Behavioral Approaches to the Treatment of Autism and Severe Behavior Disorders

Behavioral Approaches to the Treatment of Autism and Severe Behavior Disorders

Wayne W. FisherUniversity of Nebraska Medical Center’s Munroe-Meyer Institute

UNMC Munroe-Meyer Institute

Autism and Childhood Schizophrenia

Once thought to be a form of schizophrenia

Differs from schizophrenia in terms of symptoms, age of onset, family history, etiology, and response to treatment


Definition of Autism

markedly abnormal or impaired development in:

1. social interaction

2. Communication

and markedly restricted repertoire of activities and interests.


Definition of Autism

Definitions are cheap, but explanations are dear, and we must be careful not to confuse them.

» David Palmer, 2004


Prevalence of Autism

Typically diagnosed within first three years

Recent estimate indicate that the prevalence of ASD is between 1 in 38 and

1 in 88 Four times more prevalent in boys than

girls


Recurrence Risk for Siblings

If an older sibling has and autism spectrum disorder, the risk for a

Younger brother is 1 in 4

Younger sister is 1 in 11


NIH Research Dollars Devoted to Autism

When Compared with Other Serious Childhood Conditions, Autism is Much More Common, but Fewer Dollars Per Case are Spent on Autism.


0

10

20

30

40

50

60

70

Autism Juvenile DiabetesMuscular Dystrophy Leukemia Cystic Fibrosis

Prevalence of Autism and Other Conditions

(Number of Cases per 10,000 Children)


$-

$20,000

$40,000

$60,000

$80,000

$100,000

$120,000

$140,000

Autism Juvenile Diabetes Muscular Dystrophy Leukemia Cystic Fibrosis

NIH Research Dollars for Autism and Other Conditions

(Number of Dollars per Case)


Unfortunately, you have what we call “no insurance.”


Demographics of Autism

Affects all racial, ethnic, and national groupsFamily income, lifestyle, and educational levels

do not affect the chance of autism's occurrence Diagnosis of autism is growing at a rate of 10-

17 percent per year (U.S. Department of Education, 2002)


Assessment and Diagnosis of Autism

No medical tests for diagnosing autismAccurate diagnosis is based on observation of the

individual's communication, behavior, and developmental levels.

Autism Diagnostic Interview-R (ADI-R) Autism Diagnostic Observation Schedule (ADOS) Home and/or school observation Video analysis of behavioral observation


Social Behavior Generally Requires Little or No Explicit Training


Assessment and Acquired Autism

Autism is most often diagnosed between 2 and 5 years of age.

Thus, it is natural for parents to look for environmental events occurring shortly before this time that may have caused the autism, such as childhood vaccines.


MMR Vaccines and Autism10 of the original 13 authors of the

investigation that started the controversy have retracted the study’s interpretation, as has the journal, The Lancet

Prevalence rates of autism are equivalent in children who have and have not been vaccinated.

Increases in the prevalence of autism did not abate when thimerosal was removed from vaccines.

Regression in autism is no more likely in the months after the MMR vaccine than in the months before the vaccine.


Identifying the Genetic Bases of Autism Spectrum Disorders

Etiologic Workups Identify Specific Genetic Causes for Autism in About 20% of Cases.

At the Munroe Meyer Institute, Shaefer and Colleagues (2006) have developed a 3-Tiered Approach that Identifies Genetic Causes in 40% of Cases.


Associated Disorders

Mental Retardation 50% 15%Seizure Disorder 35% 10% Self-Injury, Aggression 50% Tourette DisorderBipolar Disorder

Autism ASD


Associated Etiologic Diagnoses

Fragile-X syndromeTuberous SclerosisWilliams syndromeLandau-Kleffner syndromeCongenital RubellaSmith-Magenis syndromeNeurofibromatosis


Genetics and Twin Studies

Autism runs in familiesHeritability for autism is about 90%Monozygotic twin concordance, 60%-

100%Dizygotic twin concordance, 10%Associated with abnormalities on

chromosomes 7q, 2q, and 15q


Applied Behavior Analysis (ABA)

What is ABA?How is it different from other

approaches?How is it Done?


