Kramer writes with a measure of guilt in Ordinarily Well about his desire, after publishing Listening to Prozac, to escape his association with the celebrity drug. He wrote a novel, investigated other matters, saw patients, taught (he is a professor at Brown University’s medical school). But Kramer’s wagon was hitched to a psychopharmacological star 25 years ago, and in the past decade, he found that he could no longer ignore what he felt had become a veritable assault on antidepressants. His thankless but bracing challenge in his new book is to explain how “evidence-based medicine”—a movement that took off in the 1990s with the laudable goal of grounding clinical recommendations in objective studies—has facilitated the opposite of its original intent. It has, he believes, shrunk the circle of allowable evidence and encouraged a blinkered certitude inimical to good medical science.
Determined to meet Prozac’s critics on their own terms, Kramer curbs his natural storyteller’s inclination, corralling personal reflections and case histories into discrete sections called “interludes” as he delves into serious lab work. The nether region of Ordinarily Well is less fun than the Freudian underworld where bird-headed gods dance with your mother; it is also more frightening—at least for humanities majors. It is the dwelling place of randomized trials, effect sizes, blind studies, confounds, meta-analysis, placebo effects, and the Hamilton scale. But Kramer is an excellent guide as he subjects evidence-based purism—“a near-exclusive reliance on the findings of randomized, controlled, double-blind outcome trials”—to the scrutiny he believes it needs.
Randomized trials were a huge medical advance in the 1940s, but Kramer explains that Sir Austin Bradford Hill, the statistician who pioneered the new design, later warned about their limitations. When it came to testing disorders like anxiety and depression, with subjective symptoms, Hill saw that “to optimize outcomes, doctors would need to adjust doses and observe responses,” Kramer writes, and “the clinician’s perception might be the most accurate gauge of results.” This is precisely what the Swiss psychiatrist Roland Kuhn—who discovered imipramine—did in 1956. Both a researcher and a clinician, he was able to prove the efficacy of an antidepressant that even 60 years later remains a standard, because he saw the patients in his asylum every day and carefully calibrated his trials to his patients’ needs, and to their responses to his recalibrations.
To demonstrate how far we have strayed from Kuhn’s hands-on, informed investigations, Kramer undertakes a fine piece of up-to-date reporting—a visit to a private drug-testing center that conducts trials, primarily with antidepressants, for pharmaceutical companies. He turns to storytelling not to dismiss the value of medical statistics, but to place before the reader’s eyes the main ingredient that never makes it onto the package label: people. It is one thing to be told just how variable placebo effects can be—changing with the weather, the economy, the nature of the “minimal supportive psychotherapy” supplied along with the sugar pill. It is very different to eavesdrop on a community of drug-trial subjects who often help recruit one another—generally from a population very different from the one the drugs are intended for. Some are homeless, others merely unemployed and eager for the social environment of the testing center and the $50 or so they are paid per visit. Many have an incentive to exaggerate the severity of their depression; they can then remain in the study and receive free health care. Others underreport the nature of intractable afflictions in order to qualify for the trial. Listen, along with Kramer, as a subject supplies answers for a checklist, and you’ll understand his “despair of rating scales.”
The rater asks, “When you are in a stressful situation, does your voice quake?”
“Under stress?” Albert’s voice goes up a pitch, cracks, and wavers. “I would say no.”
Kramer is not denying the usefulness of drug trials, but showing how context matters. He also parses the complex reasons FDA tests fulfill their intended purpose (to establish minimal effectiveness and drug safety), yet are a lousy source of information about a drug’s true capabilities, whatever they may be. In other words, extrapolating from FDA data—which critics of medication often do—rather than devising new studies is like using a pass/fail course to determine a student’s caliber.