USF Libraries
USF Digital Collections

Pediatric healthcare providers' screening practices

MISSING IMAGE

Material Information

Title:
Pediatric healthcare providers' screening practices impact of training on early identification of autism
Physical Description:
Book
Language:
English
Creator:
Meyer, Aja M
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Pediatrician
Screening
Barriers
Early intervention
Training evaluation
Dissertations, Academic -- Interdisciplinary Education -- Specialist -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This study explored the effectiveness of the Autism System of Care (ASC) trainings by measuring change in pediatric healthcare providers' method of identifying young children at-risk for autism spectrum disorders. The majority of participants were pediatricians working in either hospitals or clinics who voluntarily participated in the training. A pretest-posttest nonequivalent-groups design was used in this study. Pre- and post-test questionnaires were used to measure change in participants' screening practices. Due to a small number of participants, most findings from the study were not statistically significant. The small number of healthcare providers who participated in the ASC training was a major limitation to this study. Therefore, although results revealed that there were minimal gains between pre- and post-test administrations, this may be because of the small number of participants and does not necessarily indicate that the ASC training was not effective. Implications for future research in this area also are addressed.
Thesis:
Thesis (Ed.S.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Aja M. Meyer.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 118 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001910193
oclc - 173276418
usfldc doi - E14-SFE0001687
usfldc handle - e14.1687
System ID:
SFS0026005:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001910193
003 fts
005 20070925124728.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 070925s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001687
040
FHM
c FHM
035
(OCoLC)173276418
049
FHMM
090
LB1555 (ONLINE)
1 100
Meyer, Aja M.
0 245
Pediatric healthcare providers' screening practices :
b impact of training on early identification of autism
h [electronic resource] /
by Aja M. Meyer.
260
[Tampa, Fla] :
University of South Florida,
2006.
3 520
ABSTRACT: This study explored the effectiveness of the Autism System of Care (ASC) trainings by measuring change in pediatric healthcare providers' method of identifying young children at-risk for autism spectrum disorders. The majority of participants were pediatricians working in either hospitals or clinics who voluntarily participated in the training. A pretest-posttest nonequivalent-groups design was used in this study. Pre- and post-test questionnaires were used to measure change in participants' screening practices. Due to a small number of participants, most findings from the study were not statistically significant. The small number of healthcare providers who participated in the ASC training was a major limitation to this study. Therefore, although results revealed that there were minimal gains between pre- and post-test administrations, this may be because of the small number of participants and does not necessarily indicate that the ASC training was not effective. Implications for future research in this area also are addressed.
502
Thesis (Ed.S.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 118 pages.
590
Adviser: Kelly A. Powell-Smith, Ph.D.
653
Pediatrician.
Screening.
Barriers.
Early intervention.
Training evaluation.
690
Dissertations, Academic
z USF
x Interdisciplinary Education
Specialist.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1687



PAGE 1

Pediatric Healthcare Providers’ Screeni ng Practices: Impact of Training on Early Identification of Autism by Aja M. Meyer A thesis submitted in partial fulfillment of the requirements for the degree of Education Specialist Department of Psychological and Social Foundations College of Education University of South Florida Major Professor: Kelly A. Powell-Smith, Ph.D. Kathy Bradley-Klug, Ph.D. Rose Iovannone, Ph.D. Anthony Onwuegbuzie, Ph.D. Date of Approval: April 17, 2006 Keywords: pediatrician, screening, barriers, early intervention training evaluation Copyright 2006, Aja M. Meyer

PAGE 2

Acknowledgements The experience of proposing and defending my thesis has provided me with tremendous growth opportunities both personally and acad emically. I have been continuously challenged throughout the process as I develope d into a competent researcher. However, persevering through the thesis process would have not been possible without the support of many wonderful people. Fi rst and foremost, I would lik e to thank my committee for supporting me throughout the stages of this proj ect. Their support, time, and interest in making this research project successful were invaluable. I also would like to thank the Autism System of Care Work Group for giving me the opportunity to work on an important piece of research that will have im plications for the early identification of Autism Spectrum Disorders through the use of screening instruments with young children. Specifically, I would like to thank Dr. Quimby McCaskill for offering his support throughout the process of conducting this study. Fina lly, I would like to thank Jayme Alberssen, Adam Chambers, Kyle Popka ve, Tara-Lynn Reidy, and my family for constantly reassuring me that there was light at the end of the tunnel.

PAGE 3

i Table of Contents List of Tables iv Abstract vi Chapter 1: Introduction 1 Statement of the Problem 1 Theoretical Framework 3 Purpose of the Study 4 Research Questions 5 Hypotheses 6 Significance of the Study 6 Definition of Terms 7 Organization of Remaining Chapters 8 Chapter 2: Review of the Related Literature 9 Overview 9 Autism Spectrum Disorders 9 Prevalence/Incidence 11 Symptoms/Indicators 13 Potential Causes 15 Importance of Early Identification a nd Intervention 16 Challenges to Early Identification an d Intervention 21 Screening Instruments and Procedures 23 Developmental Screening Instruments 24 Ages & Stages Questionnaire (ASQ) 24 Parents’ Evaluations of Developmental St atus (PEDS) 25 Communication and Sym bolic Behavior Scales Developmental Profile Infant Toddler Checklist (CSBS DP) 25 Autism-Specific Screening Instruments 27 Checklist for Autism in Toddlers (CHAT) 27 The Modified Checklist for Autism in Toddlers (M-CHAT) 28 Pervasive Developmental Disorder Screening Test (PDDST) 29 Summary 29 Practice Parameters 30 Barriers to Pediatric Healthcare Providers’ Use of Screening Instruments 36 Importance of Training for Providers in Id entifying Children with ASD 41

PAGE 4

ii Summary 43 Chapter 3: Methodology 46 Participants 46 Selection of Participants 47 Research Design 50 Instrument 51 Pediatric Healthcare Provider Self -Report Questionnaire 51 Procedures 55 Analyses 59 Pre-Test Analyses 59 Post-Test Analyses 60 Research Question 1 60 Research Question 2 61 Research Question 3 61 Research Question 4 62 Research Question 5 62 Research Question 6 62 Research Question 7 63 Chapter 4: Results 64 Treatment of the Data 64 Missing Data Analysis 64 Pre-Test Analyses 65 Exploratory Factor Analyses 69 Score Reliability of Pre-Test Measures 72 Assessing Group Equivalence 73 Urban versus Rural 73 Experimental versus Control 76 Check of Normality Assumptions for Post-Test Scores 78 Post-Test Analyses 81 Score Reliability of Measures 81 Chapter 5: Discussion 89 Summary of Study 89 Summary of Results 89 Notable Findings from the measures 90 Implementation Integrity 92 Effectiveness of the interven tion 93 Implications of the Results 93 Limitations 95 Considerations for Future Research 99 Final Thoughts 100 References 102

PAGE 5

iii Appendices 111 Appendix A: Title 112 Appendix B: Title 115 Appendix C: Title 117

PAGE 6

iv List of Tables Table 1 Descriptive Statis tics for Pre-Test and Post-Test Data 66 Table 2 Demographi cs Characteristics of Sample at Pre-Test 68 Table 3 Exploratory Factor Analys is for General Knowledge Scale Items (Pre-Test) 70 Table 4 Exploratory Factor Anal ysis for Screening Scale Items (Pre-Test) 71 Table 5 Exploratory Factor Analysis for Potential Barriers Items (Pre-Test) 72 Table 6 Score Reliabilities (Cronbach’s Alpha) for all Measures by Treatment Group: Pre-Test 73 Table 7 Skewness and Kurtosis Coeffi cients for Pre-Test Scales: Urban Group 88 Table 8 Skewness and Kurtosis Coeffi cients for Pre-Test Scales: Rural Group 89 Table 9 T-Tests Comparing Participan ts’ Scores Based on Geographic Region: Pre-Test 90 Table 10 Skewness and Kurtosis Coeffi cients for Pre-Test Scales: Experimental Group 77 Table 11 Skewness and Kurtosis Coeffi cients for Pre-Test Scales: Control Group 78 Table 12 T-Tests Comparing Particip ants’ Scores Based on Treatment Group: Pre-Test 78 Table 13 Skewness and Kurtosis Coeffi cients for Post-Test Scales: Experimental Group 80 Table 14 Skewness and Kurtosis Coeffi cients for Post-Test Scales: Control Group 80

PAGE 7

v Table 15 Score Reliabili ties (Cronbach’s Alpha) for all Measures by Treatment Group: Post-Test 81 Table 16 Wilcoxon Test for Screening Patterns: Age of Patients Scale Scores: Preand Post-Test 83 Table 17 Spearman Rank Correlation Coeffici ents: Pre-Test 88 Table 18 Spearman Rank Correlation Coefficients: Post-Test 88

PAGE 8

vi Pediatric Healthcare Providers’ Screeni ng Practices: Impact of Training on Early Identification of Autism Aja M. Meyer ABSTRACT This study explored the effectiveness of the Autism System of Care (ASC) trainings by measuring change in pediatric healthcare providers’ me thod of identifying young children at-risk for autism spectrum disord ers. The majority of participants were pediatricians working in either hospitals or clinics who voluntarily participated in the training. A pretest-posttest nonequivalent-groups design was used in this study. Preand post-test questionnaires were us ed to measure change in part icipants’ screening practices. Due to a small number of participants, most findings from the study we re not statistically significant. The small number of healthcare providers who participated in the ASC training was a major limitation to this study. Th erefore, although results revealed that there were minimal gains between preand pos t-test administrations, this may be because of the small number of part icipants and does not necessa rily indicate that the ASC training was not effective. Implications for futu re research in this area also are addressed.

PAGE 9

1 Chapter 1 Introduction Statement of the Problem Autism is a lifelong developmental disabil ity that affects the functioning of the brain and typically appears dur ing the first three years of life. Autism falls under the category of Autism Spectrum Disorders (ASD), which refers to the broad continuum of cognitive and neurobehavioral difficulties present in th ese individuals (American Psychiatric Association, 2000). The fundamental features of autism are the presence of markedly abnormal or impaired development in communication and so cial interaction, as well as a distinctly restricted repertoire of behaviors and interests (Chakrabarti & Fombonne, 2001; Klinger, Dawson, & Renner, 2003; Oser & Shaw, 2001). Appearance of the disorder varies greatly depending on the developmental level and chronological age of the individual (Cha krabarti & Fombonne, 2001). The occurrence of autism is thought to be on the rise, with the latest st udies finding higher rates than what was found in studies conducted in the 1980s and ear ly 1990s. Earlier studies found that approximately 4 per 10,000 children had auti sm, while a study in 1998 found that 40 per 10,000 children have autistic disorder, with the number increasing to 67 per 10,000 if all types of autism-like behaviors are in cluded (Yeargin-Allsopp et al., 2003). Although the exact cause of autism sp ectrum disorders is still unknown, the literature reports that children with au tism spectrum disorders (ASD) show more significant gains when they receive supports a nd services early on in their development.

