“Medicine Is Awesome”: A Critical Look at the Factors That Shape Thinking About Depression and Its Treatment

Historical Roots of Modern Day “Depression”

Early conceptualizations of depression in the 1900s considered the symptoms of “melancholia” to result from complex intrapsychic factors (Figure 1). Depression and anxiety symptoms signaled a disconnect between intrapsychic needs and the outside world. The idea that there could be “depressive states” played a role in how organized psychiatry would later conceptualize clinical depression.

• Download figure
• Open in new tab

Figure 1.

Progression of conceptualization of depression. Abbreviation: DSM, Diagnostic and Statistical Manual of Mental Disorders.

The first Diagnostic and Statistical Manual of Mental Disorders (DSM), published in 1952, codified the appeal to more complex psychological factors to explain depressive symptoms. Clinical depression was believed to be either endogenous or reactive; endogenous depression, understood to be biologically based, was rare, “with a low lifetime prevalence rate of 1-2%.”¹ Thus, the DSM-I used the phrase “depressive reaction,” characterized by anhedonia, and understood it as a (conscious or unconscious) response to perceived or actual loss.² In the DSM-II, published in 1968, the wording changed slightly to reflect the psychoanalytic framework (and the broader zeitgeist) of depression; the term “neurosis” replaced “reaction.”³ Still, both the 1952 and 1968 manuals conceptualized depression as a response to something happening, either in someone’s environment or internally.

Psychiatry Becomes “Scientific”

By the 1980s, there was significant pressure on the profession of psychiatry to legitimize itself as a bona fide medical endeavor.⁴ The DSM-III, published in 1980,⁵ was a response to this threat and marked a paradigm shift in the conceptualization of psychiatric diagnosis, a shift that helped the specialty establish its legitimacy and “don the white coat.” Robert Spitzer, MD, who chaired the DSM-III, wanted to replace the “pseudo-scientific” framework of earlier editions with a psychiatric nosology that was reliable, valid, and anchored in pathophysiology in the same vein as other medical illness. The new framework explicitly incorporated the medical model and checklist approach, with a seemingly replicable list of specific symptoms and a timeframe for most disorders.

But in addition to DSM-III being seen as a reputation enhancer for the profession of psychiatry, something else happened. “Depressive neurosis” now became “major depressive disorder.” Instead of the general description in earlier editions of the DSM, clinicians were given diagnostic threshold criteria, instructed to count symptoms and assess their duration. If a patient exhibited 5 or more depressive symptoms over a 2-week period, they should be diagnosed with Major Depressive Disorder (MDD).

This change provided fertile ground to market a new crop of antidepressants, ones that were rooted in pathophysiology rather than just treating symptoms (amphetamines were marketed to treat “sadness” in the 1960s). Fluoxetine was marketed in 1988; sertraline in 1991. Pharmaceutical marketing emphasized the “chemical imbalance theory” of depression to the public. Many contemporary thought leaders in the field reinforced the idea that depression was a discrete physiologic disorder and that medication treatment for it fit hand-in-glove, like insulin for type 1 diabetes.⁶

This disease-centered model regarding the etiology and treatment of depression led sociologist Nikolos Rose to note that we have become “neurochemical selves.”⁷ He asks, “How did we come to think about our sadness as a condition called ‘depression’ caused by a chemical imbalance in the brain and amenable to treatment by drugs that would ‘rebalance’ these chemicals?”

Indeed, for over three decades, antidepressants have been considered—and promoted as—a first-line intervention for individuals with a MDD diagnosis of any severity.⁸ However, the literature is clear that, especially for mild to moderate depression, antidepressant treatment does not produce a symptom response greater than that of placebo. Approximately 50% of patients have some response to the first medication tried,⁹ but almost the same percentage of patients receiving placebo will have a decrease in symptoms of depression over the same period, no matter what type of antidepressant is studied. For over 15 years, meta-analyses have consistently found “minimal or non-existent” differences between antidepressants and placebo for mild to moderate depression.^10–12 In addition, up to 70% of treated patients will not achieve remission of symptoms.¹³

The difficulty showing a difference between active treatment and placebo could have several causes, including an expectancy effect, conditioning, the natural history of depression, or regression to the mean (i.e., in which extreme results, measured over time, tend to return to average).¹⁴ It may also be that a substantial proportion of patients wait to seek care for depression symptoms until late in the natural course of the syndrome. Without antidepressant therapy, episodes of clinical depression last from 2 months to several years, with an average of around 5 to 6 months. One-third of the patients recover within a year.¹⁵

The concept that antidepressants affect brain serotonin levels is based on a post hoc, ergo propter hoc fallacy (after this, therefore because of this) that has little pharmacologic support. Selective Serotonin Reuptake Inhibitors (SSRIs) were initially investigated for treatment of hypertension (they don’t work) and some study participants noticed improved mood during treatment. From these early observations spawned the “serotonin hypothesis” that antidepressants work by increasing synaptic levels of serotonin.

