Challenging Behaviors: Effective service provision models

1
Challenging Behaviors Effective service
provision models

• CATHLEEN C. PIAZZA WAYNE W. FISHER
• Munroe-Meyer Institute and
• University of Nebraska Medical Center

2
Three Important Components of Effective Service
Delivery

• Building and Effective Organization
• Building an Effective Team
• Staff Training and Management

3
Mission of the Center for AutismSpectrum
Disorders

• provide comprehensive, state-of-the-art clinical
  services
• advance knowledge about the causes of, and
  treatments for ASD through systematic research
• disseminate information about effective
  assessment and treatment through education,
  professional training, and consultation

4
Feeding Disorders Program
Asperger Syndrome Program
School Consult Program
Future Programs
5
Growth of Center Staff
40
35
30
25
20
Number of Staff
15
10
5
0
0
Jul-05
Jul-06
Jul-07
Jan-06
Jan-07
Jan-08
6
Number of Clients Served in 2007

• Autism Diagnostic Clinic 202
• Severe Behavior Evals 56
• Severe Behavior Tx 34
• Feeding Evals 42
• Feeding Tx 44
• Early Intervention 10
• Total
  388

7
Research by the CASD Faculty and Staff

• In the past year
• 20 clinical research studies have been published
  or accepted
• 27 Professional Presentations have been completed

8
Samples of Research by the CASD Faculty and Staff

• Lomas Mevers, J., Fisher, W. W., Kelley, M. E.
  (in press). Evaluation of variable-time delivery
  of food items as treatment for problem behavior
  reinforced by escape. Journal of Applied
  Behavior Analysis.
• Bouxsein, K. B., Tiger, J. H., Fisher, W. W.
  (in press). A comparison of general and specific
  instructions to promote task engagement and
  completion. Journal of Applied Behavior Analysis.
• Roscoe, E. M., Fisher, W. W. (in press).
  Evaluation of an efficient method for training
  staff to implement stimulus preference
  assessments. Journal of Applied Behavior
  Analysis.

9
Samples of Research by the CASD Faculty and Staff
(Cont.)

• Grow, L. L., Kelley, M. E., Roane, H. S. (in
  press). The emergence of mands during extinction
  for problem behavior. Journal of Applied Behavior
  Analysis.
• Roane, H. S., Falcomata, T. S., Fisher, W. W.
  (in press). Basing differential reinforcement
  schedule thinning on the behavioral economics
  principle of unit price. Journal of Applied
  Behavior Analysis.
• Bouxsein, K. J., Tiger, J. H., Fisher, W. W.
  (in press). Assessing the influence of
  instructional control of behavior using goal
  setting to increase the task completion of a
  young man with Asperger syndrome. Journal of
  Applied Behavior Analysis.

10
Submitted Grant Applications

• NIH grant to develop and test an internet-based
  screening and referral system for use in rural
  states
• NIH grant to conduct a randomized trial of
  functional analysis and treatment of destructive
  behavior among children with autism in school
  settings
• IES grant to develop function-based interventions
  for individuals with Asperger syndrome who
  display destructive behavior
• IES grant to test a model we have developed for
  designing individualized treatments for teaching
  conditional discriminations to young children
  with autism
• Intestinal failure and feeding disorders
  Prevalence and treatment

11
Training Opportunities at the Center for Autism
Spectrum Disorders

• Pre-doctoral internship approved by the American
  Psychological Association
• Postdoctoral fellowship
• Masters program in educational psychology with
  specialization in applied behavior analysis
• Practicum experiences for masters, bachelors,
  and AmeriCorp students

12
Building an Effective Team

• The Pediatric Feeding Disorders Program as an
  Exemplar

13
Feeding Disorders

• Feeding Disorders may be characterized by one or
  more of the following
• Symptoms associated with nutritional status
  (e.g., Failure to Thrive)
• Symptoms related to oral motor aspects of eating
  (e.g., oral motor dysfunction)
• Inappropriate behavior during meals (e.g.,
  tantrums)

14
Feeding Disorders

• Clinical presentation of a feeding disorder may
  include
• Failure to thrive
• Sensory defensiveness
• Texture selectivity
• Tube dependence
• Behavior problems

15
(No Transcript)
16
Patient Demographics

• Patient Demographics
• Mean Age 3 years (39 months)
• Gender 68 male, 38 female
• Developmental level
• 53 Developmental Delays
• 47 Typical Cognitive Development

17
Patient Demographics

18
Feeding Disorders

• Etiologies
• Medical
• Oral Motor
• Behavioral

19
Feeding Disorders

• Complex interaction of multiple factors
• Medical Mom feeds me I get sick
• Behavioral I Fight/Refuse Food goes away
• Physiological Decreased intake Increased
  tolerance
• Oral Motor No practice eating Poor eating
  skills

20
Pediatric Feeding Disorders

• Behavioral
• Does the parent end the meal, coax or provide
  increased attention, or give the child preferred
  food or toys following inappropriate mealtime
  behavior?

