Impact of Hurricane Katrina on Gulf Coast veterans with and without pre-existing mental illness: South Central MIRECC Hurricane Study

1
Impact of Hurricane Katrina on Gulf Coast
veterans with and without pre-existing mental
illness South Central MIRECC Hurricane Study

• Joseph Constans, Ph.D.
• South Central MIRECC
• May 15, 2007

2
Schizophrenic

• 2  contradictory or antagonistic qualities or
  attitudes ltboth partieshave exhibited
  schizophrenia over the desired outcome
  Elizabeth Drewgt

3
Schizophrenic

• 2  split personality

4
Impact of Hurricane Katrina on Gulf Coast
veterans with and without pre-existing mental
illness South Central MIRECC Hurricane Study

• Joseph Constans, Ph.D.
• South Central MIRECC
• May 15, 2007

5
Survey ResearchNavigating VA Data Security and
Privacy Directives

• Joseph Constans, Ph.D.
• South Central MIRECC
• May 15, 2007

6
Impact of Hurricane Katrina on Gulf Coast
veterans with and without pre-existing mental
illness.

• Principal Investigator Joseph Constans, Ph.D.
• Co-Investigators
• Greer Sullivan, M.D.
• Jennifer Vasterling, Ph.D.
• Project Director Elizabeth Deitch, Ph.D.
• VA Database managers
• Teresa Hudson, PharmD
• Mark Austen
• Survey Development and Administration Consultant
  
• Dana Perry, M.S.

7
Hurricane Katrina

• The costliest the most destructive hurricane .
• Massive size
• devastated Gulf Coast up to 100 miles from
  center.
• Over half of New Orleans flooded.
• Mississippi coastal communities were wiped out.
• Over 1700 deaths due to storm / its aftermath.

8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Impact of Natural Disasterson Health

• Increased general distress
• Lasting gt30 months
• Increased incidence of psychiatric disturbance
• PTSD (31)
• Depression (30)
• GAD (11)
• Panic disorder (10)
• Hurricane Andrew estimates (Davis et al., 1996)
• Poorer physical health outcomes
• Lower immune function
• Impaired sleep

15
Risk factors for Poor Health Outcomes

• External Factors (outside of individual)
• Severity of natural disaster
• Similar to relationship between combat /
  emotional disorders
• Severe disasters more psychopathology
• Loss of property / resources
• Poor social support
• Primary moderator of negative mental health
  outcomes

16
Risk factors for Poor Health Outcomes

• Internal Factors (characteristics of the
  individual)
• Demographic characteristics
• Gender
• Ethnic minority status
• Lower SES
• Pre-existing mental illness
• Mostly using retrospective assessment
• Prospective studies emphasized depression
• Role of PTSD / Schizophrenia ????
• Unclear how PMI interacts with other risk factors
• Cognitive appraisal style
• Beliefs about meaning of the trauma

17
Hurricane Katrina Risk FactorsImmediate
Stressors

• Fear for life of self others
• From storm itself, floodwaters, hunger, thirst
• Fear of crime and violence
• Looting, assaults, breakdowns of civil authority
• Breakdown of civil authority
• Exposure to bodies / potential contaminates
• Uncertainty Ambiguity
• N.O. evacuation stretched to a month status of
  homes, jobs, friends/family/pets uncertain.

18
Hurricane Katrina Risk FactorsProlonged
Stressors

• Loss of housing
• Temporary housing in trailers
• Permanent move
• Loss of employment / income
• Loss of belongings
• Furniture, photographs, sentimental items

19
Hurricane Katrina Risk FactorsDisruption of
Social Networks

• Families separated
• Loss of neighbors
• Loss of church organization
• Changes in family structures
• Increased divorce, marriage, births

20
Hurricane Katrina Risk FactorsDemographic
Characteristics

• Loss of housing
• Temporary housing in trailers
• Permanent move
• Loss of employment / income
• Loss of belongings
• Furniture, photographs, sentimental items

21
Reactions to KatrinaEmpirical Studies to Date

• Kessler, Galea, Jones, Parker (2006)
• Use of cell phones for contacting residents
• Cell phone numbers obtained from service
  organization
• Access to mental disorder incidence prior to
  storm
• From NCS-R
• Documented increases in mental disorders
• PTSD, depression most common
• Frequency of disorders doubled
• Found suicidal ideation did not increase
  proportionately
• Beliefs in post traumatic growth important
  moderator

22
Challenges

• Diaspora
• Door-to-door interviews strategy - impossible
• Hurricane Andrew studies
• Changes in contact information
• Those experiencing greatest loss / trauma may be
  the most likely to have moved.