How Effective is ABA for Autism?

About 50% of Children with Autism and no More than Mild Mental Retardation who Receive Early Intervention with ABA Attain Normal IQs and are Educated in Regular Classrooms with Minimal Assistance.


Outcomes of ABA for Autism

0

5

10

15

20

25

30

35

0 5 10 15 20 25 30 35 40 45

Hours per Week of Treatment

Incr

ease

s in

IQ S

core

s

r = .79

p < .02

UNMC Munroe-Meyer InstituteCost-Benefit Analysis of Early, Intensive ABA for Autism

Average Lifetime Cost for a Person with Autism is over $4 million Average cost of Early, Intensive

ABA is $150,000 over about 3 yearsAverage Lifetime Savings from

ABA Treatment is Between $1.6 and $2.7 million


Cost-Benefit Analysis of ABA treatment for Severe Behavior Disorders

Children with Autism and Severe Destructive Behavior Cost $8 million over a lifetime.

Keeping just one child out of chronic care pays for treatment of hundreds more.

Our treatment approach has over an 80% success rate.

Residential CostsOver 25 Years

Case Example(Costs since 1986)

0

500

1000

1500

2000

2500

3000

Dol

lars

in T

hous

ands

$4,050,000

$85,924

Inpatient and Outpatient Costs

3500

4000


Autism and Problem Behavior

98% of children with autism have one or more of the following problems:

1. Feeding problems

2. Sleeping problems

3. Tantrums

4. Self-injurious behavior

5. Aggression

Dominick et al. (2007)



Variables Related to Problem Behavior in Autism

Lower IQ scores were associated more problem behaviors

Lower expressive language scores were associated with more problem behaviors

Increased social deficits were associated with more problem behaviors

Dominick et al. (2007)






Functional Analysis and Treatment of Aberrant Behavior


Structural vs. Functional Diagnosis

Structural Approach

1. How often a particular set of symptoms or responses cluster or covary.

Functional Approach

1. Whether and which environmental variables influence the response.


Functional AnalysisIdentifies the environmental contexts

in which aberrant behavior is likely and unlikely.

Identifies the consequences that reinforce and maintain the behavior.

Used to prescribe effective treatments.


Common Functions of SIB

Social Positive Reinforcement (Attention, Tangible items)

Social Negative Reinforcement (Escape)

Automatic Reinforcement (e.g., Sensory Stimulation)


Essential Features of Functional Analysis Conditions

Unique discriminative stimuli that signal the available of a specific reinforcer

Establishing operation (EO) that increases motivation for the specific reinforcer

1. An EO is an environmental condition that momentarily increases the effectiveness of a reinforcer and that evokes responses that have produced that reinforcer in the past.

Contingency between the target behavior and the specific reinforcer


Identifying the Essential Features of Functional Analysis Conditions

See if you can identify the discriminative stimulus, the establishing operation, and the reinforcement contingency in each of the following functional analysis conditions.


Attention Condition

Adult is busy reading. Child is expected to play quietly with toys.

Adult attention shifts to child following SIB (e.g., “Please don’t hit yourself”).

Determines whether adult attention functions as reinforcement for SIB.


Essential Features of the Attention Condition

Discriminative Stimulus: Adult is seated in a chair reading a book.

EO: Attention is unavailable.

Contingency: SIB produces attention.


Tangible Condition

Adult takes preferred toys or leisure materials from the child and returns them following SIB.

Determines whether access to preferred items functions as reinforcement for SIB.


Essential Features of the Tangible ConditionDiscriminative Stimulus: Adult takes and

holds the preferred tangible item at the start of the session.

EO: Tangible item is unavailable in the absence of SIB.

Contingency: SIB produces the tangible item.


Demand ConditionChild is prompted to complete non-

preferred tasks by an adult.

The task is removed and the child is given a short break following SIB.

Determines whether termination of non-preferred activities functions as reinforcement for SIB.


Essential Features of the Demand Condition

Discriminative Stimulus: Instructional materials and demands are presented.

EO: Nonpreferred demands are presented.

Contingency: SIB results in temporary removal of the demands.


Alone ConditionChild is placed in a room alone

without toys or materials.