21
Pediatric Feeding Disorders

• Physiological
• Does the parent report that the child will go for
  long periods of time without eating or drinking?

22
Pediatric Feeding Disorders

• Medical
• Does the child have a history of a medical
  problem that may have caused eating to be painful
  or unpleasant?
• Medical problems sometimes are masked as a
  result of the childs refusal or low oral intake.

23
Pediatric Feeding Disorders

• Oral-motor
• Weak suck,
• Choking or gagging during meals,
• Tongue thrusting or inability to lateralize the
  tongue,
• Wet vocal sounds during or after meals,
• Preferences for smooth or creamy textures.

24
Interdisciplinary Approach

• Interdisciplinary team evaluation
• Medicine Rule out physical causes of feeding
  problem

25
Interdisciplinary Approach

• Nutrition Evaluate adequacy of current intake

26
Interdisciplinary Approach

• Social Work Evaluate Family Stressors

27
Interdisciplinary Approach

• Speech/Occupational Therapy Evaluate Oral Motor
  Status and Safety

28
Interdisciplinary Approach

• Psychology Assess contribution of environmental
  factors

29
Interdisciplinary Approach
Interdisciplinary Medicine, Nursing, Nutrition,
Occupational Therapy, Psychology, Social Work,
Speech Therapy
30
Feeding Disorders
Identify goals


increase acceptance of food


increase volume of food


increase variety


change texture


reduce inappropriate mealtime
behaviors
31
Outcome Oriented
Measurable goals are set for each patient.
32
Data Based
Feeding behavior is quantified and measured
precisely.
33
(No Transcript)
34
Establish a Baseline
BL
Treatment
BL
Treatment
Follow-up
50
40
30
GRAMS OF INTAKE
Mom
20
Jenny
10
0
2
8
14
20
26
32
38
44
50
56
62
68
74
80
86
92
98
104
110
116
122
128
134
SESSIONS
Establish a baseline to document current level of
behavior.
35
Feeding Disorders

• How do we determine which treatment is
  appropriate for any given child?
• For example, thickening liquids may reduce the
  risk of aspiration for one child but increase the
  risk of aspiration for another.

36
Linking Assessment to Treatment

• Function-based treatments
• Assessment and Treatment are linked
• Treatment evaluated for efficacy
• Is behavior improving or worsening?
• Is treatment necessary?
• Is treatment acceptable for family?

37
Treatment of Feeding Disorders
38
Treatment of Feeding Disorders
39
Cost of Traditional Treatment for Year 1
COST FOR 1st YEAR ON FEEDING TUBE
per year
cost
total
Initial PEG (or Tube Plug)
1
10,000
10,000
Medical Procedures
4
2,400
9,600
New Buttons
4
62
248
Replace G-tube
1
400
400
Feeding Pump
1
903
746
IV Pole
1
40
40
Portable Pump Bag
1
2,006
2,006
Tubes and Bags
1
522
522
GI visits
12
300
3,600
Formula
17885oz
.23/oz
4113.55
OT/SLP 2xweek
104
150
15,600
46,875.55
3
40
Cost of Traditional Treatment for Years 2-5
YEARLY COST FOR MAINTANCE ON FEEDING TUBE
per year
cost
total
Tubes and Bags
1
522
522
New Buttons
4
62
248
Medical Procedures
4
2,400
9,600
GI visits
12
300
3,600
Replace G-tube
1
400
400
Formula
17885oz
.23/oz
4113.55
OT/SLP 2xweek
104
150
15,600
34,083.55
4
41
Intensive Interdisciplinary Treatment vs.
Traditional Medical Care
Traditional Medical Care Total Cost for 5 years
183,209.80
Interdisciplinary Treatment Total Cost for 5
years 48,000
Savings 135,209.80
42

• Research for Improving Visual Inspection and
  Interpretation of Single-Case Designs

43
Research on Effective and Efficient Staff
Training Procedures

• Fisher et al. (1993) trained 87 participants to
  do visual inspection of A-B graphs in 15 minutes.
• Interpretation accuracy increased from 55 to 94
• Roscoe Fisher (2008) trained 16 individuals to
  conduct stimulus preference assessments in a
  single session.
• Implementation accuracy increased from 43 to 95
• Moore Fisher (2007) developed a video-modeling
  procedure for training staff members to conduct
  functional analysis sessions.
• Implementation accuracy increased to over 90