23
VA opportunities

• Contact information
• At time of storm
• Updated since Katrina??
• Prescription
• Health care
• Benefits
• Veteran population
• Prospective data
• Mental illness diagnoses
• Medical conditions

24
Examining impact of Katrina on Gulf Coast
veterans

• What is the health impact on veterans?
• Veterans have not been specifically studied.
• It is unclear how veterans, especially those with
  combat stress, will react to disasters.
• What is the specific impact on veterans with
  psychiatric disturbances such as PTSD/
  schizophrenia?
• Use of VA databases to identify /- for
  pre-existing mental illness
• Allows for a prospective assessment
• Allows to assess possibly unique features of PTSD
  / schizophrenia
• Examine fit of current risk models
• Does psychiatric status impact other risk
  factors?

25
Study Aim 1

• How does pre-storm psychiatric status impact
  post-disaster health functioning
• Additional negative mental health outcomes
• Physical health status
• General functioning

26
Study Aim 2

• How does pre-storm psychiatric status impact
  known predictors, mediators, and moderators of
  post-disaster health outcomes?
• Increased exposure to trauma
• Lower post-disaster social support
• Pathological cognitive styles
• Increased negative appraisals
• Decrease belief in growth potential
• Decreased capacity for associated with negative
  mental health outcomes

27
Study Aim 3

• Are models used to predict post-disaster negative
  health outcomes comparable for those with and
  without mental illness?
• Test separate models for those with and without
  pre-storm mental illness

28
Overview of Study Design

• Telephone survey
• experiences during after Katrina
• Loss, social support, cognitive factors
• Current health / mental health functioning
• Administered by professional survey firm.
• 500 participants
• 250 with pre-existing mental illness
• 250 with no PMI.

29
Participants

• 500 male veterans
• Age 18-60 as of 8/29/05
• Received services at N.O. or Biloxi VAMC
• Resided in affected area at time of Katrina
• LA parishes Jefferson, Orleans, Plaquemines, St.
  Tammany, St. Bernard
• MS counties Hancock, Harrison, Jackson
• Further restricted by zip codes within
  parish/county to increase probability of direct
  exposure to hurricane stress.

30
Participants

• Grouping variable Pre-storm mental illness
• MI positive cohort (n250)
• 2 visits to mental health clinic between
  8/1/04-8/1/05
• Diagnosis of depression, PTSD, or psychotic
  disorder
• MI negative cohort (n250)
• 2 visits to primary care clinic between
  8/1/04-8/1/05
• No mental illness diagnoses

31
Survey Measures

• Measuring internal characteristics - prestorm
• Demographic Information
• Age, sex, ethnicity, marital status, education
• Military History
• Service branch, combat zone experience
• Traumatic life events
• Checklist of lifetime traumatic events other than
  Katrina (e.g., serious accident, assault).
• Modified version of Diagnostic Interview Schedule
  (CDC, 1989)
• Measuring internal characteristics post storm
• Cognitive appraisal style
• Posttraumatic Cognitions Inventory (Foa, Clark,
  Tolin Orsillo, 1999)
• Posttraumatic growth (Kessler, Galea et al.,
  2006)

32
Survey Measures (contd)

• Measuring external factors
• Katrina trauma severity
• Emotional events experienced (e.g., witnessing
  deaths, loss of friends/family/pets in storm
• Other traumas during storm aftermath (e.g.,
  assault, robbery)
• Loss - Household Information
• Living situation pre-K and now (location,
  structure, household members) extent of damage
  to home/possessions
• Loss - Employment/Income
• Status at time of Katrina, current status,
  changes
• Social Support
• Actual changes in family, neighborhood, and
  friendship networks.
• Perceived social support.

33
Survey Measures (contd)

• Health Outcomes
• Post-traumatic stress disorder
• Short PTSD Rating Interview (SPRINT Connor
  Davidson, 2001)
• Depression
• 9-item screen from Patient Health Questionnaire
  (PHQ-2 Kroenke , Spitzer Williams, 2003)
• Panic disorder
• PD screen from Brief Patient Health Questionaire
  (Spitzer, Kroenke Williams, 1999)
• Generalized Anxiety Disorder
• GAD-7 (Spitzer, Kroenke, Williams, Lowe, 2006)
  assesses frequency of anxiety symptoms