PAGE 10

2 However, many children are not being identifi ed as early as possible to obtain the benefits of early intervention (Baird et al., 2000; Scambler, Rogers, & Wehner, 2001). It is estimated that only 50% of children w ith ASD are diagnosed before kindergarten (Strock, 2004). The diagnosis of ASD may be delayed due to concer ns about labeling a child or incorrectly diagnosing a child (Filipek et al., 2000; Oser & Shaw, 2001). Although approximately 25% of children in primary care practice have developmental delays, less than 30% of prim ary care providers routinely c onduct screening tests at wellchild visits (Dworkin, 1989; Filipek et al., 2000). Research indicates that early diagnosis is associated with dramatically better outcomes for individuals with autism b ecause an accurate diagnosis and early identification can provide the basis for buildi ng an appropriate and effective educational and treatment program. In addition, early in tervention facilitates earlier educational planning, provisions for family supports and ed ucation, management of family stress, and the distribution of appropriate medical car e (Filipek et al., 2000). Because early educational intervention is the key to he lping children with autism develop into competent and productive adults, routine earl y screenings of child ren are imperative so that they receive the various services needed in a timely manner. While no one behavioral or communications test can detect autism, several screening instruments such as the Parents’ Evaluations of Developmental St atus (PEDS), the Checklist for Autism in Toddlers (CHAT), and the Pervasive Developm ental Disorder Screening Test (PDDST) have been developed that are now used to identify young children w ho may be at-risk for ASD (Prater & Zylstra, 2002).

PAGE 11

3 Given that typically developing children demonstrate eye contact, orienting to one’s name, joint attention, pretend pl ay, imitation, nonverbal communication, and language development by 18 months of age, experienced professionals can reliably diagnose autism in children as young as 18 months of age (Filipek et al., 1999). In addition, autism-specific screening instrume nts have been developed for use with children at 18 months of age (e.g., CHAT). Pediatricians generall y see young children on a regular basis throughout the first two years of life; therefore, they typically are involved in screening, identifying, and referring patie nts who are suspected of having an ASD for further evaluation. Unfortunately, pediatric healthcare providers perceive a number of barriers to the utilization of screening instruments with young children. Several frequently reported barriers include providers’ unfamiliarity wi th the early warning signs of autism, inadequate time to perform developmental scre enings during typical well-child visits, and unfamiliarity with screening instruments (Halfo n et al., 2001). Therefore, it is imperative that pediatric healthcare provi ders’ knowledge-base of screening instruments and ASD be improved (Filipek et al., 1999). Professionals need to be kno wledgeable about the early symptoms of autism as well as the available, score-validated screening instruments so that appropriate screening and re ferral procedures may occur. Theoretical Framework To be most successful in identify ing young children with autism spectrum disorders, it is important to use an ecological model of child development, such as Urie Bronfenbrenner’s framework, which takes into account biological sociological, and psychological domains (Sontag, 1996). When us ing an ecological model, a variety of

PAGE 12

4 measures are utilized in assessing the diso rder. From a developmen tal perspective, the disorder is viewed within a conceptual fr amework that considers the expectations of children at particular ages. Utilizing the eco logical model, the pediatric healthcare provider obtains a developmental hist ory, a medical evaluation, behavioral observation(s), and information related to c ognitive functioning and language ability to identify children at-risk for ASD. In addition, when making decisions th at will impact children’s continued development, it is of the utmost importan ce to utilize data-based decision making. The general steps used in data-bas ed decision making are: (a) es tablish a team, (b) develop a hypothesis, (c) gather data to assess needs, (d) use data to formulate goals, (e) develop a data-based plan, and (f) monitor progress and document success (Yang & Goldstein, 1999). When pediatric healthcare providers ut ilize an ecological framework to enhance their understanding of child development a nd employ data-based decision making, their young patients are more likely to receive the early intervention supports and services they need to maximize their development (Filipek et al., 2000). Purpose of the Study Although a great deal of research supports the notion that earl y identification of autism spectrum disorders leads to better outco mes, a large number of children with ASD still are not identified as early as possible. Furthermore, the recent increase in the number of individuals diagnosed with ASD heightens the importance of early identification. To this end, this study attempted to discover the effectiveness of the Au tism System of Care (ASC) trainings by measuring change in pediatric healthcare providers’ method of

PAGE 13

5 identifying young children at-risk for autism spectrum disorders. Preand post-test questionnaires were used to measure change in participants’ screening practices. Research Questions The following research questions were addressed in this study: 1. What is the effect of the Autism System of Care (ASC) training on use of developmental and autism-specific screen ing instruments by pediatric healthcare providers? 2. What is the effect of the ASC training on the use of developmental screening instruments in regard to age of patient? 3. What is the effect of the Autism System of Care training on pediatric healthcare providers’ perceived barrier s to increasing the use of screen ing instruments and/or referring patients? 4. What is the effect of the Autism System of Care training on pediatric healthcare providers’ perceived levels of knowle dge related to Autism Spectrum Disorders? 5. What is the effect of the Autism System of Care training on the self-efficacy of pediatric healthcare providers regarding th e ability to screen accurately and refer a child suspected of having an Autism Spectrum Disorder? 6. What is the relationship between pediatric healthcare providers’ perc eived barriers to utilizing screening instruments a nd their actual use of developmental and autism-specific screening instruments be fore and after completion of the training? 7. What is the relationship between percei ved barriers to util izing screening instruments and the use of developmental screening instruments in regard to age of patients before and af ter completion of the training?

PAGE 14

6 Hypotheses The following research hypotheses were tested in this study: (a) Autism System of Care (ASC) training increases pediatri c healthcare providers ’ routine use of developmental screening instruments and auti sm-specific screening instruments, (b) ASC training increases pediatric healthcare provide rs’ routine use of de velopmental screening instruments with patients at younger ages than the ages of patients at screening prior to completion of the training, (c) ASC training decreases pediatric healthcare providers’ perceived barriers to the use of screening in struments and/or referring patients, (d) ASC training increases pediatric he althcare providers’ ge neral knowledge related to ASD (e.g., early warning signs and scor e validated screening instruments), (e) ASC training increases pediatric healthcare providers’ percei ved self-efficacy regarding their ability to screen and refer children suspected of AS D, (f) ASC training decreases pediatric healthcare providers’ perceived barriers to utilizing screening instruments while increasing their use of developmental and auti sm-specific screening instruments, and (g) ASC training decreases pediatric healthcare providers’ perceived barriers to utilizing screening instruments while increasing their use of developmental screening instruments with patients at younger ages than the typical age at screening prior to completion of the training. Significance of the Study This study provides valuable information about the effectiven ess of the Autism System of Care trainings in changing pediat ric healthcare providers’ method of the early identification of children at-risk for ASD. Becau se the benefits of early intervention have been well documented in the literature, the early identification of ASD is crucial for

PAGE 15

7 optimal outcomes for these children (Filipek et al., 1999). Young children with ASD and their families will benefit greatly from early intervention services, and pediatric healthcare providers play a critical role in th e early identification of these disorders. The Autism System of Care trainings also may pl ay a significant role in enabling pediatric healthcare providers to identify children w ith ASD early in their development. Definition of Terms Autism Spectrum Disorders Autism Spectrum Disorders (ASD) also are known as Pervasive Developmental Disorders (PDDs). Th ese disorders are typi cally diagnosed in early childhood and cause pervasive impairme nt in thinking, feeling, language, and the ability to relate to others (Strock, 2004). Th ere are five disorders, each with different levels of severity, that fall under ASD: (a) autis tic disorder (a severe form), (b) pervasive development disorder not otherwise specified (PDD-NOS) (a mild form), (c) Asperger syndrome (a milder form), (d) Rett syndrome (a rare, severe form affecting females), and (e) childhood disintegrative disorder (a rare, seve re form) (Strock, 2004). Early identification Early identification refers to th e detection of ASD and/or other disabilities early on in children’s de velopment (Filipek et al., 2000). Evaluation An evaluation is the process of determ ining whether an individual is eligible for early intervention or special e ducation services (Oser & Shaw, 2001). Screening A screening is a brief, point-in-time procedure for deciding which individuals need a referral for further assessment (Oser & Shaw, 2001). Organization of Remaining Chapters The remaining chapters present informati on that is pertinent to this study. More specifically, Chapter 2 provides a thorough revi ew of the related l iterature, discussing

PAGE 16

8 ASD and the process of identifi cation, screening, a nd diagnosis. Furtherm ore, the role of pediatric healthcare providers in the early identification of ASD is reviewed, including a discussion of the perceived barriers to early identification and the utilization of screening instruments. Chapter 2 concludes with a disc ussion of the importance of training to facilitate change in service delivery for pe diatric healthcare providers so they are better able to identify young children with ASD. Chapter 3 details the methodology that was used in this study, including sampling, instru mentation, procedures, and data analysis.

PAGE 17

9 Chapter 2 Review of the Related Literature Overview This chapter provides a review of the literature relevant to this study. Autism spectrum disorders (ASD) are discussed, including the prevalence/incidence, symptomatology, and potential causes. The importance of early identification and intervention is discussed, as well as the screening and identification processes for ASD, including a review of screening instrument s and procedures. The role of pediatric healthcare providers in this process is presen ted, and both supports for and barriers to the developmental screening process are presented. This chapter concludes with a discussion of the importance of training pediatric h ealthcare providers in relation to changing practices effectively, thereby better enabling practitioners to identify children with ASD as early as possible. Autism Spectrum Disorders Autism, a complex neurodevelopmental disorder that affects the functioning of the brain, is the most prevalent disorder that falls under the category of “Autism Spectrum Disorder” (ASD). Autism is considered a spectrum disorder because the symptoms and characteristics can present them selves in a wide variety of combinations, from mild to severe. Although ASD is defined by a certain set of behaviors, individuals can exhibit any combination of the behaviors in any degree of severity. The diagnostic category of ASD includes five disorders with different levels of severity: (a) autistic

PAGE 18

10 disorder, (b) pervasive developmental disord er--not otherwise specified (PDD-NOS), (c) Asperger syndrome, (d) childhood disintegrati ve disorder (CDD), and (e) Rett syndrome (American Psychiatric Association [APA], 2000). In the Diagnos tic and Statistical Manual of Mental Disorders, Fourth Ed ition (Text Revision) (DSM-IV-TR), “autistic disorder” is listed under the heading of “P ervasive Developmental Disorders” (APA, 2000). Autistic disorder is diagnosed when an indi vidual displays a tota l of 6 or more of 12 symptoms listed across three major area s: social interac tion, communication, and behavior (APA, 2000). Specifically at least two symptoms mu st fall under the category of qualitative impairment in social interaction, such as marked impairment in the use of multiple nonverbal behaviors and/or lack of so cial or emotional reciprocity. At least one symptom must fall under the category of qua litative impairments in communication, such as marked impairment in the ability to initiate or sustain a conversati on with others and/or lack of varied, spontaneous make-believe pl ay. Finally, at least one symptom must fall under the section of restricted, repetitive, and stereotyped patt erns of behavior, interests, and activities, such as appare ntly inflexible adherence to specific, nonfuncti onal routines or rituals and/or st ereotyped and repetitive motor manne risms (APA). In addition, there must be delays or abnormal functioning in at least one of the above mentioned areas (social interaction, la nguage as used in social communication, and symbolic or imaginative play) with onset prior to age 3 years. A diagnosis of Pervasive Developmental Disorder--Not Otherwise Spec ified (PDD-NOS) is given when children display similar behaviors but do not meet th e criteria for autistic disorder (APA, 2000).