44
Prior Research on Visual Inspection

• Behavior analysts have suggested that visual
  inspection is generally reliable and
  conservative, but findings from empirical studies
  have suggested otherwise
• DeProspero and Cohen (1979)
• interrater-agreement coefficient of 0.61
• Other studies have found similar levels

45
Prior Research on Improving Visual Inspection

• Hagopian et al. (1997) used structured criteria
  to increase inter-rater reliability to .94, but
• Criteria are for multielement designs
• Criteria are somewhat cumbersome

46
Prior Research on Improving Visual Inspection
(cont.)

• A simple and efficient method of visual
  inspection involves providing visual aids in the
  form of trend lines
• Bailey (1984) increase the reliability of visual
  inspection
• Rojahn Schulze (1985) increased agreement
  levels between visual inspection and statistical
  analyses
• However, reliability coefficients were still at
  unacceptable levels

47
Split-Middle Method

• Kazdin (1982) recommended using trend lines in
  combination with a statistical procedure (the
  binomial equation) to improve visual inspection,
  a technique call the Split-Middle Method.

48
9
8
7
6
5
Responses Per Minute
4
3
2
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Sessions
49
9
8
7
6
Responses Per Minute
5
4
3
2
1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Sessions
50
Split-Middle Method (cont.)

• The S-M method is simple and efficient
• It would probably improve the reliability of
  visual inspection
• However, the binomial test may produce an
  unacceptable level of Type-I errors when the data
  series contains serial dependence (Crosbie, 1987).

51
Improving the Split-Middle Method

• We proposed a refinement of the S-M technique
  that was designed to be more conservative and
  hopefully produce reliable visual inspection and
  better guard against Type-I errors.
• This refinement involved two criterion line, the
  trend and the mean line from baseline and we
  named it the Dual-Criteria (DC) method.

52
9
8
7
6
Responses Per Minute
5
4
3
2
1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Sessions
53
Experimental Sequence

• Study 1 Monte-Carlo to test the accuracy and
  power of S-M method, our CD method, and two
  statistical methods.
• Study 2 Used a multiple-baseline design to
  evaluated the extent to which the CD method
  improved the accuracy of visual inspection with 5
  behavior therapists.
• Study 3 Used the CD method in a PowerPoint slide
  presentation to rapidly train a large group (n
  84) accurately visually inspect graphs.

54
STUDY 1 Monte Carlo Simulation

• 600,000 graphs were generated using computer
  simulation
• Graphs varied in terms of effect size, ranging
  from 0 to 3 SD
• Graphs varied in terms of level of
  autocorrelation, either 0 , 0.1. 0.3, or 0.5
• Graphs varied in terms of the length of the data
  set, 10 or 20 points

55
Interpretation Methods

• The 600,000 graphs were interpreted using 3
  visual inspection methods and 2 statistical tests
• Visual Inspection Methods
• SM (Split Middle)
• DC (Dual Criteria)
• CDC (Conservative Dual Criteria)
• Statistical Tests
• GLM (General Linear Model)
• ITSE (Interrupted Time Series)

56
Interpretation Methods

• The rates of Type-I errors and the power levels
  of the 5 interpretative tests were then compared.
  
• Due to the large number of graphs, a computer was
  used to apply The SM, DC, and CDC visual
  inspection criteria in Study 1, whereas human
  visual inspectors participated in Studies 2 and
  3.

57
(No Transcript)
58

• Using Computer-Based Visual Aids to Inspect
  Functional Analysis Graphs

59

60

61

62

63

64
Type-I Error Results

• Only CDC and ITSE produced tolerable error rates
  at all levels of autocorrelation.
• CDC produced lower rates of Type-I errors than
  the other four interpretive procedures at all
  levels of autocorrelation.

65
Power Results

• Of the two interpretive method that controlled
  Type-I error rates
• CDC showed considerably greater power than ITSE.

66
Study 1 Conclusions

• Visual Aids combined with structured criteria can
  improve the accuracy of visual inspection to a
  point where
• Visual inspection produced fewer Type-I errors
  than statistical methods, and
• Visual inspection produced greater power than
  statistical methods.

67
Baseline
Treatment
100
80
60
40
Participant 1
20
100
80
60
40
Participant 2
20
100
80
60
Percentage Correct
40
Participant 3
20
100
80
60
40
Participant 4
20
100
80
60
40
Participant 5
20
1
2
3
4
5
6
7
8
9
10
Sessions
68
RESULTS

• Participant BL (M) Tx (M)
• 1 51.7 95.0
• 2 47.5 88.3
• 3 57.0 93.8
• 4 65.0 93.0
• 5 55.8 97.5

69
STUDY 3 METHODGroup Training of Visual
Inspectors in the DC Method

• Materials
• 3 20-graph packets containing graphs similar to
  those used in Studies 1 2
• Graphs were presented in a PowerPoint
  presentation

