The Treatment for Adolescents with Depression Study (TADS) set the stage for a rational acute treatment intervention strategy for moderately to severely depressed adolescents. Major depressive disorder in adolescents is common and causes significant morbidity and family burden (1). Mortality is also a concern, since depression is strongly associated with suicidal events in as many as 50% of adolescents with the disorder (2). Hence, improvement in the outcome of treatment for major depressive disorder in adolescents should bring significant public health value, which is why the National Institute of Mental Health (NIMH) funded TADS in 1999.

TADS is a randomized placebo-controlled trial intended to evaluate the short- (12 weeks) and longer-term (36 weeks) effectiveness as well as the durability (1-year naturalistic follow-up) of the following four treatments among adolescents with major depressive disorder: clinical management with fluoxetine, cognitive-behavioral therapy (CBT), combined fluoxetine plus CBT, and clinical management with an image-matched placebo. As an effectiveness trial, masking was uneven, by design, for the first 12 weeks, with fluoxetine and placebo administered under double-blind conditions, while CBT and the combination of fluoxetine plus CBT were open (i.e., both the patient and treating clinician knew the treatment assignment). Subjects who were assigned to placebo were treated openly after week 12, and the placebo group was not included in longer-term treatment analyses by intention. All primary outcome assessments were made by an independent evaluator blinded to treatment assignment.

Methods used in TADS have been extensively documented in previous reports (3, 4). Despite the inevitable methodological concerns that arise in effectiveness research regarding, for example, construction of the CBT condition or lack of a CBT plus placebo comparison group (see reference 5 for more detailed discussion), TADS, by most accounts, represents the state of the art in comparative treatment trials for adolescents (6—9). More importantly, TADS represents the dedication and hard work of hundreds of collaborating scientists and site personnel and, most critically, patients and their families.

As a result of the publication of TADS methods (4), sample characterization (10, 11), acute treatment (12), long-term treatment, and TADS stage IV (naturalistic follow-up) findings (13), now is an appropriate time to ask the question, "What are the take-home messages from TADS?" In making recommendations, we restrict ourselves (with several carefully chosen exceptions) to the data reported in TADS. Given the absence of the placebo group and gradually increasing subject attrition beyond week 12, we acknowledge that the results from the acute treatment stage are more robust than those from later stages. Similar to the overwhelming majority of clinical trials, findings from TADS represent outcomes for the group-average patient. Consistent with the basic tenets of evidence-based medicine (14), it is the role of the doctor working with the patient and family to determine the applicability of TADS findings to specific patient care decisions.

Compared with epidemiological, CBT, and some medication trial samples, TADS patients are broadly representative of treatment-seeking indisputably ill depressed adolescents (10). The great majority of TADS patients (96%) fell into the moderate to severe range of illness, and more than one-half of the sample (51.9%) was comorbid for at least one other psychiatric disorder. Thus, given the mix of younger and older adolescents and both genders, excellent minority representation, and wide variability in socioeconomic circumstances, results from TADS should be broadly applicable to those depressed adolescents who should be evaluated and treated as a part of routine clinical practice (8).