PAGE 19

11 This study will focus on two of the disorders, autism and PDD-NOS, because they are two of the more prevalent disorders under the diagnostic categ ory of ASD (Oser & Shaw, 2001). In addition, autistic disorder and PDD-NOS have symptomatology that allow for earlier identif ication and intervention (Oser & Shaw, 2001). For the purpose of this review of the literature, the author will use the term “autism” to encompass both autistic disorder and PDD-NOS. Additionall y, the term “Autism Spectrum Disorder” (ASD) will be used in place of “Pervasive Developmental Disorder” because it is considered to describe more fully th e continuum of symp toms presented by young children (Oser & Shaw, 2001). Prevalence/Incidence It is important to differentiate preval ence from incidence when discussing the increase in the reported cases of autism. Prev alence refers to the pr oportion of individuals in a population who suffer from a defined di sorder, whereas incidence refers to the number of new cases occurring in a populat ion over a period of time (Fombonne, 2003). It should be noted that bo th prevalence and incidence estimates will be inflated when the definition of ASD is broadened and diagnos tic instruments are improved. Two recent studies ( Kaye et al., 2001; Powell et al., 2000 ) have provided incidence estimates that showed an increasing trend over a brief peri od of time; however, neither study examined changes in diagnostic criteria or sensitivity of case detection procedures during this time period (Fombonne, 2003). Therefore, the recent in crease in rates of pr evalence cannot be directly attributed to an increase in incidence of ASD. Further research is needed to test hypotheses accurately on changes in the incidence and prevalence of ASD ( Fombonne, 2003 ).

PAGE 20

12 Although autism was once thought to be a fair ly rare disorder, it is more prevalent in the pediatric population than cancer, di abetes, spina bifida, and Down syndrome (Filipek et al., 1999). The apparent increase in the incidence and prevalence of autism spectrum disorders has led to increased c oncern about the disorder (Chakrabarti & Fombonne, 2001; Yeargin-Allsopp et al., 2003). Early studies conducted by Lotter (1966) and Wing and Gould (1979) found that approx imately 4 per 10,000 children had autism, while a study by Bertrand et al. (2001) found that 40 per 10,000 ch ildren had autistic disorder, with the number increasing to 67 pe r 10,000 if all types of autism-like behaviors are included. Similarly, Bair d et al. (2000) found a rate of autism of 30.8 cases per 10,000; however, the rate increased to 57.9 cases per 10,000 for all autism spectrum disorders. A number of recent studies have examined the prevalence of ASD, with considerable variability in thei r results. For instance, Filipe k et al. (2000) estimated that ASD occur at a rate of 20 in 10,000 children, whereas Chakrabarti and Fombonne (2001) reported that ASD are estimated to occur in as many as 60 in 10,000 individuals. Gillberg and Wing (1999) in their meta-analysis found an increase in prevalence, from 4.7 cases per 10,000 in children bor n prior to 1970, to 11.2 cases per 10,000 in children born in 1970 and later. Although the rise in the numbe r of individuals diagnosed with autism is supported by the l iterature, it is still unclear whether the increase in autism is strictly due to an increase in prevalence, or if the increase re flects improved awareness and diagnostic instruments available for ASD (Klinger, Dawson, & Renner, 2003). The observation of these noticeably increasing pr evalence rates supports the necessity for improved early screening and diagnostic procedures (Filip ek et al., 1999).

PAGE 21

13 Autism is approximately four times more prevalent in males than in females, with a male/female ratio of 4.3:1 (Fombonne, 2003). However, the ratio appears to vary with IQ, ranging from 2:1 in those with severe dys function to more than 4:1 in those with average IQ scores (Filipek et al., 1999). Ther e are no significant differences in prevalence or symptomatology of ASD when comparing diverse racial, ethnic, and social groups. Furthermore, socioeconomic factors, lifesty le choices, and educational levels do not appear to affect the chances of ASD occurrence, making it an equal-opportunity disorder (Fombonne, 2003). Autism is considered a unive rsal disorder, as studies throughout the world have reported consistent sympto matology, intellectual functioning, gender differences, and socioeconomic factor s (Fombonne, 2003; Klinger et al., 2003). Symptoms/Indicators Autism is characterized by pervasive im pairment in thinking, feeling, language, and the ability to relate to others. More specifically, impairments in reciprocal communication skills, atypical language develo pment, and a restricted and repetitive range of behaviors are commonly present. It is unclear whether these different areas of development are intrinsically linked, whereby im pairment in one area leads to difficulties in other areas. However, Klinger et al. (2003) report that it is pr obable that a group of deficits, rather than one primary deficit, affect these areas of development in young children. The social symptoms that are commonly impaired in ASD include the ability to share attention with another individual, to understand another person’s emotions (this concept is termed “theory of mind” in the l iterature), and to enga ge in pretend play. Because the development of early social abil ities is considered to be a precursor to

PAGE 22

14 language development, children with autism al so tend to experience a significant delay in this area (Klinger et al., 2003). As both verb al and nonverbal communication skills may be impaired by autism, greater understandi ng is needed of both normal and abnormal development in this area (Bri stol-Power & Spinella, 1999). Wetherby et al. (2004) examined warning si gns of ASD in the second year of life. They found that young children with ASD are lik ely to be delayed in using words and their vocalizations are likely to lack consonants and to ha ve atypical prosody. In addition, children with ASD are not likely to respond to their name or to instructions even with contextual cues (Wetherby et al.). These childr en are likely to be de layed in using objects conventionally in play and also are likely to display repetitive m ovements with their body and/or objects. Moreover, young children w ith ASD are typically delayed in sharing attention with eye gaze, sharing affect, and dr awing others’ attention to objects or events of interest (Wetherby et al.) Additionally, gestures of poi nting and showing, and a lack of coordination of gestures with eye gaze, facial expression, or vocalizations is evident in children with ASD. However, it is important to note that some of these warning signs also are seen in children with de velopmental delay (Wetherby et al.). Numerous studies have demonstrated defi cits in joint attention skills of children with ASD. These deficits include difficultie s using eye gaze to coordinate attention, following the attentional focus of another pers on, and drawing anothe r’s attention to an object or event of intere st (Mundy, Sigman, & Kasari, 1990; Stone, Ousley, Yoder, Hogan, & Hepburn, 1997; Wetherby, Prizant, & Hutchinson, 1998). Longitudinal research findings suggest that the failure to acquire gestural joint a ttention may be a core

PAGE 23

15 deficit in ASD and a critical milestone th at impairs language development (Mundy et al., 1990; Sigman et al., 1999). Repetitive behaviors are commonly seen in children with autism, and these behaviors typically fall into one of two categories. The fi rst category comprises lowerlevel behaviors that present repetitive motor movements; the other category consists of higher-level behaviors in whic h an individual is insistent on following a specific routine or holds a very narrow range of interest s (Turner, 1999). Several other behavioral symptoms are often related to autism. Self-inj urious behavior, such as head banging, hair pulling, and hand biting, is typically seen in lower-functioning individuals with autism. In addition, sleep disturbance, eati ng disturbance, and excessive anxiety also can occur with autism (Klinger et al., 2003). Potential Causes Currently, the etiology of autism spect rum disorders is unknown. Therefore, interventions are structured to reduce the interfering symptoms of ASD. Clinicians initially believed autism was caused by cold, rejecting parents from wealthy families. In particular, mothers were often blamed for th e child’s condition; therefore, the term “Refrigerator Mom” was used to describe these mothers (Bettelheim, 1967). However, this notion does not hold merit in the current literature. Autism was once viewed as a psychogenic disorder; however, compelling evidence now suggests that autism is a disorder of abnormal brain development that is largely genetic. A number of family and twin studies has revealed th at genetic factors play a ro le in the occurrence of ASD (Rutter, 2000).

PAGE 24

16 Nicolson and Szatmari (2003) reviewed th e findings from a number of genetic and brain-imaging studies of autism ove r the past 15 years. The findings were synthesized, and overwhelming evidence was f ound to support a neurobiological basis for autism. The risk to siblings of children with autism is approximately 50 to 100 times greater than the risk to the general population. However, th ese statistics only provide evidence that the disorder runs in families. To determine whether the basis of the familial aggregation is environmental or genetic, twin studies must be conducted. Several twin studies have revealed much higher conc ordance rates for monozygous than dyzygous twins. These findings indicate the presence of si gnificant genetic factors, with heritability estimates greater than 90%, which make AS D the most heritable of the psychiatric disorders (Szatmari, Jones, Zwaigenbaum, & MacLean, 1998). Nicolson and Szatmari (2003) concluded that the likely cause of autis m is a genetic defect in the control of neurodevelopment, resulting in structural and functiona l changes predisposing an individual to autism. Although evidence is co ntinuing to accumulate for an underlying genetic cause for ASD, more research must be conducted in order to determine its etiology. Given that there is currently no biological marker for ASD, screening and diagnosis must be based on behavioral featur es (Filipek et al., 1999). The consistent use of screening instruments that yield valid info rmation for the detecti on of children at risk for ASD likely will lead to earlier and improved interventions for children with ASD (Filipek et al., 1999). Importance of Early Identif ication and Intervention The early identification of autism spectrum disorders leads to better gains for these children if suppor ts and services are initiated ear ly on in development. Although

