SUICIDALITY CLASSIFICATION PROJECT

1
SUICIDALITY CLASSIFICATION PROJECT

• Kelly Posner, Ph.D.
• Maria Oquendo, M.D.
• Barbara Stanley, Ph.D.
• Madelyn Gould, Ph.D, M.P.H
• Statistical Consultant Mark Davies, M.P.H.

2
Rationale for Reclassification

• The Problem
• Field challenged by lack of conceptual clarity
  about suicidal behavior and corresponding lack
  of well-defined terminology
• Lack of systematic or standardized language used
  to define suicidal behavior in the 25 industry
  antidepressant trials
• Difficulty in interpreting the meaning of
  reported adverse events that occurred in these
  trials

3
The Problem

• AEs that should have been called suicidal may
  have been missed
• AEs may have been inappropriately classified as
  suicidal

4
Examples of Difficulties in Adverse Event Labeling
Original Label Narratives
Personality Disorder Age 10 The patient exhibited symptoms of personality disorder, assessed as moderate in severity, and was discontinued from the study due to this event. One day later (date), the patient attempted to hang himself with a rope after a dispute with his father. The investigator did not consider this event to be a suicide attempt, but rather to be part of the personality disorder, which was still ongoing.
Suicide Event this patient is reported to have engaged in an episode of automutilation where she slapped herself in the face.
Medication Error Age 14 The patient took 11 tablets impulsively and then went to schoolthe patient denied that it was a suicide attempt.
Hostility Age 10 Before his mothers call to the site and again after arguing with his stepfather, he wrapped a cord from the miniblinds around his neck, threatening to kill himself.
5
How to Address this Problem?

• a common set of guidelines needed to be applied
• Data needed to be examined consistently across
  trials
• Used research-supported definitions/concepts with
  reliability and validity

6
Broaden Range of Adverse Events

• To avoid bias in ratings (would not want raters
  to only have what sponsors had identified as
  possibly suicidal)
• To identify suicidal events that may have been
  missed

7
What Was Included to Broaden Range

• Events originally identified by sponsors as
  possibly suicide related
• Accidental Injuries (including accidental
  overdoses)
• Serious Adverse Events (this includes
  life-threatening events and all hospitalizations)

8
Why Were Experts in Suicide Needed?

• Limited information provided in narratives,
  particularly frequent lack of stated suicidal
  intent
• Allowed for inference based on details of
  behaviors and related clinical information

9
Expert Rater Panel

• Annette Beautrais, Ph.D
• David Brent, M.D.
• Greg Brown, Ph.D.
• Kees van Heeringen, M.D., Ph.D.
• Cheryl King, Ph.D., ABPP
• Peter Marzuk, M.D.
• Patrick O'Carroll, M.D., M.P.H
• David Rudd, Ph.D., ABPP
• Anthony Spirito, Ph.D., ABPP
• Alternate Alec Miller, PsyD.

10
Columbia Suicidality Classification Rating Scale
Behavior Towards Self Injury Suicidal Thoughts Intent or Wish to Die Event category Rating
Suicide Attempt __ Check if intent is inferred ___ Check if Suicide Attempt is only information provided
Preparatory Actions Towards Imminent Suicidal Behavior
? Self-Injurious Behavior, Intent Unknown
_ Self-Injurious Behavior, No Suicidal Intent, (To Affect Internal State /Circumstance)
_ Suicidal Ideation __ passive __ active__ active w/ plan __type unknown
_ _ Other No indication of deliberate self-injury or suicidal behavior or ideation Accident________ Medical________ Psychiatric________
? ? Not Enough Information
Please specify
Comments
11
Definitions for the Columbia Suicidality
Classification Scale
Suicide Attempt Self- injurious behavior associated with some intent to die. Intent can be stated or inferred by rater. No injury needed.
Preparatory Acts Towards Imminent Suicidal Behavior Person takes steps to injure self but is stopped by self or other. Intent to die is either stated or inferred.
Self-Injurious Behavior, Intent Unknown Self- injurious behavior where associated intent to die is unknown and cannot be inferred.
Self-Injurious Behavior, No Intent, Primarily to Affect Circumstance Self- injurious behavior associated with no intent to die behavior is intended to effect change in others or the environment.
Self-Injurious Behavior, No Intent, Primarily to Affect Internal State Self- injurious behavior associated with no intent to die intended to relieve distress. Typical examples are superficial cuts or scratches, hitting/banging, or burns.
Suicidal Ideation Passive thoughts about wanting to be dead or active thoughts about killing oneself, not accompanied by preparatory behavior.
Other Accident Passive thoughts about wanting to be dead or active thoughts about killing oneself, not accompanied by preparatory behavior.
Other Psychiatric_____________ Psychiatric symptoms only (when no evidence of any type of suicidality)
Other Medical______________ Medical symptoms or procedure only
Not Enough Information Insufficient information to classify the event
Infer intent if the behavior is clinically
impressive or there is more than one piece of
evidence suggesting suicidal intent
12
What is the Classification Scheme?
Suicidal
Non Suicidal
Indeterminate
Suicidal Ideation Code6 N62
Suicide Attempt Code 1 N 36
Self-Injurious Behavior Without Suicidal Intent
Codes4,5,,11 N17
Non- Consensus N 0
Other -Accidental -Psychiatric -Medical Codes7,8
,9,12 N 260
Not Enough Information Unable to Classify
Whether Deliberate Self-Injury or other Code
10 N 9
Preparatory Actions Towards Imminent Suicidal
Behavior Code 2 N 8
? Suicidal
Self-Injurious Behavior With Unknown
Intent Code3 N35
13
What Was Done? Classification Methodology