Response rates for most CBT or medication studies are approximately 60% (i.e., approximately 40% of treated patients do not respond well with either monotherapy) (15). Consequently, expert clinicians often recommend combined treatment as the treatment of choice for major depressive disorder (16). With the caveat that time to maximum medical benefit rather than response at a specific time point is the key metric, TADS confirms the wisdom of clinician experience. After 12 weeks of TADS treatment, CBT alone was far less effective than the combination of fluoxetine with CBT or fluoxetine alone, with response rates of 71% for combined fluoxetine plus CBT, 61% for fluoxetine alone, 43% for CBT alone, and 34% for placebo (11). Not surprisingly, since the typical endpoint in most CBT trials is 18 weeks rather than 12 weeks, CBT "caught up" with fluoxetine (response rates: fluoxetine, 69%; CBT, 65%), while combined fluoxetine plus CBT reached maximum medical benefit at week 18 (response rate 85%), 3 months earlier than fluoxetine alone or CBT alone, with all treatments converging by week 36 (response rates: combined fluoxetine plus CBT, 86%; fluoxetine, 81%; CBT, 81%) (12). Since patients inquire about how soon they will begin to feel better, the finding that combined fluoxetine plus CBT accelerates early response (17) and maximum medical benefit (12) relative to fluoxetine alone or CBT alone is particularly meaningful. Likewise, secondary analyses of outcomes at week 12 showed that the combination of CBT and fluoxetine but not fluoxetine alone proved superior to CBT alone and placebo with respect to function and quality of life (18), remission (19), acceptability, tolerability, and safety (20). Reflecting the order of effect sizes at week 12 (combined fluoxetine plus CBT [effect size=0.98] > fluoxetine [effect size=0.68] > CBT [effect size=0.03]), combined fluoxetine plus CBT proved superior to CBT and placebo on >90% of the 16 possible week-12 endpoints and to fluoxetine on one-half of the 16 possible week-12 endpoints, whereas fluoxetine was superior to CBT and placebo just one-half the time (5). Thus, TADS clearly shows that the combination of CBT and fluoxetine significantly accelerates benefits across a wide range of outcomes relative to CBT alone or fluoxetine alone.

Patients care about being well, not simply being better, and the tendency in depression is to maximize response by treating the patient to remission if at all possible (21). Although most TADS patients exhibited meaningful improvements acutely, only 23% reached remission at week 12 (odds ratios: combined fluoxetine plus CBT versus fluoxetine=2.1; combined fluoxetine plus CBT versus CBT=3.3; and combined fluoxetine plus CBT versus placebo=3.0) (19). By design, TADS asked if continued treatment would improve response and remission rates at the end of 36 weeks of treatment. As hypothesized, the remission rate improved to 60% by week 36, with remission rates of 60% for the combination of fluoxetine with CBT, 55% for fluoxetine, and 64% for CBT, indicating that longer-term treatment is better than short-term treatment (22). Similarly, among patients who had not achieved sustained response by week 12, sustained response rates increased by week 36 from an average of 39.3% at week 12 to 80.0% for combined fluoxetine plus CBT, 61.5% for fluoxetine, and 77.3% for CBT (23). Of those achieving sustained response, more patients assigned to combined fluoxetine plus CBT and CBT alone maintained this response through week 36, indicating that treating patients with either CBT or CBT plus fluoxetine yields important additional benefits over medication alone (23). Given enhanced relapse rates among patients who discontinued treatment early (24) and data from TADS demonstrating that placebo-treated patients who received open treatment after week 12 "caught up" to the active treatment conditions by month 9 (25), it appears that 6 to 9 months of treatment likely will be adequate for the average patient.

In the TADS sample, the mean episode duration at baseline was greater than 1 year, and more than one-half of the subjects had received previous treatment within their current episode, suggesting that depression persists absent intensive evidence-based treatment (10). In epidemiological studies, most children and adolescents who experience an episode of depression typically recover within 1 to 2 years from baseline (26), and approximately one-half of those who recover subsequently relapse (27). Likewise, in the follow-up of pharmacotherapy or psychotherapy randomized controlled trials, 25% to 50% of treated subjects experience a relapse within 6 months to 2 years posttreatment (see references 28, 29). Taken together, these findings strongly suggest that depression is a waxing and waning illness absent persistent evidence-based treatment and new intervention strategies are required to promote long-term reductions in depressive symptoms (30). In this context, results from TADS stage IV analyses (13) (response rates 1 year from active treatment discontinuation: fluoxetine with CBT combined, 82.2%; fluoxetine, 75.2%; CBT, 70.3%) suggest that longer-term treatment results in sustained clinically meaningful improvement even when active treatment is discontinued.