PAGE 25

17 substantial literature provides support for th e positive effects of ear ly identification, many children with ASD are not identified nor suppor ted as early as possi ble to benefit from early intervention services (Oser & Shaw, 2001). Profession als such as developmental pediatricians, child neurologists, and child psychiatrists are ty pically knowledgeable about ASD and have experience working w ith children who have these disorders. Therefore, these clinicians are frequently involved in assessing, diagnosing, and treating children with ASD (Oser & Shaw, 2001). Evidence is growing that demonstrates the effectiveness of intensive early intervention with a signif icant proportion of young children with ASD (Dawson & Osterling, 1997; Filipek et al., 2000; Oser & Shaw, 2001). Dawson and Osterling (1997) reviewed eight model preschool intervention pr ograms for children with autism that have been operating since the 1980’s. The findings suggest that many children with autism who receive early interventi on services make significant developmental gains. These gains were measured by the programs in a vari ety of ways (e.g., IQ scores, developmental scores on standardized tests, observational m easures taken in the classroom). Because of the variation in measures used, it is difficult to compare the outcomes of these different programs; therefore, a genera l analysis of the overall progr ess of the 150 children in the early intervention programs was completed (Dawson & Osterling, 1997). All of the programs were effective in fostering significan t developmental gains, as well as positive school placements (e.g., these child ren are frequently able to be included in general education classrooms by the time they begin elementary school). Dawson and Osterling (1997) discovered that as long as certain fundamental program features are present, children tend to have favorable outcomes regardless of the

PAGE 26

18 specific philosophy of the intervention progr am. Although the majority of children with autism who receive early intervention services make gains, it is still unclear whether the rate of progress is related to child charact eristics such as IQ and language ability (Dawson & Osterling, 1997). Children from all eight preschool programs made, on average, an IQ gain of approximately 20 points. Although the majo rity of the children participating in the program had an IQ score in the mental retardation range (< 70) at the beginning of the program, most of the children res ponded positively to early intervention, making considerable progress. Dawson and Osterling (1997) concluded that further research must be conducted to determine wh ether one intervention approach is more effective than another, and to ascertain th e most appropriate early intervention program intensity level. The contention that early experience is important for promoting the most favorable long-term outcomes for children w ith developmental disabilities has been supported by studies of behavi oral outcomes and early inte rvention in various at-risk populations (Dawson, Ashman, & Carver, 2000). The growing literature in the area of biological research indicate s brain development begins prenatally and continues throughout the first few years of life. This information suggests that there may be a “sensitive period” whereby early intervention se rvices would have a significant impact on behavior outcomes for children with ASD (Daw son et al., 2000). As research and policy have emphasized the significance of early experience in the development of young children, new techniques for studying infant behavior and brain activity have been developed. These latest procedures have al lowed researchers to learn more about the relationship between biology and behavior in infants and young children. Early

PAGE 27

19 development in children consists of many f undamental “experience expectant” processes, whereby children are anticipated to meet certain milestones (Dawson et al., 2000, p. 706). However, genetic or acquired brain abnormalities in ASD preclude these children from obtaining normal experiences in otherwise nor mal environments (Dawson et al., 2000). Dawson et al. (2000) concluded that b ecause the prenatal and early postnatal years represent a sensitive period with respect to the long-term benefi cial effects of early intervention on brain and behavior al development, increased e fforts at early identification are needed. Although prevention and early intervention effort s should not focus only on the earliest years of development, it is appa rent from the extensive research that these efforts should begin as early as possible. B ecause long-term negative consequences have their greatest influences dur ing early development, with the promotion of optimal prenatal and infant–toddler development, th ese negative consequences can be minimized or avoided completely. In a ddition, greater public awareness and education of healthcare providers in regard to the ear ly detection of developmenta l disorders and how to access appropriate interventions are n eeded. Providers need to be pr oficient in the identification of early symptoms of autism so that appropr iate screening and re ferral procedures can occur. Research indicates that intervention provi ded before 3.5 years of age has a greater impact than interventions begun after five years of age (Filipek et al., 2000; Harris & Handleman, 2000). Harris and Handleman (2000) conducted a study examining the predictive power of age and IQ at the begi nning of an early intervention program using applied behavior analysis. The children who pa rticipated in the intervention program at the Douglass Developmental Disabilities Center were examined in a 4to 6-year follow-

PAGE 28

20 up after they left the preschool. At the st art of the program, 27 children with autism between the ages of 31 and 65 months had IQ scores between 35 and 109 on the Stanford Binet. Harris and Handleman found that children with both higher IQ scores ( M = 78) and younger age at intake ( M = 42 months) were predictive of being in a general education class after completion of the program. Children who had lower IQ scores ( M = 46) and were older at intake ( M = 54 months) were strongly associated with placement in special education classrooms. These results support th e necessity for early intervention services for children with ASD. However, Harris a nd Handleman emphasized that both children with lower IQ scores and olde r children also showed measurab le gains in IQ scores from treatment. Harris and Handleman concluded that although receiving in tervention services at a very young age is most beneficial, olde r children also respond quite favorably to intervention services. Research on social communication ha s important implications for earlier identification and intervention in young child ren with ASD because the skill deficits identified are skills that typically develop during the first 12 to 18 months of life. These findings suggest that there may be a set of pre-linguistic behaviors (e.g., gaze/point following, shared affect, gestures, communicativ e vocalizations, symbo lic play) that are important early indicators of ASD. These beha viors also may help to distinguish children with ASD from both typically developing chil dren and children with other developmental delays (Wetherby et al., 2004). The substantial effect of early intervention has been dramatically demonstrated in the case of autism spectrum disorders. If inte nsive behavioral interv entions are initiated by 2 years of age, a substantial number of children with autism show remarkable

PAGE 29

21 improvements in their development (Dawson & Osterling, 1997). These findings suggest there is an urgent need to improve early iden tification so that children with ASD are able to access interventions as early as possible (Wetherby et al., 2004). The Individuals with Disabilities Edu cation Act (IDEA) was created to ensure that young children with disabili ties receive early supports an d services. IDEA is a law that guarantees all children with disabilities access to a free and appropriate public education. However, according to the 23rd annual report to congress on the implementation of the IDEA (U.S Department of Education, 2001) young children with developmental delays, includi ng those with ASD, appear to be under-identified and underserved. In the United States from 1999 to 2000, approximately 1.8% of children under the age of 3 year s received early inte rvention services under the Individuals with Disabilities Education Act (IDEA) Part C; however an estimated 5% of preschoolers were served under Part B of IDEA. These data indicate that a cons iderable proportion of children under the age of 3 years with de velopmental delays such as ASD are not identified or fail to receive early intervention services. Challenges to Early Identif ication and Intervention The National Early Childhood Technical Assistance System (NECTAS) has assisted states in identifying and addre ssing the challenges related to the early identification of children with ASD, incl uding the importance of building the knowledge base on effective practices (Oser & Shaw, 2001). To attend to these challenges, the NECTAS Forum on ASD was created. This group of po licy-makers identified national issues and promising practices in state early intervention and presc hool special education systems. Through the use of focus groups, conference calls, web-based discussion

PAGE 30

22 forums, and survey research, NECTAS iden tified promising practi ces in state early intervention and preschool special educa tion systems (Oser & Shaw, 2001). In 1999, a survey to identify challenges to the early identification of ASD was mailed to 126 statelevel policy makers. Thirty-fiv e coordinators responded (27. 8%), identifying challenges such as developing policies for public awaren ess and early intervention, involving parents in the identification process, and providing information to parents regarding the process of evaluations. The lack of appropriate tool s and techniques available to identify young children with ASD was reported to be a challenge to early identification as well. Information derived from the NECTAS Forum on ASD activities will aid in the development of future strategies in early intervention and preschool special education systems (Oser & Shaw, 2001). The NECTAS Forum on ASD discussed specific challe nges and strategies for earlier identification, including (a ) raising public and profession al awareness, (b) tools for screening, (c) determining eligibility for services, and (d) transition. Raising public and professional awareness involves increasing the awaren ess of warning signs of ASD among primary healthcare providers as we ll as the public. This awareness can be accomplished by developing an early identification campaign that includes ASD, providing resources and traini ng for primary healthcare provi ders as well as recent practice parameters, and extending awareness e fforts to include places such as schools, child care centers, and child welfare agencies. Tools for screening refers to the use of a multi-stage process for early identification, the routine screening for early language development, and the distribution of info rmation on early warning signs for ASD to primary referral sources. The NECTAS Forum on ASD also discussed the importance of

PAGE 31

23 awareness and training regarding screening tools (e.g., Checklist for Autism in Toddlers [CHAT]), and the disseminati on of information about screen ing instruments available for milder disorders in the spectrum. Determining eligibility for services refers to the development of guidelines for evaluation a nd assessment procedures. In addition, the NECTAS Forum on ASD recommended more fre quent re-evaluations and follow-up of children with ASD, with child ren diagnosed with PDD-NOS being re-evaluated before the age of 3 years. Finally, transition refers to the planning of transitions (e.g., from early intervention program into preschool classr oom) as soon as possible, and collectively addressing assessment and evaluation issues among various personnel (e.g., Part C and Part B of IDEA) (Oser & Shaw, 2001). Screening Instruments and Procedures Developmental screening is intended to identify young child ren who may need more comprehensive evaluations to asse ss their development; therefore, it is recommended by the American Academy of Pediatrics (AAP) that all infants and children are screened for deve lopmental delays or disabilities (AAP, 2001). The use of developmental screening instruments is an e fficient way to record observations and help providers identify more children with deve lopmental delays (AAP). Some research suggests that although a number of screening tools are availa ble for identifying ASD in young children, the disorders may often rema in unrecognized and undiagnosed because suitable tools for routine developmental screening and autism-specific screening remain unavailable (Filipek et al., 1999). However, the National Research C ouncil Report on Educa ting Children with Autism (2001) reviewed severa l screening instruments for the detection of ASD. The

PAGE 32

24 Checklist for Autism in Toddlers (CHAT) has been score validated, and the Modified Checklist for Autism in Toddlers (M-CHAT), th e Ages and Stages Questionnaire (ASQ), and the Pervasive Developmental Disorders Sc reening Test (PDDST) are in the process of being score validated. In addition, NECT AS reported that developmental screening instruments, such as the Parents’ Evalua tion of Developmental Status (PEDS), can accurately provide information about their child’s development. Developmental Screening Instruments Both general developmental and auti sm-specific score-validated screening instruments can play a significant role in th e earlier identificati on of young children with ASD (AAP, 2001; National Research Council, 2001). General devel opmental screening instruments have a wide application with chil dren of varying ages, allow flexibility to obtain parental report with minimal assistance, ask more universal qu estions of parents, and coordinate with typical developmental m ilestones. However, due to their broad use, these instruments often lack the sensitivity to screen specifically for autism. Therefore, when results of general developmental sc reening tools raise co ncern, follow-up with autism-specific screening instruments is re quired. General deve lopmental screening instruments that were reviewed in the ASC trainings include the Ages & Stages Questionnaire (ASQ; Bricker & Squires, 1999), the Parents’ Evaluations of Developmental Status (PEDS; Glascoe, 1998), and Communicat ion and Symbolic Behavior Scales Developmental Pr ofile Infant Toddler Checklist (CSBS DP Infant Toddler Checklist; Wetherby & Prizant, 2002). Ages & Stages Questionnaire (ASQ). The ASQ uses parental report for children birth to five years of age. The questionnaire can be administered at a number of age