• Chose expert panel
• Expertise in adolescent suicide and suicide
  assessment
• Based on reputation and publications
• No involvement in industry youth depression
  trials in question
• No expert rater was employee of Columbia
  University
• Training teleconference to review classification
  parameters
• Training reliability exercises (to ensure
  appropriate application of classification)

14
Design continued

• All case narratives blinded to any potentially
  biasing information
• Random distribution of 427 events to 9 experts
• PI blind to randomization procedures
• Each case independently rated by 3 raters
• Each rater received approx. 125 events to rate
• Any group of 3 raters shared only 5 cases
• Review of all ratings for QA and identification
  of non-agreement cases
• Consensus teleconferences for any disagreement
  cases
• Double data entry for quality assurance

15
Consensus Process

• If ratings did not have unanimous agreement,
  consensus discussion held
• Each case discussed by 3 raters involved
• Discussion of each case led by an expert other
  than those originally assigned the case
• Goal of discussion was to reach 100 agreement
• If 100 agreement could not be reached, case
  became indeterminant
• Original majority opinion did not always end up
  as the final consensed classification

16
What Was Rated? Blinding of Event Narratives to
Avoid Bias

• Received from FDA blind to all potential drug
  identifying information
• Drug name
• Company/sponsor name
• Patient identification numbers
• Active or placebo arm
• Any and all medication names and types (e.g. tx
  with other meds may be associated with a
  particular antidepressant side effect profile and
  thus could potentially bias)
• Primary Diagnosis
• Additional Blinding of potentially biasing
  information
• Original label of event given by investigator or
  sponsor
• serious or non-serious labels

17
Rating Guidelines How Was Classification Scheme
Applied?

• Applied concepts using their clinical expertise
  and judgment
• Used experience to integrate clinical information
  and infer when appropriate
• Reasonable certainty in order to commit to a
  rating
• Rating based on what was probable or likely, not
  what was possible

18
Guidelines for Intent Inference

• Infer if clinically impressive or
• Using 2 smaller pieces of clinical information
• Clinical information that could inform inference
  of intent included
• Clinical circumstances (method used, number of
  pills)
• Past history of suicide attempt
• Past history of self-injurious behavior/self-mutil
  ation
• Family history of suicide/suicide attempts

19
Case Example of Inferred Intent Clinically
Impressive Circumstances

• Clinical impressiveness overrules stated intent

The subject attempted suicide by immolation.
Her siblings doused the flames immediately. She
was left with minor burns on her abdomen and one
on her left shoulder that were treated. The
subject admitted that she was angry with her
parents for going away and leaving her alone at
home, because she was fearful. The subject
admitted that she had acted impulsively and had
not intended to kill herself.
20
Another Examples of Clinically Impressive
Circumstance Suicide Attempt

• CASE 1 The patient, age 16, claimed to have
  ingested 100 tablets of the taper study
  medication after a fight with her mother.
• The patient informed her mother, who then brought
  the patient to an emergency room. The patient
  reportedly felt shaky
• The emergency room physician stated that the
  patient looked okay, but was slightly
  tachycardic with a pulse of 100
• Tox screen negative
• The patient remained in the emergency room for
  several hours until she was completely
  asymptomatic
• The patient was later admitted to psych unit

21
Examples of Suicide Attempt, cont.

• CASE 2 After a conflict with her father, the
  patient, age 17, took an overdose of 20 (several)
  tablets. In her fathers opinion, the overdose
  was 5 tablets. The patient did not have any
  symptoms of an overdose, not even nausea.

22
Examples of Suicide Attempt, cont.

• CASE 3 Following a disagreement with her mother,
  the patient, age 15, intentionally overdosed.
  She consumed 12 tablets of study meds, 23 __, 12
  __, 23 ___, 29 ___, 4 ___ and 10 __ tablets.
  Consumed 113 tablets.
• CASE 4 The patient, age 15, impulsively slit her
  wrists following an altercation with her mother.
  The wounds were superficial and were not
  stitched.
• CASE 5 Age 17, she attempted suicide by taking 8
  tablets of ____ after a fight with her father,
  whom she considered harsh and rejecting.

