NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Forum on Neuroscience and Nervous System Disorders. CNS Clinical Trials: Suicidality and Data Collection, Workshop Summary. Washington (DC): National Academies Press (US); 2010.

Cover of CNS Clinical Trials

CNS Clinical Trials: Suicidality and Data Collection, Workshop Summary.

Show details

4Partnerships, Opportunities, Collaboration


Data sharing and collaboration among experts is a time-tested guide to help practicing clinicians abide by the Food and Drug Administration’s (FDA’s) black box warning. Practicing clinicians, whether in primary care or specialty care, need guidance about what warning signals and adverse events to look for during the course of antidepressant treatment. The necessary guidance critically depends on two key issues for researchers and practitioners alike: deployment of common outcome measures and understanding of the evolution from suicidal ideation into suicidality. Addressing these issues is no small task. Measures and questionnaires used in randomized controlled trials (RCTs) are too lengthy and elaborate and thus do not readily lend themselves for use in everyday medicine. RCTs also typically exclude patients with suicidal ideation or suicidality. Furthermore, it is key to understand longitudinal trends and their predictive relationship to suicidality.

Madhukar Trivedi of the University of Texas Southwestern Medical Center at Dallas focused his presentation on the lessons learned from a 5-year depression trial, enrolling about 4,000 adults and sponsored by the National Institute of Mental Health (NIMH, 2009). This clinical trial had several unique features. It studied real-world depression treatments in everyday clinical practice, as opposed to the rarified circumstances of a typical RCT with its strict inclusion and exclusion criteria. RCT patients are typically in better health and have fewer comorbidities than real-world patients. The study sought to find out how patients fared over the long term with depression treatment; its focus was on patients who are hard to treat, considering that the majority of depressed patients do not respond significantly enough to the first antidepressant they try (Little, 2009). It also sought to identify the comparative effectiveness of the several tiers of pharmacological therapies. The trial was conducted in a network of primary and specialty care settings across the country. Simply put, its goal was to help practicing clinicians sort out treatment recommendations in everyday practice. Until now, no studies have given guidance essential for patient management over the course of antidepressant treatments. With nearly 20 medications to choose from, this is no easy feat.

Formally known as Star*D, the Sequenced Treatment Alternatives to Relieve Depression trial used a common set of outcome measures, including one three-part question covering suicidal ideation and behavior, which was the centerpiece of Trivedi’s presentation. The longitudinal nature of this study and its real-world setting helped his team discern the evolution of suicidal ideation into suicidality. That path rarely has been traced because such patients with ideation are normally excluded from clinical trials.

Suicidal ideation and suicidality were measured by a three-part question of the QIDS questionnaire (Quick Inventory of Depressive Symptomatology—Self-Report); (Zisook et al., 2009). The level of severity ranges from 0 to 3 (Box 4-1), with a score of 1 meeting the definition of mild suicidal ideation and 3 meeting the definition of a suicidal attempt. Designed as an open trial, there was no comparison group, so patients were assessed in relation to their baseline visit.

Box Icon

BOX 4-1

Question About Suicidal Ideation in the Quick Inventory of Depressive Symptomatology—Self-Report. Thoughts of death or suicide: I do not think of suicide or death.

The study participants were treated with Citalopram, and nearly half of them had suicidal ideation at baseline. Of that half with suicidal ideation, 74 percent showed improvement at their first post-baseline visit 2 weeks later, 22 percent remained the same, and 4 percent worsened. Of the latter, 1 percent still had suicidal ideation by the last visit at 12 weeks from baseline. These patients would have been excluded from clinical trials. Trivedi said it is important to point out that the small percentage of patients who worsen take longer to respond to treatment. Among the other half of the sample, those without suicidal ideation at baseline, 7 percent experienced the emergence of suicidal ideation by their first post-baseline treatment visit, a small upsurge that the investigators attributed to early increases in energy, activation, and anxiety (Szanto et al., 2007). But the majority of that 7 percent, or 63 percent, did not report suicidal ideation by their final visit 12 to 14 weeks later. Fifteen of the total sample attempted suicide by the end of the trial, but there were no completed suicides.

The authors concluded that although there was an early increase in suicidal ideation in a small group of depressed patients, the majority of such cases were fleeting, suggesting that most cases of “emergent” suicidal ideation are more likely to be tied to natural fluctuations of suicidal ideation than to treatment. The major risk factors for developing treatment-emergent suicidal ideation were drug abuse, severe depression, and melancholic features. Demographic risk factors for suicide attempts were being less than 19 years old and being an African American. These findings suggest the importance of tracking suicidal ideation because a small percentage of patients do worsen and their risk factors are becoming clear.


Lack of progress in studying and treating suicidality is part of a broader trend in the United States. The country has undergone a decade-long decline in innovation of new medical products, according to ShaAvhree Buckman, acting director of the Office of Translational Sciences at the FDA. Innovation in the form of new product approvals has slowed strikingly, while pharmaceutical company research and development spending has progressively increased (Figure 4-1). This so-called innovation gap galvanized the FDA several years ago to spearhead collaborative efforts to facilitate and modernize product development.

FIGURE 4-1. Innovation gap between new drug approvals and spending on research and development by year.


Innovation gap between new drug approvals and spending on research and development by year. NOTE: NME = new molecular entitites; PhRMA = Pharmaceutical Research and Manufacturers of America; R&D = research and development. SOURCE: Buckman, 2009. (more...)