Indirectly assesses whether SIB may be maintained by automatic reinforcement (e.g., sensory stimulation).


Essential Features of the Alone Condition

Discriminative Stimulus: Absence of another individual or materials.

EO: Alternative sources of stimulation are unavailable.

Contingency: SIB produces self-stimulation.


Toy Play Condition

Child and adult play together with preferred toys or leisure items.

Adult delivers praise about once every 30 seconds for the absence of SIB.

Designed to be an analogue of an “enriched environment”, which serves as a control condition.

UNMC Munroe-Meyer InstituteEssential Features of the Toy Play Condition

Discriminative Stimulus: Adult and toys are near the individual.

EO (or AO): Attention and tangible items (toys) are freely available and no demands are presented.

Contingency: SIB produces no consequence.

0

1

2

3

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Agg

ress

ive

Res

pons

es

Per

Min

ute

Sessions

DemandTangible

Attention

Ignore

Play

00

22

44

66

88

11 22 33 44 55 66 77 88 99 1010 1111 1212 1313 1414 1515 1616 1717 1818 1919 2020 2121

Res

pons

e pe

r Min

ute

KirkKirk

Session

Baseline Extinction BaselineFCT +Extinction

FCT +

DestructiveBehavior

Communication


Functional Analysis of Covert Drug Ingestion

Three Operant Hypotheses Generated After Caregiver Interview and Chart Review

1. Attention/Excitement From Medical Procedures

2. Attention From Mother

3. Escape From Work Activities


Functional Analysis of Covert Drug Ingestion (cont.)

Sessions Conducted in a Classroom and an Adjoining Medication Room Baited With Placebos in a Pillbox.

Patient was Left Unsupervised in the Classroom With a Schoolwork Assignment.



In the Pillbox Were Four Pill Bottles Containing Placebos.

Each Pill Bottle had a Uniquely Colored Label.

Consuming Pills From Each Pill Bottle Produced a Specific Consequence.



Red = Medical Attention

Orange = Attention From Mother

Blue = Rest Period (Escape)

Yellow = Control (Ignore)

0

5

10

15

20

1 5 10 15

PIL

LS IN

GE

STE

D P

ER

MIN

UTE

Escape From Work

Parent Attention

SESSIONS

ControlMedical Attention

LYLE

FUNCTIONAL ANALYSIS


Treatment of Escape-Maintained Drug Ingestion

Lyle Earned Preferred, Nonwork Activities for Completing Scheduled Activities and Turning in Pills.

He was Required to Complete His Least Preferred Work Activity (Shoe Polishing) if He Ingested Pills.


15

20 Baseline Treatment Baseline Treatment

30 40 60 80

TREATMENT ANALYSIS

LYLE

0

5

10

10

SESSIONS

20 50 70

PIL

LS IN

GE

STE

D P

ER

MIN

UTE


Treatment Extension of Escape-Maintained Drug Ingestion

Lyle was Gradually Exposed to Different Settings where He was Observed and Supervised Less.

Detection Methods Were Faded From Direct Observation to Pill Bottles With Residue and Then to Weekly Tox Screens.


Functional Communication Training (FCT)

Functional communication training (FCT) is a treatment commonly prescribed when a functional analysis has shown that an individual’s problem behavior is maintained by social consequences (e.g., Carr & Durand, 1985; Fisher et al., 1993; Horner, Day, Sprague, O’Brien, & Heathfield, 1991; Lalli, Casey, & Kates, 1995; Wacker et al., 1990).


Functional Communication Training (FCT)

With FCT, the individual is taught a communicative response that produces access to the reinforcer responsible for maintenance of the problem.

For example, an individual whose problem behavior is maintained by escape from tasks might be taught to request a break by signing ‘‘finished’’ (e.g., Hagopian, Fisher, Sullivan, Acquisto, & LeBlanc, 1998).


Unique Features of FCT

FCT is a DRA procedure that:1. specifies its reinforcer (i.e., a mand specifying the

reinforcer that previously maintained problem behavior),

2. requires minimal response effort,

3. is initially reinforced on a dense schedule (e.g., FR 1),

4. can recruit reinforcement across environmental contexts.



