National Academies Press: OpenBook

CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary (2010)

Chapter: 2 Data Collection and Optimization

« Previous: 1 Perspectives from the FDA, Academia, and Patients
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 17
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 18
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 19
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 20
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 21
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 22
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 23
Suggested Citation:"2 Data Collection and Optimization." Institute of Medicine. 2010. CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary. Washington, DC: The National Academies Press. doi: 10.17226/12829.
×
Page 24

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Data Collection and Optimization NEW APPROACHES TO STUDYING THE PREDICTIVE POWER OF SUICIDAL IDEATION FOR SUICIDALITY The relationship between suicidal ideation and behavior (e.g., at- tempts or completions) was the focus of several presentations. The first was by Matthew Nock of Harvard University, who spoke about his epi- demiological findings. Similar to the clinical trial data presented by Thomas Laughren, Nock’s epidemiological findings revealed that sui- cide ideation is not a strong predictor of a suicide attempt (Nock et al., 2009a). Still, most of those who make a suicide attempt do express ideation ahead of time. His presentation covered new methodologies to obtain greater specificity about the relationship between ideation and suicidality. The types of studies Nock reported are epidemiological cohort stud- ies, usually of previous ideation and attempts, and psychological autop- sies for completed suicides. The latter analyze the cause(s) of death by examining the body and the circumstances that led or contributed to death. Such studies contain a set of structured questions, administered in face-to-face interviews of close friends and family, to infer the dece- dent’s intent, risk factors, and related contributors. Psychological autopsy studies have found frequent expression of suicidal thoughts before com- pleted suicide. Overall, about 50 to 66 percent of people who complete 17

18 CNS CLINICAL TRIALS: SUICIDALITY AND DATA COLLECTION suicide disclose their ideation or intent to those around them, according to one meta-analysis (Cavanagh et al., 2003). These studies have one critical limitation, however: The definitions of ideation and intent vary, leading to variations in prevalence. One of the most recent and thorough review articles found that 75 percent of those who subsequently die by suicide visited their primary care provider in the last year of life and 45 percent in the last month (Luoma et al., 2002). An easy conclusion from these findings is that more attention should be given to suicidal ideation and threats. However, the major problem is that expressions of ideation or threats are highly common. A large, na- tionally representative survey of U.S. adults found that 15 percent of them report seriously considering suicide at some point in their life (Nock et al., 2009a). Approximately one-third of those who think about suicide at some point in their life later make a suicide attempt. Narrow- ing the time frame to the past 12 months, 15 percent of ideators proceed to make a suicide attempt (Nock et al., 2009a). What these findings sug- gest is that it is exceedingly difficult to identify from the large number of ideators the small percentage of those who will progress to a suicide at- tempt. While suicide has general risk factors, as well as chronic versus short-term risk factors, no combinations of risk factors currently have sufficient sensitivity and specificity to predict who among the groups at risk will make an attempt or completion, under what circumstances, and at what time (Goodwin and Jamison, 2007). One surprising finding mentioned by Nock was that depressed peo- ple with suicidal ideation, although a risk group, were not the leading group at risk for progression to a suicide attempt. Ideators with anxiety disorders and disorders of impulse control, such as conduct disorder and posttraumatic stress disorder, exceeded depression in their predictive power for making suicide attempts, according to a large, nationally rep- resentative epidemiology study of U.S. adults, the National Comorbidity Survey Replication (Nock et al., 2009a, 2009b). Having an anxiety dis- order is an independent risk factor for suicidal ideation and suicide at- tempts. Anxiety and mood disorders, when comorbid, heighten the risk as compared with a mood disorder alone, according to a prospective, population-based study of adults in the Netherlands (Sareen et al., 2005). Two new alternatives to the general measure “suicidal ideation” show promise: real-time, electronic monitoring systems of suicidal think- ing and behavior at different time points, and a new psychometric meas- ure asking patients to rate suicidal intention “at its worst point in time.”

