Computerized Predictive Model of DistressDistraction - PowerPoint PPT Presentation

Title: Computerized Predictive Model of DistressDistraction


1
Computerized Predictive Model of
Distress/Distraction

• Ann Marie McCarthy
• The University of Iowa
• MNRS, 2009
• Pediatric Nursing Preconference
• NINR 1 R01 NR05269-01 Funding from 2002-2011

2
Translation Science Research into Practice
3
Research Team

• University of Iowa, Iowa City IA
• Ann Marie McCarthy, Primary Investigator
• Charmaine Kleiber, Co-PI
• Deb Schutte, Investigator
• Kirsten Hanrahan, Program Associate
• Nancy Weathers, Project Director
• Brenda Gordley, Research Assistant
• Bridget Zimmerman, Nick Street, Jeff Murray,
  Consultants
• Blank Childrens Hospital, Des Moines IA
• Susan Allen, Investigator
• Julia Krapfl, Research Assistant
• St. Louis University, St. Louis MO
• Nina Westhus, Investigator

4
Background

• Procedural pain is a major source of distress in
  children.
• Distraction- diverting a childs attention away
  from the medical procedure and focusing on other
  things- can decrease child distress with
  procedures.
• We can teach parents to use distraction to
  decreases child distress.

5
Long Term Goal

• To develop a clinically useful predictive
  distress profile that can be used to identify
  interventions appropriate to individual family
  needs

6
Predicting Childrens Response to Distraction
from Pain orDistraction Intervention Grant (DIG)
NIH/NINR 1 R01 NR05269 Funding from 2002-2007
7
DIG Purpose

• To identify predictors of child response to a
  peripheral IV insertion when a parent provides
  distraction

8
DIG Methods

• Design Random assignment to control or
  experimental groups
• Sample 542 children, 4-10 years old, planned IV
  insertion
• Intervention
• Educational material discussion
• Video
• Distractors

9
Participants -All Site
10
(No Transcript)
11
Predictors and Outcomes

• Predictors Child, parent, procedure variables
• Genotype
• Child Outcomes
• Behavioral Observation Scale of Behavioral
  Distress-Revised (OSBD-R) (Elliott et al, 1987
  Jay et al, 1983 Jay Elliott, 1984)
• Biological Salivary cortisol levels (Kirschbaum
  Hellhammer, 1994)
• Self Perception of Pain Oucher Scale (Aradine et
  al., 1988 Beyer Aradine, 1988)
• Parent Perception of Child Distress (Kazak et al,
  1996)

12
DIG Predictors of High Distress
13
Developing and Testing the Computerized
Predictive Model for Distress/Distraction
NIH/NINR 1 R01 NR05269 Funding from 2007-20011
14

Next Step
Model of Child Responses to Painful Procedures
when Distraction is Provided by a Trained
Parent (McCarthy Kleiber, 2006).
Customized Intervention
Extraneous Procedural Variables

• Difficulty of Procedure

• Use of EMLA

Child's Response
Child Characteristics
Child Perception of Psychologic Physical
Discomfort

• Behavioral

• Biological

• Age

• Anxiety

• Self Report of Pain

• Gender

• Coping Styles

• Ethnicity

• Ability to Attend

• Parent Report of Child Distress

• Experience

• Genotype

• Temperament

• Illness Severity

• Diagnosis

• Pain Sensitivity

Parent Response
Parent Characteristics

• Gender

• Distraction Performance

• Ethnicity

• Experience

• Perception of Success and

Satisfaction with Procedure

• Anxiety

Intervention

• Parenting Style

• Credibility and Efficacy of Intervention

15
CPMD Study Purpose

• The purpose is to use the Computerized
  Predictive Model for Distress/Distraction (CPMD)
  to predict high, medium, or low risk for distress
  groups and to test the level of distraction
  intervention appropriate for each risk group.

16
Computerized Predictive Model of Distraction
17
Data Mining

• searching for consistent patterns and
  relationships among data points based on analysis
  of a large set of data and turning those
  discovered relationships into a model. The
  model is then applied to other test cases to
  determine if the model will hold on new cases

Abbott, P. A., Zytkowski, M. E. (2002).
Supporting clinical decision making. In S.P.
Englebardt R. Nelson (Eds.), Health Care
Informatics An Interdisciplinary Approach. St.
Louis, Missouri Mosby.
18
Predictive Modeling Using Data Mining

• All items identified with Data Mining were also
  identified by traditional regression
• Identified specific predictive items rather than
  global instruments. For example
• I feel anxious item rather than anxiety
  subscale of the STAI
• My child never stops moving item rather than
  activity subscale on DOT
• Items from DIG study244
• Items in CPMD47

19
Computerized Predictive Model of Distraction
(CPMD)
20
Designing the C Computer Application

• Predicts risk for child distress with medical
  procedures
• Tailors evidenced based interventions to the
  parent and child based on their responses to
  questions
• Randomizes high, medium and low risk dyads to
  test interventions in next study
• Links to central secured server

21
Computerized Predictive Model of Distraction
21
22
CPMD Methods

• Design Randomized Clinical Trial
• Subjects Children ages 4-10 having a scheduled
  IV (n 582)
• Settings
• UI Childrens Hospital
• Blank Childrens Hospital
• Cardinal Glennon SLU

23
CPMD Methods

• Instruments
• CPMD that includes elements of child, parent, and
  procedural variables
• Genotyping
• Outcomes
• Behavioral (videotaping)
• Biological (cortisol)
• Oucher (included in CPMD)
• Parent report (included in CPMD)

24
Login Instructions

• Enter the CPMD site http//beedance.biz.uiowa.ed
  u/cpmdpro/login.php
• Login ID demo
• Subject ID enter a 4 digit number
• Password qwe123
• Click on menus and follow the prompts.

25
Translation Science Research into Practice
26
DIG Related Analyses

• Distraction Coaching Index (DCI)
• Genetics
• EDNRA
• Candidate genes for pain anxiety
• Cortisol
• Norms
• ADHD

27
Distraction Coaching Index (DCI)

• Treatment Integrity Is the intervention carried
  out as expected?
• Development of the Distraction Coaching Index
  (DCI) (Kleiber et.al., Childrens Health Care,
  2007)
• Frequency
• Quality
• Video taped and scored later

28
Genetics Results

• Forty percent of subjects reported high pain
  (Oucher score 4-10) despite the appropriate
  application of topical analgesic.
• Pain response did not differ by
• Gender (?2 0.28, .5 lt p lt .7)
• Type of topical anesthetic (?21.07, p0.301)
• Ethnicity ( ?2 0.37, p0.54)

29
Logistic Regression for Likeliness of High Pain
Kleiber, C., Schutte, D.L., McCarthy, A.M.,
Floria-Santos, M., Murray, J.C., Hanrahan, K.
(2007). Predictors of topical analgesic
effectiveness in children. Journal of Pain,
8(2)168-174.
30
Salivary Cortisol Norms for Children
31
Cortisol Responsivity ADHD
Note Expect cortisol to be higher on the clinic
(stressful) day as it was in the Non-ADHD sample.

32
Implications for Practice

• We can teach parents to use distraction.
• Higher distraction quality is associated with
  less behavioral distress.
• Children have pain despite topical anesthetics.
  There may be a genetic cause related to the EDNRA
  gene.
• Normal cortisol levels and responsivity follow
  the pattern of adults. Children with ADHD have
  atypical cortisol responses.
• Predictors of parent child responses can be
  used in practice

33
Thank You for Your Interest!