70
STUDY 3 METHODGroup Training of Visual
Inspectors in the DC Method

• Participants/Setting
• 87 adults attending a workshop on behavior
  analysis at an annual meeting of a state chapter
  of the Association for Behavior Analysis
• Participants who received an answer sheet with a
  large A at the top were assigned to Group A
  those that received an answer sheet with a large
  B at the top were assigned to Group B
• Each group received 15 min of verbal instruction
  and modeling

71
(No Transcript)
72
DISCUSSION

• Study 1 answered basic questions about the
  accuracy of several methods of interpreting A-B
  designs
• Study 2 showed how the information from Study 1
  could
• Increase the accuracy of human visual inspectors
• Bias the inspectors toward more conservative
  interpretations (i.e., decreasing the ratio of
  Type-I to Type-II errors)

73
DISCUSSION

• Study 3 showed that procedures in Study 2 could
  be translated into a format that would facilitate
  rapid training of large groups of individuals to
  interpret single-case designs

74
The Effects of Videotaped Modeling on Staff
Acquisition of Functional Analysis Methodology
James W. Moore and Wayne W. Fisher
75
Introduction

• Recent studies have focused on staff and teacher
  acquisition of FA skills (Iwata et al., 2000,
  Moore et al., 2002, Wallace et al., in press)
• Each study introduced multiple training
  strategies simultaneously

76
Training Strategies

• Iwata et al. (2000)
• Method section of Iwata et al. (1982/1994)
• Written protocols
• Lectures
• Videotape examples (simulated)
• Feedback

77
Training Strategies

• Moore et al. (2002)
• Method section of Iwata et al. (1982/1994)
• Written protocols
• Face-to-face consultation
• Videotape examples (in vivo)
• Modeling
• Rehearsal
• Performance Feedback

78
Purpose

• In the current study we attempted to
• Replicate previous research on staff acquisition
  of FA methods
• Compared videotape examples that contained models
  of all therapist behaviors to examples that
  demonstrated limited examples

79
Participants

• Each participant held a BA in psychology
• One had limited experience in FA methods, the
  others had none
• One had extensive training in other areas of ABA,
  the other two had limited or no experience

80
General Procedures

• Multiple Baseline with Mutli-Element features
• Simulated FA conditions (Iwata et al.,2000 Moore
  et al., 2002)
• Sessions lasted 5 minutes
• Participants trained on Play, Attention, and
  Demand conditions

81
Dependent Variable

• Correct therapist responses
• Scored correct based on occurrence,
  nonoccurrence, or termination of therapist
  behavior relative to client behavior and/or
  protocol

82
Dependent Variable

• Percentage of Correct Responses
• (Number of correct responses divided by total
  opportunities for response) X 100
• Target client behavior self-injurious behavior
• Forceful striking, scratching, rubbing, poking,
  or biting down on own body parts

83
Independent Variables

• Written materials
• Lecture
• Complete Videomodel (CVM)
• Partial Videomodel (PVM)
• Post-Session Feedback

84
Naturalistic Baseline

• Therapist given Written Materials the day before
  and asked to read
• Comprehension tests
• FA conditions run with an actual client
• No feedback was provided

85
Simulated Baseline

• Therapist asked to read the Written Materials
  again
• No feedback was given from the naturalistic
  baseline
• Simulated FA conditions
• No feedback was provided

86
Training Phase I

• All participants received Lecture training
• CVM training
• PVM training
• Lecture Only
• Simulated FA conditions
• Mastery Criteria 80 or greater accuracy on
  three consecutive conditions

87
Training Phase II

• The condition trained via the CVM was omitted if
  mastery criteria were reached
• The condition trained via Lecture Only received
  training with the CVM
• The condition trained via PVM remained in that
  training mode if mastery criteria were not
  reached

88
Training Phase III

• The condition trained via the CVM in Phase II was
  omitted if mastery criteria were reached
• The condition trained via the PVM received
  training with the CVM

89
Training Phase IV

• Only used with Participant 3
• One condition received CVM-only training
• The other condition received CVM with
  post-session feedback training

90
(No Transcript)
91
(No Transcript)
92
(No Transcript)
93
(No Transcript)
94
Training Results

• CVM training was superior to PVM training or
  Lecture-only training
• Even when mastery criteria were not met for CVM
  training, accuracy still demonstrated significant
  gains
• For two participants, PVM resulted in declining
  accuracy over time

95
Conclusions

• Videotape examples are adequate only if the
  examples contain samples of all potential
  therapist behaviors
• Indirect forms of training (readings, lectures)
  provide only minimal competence in FA protocols

96
Concluding Comments

• Applied Behavior Analysis has well-defined and
  effective procedures for treating challenging
  behavior and for training parents and staff to
  accurately implement those procedures.

97