DATA COLLECTION AND OPTIMIZATION 19 REAL-TIME ELECTRONIC MONITORING SYSTEMS Given the poor predictive power of current measures of suicidal ideation, it would be tempting to conclude that suicidal ideation should be discarded in favor of more robust measures. Instead, new techniques are being harnessed to study ideation and to understand the transition from ideation to suicidal behavior, especially with new methods investi- gated prospectively. One novel methodological approach borrows techniques from the social sciences. It uses methods that do not rely on verbal and retrospec- tive self-report, relying instead on systematic weekly monitoring of idea- tion, said Nock. The key is to collect data closer to these events using real-time methods and/or patients’ daily assessment of self-injurious thoughts and behaviors. This methodology draws from a concept known as ecological momentary assessment, a collective term referring to de- tailed investigation of several mental health features in real-time, such as affect, mood, and interpersonal behavior. A study participant could be asked to report on these features around the same time that the experi- ence occurs (Moskowitz and Young, 2006). Nock reported on a preliminary study giving electronic diaries or Palm Pilots to a small sample of 30 adolescents. Over a period of 2 weeks, the subjects were asked to record evidence of self-injurious thoughts and behaviors. For example, the subjects recorded 100 episodes of non-suicidal, self-injurious behavior, such as cutting and burning, and 26 episodes of suicide ideation. Furthermore, these methods offer collection of data on the intensity, duration, and triggers of each episode. While Nock did not propose that every clinical trial collect such detailed meas- urements, he would like to see the methods used to understand the emer- gence of suicidal behaviors from suicidal ideation. One practical study, he noted, would be to equip adolescents with electronic diaries that beeped once or twice daily, prompting them to answer a range of ques- tions about adverse effects, mood, and other data. The detailed informa- tion the subjects provide might shed light on the emergence of ideation over time. Simply put, Nock noted, these new techniques offer the virtues of being prospective, being recorded near the time of each suicidal event, and being more informative about the meaning and severity of ideation. With that information, it may be possible to determine those features of suicidal ideation with better predictive power for suicidality.

20 CNS CLINICAL TRIALS: SUICIDALITY AND DATA COLLECTION THE PREDICTIVE VALIDITY OF SUICIDE IDEATION AT ITS WORST POINT IN TIME FOR SUICIDE ATTEMPTS Evaluating the predictive validity of existing measures of suicide ideation was the focus of a presentation by Gregory Brown of the University of Pennsylvania. He and his colleagues sought to evaluate measure(s) of ideation for predicting suicide behavior and completed suicide with high sensitivity and specificity. Analyses relied on data from epidemiological studies and clinical trials. The investigators defined ideation as including suicide intent, the urge to kill oneself, or a specific plan to kill oneself. Of several commonly used psychometric scales of suicide ideation in clinical trials, Brown’s team began with the 20-year-old Scale of Suicide Ideation (SSI-C), a scale for measuring the severity or intensity of sui- cide ideation. The scale has excellent internal reliability, interrater reli- ability, current validity, and long-term predictive validity for completed suicide, noted Brown. The SSI-C is a 21-item, interviewer-administered rating scale that measures the current intensity of patients’ specific atti- tudes, behaviors, and plans to commit suicide on the day of the inter- view. The ratings for the first 19 items are summed to yield a total score, ranging from 0 to 38. The SSI-C consists of five screening items. Three items assess the wish to live or the wish to die and two items assess the desire to attempt suicide. If the respondent reports any active or passive desire to commit suicide, then 14 additional items are administered. Indi- vidual items assess suicidal risk factors such as the duration and fre- quency of ideation, sense of control over making an attempt, number of deterrents, and amount of actual preparation for a contemplated attempt. Two additional items record incidence and frequency of previous suicide attempts. However, the SSI-C does not measure previous suicide ideation “at its worst point in the patient’s life,” which is the subject of a separate scale (SSI-W) (Beck et al., 1997, 1999). Studying nearly 4,000 psychiatric outpatients, Beck and collabora- tors found that the SSI-W had the strongest odds ratio for outpatients later completing suicide, nearly 14 times higher (OR = 13.84, CI, 5.6–34) than for outpatients who subsequently did not complete suicide, using the National Death Index to verify deaths of subjects in the sample (Beck et al., 1999). The SSI-W surpassed the odds ratios for the SSI-C and the Beck Hopelessness Scale. To determine optimal cutoff points for maxi- mizing the sensitivity and specificity of the SSI-W, Beck and colleagues compared its score against completed suicide by examining the Receiver Operating Characteristics (ROC) curves. The ROC is a graphical plot