34
Survey Measures (contd)

• Health Outcomes
• Anger
• Anger subscale of the Buss-Perry Aggression scale
  (Buss Perry, 1992)
• Alcohol Use
• 3 items from Alcohol Use Disorders Identification
  Test (Babor, de la Fuente, Saunders Grant,1989)
• Also, compare consumption pre- post-Katrina
• Physical Health
• 9-item checklist for specific health problems
  (e.g., diabetes, cancer)
• 16-item measure for respondent evaluation of
  state of health some original items and some
  from SF-12 Health Survey (Ware, Kosinski
  Keller, 1996)

35
Procedure Participant Identification

• Using info from VA data warehouse
• 250 MI veterans will be randomly selected.
• PTSD, depression, psychotic disorder
• 250 MI- veterans
• Groups match MI cohort on age, ethnicity, zip
  code at time of Katrina, number of clinic visits.
• Names contact info supplied to survey
  administration firm.

36
Procedure Contact Protocol

• Pre-notification letter
• sent to potential participants, allows opt-out
  by postcard or phone (wait 3-4 weeks).
• Interviewer call
• obtain verbal informed consent.
• Survey as telephone interview
• (about 1 hour long).
• Address recorded
• 20 compensation check.

37
Procedure Follow-Up Location

• For veterans
• Undeliverable mail
• Invalid phone numbers
• Use of locator service (possibly using SSN).
• Second iteration of identification/enrollment
• Generate list of names for 120 of remaining
  veterans needed to reach n250 for each group.
• Follow contact protocol
• Repeat if necessary

38
Results

• Overcoming obstacles.

39
Data Security Issues

• Stolen laptop with veteran data
• Washington, DC
• Laptop recovered no PHI compromised
• Loss of veteran data external storage device
• Birmingham, AL

40
Security Requirements for VA Research Information

• Memorandum February, 2007
• VA research data may not be stored outside VA
  (unless applicable permission obtained).
• Data transfer to non-VA computer must comply with
  VA directive 6504.
• When VA data is stored on non-VA system, system
  must meet requirements set for in FISMA.
• Research subjects names, addresses, and Social
  Security Numbers may be stored only on VA
  servers.

41
Challenges Privacy/Security Issues

• New VA security directives.
• Data Security checklist
• Required for all research
• protocols
• Researchers must indicate
• compliance with several
• statements about data security
• Research may only proceed
• without delay if all statements
• can be answered yes.
• Two statements are particularly
• problematic for this study ?

42
(No Transcript)
43
Problematic Security Checklist Items

• Names, addresses, and Social Security Numbers
  (real and scrambled) have been replaced with a
  code. Note Names, addresses, and Social
  Security Numbers (real or scrambled) may only be
  maintained on a VA server and documentation of
  the procedure by which the data were coded must
  remain within the VA
• We cannot answer yes to this statement ?

44
Problematic Security Checklist Items

• Contact info is not useful in coded/scrambled
  form.
• we need to actually contact veterans
• Social Security numbers required
• aid in location of displaced individuals.
• If SSN not used, possible under-representation of
  those most disrupted by Katrina
• The info must be removed from VA servers
• Survey firm needs to generate letters,
  contact/callback protocols incentive checks.
• Need to use info for location services
• Need to interface with proprietary CATI software

45
Problematic Security Checklist Items

• For data that will reside on a non-VA server
  The server has to be certified and accredited as
  required by Federal Information and Security
  Management Act of 2002 (FISMA). Note your
  facilities ISO should be consulted.
• We cannot answer yes to this item either ?

46
Problematic Security Checklist Items

• Survey administration firm, SRBI, has provided
  extensive data security protocol, and certifies
  that they meet the standards set forth in FISMA.
• However, they are not officially FISMA
  accredited.
• FISMA accreditation would take at least a year,
  and SRBI maintains it would cost more than all
  their VA contracts combined.

47
Unclear Procedures

• Directives are vague at this point regarding what
  can be done if items must be answered no
• Checklist states If you answer NO to any one of
  the statements, you may not remove or transmit
  the data outside the VA and you must consult with
  your supervisor, ISO and Privacy Officer.
• Can supervisor, ISO Privacy Officer approve an
  exception to checklist elements?
• Meaning of consult is unknown
• No clear placement of responsibility for
  approval/exception.
• No forms exist for approval/exception

48
Unclear Procedures (cont)

• National Procedures also vague
• No clear written information about how to obtain
  names, addresses and phone numbers.
• Web site and other instructional materials only
  address SSN

49
Plan for Action

50
South Central MIRECC Hurricane Study A
Metaphor for the Rebuilding of New Orleans?

• Takes longer than expected
• Systems become dysfunctional
• IRB
• Personnel
• PI becomes dysfunctional?
• Hurdles will emerge.
• We will ultimately succeed!