Bridging the Gap Between Research and Real World

Ksenija Hotic

Over the past 25 years, the rate of antidepressant use in the United States has climbed steadily (Pratt et al., 2011). Despite more than 14 percent of Americans 60 and older being prescribed antidepressants (Pratt et al., 2011; Motjabai and Olfson, 2015), many older patients continue to experience significant depressive symptoms (Kessler et al., 2010).

Why do some older patients continue to experience symptoms of depression while on antidepressants?

Meta-analyses of placebo-controlled trials consistently demonstrate the efficacy of antidepressants in both the short- and long-term treatment of late-life depression (Kok et al., 2011; Nelson et al., 2014).

Some clinicians have suggested that antidepressants are effective in older patients who participate in randomized clinical trials but not in “real-world patients” seen in clinical settings. Typically, patients seen in clinical settings are more difficult to treat than those who enroll in clinical trials. However, this explanation is unlikely to account for the gap between the efficacy of antidepressants in clinical trials and their much poorer effectiveness when they are prescribed under usual care conditions.

Several studies point to the role that the delivery of care may play in how older patients respond to antidepressants (Mulsant et al., 2014). Two large randomized studies have shown that older patients treated following measurement-based algorithms are more likely to respond to antidepressants than those treated under usual care conditions (Unutzer et al., 2002; Alexopoulos et al., 2009).

Why might this be? Patients who participate in clinical trials are treated according to a well-defined treatment protocol that predetermines the antidepressant they should receive and its titration. This is based on depression and side-effect scores obtained with validated instruments. In contrast, under usual care conditions antidepressants are selected idiosyncratically to “match” patients’ specific factors and titrated based on clinical impression. These differences in process of care contributed to the finding that antidepressants prescribed under usual care conditions in these two large studies were half as effective as a placebo used in research clinics (Mulsant et al., 2014; Huang et al., 2015).

Other studies show that the number and frequency of visits can predict the efficacy of both antidepressants and placebo in the treatment of depression (Posternak and Zimmerman, 2007; Rutherford et al., 2014). In the absence of a structured process of care including regular visits to monitor symptoms and side effects, older, depressed patients who are dissatisfied with their antidepressants are likely to stop taking them.

Another likely explanation for the antidepressant efficacy-effectiveness gap is a partial mismatch between older patients who receive antidepressants and those who could benefit from them. The number of American adults who are prescribed an antidepressant is at least double the number who meet the criteria for a depressive or anxiety disorder and are most likely to benefit from antidepressant pharmacotherapy (Fournier et al., 2010; Kessler et al., 2010; VɆhringer and Ghaemi 2011; Wiechers et al.., 2014; Motjabai and Olfson, 2015). At the same time, other studies show that at least half of older patients with a major depressive disorder do not get treated (Kessler et al., 2010; Pratt et al., 2011). Depressed patients who are older, male, poor, Hispanic, or African American are among those least likely to be diagnosed as having depression and be treated for the disorder (Harman et al., 2001; Olfson and Marcus, 2009; Pratt et al., 2011).

Clinicians in general and psychiatrists in particular can improve the outcomes of older, depressed patients by following a few key principles:

• Be careful when selecting the patients to whom you prescribe antidepressants. Some older patients who are asking for an antidepressant probably do not need one and some patients who are reluctant to take an antidepressant may need it.

• Use a measurement-based algorithm. Instead of trying to match specific patients with specific antidepressants (Simon and Perlis 2010; Mulsant et al., 2014), use only a small number of preselected antidepressants in your practice. Start older patients on a low dose, and titrate systematically based on the scores of a depression scale (Mulsant et al., 2001).

• See older patients more frequently. Plan on frequent follow-ups with your older patients, during which you can offer consistent education and support. Typically, older patients should be seen four to six times during the first 12 weeks of treatment (including a visit within the first two weeks after initiating a new antidepressant). Titrate antidepressants that are tolerated but do not seem to work initially; discontinue antidepressants that are not tolerated or do not seem to work after six to 12 weeks (including at least four weeks at the maximum recommended and tolerated dosage). If the patient does not seem to respond to an antidepressant after 12 weeks, consider switching to a different antidepressant rather than combining several medications (Mulsant et al., 2014; Schiff et al., 2011). ■