PAGE 33

25 intervals, from 4 to 60 months. The questi onnaire takes approximately 10-15 minutes for parents or caregivers to co mplete, and 2-3 minutes to score. Developmental areas including communication, gross motor, fine motor, problem solving, and social are addressed. The ASQ provides clear drawings and directions fo r eliciting thoughtful responses, and separate forms for each age ra nge of 10 to 15 items are tied to the well child visit schedule. The ASQ provides pass or fail scores, and has been wellstandardized and score validated with good sensitivity and excellent specificity (Filipek et al., 1999). Parents’ Evaluations of Developmental Status (PEDS). The PEDS is a screening and surveillance tool used with children from birth to eight years of age. It allows clinicians to make evidence-based decisions and is designed to detect a wide range of developmental issues as well as various type s of parental concerns. The PEDS identifies when to refer for additional screening or monitor developmental progress. The tool promotes collaboration between parents and pr oviders by eliciting pare nts’ concerns. The parents respond to 10 carefully constructed questions, with 90% of parents completing the written questionnaire while waiting for their appointment. Approximately two minutes are needed to score and interpre t the results (Filipek et al., 1999). High, moderate, or low-risk scores are obtained for developmenta l and behavioral problems. The sensitivity for the PEDS ranges from 74 %-79% and the specificity ranges from 70%80% across age levels (Filipek et al., 1999). Communication and Symbolic Behavior Sc ales Developmental Profile Infant Toddler Checklist (CSBS DP). The Communication and Sym bolic Behavior Scales Developmental Profile (CSB S DP; Wetherby & Prizant, 2002) is a standardized

PAGE 34

26 instrument designed for routine screening a nd evaluation of communication and symbolic abilities. This tool was designed for childre n between 12 and 24 months of age to assess typical communication milestones and parent al concern regarding development. The CSBS DP was developed based on the Commun ication and Symbolic Behavior Scales (CSBS; Wetherby & Prizant, 1993), which is a more in-depth tool designed for program planning. The CSBS DP is a br ief questionnaire consisting of a 24-item Infant-Toddler Checklist for screening that can be complete d by a parent. A longer follow-up Caregiver Questionnaire is available as well as a Behavior Sample. Th e Behavior Sample consists of a face-to-face evaluation of the child interacting with a parent and physician that is videotaped for later analysis. These three components (Checklist, Caregiver Questionnaire, and Behavior Sample) were designed to measure seven pre-linguistic skills. These skills are organized into three composites: the Social composite, including Emotion and Eye Gaze, Communication, and Ge stures; the Speech composite, including Sounds and Words; and the Symbolic compos ite, including Unders tanding and Object Use (Wetherby et al., 2004). The CSBS DP has been field-tested nationally, and the findings provide good evidence for score reliability and validity and support the use of the Checklist as a firstlevel screening and the Behavior Sample as a second-level evaluation following the Checklist (Wetherby & Prizant, 2002; Wether by et al., 2002). The CSBS DP appears to be effective for the early id entification of young children w ith ASD as it measures prelinguistic skills that have been identified as de ficits in preschoolers with ASD. Therefore, the CSBS DP Checklist and Behavior Sample are appropriate scr eening and evaluation tools for identifying children with developmen tal delays at 12 to 24 months of age

PAGE 35

27 (Wetherby et al., 2004) Use of a parent report tool, such as the Infa nt Toddler Checklist, minimizes the time required of healthcare providers while maximizing the role of the family. In addition, the Checklis t provides reasonably accurate information regarding the need to refer a child for a developmental evaluation (Filipek et al., 1999). Autism-Specific Scr eening Instruments Autism-specific screening instruments have been developed exclusively to screen for autism spectrum disorders. In addition, most of these instruments have been designed to concentrate on social and communication im pairment in children aged 18 months and older and focus on all three DSM-IV-TR criteria for autism. At this time, there is a lack of highly score-validated autism-specific sc reening instruments available for children under the age of 18 months. Autism-specific scre ening instruments that were discussed in the ASC trainings include the Checklist for Autism in Toddlers (CHAT; Baron-Cohen, Allen, & Gillberg, 1992), the Modified Ch ecklist for Autism in Toddlers (M-CHAT; Robins et al., 2001), and the Pervasive Deve lopmental Disorder Screening Test (PDDST; Siegel, 1998). Checklist for Autism in Toddlers (CHAT). The Checklist for Autism in Toddlers (CHAT) was developed by Baron-Cohen a nd colleagues in 1992. The CHAT was developed in an effort to move toward earlier screening and identification of young children, 18 months of age, at risk for au tism spectrum disorders. The questionnaire comprises two components; the first of whic h contains nine items reported by parents, such as whether the child ever demonstr ates pretend play. The second component includes five items that require a brief, semi-structured observation by a primary care provider at the well-child vi sit. These components assess parallel functioning in three

PAGE 36

28 main areas: (a) protodeclarativ e pointing, (b) gaze monitori ng, and (c) pretend play. The CHAT takes approximately 10 to 15 minutes to complete. If a child fails the CHAT, it is recommended that the child be re-screened ap proximately one month later. If the child fails the CHAT for a second time, the child s hould be referred to a specialist for further evaluation as the CHAT is not a diagnostic tool. The ease of administration and its demonstr ated specificity to symptoms of autism in children 18 months of age are two st rengths of the CHAT (Filipek et al., 1999). Findings from Baird et al.’s (2000) study demons trated that the sens itivity for the CHAT (number of children identified by the CHAT/number of children with autism in the entire sample) was low (e.g., from 20%-35%). However, the specificity (number of children without autism in the group who were not identified by the CHAT/number of children without autism in the sample) was very hi gh (e.g., from 98%-99.8%) (Scambler, Rogers, & Wehner, 2001). Filipek et al. (1999) conclude d that the CHAT appeared to be a useful screening tool for identifying children 18 mont hs of age at risk for autism. However, the CHAT appears to be less sensitive to milder symptoms of autism, such as children with Asperger syndrome or PDD-NOS (Filipek et al., 1999). Baron-Cohen et al. (1996) found that the CHAT has a specificity of 98%, but a sensitivity of 38%, and missed many children at 18 months who were later diagnosed with ASD. While the score validity of the CHAT is disappointing, it indicates the need for further res earch on young children with ASD and provides important clues to earl y indicators of ASD, based on the children they were able to identify early (Baron-Cohen et al., 1996). The Modified Checklist for Autism in Toddlers (M-CHAT). Robins, Fein, and Barton (1999) developed the Modified Ch ecklist for Autism in Toddlers (M-CHAT),

PAGE 37

29 which is an extension of the CHAT. The M-C HAT contains the nine parent-report items from the CHAT, and additional items were developed based on symptoms thought to be present in very young children with autis m (Robins et al., 2001). The questionnaire consists of 23 (yes/no) items reported by pare nts, in contrast to the combined parent report and physician observation used in the CHAT. Because the M-CHAT is a parentonly screening instrument, the range of behavi ors assessed is larger than on the CHAT (Charman et al., 2001). Robins et al. s uggest that the M-CHAT will have better sensitivity at 24 months of age compared to 18 months of age. Pervasive Developmental Disorder Screening Test (PDDST). Siegel (1998) developed a parental questionnair e that consists of three stages, each targeting a different level of screening to be used in different se ttings. The first stage is a parent questionnaire aimed for use in primary care settings with ch ildren from birth to 36 months of age. The PDDST rates both positive and negative symptoms, and contains a number of items pertaining to regression. In addition, the PDDST examines temperament, sensory responses, motor stereotypies, attention, attachment, and peer interest (Filipek et al., 1999). Summary Unlike the CHAT or M-CHAT, the Infant -Toddler Checklist is not designed to screen specifically for ASD, but rather, is de signed as a first-level screen for children with a broad array of communication delays. In regard to the Checklist, findings suggest that children with ASD are likely to have low scores on the Social composite of the Checklist and this pattern could be used to in dicate the need to c onduct an autism-specific screen next, such as the CHAT or M-CHAT (Wetherby et al., 2004). However, there are

PAGE 38

30 not yet sufficient validity data on the CHAT, MCHAT, or any other pare nt report tool to support their use as a second-level screen for ASD in the second year of life, and therefore, further research is needed. Practice Parameters Pediatric healthcare providers are typical ly involved in the identification of ASD because they see young children on a regular basi s at well-child visits, which occur quite frequently throughout first two years of life (American Academy of Pediatrics [AAP], 2001). These well-child visits present numerous opportunities to identify children with developmental delays or disabi lities early in their developmen t. Therefore, physicians can play a key role in the early identification a nd subsequent early intervention of infants and toddlers with ASD. Although the physicia n's role emphasizes the monitoring and screening of the development of young ch ildren, limited information is available regarding physicians' actual monitoring and screening practices (Filipek et al., 1999; Sices, Feudtner, McLaughlin, Drotar, & W illiams, 2003). Given the importance of early identification and practitioners’ role in this process, it is problematic that less than 30% of primary care providers conduct regular st andardized screening tests at well-child appointments (Dworkin, 1989). The American Academy of Neurology and Child Neurology Society endorsed a multidisciplinary consensus panel to review th e literature on screening and diagnosis of ASD and make recommendations on practice parameters (Filipek et al., 1999). The consensus panel was developed from nominations from a variety of organizations related to ASD, such as the American Academy of Child and Adolescent Psychiatry and the American Academy of Pediatrics. The panel developed a number of recommendations for

PAGE 39

31 the screening and diagnosis of ASD, with two levels of inves tigation: (a) Routine Developmental Surveillance and Screening Sp ecifically for Autism, and (b) Diagnosis and Evaluation of Autism (Filipek et al., 1999). The first level consists of the following recommendations: 1. All professionals involved in early child ca re should be familiar with the symptoms of ASD to recognize potential social, communica tive, and behavioral indicators of the need for further diagnostic evaluation. 2. Developmental screenings should be performe d at every well-child visit, and at any age thereafter if concerns are raised (recommended screening tools include ASQ, PEDS, and BRIGANCE). 3. Failure to meet the nearly universally pr esent developmental m ilestones (no babbling by 12 months, no gesturing by 12 months no single words by 16 months, no 2-word spontaneous phrases by 24 months, any loss of any language or social skills at any age) is an absolute indication to proceed with further evaluations. 4. Level 1 laboratory investigations, such as audiological assessments and lead screens should be conducted. 5. Professionals should be familiar with and use one of the screening instruments for children with autism (e.g., CHAT, PDDST). 6. The social, communication, and play developmen t and behavior of siblings of children with autism need to be carefully monitored. 7. As mandated by IDEA, a referral for early intervention should be initiated by the primary care practitioner, with children under 36 months of age referred to zero-tothree service systems, and children 36 m onths of age and older referred to the local