Because of the ease and consistency with which reinforcement can be obtained during FCT, some authors have suggested that the client ‘‘controls’’ the delivery of reinforcement (e.g., Carr & Durand, 1985).

In addition, Carr and Durand suggested that “control over reinforcement,’’ contributed to the effectiveness of FCT.


Unique Features of FCTTwo investigations found that noncontingent

reinforcement (NCR), which does not allow the client to control the schedule of reinforcement, and FCT, which does, produced equivalent reductions in problem behavior (Hanley, Piazza, Fisher, Contrucci, & Maglieri, 1997; Kahng et al., 1997).

Nevertheless, we found that participants preferred FCT over NCR when given a choice (Hanley et al., 1997).



FCT may promote generalization and maintenance because the communication response may prompt both trained and untrained caregivers to deliver differential reinforcement appropriately (e.g., Durand & Carr, 1991).


Limitations of FCT

Teaching the FCT response may evoke problem behavior (particularly if it is maintained by escape).

Individuals may display the FCT response at exceedingly high rates (e.g., requesting a break from every school task).

Individuals may request reinforcement at times when it is impossible or inconvenient to deliver (e.g., caregiver tending to an infant sibling).


Selecting the Functional Communication Response (FCR)

The FCR should be simple.The FCR should produce the reinforcer

identified during the functional analysis. The FCR should quickly remove the

establishing operation for problem behavior.


Training the Communication Response

The communication response matches the function of the child’s problem behavior.

1. Demand -> “Break please.”

2. Attention -> “Play with me, please.”

3. Tangible -> “Toy please.”


Training the Communication Response

Children who do not speak are often taught to use a picture-exchange communication response.

1. Attention -> Child hands over a picture of the adult and child playing together.

2. Demand -> Child hands over a picture of the child leaving a work table.


Function-Based Extinction

EXT (Att): Attention no longer follows the target behavior

EXT (Tang): Tangible item is longer presented following the target behavior

EXT (Esc): Demands continue following the target behavior

EXT (Auto): The sensory consequences of the target response are eliminated or the response is prevented.


Stimulus Control Refinements of FCT

Controlling the establishing operation for problem behavior is particularly important during the early stages of treatment.

Bringing the functional communication response under tight discriminative control is particularly important for increasing the practicality and generality of FCT.


Study 1: Does strict control of exposure to the establishing operation (EO) for problem

behavior facilitate the effectiveness of FCT?We compared FCT using a picture exchange (or card touch) as

the FCR with a vocal FCR.The picture exchange version of FCT (FCT-card) allows strict

control of the EO, because the therapist can immediately guide the FCR and deliver the reinforcer, which removes the EO.

The vocal version of FCT (FCT-vocal) does not allow strict control of the EO, because one cannot guide a vocal response.


Study 1: ProceduresFollowing baseline, the two FCT treatments (FCT-card

and FCT-vocal) were both introduced using a progressive time-delay procedure to teach each FCR.

The two treatments were compared using a multielement phase.

Following a return to baseline, the more effective treatment was re-introduced in the final phase.

t = 3.0; p = .01


Study 1: ConclusionsThe FCT-card card condition allowed better control of

the EO for problem behavior.The FCT-card condition produced more rapid

reductions in problem behavior. The FCT-vocal condition was associated with an

extinction burst for one participant, which was not observed in the FCT-card condition.


Controlling the EO While Training the Communication Response

When FCT is initiated, the communication response should always produce the reinforcer.

If the child cannot do the response independently, we help them and then immediately deliver the reinforcer.


Stimulus Control Refinements of FCT

Controlling the establishing operation for problem behavior is particularly important during the early stages of treatment.

Bringing the functional communication response under tight discriminative control is particularly important for increasing the practicality and generality of FCT.