23
Examples of Self-Injurious Behavior, Intent
Unknown

• CASE 1 The patient, age 10, had superficial
  scratches, left arm, scratched self with scissors
  
• CASE 2 The patient, age 14, ingested or
  simulated ingestion of 2-3 cigarettes. The
  patient was reported as feeling tired and playing
  a theatrical role
• CASE 3 Subject, age 9, reported he had ingested
  four of his brothers tablets on a dare
• CASE 4 The patient, age 10, swallowed a small
  amount of aftershave lotion while angry

24
Examples of Preparatory Actions

• CASE 1 Age 16 on day 63, she tried to hang
  herself and was prevented from doing so by her
  family.
• CASE 2 The voice commanded the patient, age 18,
  to jump from the roof. Although the patient went
  to the roof, he did not jump.
• CASE 3 The patient, age 10, experienced suicidal
  ideation with plan. Reportedly, the subject held
  a kitchen knife to her neck while alone but did
  not cut herself and the event was not witnessed.
  At her next scheduled visit, the subject reported
  the suicidal ideation described above.
• CASE 4 The patient, age 18, was noted to be
  hostile, hopeless, and helpless and had written
  suicide notes.

25
Examples of Self-Injurious Behavior, No Intent

• CASE 1 The patient stated that there is
  increased family tension over the past six days
  and that she made superficial cuts on her wrist
  with an Exacto Knife. The patient and mother
  reported that the cuts werent deep and that they
  looked like a cat scratched her. The patient
  adamantly denied any suicidal gestures or intent.
  The patient stated that she only wanted a
  release and that cutting and hitting her legs
  offers a release.
• CASE 2 Denies suicidal thoughts. The first time
  she cut herself was at age 16 and stated that she
  did it for attention. Today her cutting was more
  spontaneous. She reported that cutting gives her
  a good weird feeling.

26
Results Number of Events and Cases
FDA Severity Hierarchy

• 427 Events

378 Cases
Suicide Attempt Preparatory Actions Suicidal
Ideation SIB, intent unknown Not enough
Information
27
Results Expert Rater Consensus

• Only 2/427 cases had no agreement among the 3
  raters
• 59 cases had agreement among 2 of 3 raters, had
  to go to teleconference
• No cases in which consensus not able to be
  reached during teleconference

28
Discordant Cases Between Sponsor and Columbia
Classifications

• 40/427 cases in which sponsor and Columbia
  classification differed
• 26 new cases identified that had not been
  identified by sponsors as possibly
  suicide-related
• 2 new cases of self-injurious behavior without
  suicidal intent that had been labeled something
  other than deliberate self-harm
• -12 cases originally called possibly suicidal
  changed to something other than possibly suicidal
  

29
Discordant Cases Newly Identified

• 26 new possibly suicide-related cases identified
  among expanded body of events
• 26 possibly suicide-related
• 1 Suicide Attempt
• 1 Preparatory Act
• 13 Suicide Ideation Events
• 4 Self-Injurious Behavior Intent Unknown
• 7 Not Enough Information to say whether
  deliberate self-harm

30
Example of Newly Identified Suicidal Event
Preparatory Act

• The patient, age 11, held a knife to his wrist
  and threatened to harm himself. The patient was
  hospitalized with an acute exacerbation of major
  depressive disorder. The patient was treated
  with __ and discharged in stable condition.
• Original Adverse Event Label Exacerbation of
  Major Depressive Disorder

31
Discordant Cases Events Changed from Suicidal to
Something Other

• 12 Events
• 2 changed to psychiatric
• 1 changed to accident
• 9 changed to self-injurious behavior no suicidal
  intent
• Example Event
• this patient is reported to have engaged in an
  episode of automutilation where she slapped
  herself in the face. The event resolved the same
  day without any intervention.

32
Results Agreement with Sponsor (I) Possibly
Suicidal (1,2,6,3,10)
Columbia Classification- Possibly Suicide Related
No
Yes
237 26
12 102
Sponsor Classification- Possibly Suicide Related
263
No
114
Yes
128
249
Kappa.77
33
Results Agreement with Sponsor (II) Definitely
Suicidal (1,2,6)
Columbia Classification- Definitely Suicide
Related
No
Yes
248 15
32 82
Sponsor Classification- Definitely Suicide
Related
263
No
114
Yes
97
280
Kappa.69
34
Results Reliability of Columbia University
Classification

• Primary Outcome ICC
• Suicide Attempt .81
• Preparatory Actions .89
• Suicidal Ideation .97
• Sensitivity Outcome
• SIB, Intent Unknown .67
• Not Enough Info .47
• Median ICC .86

35
Future Directions

• Improve adverse event reporting for
  suicide-related events by
• Developing a consistent terminology
• Developing guidelines for classification of
  suicidality (so adequate info provided by
  clinician)
• Utilization of research assessment tools (what
  questions to ask, how to ask, and what measures
  aid this)
• Improved, More Valid Identification and
• Documentation Of Suicidality

Also describes 12