Through the creation of its Critical Path Initiative,1 the FDA has taken unprecedented steps to forge partnerships with the National Institutes of Health (NIH), industry, advocacy groups, and scientific societies, among others. Under the auspices of the Critical Path Initiative, the FDA is committed to promoting development of the infrastructure and furnishing tools to make product development more efficient and streamlined. It hopes to help sponsors predict early in the development process which products are most likely to be safe and effective, thereby avoiding expensive product failures in the later stages of development. Such failures in recent years have chilled the climate for investment.

The FDA’s role, said Buckman, is carefully carved out to encourage development not of any single product or sponsor, but to encourage collaboration focused on overcoming widespread impediments to innovation, such as the lack of animal models or the lack of biomarkers. The purpose of fostering collaborations is to pool knowledge and resources and thereby help the FDA develop relevant data standards and regulations and build support for relevant academic science. The public–private partnerships and collaborations arising from the FDA’s activities are focused on the “precompetitive” space, that is, areas of science and methodology that spur all partners rather than give any participating group a competitive edge.

The FDA has yet to start any specific initiatives in suicidality, although a new proposal is being developed by Charles Beasley of Eli Lilly (see the next section). Some of the FDA’s previous and ongoing initiatives might serve as exemplars for the suicidality field. Some of the most prominent ongoing initiatives have been in the following areas: (1) Serious Adverse Events, the goal of which is to detect and validate DNA variants that are clinically useful in predicting patients’ risk of experiencing drug-induced serious adverse events; (2) Clinical Trials Transformation, the goal of which is to focus on practices that, if adopted broadly, will increase the quality and efficiency of clinical trials; and (3) the Patient Reported Outcomes consortium, the goal of which is to develop and evaluate self-reported questionnaires (termed “patient-reported outcomes”) to measure safety and/or efficacy in clinical trials. For the latter, the FDA has a liaison relationship with no voting rights or other fiduciary role. Submission of dossiers to the FDA does not automatically constitute “fit for purpose,” the legal term for a method of measurement that satisfies the FDA’s standards for appropriateness and quality.

Some common themes driving the Critical Path Initiative include identifying the public health need, determining whether partners are willing to share data precompetitively, identifying needed data standards, and sharing data in the public domain as quickly as appropriate. Although the FDA considers itself a catalyst in its initiatives and is committed to their success, the actual success of any single initiative requires partners to commit to collaboration and data sharing. Pinpointing the basis of success after 4 to 5 years since creation of the Critical Path Initiative, Buckman advised, “A lot of these efforts, they do take time. They take commitment. They take a champion.”


A novel proposal for the FDA’s Critical Path Initiative was urged for suicidality studies by Charles Beasley, a distinguished scholar and chief scientific officer of Global Patient Safety at Eli Lilly. It would create a large database for safety purposes, among others. (Beasley stated clearly that the concept is his own, and represents official policy of neither Eli Lilly nor the Pharmaceutical Research and Manufacturers of America.) Beasley envisioned that the database would serve as a way of pooling RCT findings: It should include not only adverse events associated with antidepressants, but also those associated with many psychiatric and non-psychiatric medications. The imperative behind the formation of the warehouse, in his view, would be to answer many unresolved and thorny questions, such as whether and the extent to which suicidal ideation predicts suicidality, as well as many other research questions. The most immediate question of interest, from Beasley’s perspective, is whether suicidal ideation as detected by an instrument such as the Columbia Suicide Severity Rating Scale (C-SSRS) is predictive of suicidal acts or completed suicide during the short-term, index acute treatment episode. The answer to this question has broad implications for the appropriateness of the FDA’s labeling decisions. The warehouse would be open to academic, industry, and all other researchers.

Beasley predicated his idea on one of the current Critical Path Initiatives to assemble an electrocardiogram (ECG) data warehouse for studying cardiac toxicity. The warehouse is designed as a repository of more than 2 million ECGs, according to earlier remarks by Buckman. The ECG warehouse, which is mandatory under the Food, Drug, and Cosmetic Act, is intended to enable academic and industry researchers to find better biomarkers of cardiac toxicity. The purpose is to develop more efficient clinical trial outcome measures and improve patient safety, given that the utility of the Q-T interval, a measure of the heart’s electrical cycle, has been seriously questioned. The ability to undertake suicidality studies rests on having an extremely large sample size because of the rarity of completed suicides.

Beasley foresees that at least 100 completed suicides would be a crucial component of the warehouse. To stimulate discussion, he explained that the overall patient size of the warehouse, patient demographics, placebo controls, and critical data elements are inchoate, as are administrative, procedural, and funding matters. Beasley stressed that uniformity of data collection is essential. The idea generated lively discussion among participants. Most of the questions related to the most suitable outcome measures (e.g., Columbia Classification Algorithm for Suicide Assessment, C-CASA), the inclusion of biomarkers such as genetic data, the inclusion of comorbidities, longitudinal data, efficacy data, and the protection of corporate secrets.

Participants were intrigued by the proposal laid out by Beasley, although what is left to discuss is how best to move forward, who should be at the table, and under what auspices the discussions would be held and advanced. Many also agreed that any potential warehouse would need to incorporate other data elements—ideally in real time—to make the collection more robust and useful. This brings into question the issue of partnering with the proper stakeholders in order to realize such a robust warehouse. A number of participants believed that including several NIH institutes would be a good first step.

Copyright © 2010, National Academy of Sciences.
Bookshelf ID: NBK52954


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (818K)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...