Schedule Thinning During Functional Communication Training

Signaled and unsignaled delayed reinforcement (Vollmer et al., 1999)

Activities or alternative reinforcers during the reinforcement delays (Fisher et al., 1998; 2000)

Multiple schedules with reinforcement and extinction components (Betz et al., 2913; Fisher et al., 1998; Hanley et al., 2001)



Typical Multiple-Schedule Training During FCT

Quasi-random alternation between and FR-1 schedule and EXT for communication

Each component correlated with a specific signal

Initially, the duration of the reinforcement component is 3 to 4 times longer than the EXT component

Gradually, the EXT component is lengthened relative to the SR+ component

1

2

3

4

5

6

7

8

FAD

ING

STE

PS

= FR1

= Extinction

Decreases reinforcer deliveries by about 75%


Purpose

To identify the necessary and sufficient components for the reinforcement schedule thinning

1. Study 2: Evaluated the effects of contingency-correlated stimuli during FCT

2. Study 3: Evaluated the necessity of systematic and gradual fading steps during schedule thinning under multiple schedule components

3. Study 4: Evaluated the extent to which contingency-correlated stimuli facilitated generalization across therapists and environments


Functional Analysis and FCT Treatment Evaluations

Prior to Studies 2, 3, and 4, we conducted functional analyses to show that problem behavior was reinforced by social positive reinforcement for all participants.

We also conducted treatment evaluations using an ABAB design to show that FCT was an effective treatment.





Study 2: Are contingency-correlated stimuli necessary?

Compared rates of FCR and problem behavior under equivalent mixed and multiple schedules using an ABAB design

1.Mixed FR-1:60 s / EXT:60 s2.Multiple FR-1:60 s / EXT:60 s

Data on problem behavior is not shown because it remained at near-zero levels throughout this study.


Mixed vs. Multiple Assessment

General Procedures1. Sessions = 10 min

2. Contingency specifying stimuli in both conditions

3. FR1 = 60 s

4. Extinction = 60 s

5. All sessions started with reinforcement interval

6. All problem behavior was on extinction

FR1: 60 s

EXT:60 s

CASEY



Study 3: Is schedule fading necessary?

Compared rates of alternative and problem behavior when schedules were switched from rich to lean without schedule fading in a multiple baseline design

Mult FR-1:60 s / EXT:60 s to Mult FR-1:60 s / EXT:240 s

Mixed FR-1:60 s / EXT:60 s to Mult FR-1:60 s / EXT:240 s





Study 4: Do contingency-correlated stimuli facilitate generalization across therapists and

settings?Compared rates of alternative and problem behavior

when a multiple schedule was introduce across therapists or settings in a multiple baseline design

Mixed FR-1:60 s / EXT:60 s to Mult FR-1:60 s / EXT:60 s

Mult FR-1:60 s / EXT:60 s to Mult FR-1:60 s / EXT:300 s


University of Nebraska Medical Center




Conclusions

Study 2 showed that inclusion of contingency- correlated stimuli was a necessary component of FCT schedule thinning.

Study 3 showed that inclusion of contingency- correlated stimuli was sufficient to maintain appropriate responding without gradually fading the schedule density during FCT schedule thinning.

Study 4 showed that contingency-correlated stimuli also facilitated generalization of FCT effects across therapists and settings.


Study 5: How effective is Mult-FCT across a large cohort of patients with problem behavior?

Hagopian et al. (1998) found that FCT with EXT (FCTE) failed in 14 of 25 cases during schedule thinning.

They also found that FCT with punishment (FCTP) reduced destructive behavior by 90% or more in all cases, even following reinforcer-schedule thinning.

In the current study, we summarized the results of 14 applications of Mult-FCT implemented with 12 cases and compared the results with those of Hagopian et al.

212


Amount of Reinforcer-Schedule Thinning


Calculating Percentage Reduction in Problem Behavior

Percentage reductions in problem behavior were calculated using the following formula (same as in Hagopian et al. [1998]):

Mean rate during last 5 sessions 1 – Mean rate during baseline


Per

cent

Red

uctio

n in

Pro

blem

Beh

avio

r

0

10

20

30

40

50

60

70

80

90

100

110

n = 11applications

n = 17applications

n = 14applications

FCTE +Fading

FCTP +Fading

Mult-FCT +

Hagopian et al. (1998)Current

Data Set

Fading


Future Grants and Research on:Consequence Control Refinements of FCT

Using behavioral momentum theory (BMT) and accompanying equations to prevent relapse and resurgence of problem behavior.

