake and end of treatment.
Most of the stories covered the HDRS at some level (surface level to deep), but a surprising number neglected to mention a critical feature: of the six studies, three involved Paxil (an SSRI) and three involved an old tricyclic called imipramine, which is hardly used anymore. Several caveats are noted, including that neither is typically used first-line, two of the studies of imipramine were at a low dose, the study duration was only 6 to 11 weeks on average, and Paxil has some additional side effects beyond those typical of the SSRI class.
These are hardly generalizable studies. Newer antidepressants, some of which are more efficacious and/or have different mechanisms of action, may have had different outcomes. Yet this is not the outcome you would glean by skimming the headlines.
What do the news articles say?
The gold and silver medalists. The Columbus Dispatch notes that the studies considered only two drugs, and that patients shouldn't stop taking their medicine based on the results. Sources in the article also note that the decision is really patient-dependent. "Some patients with less-severe depression have a hard time because of particular stressors in their lives, not because they're biologically hard-wired for depression," one doctor said. "It's not going to change your poor financial situation; it's not going to change your terrible relationship with your spouse." And if medication seems to help, the doctor advised against second-guessing it.
The NY Times, CNN, LA Times and WebMD cautioned patients against refusing to try medicine, and raised the possibility (as did many of the news reports) that non-pharmacological treatment - attention from physicians, talk therapy, exercise, etc - could benefit a mildly or moderately depressed patient as much as a pill. They also noted that, much like any other disease process, the more severe the disease or symptoms, the more improvement with treatment.
The bronze medalists. Reuters had bits and pieces of the key points, but focused more on the idea that people should consider alternatives to pharmacotherapy without going into why. USA Today touches on the surface of most of the points, but end by bringing up a different study from 2008 (which didn't have the same criteria as Fournier's) and that "the system" failed in approving these drugs. The WSJ focused on the financial aspect (which is more of its audience, so this is to be expected) and wasn't quite clear that there were only two drugs studied and the results aren't generalizable. (And by nature, the WSJ Health Blog was more abbreviated, but did link to other stories.) The NY Daily News, as one might expect, took a rather slanted approach: "Unsurprisingly, GlaxoSmithKline spokeswoman Sarah Alspach seemed unconvinced by the report's findings." They also did not discuss alternatives or whether the results could be generalized.
Watching on TV back at home. Fierce Pharma was a bit off its game on this one. From quotes like "Yet another study is calling the efficacy of antidepressants into question," to "What about all those studies that show antidepressants work better than placebo? You know, the studies that helped these drugs get approved?" the article seems to be going more for levity than substance.
What did the JAMA researchers conclude?
The findings "indicate that the efficacy of antidepressant medication treatment (ADM) for depression varies considerably as a function of symptom severity." This is fairly logical, but this study is one of the first to actually show that. They continue, "True drug effects (an advantage of ADM over placebo) were nonexistent to negligible among depressed patients with mild, moderate and even severe baseline symptoms, whereas they were large for patients with very severe symptoms." Again, no big surprises here (with possibly the exception of results in severe, versus very severe, patients). The researchers even note that these results were essentially what they expected.
"What makes our findings surprising," they note, "is the high level of depression symptom severity that appears to be required for clinically meaningful drug/placebo differences to emerge, particularly given the evidence that the majority of patients receiving ADM in clinical practice present with scores below these levels."
Their concluding paragraph is worth quoting in full:
Prescribers, policy makers, and consumers may not be aware that the efficacy of medications largely has been established on the basis of studies that have included only those individuals with more severe forms of depression. This important feature of the evidence base is not reflected in the implicit messages present in the marketing of these medications to clinicians and the public. There is little mention of the fact that efficacy data often come from studies that exclude precisely those MDD patients who derive little specific pharmacological benefit from taking medications. Pending findings contrary to those reported here and those obtained by Kirsch et al and Kahn et al, efforts should be made to clarify to clinicians and prospective patients that whereas ADM can have substantial effect with more severe depressions, there is little evidence to suggest that they produce specific pharmacological benefit for the majority of patients with less severe acute depressions.While I agree with most of the paragraph, and most of it can indeed be extended to most medications (not just antidepressants), what they leave out is a call for research into this question of whether pills can help people with mild to moderate depression. Also, a group of 100 mildly depressed patients will have 100 different causes of their depression. If, as a physician said in the Columbus Dispatch, the depression is due to environmental stressors, then talk therapy or other non-pharmacological alternates may indeed work equally as well as a pill. But if a mildly depressed person has mild neurotransmitter imbalances, a pill may work better than alternatives.
An article from NPR's Alix Spiegel reflects on the fact that "shockingly few of the studies on antidepressant pills published over the last thirty or so years have considered how the medications affect people with only mild or moderate symptoms." The JAMA article "tries to do that and concludes antidepressants don't seem to work any better than a sugar pill for those with milder forms of depression. But the differences in how depression is measured in academic studies compared with how doctors assess it in the real world is another factor to bear in mind."
Mental health practitioners, Spiegel says, say the "severe" and "very severe" categories are
misleading and don't corresponded well with how real-life doctors see their patients. "'The Hamilton concept of severe, I think many psychiatrists would think of as moderate,' said Dr. Mark Olfson, a professor of clinical psychiatry at Columbia University."
Spiegel's conclusion is spot on, and should be one of the key takeaways from the JAMA article: "So, the conclusion of the JAMA study that antidepressants work only in severely depressed people may be missing the mark. The drugs may not be so weak. Part of the problem could be an academic scale for measuring depression that pushes people with forms of the disorder that most doctors would consider to be moderate into the severe camp."
To tie it all together - people don't often have access to evidence, raw data, or medical journal articles, so they rely on news media to convey the important information. Unfortunately, news media quality varies widely. And while it may be easier to smooth over the complexities of scientific research when writing for the lay public, sometimes the fine print contains the real story. For something as individual as your own health, remember that just because it's written, doesn't mean it's true. Except this article, of course.
1 IMS Health, U.S. National Sales Perspectives and National Prescription Audit Plus.
2 Columbus Dispatch
3 Fierce Pharma
4 Wall Street Journal
5 USA Today
6 WSJ Health Blog
8 New York Times
11 NY Daily News
12 LA Times
13 There is some controversy about this scale; it used to be the gold standard but has come under recent criticism, as the authors note.