DATA COLLECTION AND OPTIMIZATION 21 that measures true positives versus false positives using a particular cut- off score. The results led the investigators to conclude that suicide idea- tion at its worst point identified a subset of patients at relatively high risk for completed suicide with high sensitivity and specificity. The mean time to completed suicide was about 4 years after participation in the study. The investigators, seeking to evaluate the predictive validity of the SSI-C over a short-term period, studied individuals who had attempted suicide because this group has a high risk for subsequent attempts. Brown and his colleagues analyzed findings from their previously published clinical trial of a cognitive therapy intervention among suicide attempters seen at an emergency department (Brown et al., 2005). Because a high number of reattempters were found among the study group, the investigators were able to examine the short-term predictive capacity of current ideation versus ideation at its worst point. Participants in the study were evaluated at several follow-up visits (1 month, 3 months, 6 months, 1 year, and 18 months) and assessed for both current ideation and worst point ideation since the previous follow-up visit. Brown and collaborators found that the predictive validity of the SSI for current ideation on subsequent suicide attempts was poor. However, they also found that high scores on the SSI at the worst point in time since the participant’s previous visit was a significant predictor of a repeat suicide attempt recorded at the 3-month visit. The investigators concluded that suicide ideation at its worst point was a significant predictor of a subsequent attempt over the near term and a better indicator for short- term risk than current ideation. Regarding the broader question of current ideation over the course of a randomized clinical trial (RCT), several discussants expressed skepti- cism about its value as a predictor of suicidality. Robert Gibbons rec- ommended another way to examine the nearly 400 RCTs analyzed by the Food and Drug Administration. He suggested examining the weekly Hamilton Depression Rating Scale―in particular, item 3 focusing on the degree of suicidal ideation and planning according to four levels of se- verity—plus overall Hamilton weekly ratings to determine treatment re- sponsiveness and suicide attempts. That type of analysis would be more likely to improve suicidality measures, he remarked. He and Charles Beasley of Eli Lilly are conducting a reanalysis of the RCT data to de- termine whether this approach improves suicidal ideation as a predictor of suicidality.

22 CNS CLINICAL TRIALS: SUICIDALITY AND DATA COLLECTION NEUROBIOLOGICAL CONTRIBUTIONS TO PREDICTING SUICIDALITY A biological marker would be a valuable and objective contributor to predicting suicidality. J. John Mann of Columbia University covered a constellation of potential biological markers for suicidality. At least one neurobiological defect, serotonin deficiency, is significantly associated with suicidality. That defect and other prominent defects might eventually be combined with measures of ideation and suicide attempts to predict suicidality with greater specificity and sensitivity, he said. Several biological markers under study are low levels of the neurotransmitter serotonin, low norepinephrine levels, and hypothalamic- pituitary-adrenal (HPA) axis dysfunction. Their roles were at the core of Mann’s presentation. But he was keen to point out that biology does not act alone. Biology is one of several key factors that contribute to suicidality, according to his model (Figure 2-1). Multifactorial models are commonly adopted to explain the origins of psychiatric disorders (Engel, 1978). Given the heritability of completed suicide, which stands at 21 to 50 percent (Currier and Mann, 2008), it becomes clear that multiple factors must be at play. Depression or psychosis Objective state Life events Hopelessness/reasons for living Perception of depression Low norepinephrine/ Suicidal ideation HPA axis Subjective state and traits Suicidal planning Low serotonin activity Impulsivity/restraint Aggression Alcoholism, smoking, Impaired problem solving, substance abuse, head Suicidal act poor set changing, injury cognitive rigidity, negative perceptual sets FIGURE 2-1 A model of suicidal behavior. NOTE: HPA = hypothalamic-pituitary-adrenal, RFL = reason for living. SOURCE: Mann, 2009.