PAGE 40

32 school district. 8. Healthcare providers need to increase their comfort levels in talking with families about ASD. 9. Screening tools for older children with mild er symptoms of ASD need to be available in educational and recreational settings. The second level of recommendations deal s with diagnostic issues such as who should make the diagnoses (e.g., professionals who specialize in the treatment of ASD), the criteria on which the diagnoses should be based (e.g., DSM-IV-TR), and the level of sensitivity and specificity that the diagnosti c instrument should contain (Filipek et al., 1999). The panel concluded that further research is required to identi fy more precise early warning signs to differentiate accurately children with ASD from other populations. Pediatric healthcare providers are in a key position to detect co mmunication difficulties in young children earlier on by conducting r outine developmental surveillance on all patients. The panel recommended that pr oviders perform routine developmental screenings for ASD at each well child visit using standardized instruments that utilize parental report (Fi lipek et al., 1999). In addition, the consensus panel sugges ted that failure to meet any of the following five milestones is a definite indi cation for further eval uation: (a) no babbling by 12 months, (b) no gesturing by 12 months, (c) no single words by 16 months, (d) no 2word spontaneous phrases by 24 months, and (e) any loss of any language or social skills at any age. It is also important to monitor other social communicati on parameters, such as deficits in joint attenti on and symbolic communication. Limitations in communication development may be the first symptom eviden t to parents and professionals. The panel

PAGE 41

33 identified several early indicat ors that would necessitate fu rther evaluation, including the use of sounds, gestures, words, and word combinations. The panel stressed the importance of recognizing that many of these symptoms also may be evident in children with developmental disabilitie s who do not have ASD, or in children who are delayed, but naturally catch up without inte rvention (Filipek et al., 1999). Although the literature indi cates that a number of primary care providers are not routinely screening children for developmental disa bilities during well-child appointments, research findings suggest that pediatric healthcare pr oviders are aware of the important role that parental report a nd knowledge of developmental milestones can have on the early identification of children w ith ASD (Sices et al., 2003). Sices et al. (2003) found that most physicians reviewed developmental milestones and prompted parents for developmental concerns at preventive care visits. However, only approximately one-half of the physicians used a formal developmental screening instrument. Although it is important for physicia ns to be aware of the possible usefulness of parental report and knowledge of developm ental milestones, it is also critical that physicians are knowledgeable about both gene ral developmental screening instruments and autism-specific instruments. Moreover, it is essential that physicians utilize these tools to enable them to identify children w ith ASD as early as possible (Sices et al., 2003). However, primary referral sources are often unaware of early warning signs for ASD and frequently take a “wa it and see” attitude with parent s, which contributes to the delay in referral and subseque nt identification, as well as a delay in initiating supports and services (Oser & Shaw, 2001).

PAGE 42

34 A limited number of studies has exam ined physicians’ use of screening instruments to identify young ch ildren with developmental de lays (Sices et al., 2003). Shonkoff, Dworkin, Leviton, and Levine (1979) examined primary care approaches to developmental disabilities. The results reveal ed that only 19% of pe diatricians reported that their approach to a young child with a language delay would include the use of a standardized developmental sc reening instrument. Furtherm ore, 38% of pediatricians indicated that they would use a developmenta l screening instrument if parents raised a concern about possible mental retardation in their 3-year-old child. Dobos, Dworkin, and Bernstein (1994) conducted a simila r study 15 years after the study by Shonkoff et al. (1979). Dobos et al. found that 61% of pe diatricians reported use of screening instruments with children suspected of de velopmental delay (e.g., mental retardation). The pediatricians from this study also were more likely to refer these children to be assessed by specialists compared to the pa rticipants from Shonkoff et al.’s (1979) investigation. Although the use of score-validated screen ing tools is an effective way in which physicians can identify children with developm ental delays, research reveals that more than one-half of children with developmental disabilities ar e not detected before school entry. Furthermore, physicians often unde r-identify language-related delays and disabilities in young children (Sices et al., 2003). Consistent use of developmental screening tools could signifi cantly improve physicians’ abilit y to detect children with developmental delays. The use of formal instru ments to obtain parental concerns, such as the Parents' Evaluation of Developmental Stat us (PEDS), also could aid physicians in the

PAGE 43

35 early identification of delays a nd disabilities and guide the re ferral process (Sices et al., 2003). Sices et al. (2003) examined how prim ary care physicians identify young children with developmental delays. A survey was mailed to a national random sample of pediatricians and family physicians, with a to tal of 540 surveys retu rned (341 returned by pediatricians and 199 returned by family physicians). Thus, the overall response rate for the survey was 49.3%, which is similar to the average response rate to mail surveys (i.e., 54%) for physicians (Asch, Jedrziewski, & Ch ristakis, 1997). The survey inquired about the methods used during the preventive care vi sits at 2 years of age to identify children with developmental delays. Information rega rding participants’ self-reports of current developmental screening practices were obt ained, and several hypotheses were tested examining whether reported identification e fforts varied depending on physician beliefs. In addition, participants also were queried about factor s that may influence their developmental screening procedures. Fivepoint Likert-type scal es were used to determine the priority of developmental sc reening compared with other components of the preventive care visit. In addition, physicians were asked to give their opinions about seven statements concerning factors that might impact physicians' surveillance or screening practices (Sices et al., 2003). Findings from this study revealed that dur ing routine preventive care visits with 2year-old children, most physicians reported us ing a list of developmental milestones as well as the prompting of parents for specific concerns in multiple areas of the child's development. One-half of the pediatrician s and more than one-half of the family physicians reported using some form of scor e-validated instrument for developmental

PAGE 44

36 screening. Approximately one-half of all the physi cians reported that they used a specific score-validated developmental screening instrume nt as part of their routine practices with children ages 1 to 3 years. Finally, pediatri cians and family physicians reported using a similar group of available screening instrument s (Sices et al., 2003). It is problematic that even with the research suppor t for early identification a nd screening practices, many providers are not consistently screening all of their young patients dur ing well-child visits (Filipek et al., 1999). The following section de tails a number of pot ential factors that could impede the routine screening of young ch ildren for developmental disorders such as ASD. Barriers to Pediatric Healthcare Prov iders’ Use of Screening Instruments As noted earlier, given the research support for the benef its of early identification and intervention, it is crucial that children are identified at as early an age as possible. Although there is substantial evidence for symptom onset prior to 18 months of age, many children with ASD are not diagnosed until six years of age (Filipek et al., 1999). There are several hypotheses as to the reasons for the delay between presence of symptomatology and diagnosis in children with ASD. One hypothesis is that the primary referra l sources, such as pediatric healthcare providers, may be unfamiliar with the early warning signs of ASD, and therefore are hesitant to refer these young patients for se rvices. When practitione rs are unfamiliar with the warning signs of ASD, or are unfamiliar w ith the disorders in general, their selfefficacy in relation to identifying young childre n with ASD may be low. Perceived selfefficacy refers to a person’s beliefs about their ability to produce desired outcomes and exercise control over events that affect their lives (Bandura, 1994). When people doubt

PAGE 45

37 their capabilities, they are more likely to shy away from diffic ult tasks that they view as personal threats (Bandura, 1994). However, wh en self-efficacy is high, people approach difficult tasks as challenges to be overcome rather than as threats to be avoided. In addition, efficacious persons tend to become deeply engrossed in activities and set challenging goals while maintaining a st rong commitment to them (Bandura, 1994). Therefore, it is crucial that practitioners feel confident and knowledgeable in their abilities to screen and identif y young children at-risk for ASD. Another significant dilemma for healthcare providers is that identification must precede the provision of services, and they ma y be hesitant to recommend a complete evaluation for developmental disabilities for f ear it will bring about anxiety in parents. Furthermore, there is warranted concern re garding the emotional impact on the family with the diagnosis of ASD, as some continue to hold the belief th at ASD carries a poor prognosis (AAP, 2001; Oser & Shaw, 2001). This apprehension could contribute to the delayed identification of children with milder symptoms, as those with evident delays are more likely to be identified earlier (AAP, 2001). Recently, advances have been made in be havioral diagnostic criteria that have lowered the potential age of diagnosis from around 5 years of age to as early as 18 months of age (Filipek et al., 1999). The consideration of additional behaviors not previously thought to be diagnostic, such as motor behaviors, also have helped to lower the potential age of diagnosis, with some re search supporting accurate diagnosis at 8 to 12 months of age (Teitelbaum, Teitelbaum, Nye, Fryman, & Maurer, 1998). The diagnostic features that are indi cative of ASD typical ly develop throughout the first two years of life; therefore, ASD should be evident in very young children.

PAGE 46

38 Based on retrospective accounts, most careg ivers report that their children with ASD displayed symptoms within the first two years of life (Wimpory, Hobson, Williams, & Nash, 2000). Furthermore, most families express concern to their pediatrician by the time their child is 18 months of age (Howlin & Moore, 1997). Alth ough many children with ASD are not diagnosed until at least three year s of age, a diagnosis of ASD at two years of age was found to be associated with the same diagnosis at three year s of age or older in the vast majority of children. Therefore, dia gnoses of ASD in children two years of age is as reliable (and consistent) as diagnoses ma de in children three years of age or older (Lord, 1995; Stone et al., 1999). Children with ASD may not be identified during the first two years of life because although some indi cators of ASD commonly are present by two years of age (e.g., impairments in social in teraction and communication), others are not evident until later. For example, restricted and repetitive activities and interests are common indicators of ASD, yet these behaviors typically are not present until closer to 3 years of age. This delay in the onset of symptomatic behaviors could be a significant factor in the later diagnosis of ASD in a majority of ch ildren (Wetherby et al., 2004). Sices et al. (2003) also examined the barri ers to the use of screening instruments for physicians. Findings indicated that less than one-half of physic ians agreed that there is adequate time to perform developmental sc reening during a typical well-child visit. Furthermore, very few agreed that reimburse ment for well-child visits is sufficient to cover the time spent on developmental sc reenings. Pediatricians reported feeling confident in their care of a ch ild diagnosed with a developmental delay more than twice as often as family physicians. Pediatricians also were twice as likely as were family physicians to agree that sufficient resources exist in their communities to address the

PAGE 47

39 needs of children with delays. Finally, pediatricians also were two times as likely as were family physicians to report that they posse ss the clinical expertis e to identify most children with developmental delays without the use of a developmental screening instrument (Sices et al., 2003). In summary, findings from this study indica te that most physicians rely on lists of developmental milestones and/or prompting for parental concern to identify children with developmental delays. Physicians also repor ted time and reimbursement as significant barriers to the use of screening instruments. The authors from this study concluded that although the barriers to developmental screeni ng in primary care are significant, most physicians are aware of the value of early intervention services for young children with developmental delays (Sices et al., 2003). Halfon et al. (2001) also conducted a st udy to examine the barriers to the early identification of developmenta l disabilities. In 2000, a su rvey was administered to members of the American Academy of Pediatri cs (AAP) to identify relevant barriers to the timely identification of developmenta l issues in primary care practice. More specifically, they sought to ascertain the barrie rs to the use of score-validated screening instruments. In regards to children birth through 35 months of age, participants were asked to describe the barrie rs to the provision of deve lopmental assessments during pediatric health supervision as a function of practice characte ristics. Halfon et al. found that 94% of the participants agreed that pediatricians should inquire about children’s development. In addition, 80% of the participan ts felt confident in their ability to advise parents; however, 65% reporte d less adequate training, and only 36% agreed that there was adequate time for developmental assessments (Halfon et al., 2001).