Although the absolute magnitude of the medication-attributable risk is low (number needed to harm=50 to 150), treatment-emergent suicidal events are an important public health concern in children and adolescents treated with antidepressants (31, 32). Conversely, reduced rates of antidepressant prescribing after the black box warning likely increased the death rate from suicide in depressed children and adolescents (33), thereby highlighting the importance of providing evidence-based care for depressed youth (15). In this context, it is critically important to know, first, whether CBT carries the same risk for treatment-emergent suicidality as medication and, second, whether adding CBT to medication might reduce the risk of a medication-associated increase in suicidality. Findings from TADS across 36 weeks of treatment (11) clearly show that patients treated with fluoxetine alone were twice as likely as patients treated with combined fluoxetine plus CBT or CBT alone to show both clinically significant suicidal ideation (on patient self-report) and to experience treatment-emergent suicidal events (on clinician report: fluoxetine, 14.7%; combined fluoxetine plus CBT, 8.4%; CBT, 6.3%). Thus, the TADS answers to these two questions regarding treatment with CBT or combined fluoxetine plus CBT are as follows: "No, there is no increased risk of a suicidal event with CBT alone" and "Yes, there is a protective effect on suicidality from adding CBT to medication." The increased risk of a suicidal event with fluoxetine cannot simply be the result of improvement, since the combined fluoxetine plus CBT group, which began with statistically higher suicidal ideation, exhibited improvement as much or more than the fluoxetine alone group. Yet, patients who received combined fluoxetine plus CBT had a rate of suicidal events that approximated that of the CBT alone group, and suicidality did not increase in the CBT alone group during the 12- to 24-week interval when a large portion of the CBT benefit occurred. More likely, CBT, either alone or in combination with fluoxetine, mitigates against suicidality, perhaps by 1) minimizing the probability that ideation will lead to an attempt, 2) decreasing the likelihood or improving the management of stressful psychosocial events, 3) decreasing family conflict and enhancing family problem solving ability, 4) providing skills to manage negative affects, agitation, irritability, or disinhibition, and 5) enhancing the ability of the patient’s healthcare team to provide adequate monitoring. If this finding is true (which requires replication and extension), patients placed on medication likely would benefit from CBT as well.

Examining short-run outcomes and costs at week 12, we found a notable difference in cost-effectiveness favoring fluoxetine relative to the combination of fluoxetine plus CBT at week 12, with both fluoxetine and combined fluoxetine plus CBT faring better than CBT alone (34). However, following TADS participants for a longer time period revealed that both CBT and combined fluoxetine plus CBT were equally cost effective and, more importantly, the combination of fluoxetine with CBT was superior to fluoxetine alone, largely because of the healthcare costs stemming from fluoxetine-associated increases in suicidal events (35). Given the other advantages associated with combined treatment, the cost-effectiveness argument adds additional strength to the argument that CBT and medication combined should be made widely available in health systems caring for depressed adolescents. Moreover, alone among TADS treatments, combined treatment seemed largely immune from site effects rendering it preferable from the point of view of practical dissemination.

Since the original publication of TADS acute treatment results, two other important publicly funded clinical trials addressing treatment of adolescent depression have been reported. In the Adolescent Depression and Psychotherapy Trial (ADAPT), conducted in community clinical settings in the United Kingdom, depressed adolescents were first treated with nonspecific psychotherapy, and patients who did not respond were then randomly assigned to treatment as usual with an SSRI (primarily but not exclusively fluoxetine) or to an SSRI plus CBT combined (36). No differences were detected in efficacy or safety outcomes at the end of the 28-week treatment period. The apparent discrepancy between the results of TADS and ADAPT may be explained by the greater severity of the ADAPT sample, the greater intensity and possibly higher quality of TADS CBT, and the fact that all patients in ADAPT received treatment-as-usual psychotherapy as a background factor. This is an important point, since elements of CBT often found in routine medical management were not allowed in TADS medical management. Simple advice, cognitive restructuring, guidance for dealing with negative thought patterns, and linking depressive thoughts, behaviors, and emotions may positively affect both response and suicide outcomes in patients treated in clinical practice. Conversely, the results of the other trial, the Treatment of Resistant Depression in Adolescents (TORDIA) study, supported the superiority of combined treatment with antidepressant medication plus CBT versus medication monotherapy as a second step intervention for youth who did not improve on a first trial of medication (37). Perhaps because of differences in study design and patient populations, neither ADAPT nor TORDIA revealed the protective effect of CBT on suicidal ideation or events seen in TADS.