DATA COLLECTION AND OPTIMIZATION 23 For three decades we have known that low serotonin levels are asso- ciated with suicidality (Mann, 2003). The deficiency is so consistent across studies that Mann describes low serotonin as a long-standing trait. The method used to obtain these findings was by pharmacological challenge with the amphetamine derivative fenfluramine. Upon fenflu- ramine challenge, serotonin levels should normally rise. In suicidal groups, however, serotonin levels do not rise as highly as they do in control groups. The effect occurs in a matter of hours and is age depend- ent. Because most of the studies Mann described in his presentation were conducted in adults, he stressed that the findings may not apply to children and adolescents. Low serotonin output has great significance for one of the brain areas to which serotonergic pathways project: the cerebral cortex’s prefrontal cortex and its ventrolateral and dorsolateral regions. The ventral prefrontal cortex participates in rational decision making and regulates aggressive/ impulsive behaviors (Damasio et al., 1994). This area of the cortex is altered in suicide and aggression (Mann, 2003). Positron emission tomo- graphy studies confirm that the prefrontal cortex is heavily involved in intent, planning, impulsivity, and lethality of suicidal behavior, said Mann. The odds of completed suicide are 4.5 times greater among those with low serotonin―as measured by its metabolite 5-hydroxy- in-doleacetic acid in the cerebrospinal fluid―than in those with high serotonin (Mann et al., 2006). The abnormalities in serotonin output may be traced to polymorphisms in several candidate genes for the serotonin transporter, serotonin receptors, and enzymes related to sero- tonin synthesis (Currier and Mann, 2008). Of these possibilities, the promoter region of the serotonin transporter has been the focus of nu- merous investigations. A meta-analysis of more than 20 studies found a significant association with one serotonin transporter allele and suicidal- ity (Anguelova et al., 2003). Still, low serotonin is not the sole determi- nant of suicide, emphasized Mann. Serotonin also is one of the regulators of the stress response, which is the central function of the HPA axis. A recent study by Mann’s team prospectively followed several groups for a period of 2 years: depressed patients with previous suicide attempts, depression alone, and two other control groups (Keilp et al., 2008). The groups were compared in relation to fenfluramine challenge, on plasma prolactin, on cortisol (which is re- leased by the adrenal gland during the stress response), and on psycho- metric measures of mood. The study found that blunting of cortisol and the worsening of mood, plus younger age, predicted subsequent suicide attempts in depressed patients with previous attempts. Lower prolactin

24 CNS CLINICAL TRIALS: SUICIDALITY AND DATA COLLECTION was also found after fenfluramine challenge, but the effects were not as statistically robust as those with cortisol. The stress response also affects the structure and function of noradren- ergic pathways. Fewer noradrenergic neurons in a nucleus found in the brain stem, the locus coeruleous, have been found postmortem in studies of depressed patients who completed suicide (Pandey and Dwivedi, 2007). The noradrenergic system is also a target of antidepressant treatments. Adults reporting past child abuse have excessive norepinephrine release after a laboratory stress test (Heim and Nemeroff, 2001). A noradrenaline metabolite in the cerebral spinal fluid predicts the lethality of a suicide attempt (Galfalvy et al., 2009). One key hypothesis, observed Mann, is whether excessive norepinephrine release related to the long-standing stress of severe major depression in those with childhood adversity or in genetically predisposed people eventually leads to hopelessness and pes- simism. Those psychological traits, in turn, place these people at higher risk for suicidal behavior. In summary, Mann believes that biological markers such as low sero- tonin levels predict future completed suicide. Biological findings, along with other psychological and social risk factors, may eventually be used in an integrative way to shed light on the emergence of suicidality in childhood, adolescence, and adulthood (Box 2-1). BOX 2-1 Summary of Research Findings Reported by Presenters in This Session • Epidemiological and clinical trial findings reveal that suicide ideation is not a strong predictor of a suicide attempt. • New approaches show promise for strengthening the relationship between ideation and suicidality if more nuanced information about the nature of ideation is collected simultaneously. • One approach seeks real-time electronic monitoring that prompts patients to report more about ideation at the time of its occurrence. • Another approach shows that measuring “suicidal ideation at its worst point in time,” through a questionnaire, is a significant predictor of suicide com- pletion (Beck et al., 1999). • Biological measures in response to serotonergic challenge also show promise if integrated with psychological and social measures. In a prospec- tive study, blunting of cortisol response and the worsening of mood, plus younger age, predicted subsequent suicide attempts in depressed patients with a previous attempt (Keilp et al., 2008).

Next: 3 Data Analysis »
CNS Clinical Trials: Suicidality and Data Collection: Workshop Summary Get This Book
×
Buy Paperback | $21.00 Buy Ebook | $16.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The Food and Drug Administration (FDA) now requires that all clinical trials for drugs that affect the central nervous system—including psychiatric drugs—are assessed for whether that drug might cause suicidal ideation or behavior. The Institute of Medicine's (IOM) Forum on Neuroscience and Nervous System Disorders hosted a meeting on June 26, 2009, to discuss the FDA's new policy and how to analyze best whether suicidal thoughts predict actual suicidal behavior in the near future.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!