BMT makes predictions that are at odds with current clinical “best practices”.


Limitations of FCT (cont.)

When a parent is busy and fails to deliver reinforcement for the child’s FCT response, the child is likely to revert to problem behavior, a phenomena called “resurgence”.

Recent research has shown that resurgence of problem behavior is quite common, and problem behavior often occurs at pre-treatment levels when the FCT response goes unreinforced (Mace et al., 2010; Volkert et al., 2009).


Mace et al., (2010)


Volkert et al., (2009)


Behavioral Momentum TheoryBehavioral momentum theory is a quantitative

theory of behavior that employs a comparative metaphor based on the classical mechanics of the momentum of physical objects.

Behavioral momentum theory is relevant to FCT because it provides quantitative models and predictions about the persistence and resurgence of problem behavior when the FCT produces reinforcement and when it does not.


The Momentum Metaphor

In classical mechanics, the momentum of an object is a joint function of its mass and velocity.

Momentum of a moving object =

mass x velocityMomentum of a response =

rate of reinforcement x response rate


Equation 7 Predicts the Effects of Adding and Removing Reinforcement for the FCT

Response


Common Practice Guidelines for FCT Decreases may occur more rapidly during

treatment if problem behavior is reinforced on a CRF schedule during baseline (Fisher & Bouxsein, 2011).

“… we strongly recommend that the communicative response be reinforced on a CRF schedule initially (Tiger, Hanley, & Bruzek, 2008).

Clinical guidelines on “dosage” levels of FCT are not available (i.e., How many sessions of FCT are needed before parent training and discharge?).


Specific Recommendations of BMT

The magnitude of resurgence predicted by Equation 7 is minimized by:

1. maximizing the value of t conducting many FCT sessions before exposure to an EXT challenge;

2. maximizing the value of d by correlating periods of reinforcement and EXT of the FCR with discriminative stimuli;


Specific Recommendations of BMT

The magnitude of resurgence predicted by Equation 7 is minimized by:

3. minimizing the value of r by delivering the lowest possible rate of reinforcement for destructive behavior during baseline; and

4. minimizing the value of Ra by delivering the lowest possible rate of reinforcement of the FCR during FCT.


0

50

100

150

200

250

300

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Des

truc

tive

Res

pons

es p

er H

our

Session

Baseline(VI 30)

FCT Extinction Challenge

c = 1, d = .001, p = .05

(VI 20)

Long Exposure(Room 1; Therapist = Purple)

Short Exposure(Room 2; Therapist = Yellow)


0

50

100

150

200

250

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Des

truc

tive

Res

pons

es p

er H

our

Session

Baseline FCT Extinction Challenge

c = 1, d = .001, p = .05

(VI 36 s)

VI 20 s(Room 1; Therapist = Red)

VI 120 s*(Room 2; Therapist = Blue)

*This lean schedule will be individually determined based on a progressive-interval assessment and may be different from a VI 120 s.


0

50

100

150

200

250

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Des

truc

tive

Res

pons

es p

er H

our

Session

Baseline (VI 36 s) FCT Extinction Challenge

c = 1, d = .001, p = .05

VI 20 s (Room 1; Therapist = Green)

VI 180 s(Room 2; Therapist = Blue)

PI


0

50

100

150

200

250

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Des

truc

tive

Res

pons

es p

er H

our

Sessions

Baseline FCT EXT Challenge

c = 1, d = .001, p = .05Long Exposure of Alt VI 120 s*

(Room 1; Therapist = Red)

Short Exposrue ofAlt VI 20 s

(Room 2; Therapist = Blue)

VI 20 s(Room 2; Ther = Blue)

VI 120 s*(Room 1; Ther = Red)

*Lean schedules to be individually determined based on PI assessments; they may be leaner or denser than VI 120 s.


Concluding Comments

Mathematical models like behavioral momentum theory provide a method for developing new and interesting clinical research questions that have the potential to greatly improve treatments like FCT.


UNMC Munroe-Meyer InstituteUniversity of Nebraska Medical CenterUniversity of Nebraska Medical Center

Documents

Behavioral Approaches to the Treatment of Autism and Severe Behavior Disorders