TADS was designed to answer clinically important questions concerning the short- and longer-term benefit(s) of CBT, fluoxetine, and combined fluoxetine plus CBT. Across a range of outcomes, findings from TADS demonstrate clearly that combined treatment accelerates recovery relative to CBT alone and fluoxetine alone while at the same time minimizing the risk of suicidality relative to fluoxetine alone. Taking benefit and risk as well as cost-effectiveness into account, we conclude that the combination of fluoxetine and CBT appears superior to both CBT monotherapy and fluoxetine monotherapy as a treatment for moderate to severe major depressive disorder in adolescents, with the following implications for healthcare decision makers at all levels of the healthcare system. First, treating adolescents with moderate to severe depression should have a strongly positive public health benefit and should be actively encouraged. Second, as noted previously, to maximize benefits and minimize harms for individual patients and from a public health perspective, combined treatment should be the modal treatment for moderately to severely depressed youth. Third, since accelerating symptom reduction using medication is an important clinical outcome in psychiatry (as it is in other areas of medicine), treatment with fluoxetine should be made widely available, not discouraged, with the important caveat that medication monotherapy carries a small but real risk of a suicidal event. Patients on medication monotherapy must be informed of this risk and monitored carefully during treatment. Fourth, since doctor and patient preference are fundamental ingredients in the practice of evidence-based medicine, patients and providers may wish to begin treatment with CBT alone to avoid the risk of antidepressant-induced suicidality, only advancing to medication when the response to CBT is deemed inadequate. Consistent with current best-practice guidelines (16, 38—40), findings from TADS suggest that this may be a particularly good strategy for nonsuicidal, mildly ill, sociodemographically advantaged patients (41). Others, particularly when suicidality is present from the outset in a severely ill patient, may prefer to begin with combined treatment, hoping for maximum benefit while minimizing risk. Whether a stand-alone treatment, combined with medication, or added to the treatment regimen for patients who respond partially to antidepressants, modern CBT—which foreshadows the future of psychosocial treatments in neurosciences medicine (42)—needs to be made widely available to patients across the full range of mental disorders and practice settings. Fifth, to move beyond the idea of depression as a chronic waxing and waning illness (e.g., to allow the euthymic state to consolidate sufficiently to protect against relapse), it is necessary to treat most patients to remission over a minimum of 6 to 9 months. Brief treatment for the moderately to severely ill patient will not suffice. Sixth, while not yet at the level of evidence necessary to mandate a standard of care, we believe that the literature on evidence-based treatment for adolescent major depressive disorder has matured sufficiently to merit mention when informing patients about the risks and benefits of evidence-based treatments (43, 44).

TADS has made a major contribution to the knowledge base concerning the treatment of major depression in adolescents, with 45 articles published and another 30 plus articles either "in press" or "in preparation." As TADS comes to an end, we as well as other investigators (e.g., 7, 9, 30) believe that it is now time for TADS to inform programs of research focused on 1) explicating CNS mechanisms of response, 2) maximizing the extent and durability of remission, 3) replicating and extending the protective effect of CBT on suicidality in medication-treated patients, 4) disseminating evidence-based care and dynamic treatment regimens that address treatment strategies for depression in children, and 5) emphasizing quality improvement initiatives in practice settings. The TADS de-identified data set is available online via NIMH limited-access data sets (http://www.nimh.nih.gov/health/trials/datasets/nimh-procedures-for-requesting-data-sets.shtml). We encourage the field to utilize TADS data to explicate the boundaries of adolescent depression, its treatment, and clinical trials methods in pediatric mental illness.