This concept of a relative serotonin deficiency fits well with existing medical models of disease and treatment. It is an adaptation of a common approach to numerous diseases, that there is a particular physiologic perturbation causing the disorder and a specific means of treatment. For instance, per the “magic bullet” conceptualization of antibiotic treatment of infection, a chemical specifically affects the bacterium without affecting human cells.¹⁶ Similarly, antidepressants were presumed to correct a serotonin deficit, leading to symptom resolution.

The problem with the serotonin deficiency theory is that it is unclear how this deficit occurs and what specific role serotonin plays in the development of symptoms of depression. Despite 20 years and billions of dollars spent on “psychiatric precision medicine,” no genetic marker for any mental illness, including depression, has been identified.¹⁷ Serotonin reuptake inhibitors may have very little effect on serotonin levels, and, like many psychotropic medications, their in vitro binding capacity to receptors other than serotonin is much higher.¹⁸

How a “Medicine Is Awesome” Approach Needs to Change

Our willingness to embrace this hypothesis centers on a general framework many of us operate on: “medicine is awesome.” It is the idea that medical care, and medicine itself, for most, is a great thing in that it lengthens life and improves the human condition. Therefore, we should provide medical care whenever possible. This “action bias” supported the adoption of a treatment paradigm that appeared to be evidence based. But when it comes to depression, we have had to revise our beliefs about etiology; the defined physiologic cause of depression is not a simple monoamine deficit.

Although evidence-based medicine is—and should be—grounded in empirical research, study design is malleable. All aspects of the research process, including study design, execution, analysis, interpretation, publication, and marketing, involve numerous choices that can consciously or unconsciously influence study outcomes.¹⁹ For example, recruitment can be biased toward antidepressants by selecting patients who received antidepressants until a “wash-out” period before randomization. This approach is like finding ice cream lovers who prefer chocolate, randomizing them to be given either vanilla or chocolate, and asking them which they prefer.²⁰

Also, most clinical trials of antidepressant therapy last only 12 weeks, despite the usual course of depression being much longer. During this time, small changes to some symptoms can make antidepressants look more effective than they are over time. For example, a clinically relevant difference between treatment and placebo on the commonly used Hamilton Depression Rating scale is only 3 or more points out of a possible 54 points.^10,21 In short-term clinical research, just two changes, e.g., resolution of sleep disturbance and simply acknowledging one is ill, can each, without a demonstrable effect on mood or function, result in a two point change on the scale.

What Can We Do?

To move away from a Medicine is Awesome framework when helping people with depressed mood, we need to acknowledge the limitations of the current state of knowledge, acknowledge what we know and what we don’t, and avoid codifying practices based on uncertain evidence. The threat to medicine is not that we do not know things. The threat is that we create a sense of certainty out of uncertain knowledge—that we profess to know things that we do not. This approach is concordant with several calls to limit interventions in medicine, embracing a “when in doubt, don’t” ethos.^22–25

That doesn’t mean we give up, but this humility will take us far. We have other simple, time-efficient ways to acknowledge patients’ suffering and offer them tangible, effective treatments that will help them.

First, we need to recognize that everyone’s “depression” is unique and must be understood in the context of their lives: “Mental health and well-being cannot be defined by the absence of a mental health condition, but must be defined instead by the social, psychosocial, political, economic, and physical environment that enables individuals and populations to live a life of dignity, with full enjoyment of their rights and in the equitable pursuit of their potential.”²⁶ Understanding this uniqueness allows us to address—even in primary care—some of the underlying social determinants of depression. For example, medical-legal partnerships—explicit attempts to identify and address health-harming legal needs of patients by having a pro-bono attorney on site who can address these needs (e.g., unsafe housing; immigration status)—is an effective clinical model with a proven success record.²⁷

Empathy is critically important in all primary care settings. A critical factor influencing whether people believe they have experienced quality care is whether they believe their provider demonstrated an empathic understanding of their illness.²⁸ Indeed, despite issues with the DSM’s conceptualization of depression, patients have legitimate suffering that must be acknowledged.

Empathy and the development of trust can allow for watchful waiting, especially for patients with mild depression. Watchful waiting is not the absence of care. Adopting this approach first and foremost requires validating the person’s experience of emotional distress.²⁹ Appropriately framing depression, explaining the natural history of depressed mood, offering a full range of options, and eliciting and discussing patients’ values take little time in the context of a trusting relationship. Explaining depression symptoms not as a disease or illness but as a functional signal of the patient’s life situation can avoid the iatrogenic “prognostic pessimism” that can occur with the biological framings of depression.³⁰

Options for treatment can be discussed, carefully aligning choices with the patient’s agenda. Guided self-help, group or individual therapy, group exercise, and group mindfulness and meditation can be initial treatments (Table 1). These options can be discussed as appropriate for the patient’s context: not all patients will be able to afford or attend therapy; others may prefer yoga or weightlifting classes. Medications can be discussed as an eventual option with a thorough explanation of limited benefit and possible harms.