PAGE 48

40 Insufficient time to administer screen ing instruments (80%) and reimbursement issues (56%) were the most frequently cited barriers to utilizing formal developmental screening instruments (Halfon et al., 2001). Pa rticipants also reported barriers such as lack of available staff to assist with developmental assessments (51%), unfamiliarity with coding for reimbursement (46%), lack of developmental diagnostic and treatment services (34%), and lack of training (28 %) (Halfon et al., 2001). Unfamiliarity with screening instruments and lack of referral programs also were viewed by pediatricians as significant barriers to the use of developmen tal screening instrume nts (Halfon et al., 2001). Although a number of barriers to the use of screening instruments were reported by providers, there are ways in which th ese barriers can be overcome. Because insufficient time to administer screening inst ruments was reported as the most significant barrier to screening practices it is important to note th at score-validated parent questionnaires may be used to minimize th e time needed by providers to administer screening instruments (Halfon et al., 2001). In addition, parental concern about a child's development also may be a reliable predicto r of developmental de lays. It would be beneficial to ascertain the extent to whic h primary healthcare pr oviders are utilizing parental report questionnaires for devel opmental screening. As Halfon et al. found, a wide variation exists in the reported pract ice in the utilization of score-validated screening instruments in primar y care. Thus, future research needs to be conducted on the early identification of children with developmental delays. In the survey conducted by Halfon et al (2001), approximately one-half of the physicians reported that they use a score-valid ated developmental screening instrument in

PAGE 49

41 their practices. However, although physicians report that they are using a variety of methods to identify children with delays, a significant number rely only on lists of developmental milestones or prompting for parental concern. In addition, because providers do not have adequate time to admini ster screening instruments, when they are utilized it is likely that they are not used in a standardized manner, which diminishes their score validity (Hal fon et al., 2001). The increased use of parent questionnaires that yield valid scores (e.g., ASQ, PEDS) will ameliorate the time constraint providers’ face, which is a significant barrier to their use of scr eening instruments. Additionally, as providers continue to cite reimbursement as a central barrier to the utilizat ion of developmental screening instruments, this issue needs to be addressed at a policy level (Halfon et al., 2001). Importance of Training for Providers in Identifying Children with ASD The American Academy of Pediatrics ( AAP) developed practice guidelines that recommend routine developmental screen ing and surveillance to be conducted specifically for autism on all children. Routine developmental screening first would identify children at risk for any type of atypical development, and also would identify those specifically at risk for autism (AAP 2001). However, a number of healthcare providers do not feel comfor table utilizing developmental sc reening instruments because of a lack of training. In addition, they ma y not know how to implement these guidelines successfully to perform accurate developmen tal screenings with young patients (Halfon et al., 2001). Therefore, hea lthcare providers would benefit from specific guidance on how to incorporate routine developmental scre enings into their pract ices because a lack of guidance may result in the delay of identif ication and appropriate intervention services

PAGE 50

42 (AAP, 2001). When children with developmental delays or disabilities are identified and receive treatment early, the negative impact on the functioning of both the children and the families may be greatly reduced. It is critical that trainings address th e barriers that preven t healthcare providers from routinely using developmental screening instruments with their patients, such as time constraints and reimbursement issues. There are continuing efforts to increase awareness of ASD in practitioners, includi ng knowledge of developmental milestones, warning signs for development that is not following expected trajectories, and scorevalidated screening instruments (Oser & Shaw, 2001). Because of the importance of receiving appropriate training to identify young children with developmental disabilities, it is problematic that the major ity of pediatric healthcare pr oviders are not routinely using developmental screening tools. Halfon et al. (2001) found that family phys icians reported substantially lower selfefficacy to support children with developm ental delays/disabilities compared to pediatricians. In addition, family physicians also perceived community resources as being less available to support these children compared to reports from pediatricians. These findings underline the importance of providing all healthcare providers who work with children specific educational interventions These specific interventions should be tailored to improving confidence in managing children with developmental delays/disabilities, as well as increasing their awareness of available community resources. Furthermore, interventions also could help to improve the availability of resources within some communities (Halfon et al., 2001).

PAGE 51

43 Summary Autism is a neurodevelopmental disability that affects the functioning of the brain and typically appears during the first three y ears of life. The primary features of autism are the presence of abnormal or impaired development in communication and social interaction, and a restricted repertoire of beha viors and interests. Because the etiology of autism spectrum disorders (ASD) is unknown, inte rventions for individuals with ASD are developed to reduce the interf ering symptoms. Given the rece nt increases in the number of children diagnosed with ASD, it is crucial that children are identi fied early on so that they receive the services needed (Fil ipek et al., 1999; Oser & Shaw, 2001). Growing evidence demonstrates the effec tiveness of intensive early intervention with a significant proportion of young child ren with ASD (Dawson & Osterling, 1997; Filipek et al., 2000; Oser & Shaw, 2001). Daws on et al. (2000) found that as the prenatal and early postnatal years represent a se nsitive period for brain and behavioral development, increased efforts at early identification and intervention are needed. Negative consequences for individuals w ith ASD can be minimized or avoided completely with the promotion of optimal pr enatal and infant-toddler development, as long-term consequences have their greatest influence during early child development. It is recommended by the American Academ y of Pediatrics (AAP) that all infants and children are screened for developmenta l delays or disabilities (AAP, 2001). The National Research Council Report on Educa ting Children with Autism (2001) reviewed several screening instruments for the detection of ASD. The Checklist for Autism in Toddlers (CHAT) has been score validated, and the Modified Checklist for Autism in Toddlers (M-CHAT), the Ages and Stages Questionnaire (ASQ), and the Pervasive

PAGE 52

44 Developmental Disorders Screen ing Test (PDDST) currently ar e in the process of being score validated. In addition, NECTAS reported that developmental screening instruments, such as the Parents’ Evaluation of Developm ental Status (PEDS), can accurately provide information about children’s development. Greater public awareness and education of healthcare providers in regard to the early detection of developmental disorders and how to access appropriate interventions are needed. Providers need to be proficient in the identification of early symptoms of autism so that appropriate screening and referral procedures can occur. The NECTAS Forum on ASD discussed strategies for earlier identification, including raising public and professional aw areness, tools for screening, determining eligibility for servi ces, and transitioning. The early identification of children with developmental disabilities requires that he althcare providers are familiar with scorevalidated screening instruments. It is also critical that they feel comfortable discussing parental concerns, and that they are knowledgeable about referral resources in their communities (AAP, 2001). Although the physician's role emphasizes the screening of the development of young children, limited data ar e available regarding physicians' actual monitoring and screening pract ices (Filipek et al., 1999; Sices, Feudtner, McLaughlin, Drotar, & Williams, 2003). It is problematic th at less than 30% of primary care providers conduct regular standardized screening test s at well-child appointments, given the importance of early identification and healthcare providers’ role in this process (Dworkin, 1989). The American Academy of Pediatrics (AAP) developed practice guidelines that recommend routine developmental screen ing and surveillance to be conducted

PAGE 53

45 specifically for autism on all children. Routine developmental screening first would identify children at risk for any type of atypical development, and also would identify those specifically at risk for autism (AAP 2001). However, a number of healthcare providers do not feel comfor table utilizing developmental sc reening instruments because of a lack of training. In addition, they ma y not know how to implement these guidelines successfully to perform accurate developmen tal screenings with young patients (Halfon et al., 2001). Healthcare provi ders would benefit from gui dance on how to incorporate routine developmental screenings into thei r practices because a lack of guidance may result in the delay of iden tification and appropr iate intervention se rvices (AAP, 2001). When children with developmental delays or disabilities are identified and receive treatment early, the negative impact on the functioning of both the children and the families may be greatly reduced. Future resear ch is needed to examine the effectiveness of trainings for healthcare providers regardi ng screening practices and early identification of ASD.

PAGE 54

46 Chapter 3 Method Participants The participants in this study were pedi atric healthcare provi ders who practiced medicine and resided in the st ate of Florida. Sele cted participants a ttended one of three field-test training sess ions held throughout Florida. The number of pediatricians currently practicing in the state of Florida is approximately 3,423 (American Academy of Pediatrics [AAP], 2000). All geographic locatio ns throughout Florida were considered for the settings of the training, with the expectation that each of the three training sessions would be held in three different geographic ar eas, with at least one location defined as rural, and at least one loca tion defined as urban. The part icipants were recruited to participate in the training sessions from the following ge ographic regions: Clewiston, Jacksonville, and Tampa, Florida These geographic areas were selected because of availability and interest in the ASC training in these locations. More specifically, the first training session was held on Wednesday, May 4, 2005 at the University of South Florida, Florida Mental Health Institute from 6: 00 p.m. to 7:00 p.m. The second training session was held on Wednesday, May 18, 2005 at the Duval County Health Department from 3:00 p.m. to 4:00 p.m. The third training sess ion was held on Wednesday, June 1, 2005 at the Hendry Regional Medical Center from 12: 00 p.m. to 1:00 p.m. A fourth training session at the University of South Florida, Fl orida Mental Health Institute was added in an attempt to obtain more participants to complete the ASC training. Unfortunately, no

PAGE 55

47 participants took part in this last training session. The criter ia for participation in the ASC training session were: (a) residen ce in the state of Florida and (b) provision of services to the pediatric population. Therefore, variability in age, gender, race, geographic location, profession, setting of practice, years in pr actice, and number of trainings completed related to ASD was expected. Selection of Participants Participants were selected based on th eir professional role s and the geographic location where they practiced medicine. On e major goal of this training focused on reducing barriers to screening by problem so lving ways in which to change practice; therefore, the main aim was to recruit pedi atricians and pediatric nurse practitioners because they are most likely to be in the position to facilitate change within their practices. Furthermore, the primary interest of this study was in the screening practices of physicians because research demonstrates th at the majority of physicians (i.e., 86%) indicate they are predominantly responsib le for developmental screening and/or surveillance (Sices et al., 2003). However, registered nurses and other pediatric healthcare professionals also were recruited for particip ation in the ASC training. Prior to the recruitment of any participants, approval for this study was obtained from the University of South Florida Instit utional Review Board (IRB) to ensure the ethical treatment of the participants in th is study. Pediatric hea lthcare providers were recruited to participate in the ASC traini ng by contacting department chairs in the division of pediatrics via telephone. In a ddition, three flyers de tailing the learning objectives of the training sessions were developed by the ASC workgroup (see Appendix A). These flyers were electronically maile d to the workgroup for dissemination to

PAGE 56

48 practitioners via email. For th e first training, flyers also we re posted on the University of South Florida campus. Additionally, workgroup members distributed the flyers at the Hillsborough County Pediatrics Society (H CPS) Meeting on April 20, 2005, and the HCPS meeting on April 21, 2005. Fo r the second training, flyers were disseminated by a pediatrician who practices medicine in Jacks onville, Florida. For th e third training, flyers were disseminated both via electronic maili ngs and postings at the Hendry Regional Medical Center and the Hendr y County Health Department in Clewiston, Florida. Physicians who participated in the ASC training received one Continuing Medical Education (CME) credit, and nurse practit ioners who participated received one Continuing Education Unit (CEU). The pediatric healthcare providers were in vited to participate in one of the three ASC trainings based on location. The origin al goal was to train a minimum of 100 pediatric healthcare providers throughout the st ate of Florida, as stipulated by the grant that funded the ASC trainings. Unfortunately, this goal was not obtained because only 36 practitioners completed the ASC training. To obtain a statistical power of .80 for detecting a medium effect size for comparing the two experimental locations at the .05 level of significance, using a repeated measures analysis of variance (ANOVA), a minimum of 40 participants was needed (Cohen, 1988). Practitioners who agreed to participate in the ASC training were placed in one of two experimental groups based on their geographic location (i.e ., rural or urban). It was hoped that there would be approximately the same number of particip ants who attended each of the three training sessions and who participated in the study. Preand post-test analyses were conducted to determine if there were st atistically significant differe nces among the three training

PAGE 57

49 groups and whether any of the groups could be collapsed. All training participants were invited to participate in this study, with the ex pectation that virtuall y all participants in the trainings would complete th e pre-test questionnaire. Howe ver, it was expected that approximately 50 participants would comp lete both the preand post-questionnaire because research demonstrates the average res ponse rate to mail surveys for physicians is 54% (Asch et al., 1997). Practitioners in the state of Florida who had not participated in the ASC trainings were asked to take part in the control gr oup. These pediatric healthcare providers were contacted via mailings to their places of em ployment describing the goals and purpose of the study. These participants met the same criteria of the participants in the experimental group (e.g., practicing medicine in the state of Florida, and working with the pediatric population). The researcher attempted to c ontact approximately 50 providers, with the goal of obtaining a minimum of 25 participan ts for the control group. Attempts were made to recruit participants for the control group from the same three geographic regions as participants in the experimental group. Questionnaires were mailed multiple times, as necessary, to attain at least 25 participants in the control group to ensure that the sample size was large enough to obtain adequate statistical power. The sampling scheme used for this study represented non-random, convenience sampling. The participants were arranged in to one of three groups based on geographic location and participation in the ASC traini ng. The groups were as follows: (a) ruralexperimental group, (b) rural-control group, (c) urban-experimental gr oup. Participants in the control group were recruited from a region in central Florida. Because the participants

PAGE 58

50 were assigned to training groups based on th eir geographic location and participation in the ASC trainings, they were not rand omly assigned to the three groups. Research Design This study utilized a quasi-experimental research design. Specifically, the type of quasi-experimental design was a pretestposttest nonequivalent-g roups design (Best & Kahn, 2003). Quasi-experimental designs provide the researcher control of when and to whom the measurement is applied; however, participants were not randomly assigned to experimental and control treatments. Therefor e, the equivalence of the groups could not be assured (Best & Kahn). The pretest-pos ttest nonequivalent-groups design is commonly used when experimental and control groups are naturally asse mbled groups (Best & Kahn). This design was used because it was th e most feasible design for this study. The researcher did not have any influence over the assignments of participants to the experimental group and the control group; theref ore, similarity across groups with respect to important characteristics, such as knowle dge of ASD or use of screening instruments could not be controlled. Variables Several dependent and inde pendent variables were meas ured in this study. The dependent variables were the strength of ba rriers preand posttr aining, screening tools utilized (both general and autism-specific), perceived leve ls of general knowledge related to ASD, and self-efficacy of participants in relation to accurate screening practices. The independent variable in this study was the type of training: Autism System of Care training (i.e., experiment al group) versus no traini ng (i.e., control group).

PAGE 59

51 Instrument Pediatric Healthcare Provid er Self-Report Questionnaire The Pediatric Healthcare Provider Self-Report Questionnaire was developed by the principal investigator for the purpose of this study. This measure was designed to assess the effect of the Autism System of Ca re training on pediatric healthcare providers’ change in practice regarding the method of ASD early identificati on. The questionnaire contains a total of four se ctions: (a) general knowledge of Autism Spectrum Disorders, (b) use of screening instrume nts, (c) perceived barriers to utilization of screening instruments, and (d) demographic inform ation. The first three sections of the questionnaire were covere d in the ASC trainings. The first section of the Pediatric Health care Provider Self-R eport Questionnaire, entitled “General Information,” c ontains a total of six items. The first four items in this section assess perceived levels of knowledge of ASD, such as knowledge of autismspecific screening instruments and knowledge of early warning signs of ASD. These items utilize a four-point rating scale (1 = Poor, 2 = Fair, 3 = Good, 4 = Excellent), asking participants to rate their levels of knowledge by circling the most appropriate number that corresponds with their perceived levels of knowledge. An example of an item is, “How would you assess your overall knowledge of early warning signs of ASD?” The scale used for this section is the Percei ved Knowledge Scale, which is divided into two parts. The first part contains four item s utilizing a four-point rating scale (i.e., 1 = Poor, 2 = Fair, 3 = Good, 4 = Excellent). The response for each item is summed to generate a total scale score. Th e scores range from 4 to 16, with high scores indicating the participant perceives herself/himself to ha ve excellent knowledge related to ASD. The

PAGE 60

52 last two items in this section assess particip ants’ self-efficacy regarding their ability to screen and refer children suspected of having an ASD. The format of the last two items is open-ended, with participants asked to indicate the age, in months, at which they believe children can be accurately screened and referred for Autism Spectrum Disorders, and the age at which they believe they themselves can accurately screen and refer children suspected of having Autism Spectrum Disorders. The second section, entitled “Screening Patterns,” contains three subsections related to use of screening instruments. The screening instruments included in the questionnaire are the same instruments that we re chosen for review in the ASC training sessions. These instruments were selected for the training sessions based upon two factors known to impact their use: (a) time to administer the instrument and (b) cost of the instrument (Halfon et al., 2001). For all th ree subsections, particip ants were asked to indicate on a five-point rati ng scale (1 = Never, 2 = Rare ly, 3 = Sometimes, 4 = Usually, 5 = Always) how often they use each individual screening tool, and how often they use developmental screening tools with patients in different age groups. Subsection A contains items regardi ng participants’ use of three general developmental screening instruments, includ ing the Ages & Stages Questionnaire (ASQ), Parent’s Evaluation of Developmental Stat us (PEDS), and Communication and Symbolic Behavior Scales Developmental Profile : Infant-Toddler Checklist (CSBS DP). Subsection B has items regarding use of th e following three autism-specific screening instruments: the Checklist for Autism in Toddlers (CHAT), the Modified Checklist for Autism in Toddlers (M-CHAT), and the Perv asive Developmental Disorder Screening Test (PDDST). The Screening Scale correspond s with Subsections A and B in the second

PAGE 61

53 section (i.e., Screening Patterns). The deve lopmental subscale consists of three items utilizing a five-point rating scale (i.e., 1 = Never, 2 = Rarely, 3 = Sometimes, 4 = Usually, 5 = Always). The response for each item is summed to generate a total subscore. The scores for this subscale range from 3 to 15, with high sc ores indicating frequent use of developmental screening instruments. The autism-specific subscale contains three items utilizing the same five-point rating scale as the developmental subscale. The scores from this subscale also range from 3 to 15, with high scores indicating frequent use of autism-specific screening instruments. The tota l screening scale (i.e., developmental scale and autism-specific scale) ranges from 6 to 30. Subsection C contains items regarding how often participants use developmental screening instruments in relation to seven age ranges of patients. Ages were grouped in either 6or 12-month increments, based on the recommended ages for well-child visits (e.g., 0-6 months, 7-12 months, 13-18 mont hs, 19-24 months, 25-36 months, 37-48 months, and older than 48 months) ( AAP, 2001). The Age of Screening Scale corresponds with Subsection C from the Scre ening Patterns section, and each item from this scale was asse ssed individually. The third section of the instrument as sesses the perceived impact of potential barriers on participants’ use of screening instruments. Potential barriers were developed by reviewing the literature on barriers related to the use of developmental screening instruments and the referral of patients suspected of ASD for further evaluation. Participants were asked to i ndicate on a four-point Likert -type scale (1 = Unlikely, 2 = Somewhat Unlikely, 3 = Somewhat Likely, 4 = Very Likely) th e extent to which specific barriers ( n = 7) were likely to impact their ability to screen or refer patients for further

PAGE 62

54 evaluation. An example of an item in this se ction is, “Insufficient information regarding referral resources.” The Barriers Scale is the fourth scale on the questionnaire. The responses were summed to yield a total scale score. Scores from this scale range from 7 to 28, with high scores indicating the potential barriers are very likely to impede use of screening instruments. The fourth and final section of the questi onnaire contains nine items that elicit demographic information. The following inform ation was gleaned: (a) age, (b) gender, (c) race, (d) location of pract ice, (e) profession, (f) setti ng of practice, (g) years in practice, (h) number of traini ngs completed related to Autism Spectrum Disorders, and (i) number of trainings completed related to changing practice/service delivery. For number of years in practice, and number of tr ainings completed related to both ASD and changing practice/service delivery, participants were asked to write in the appropriate number. For all other demographic items, multiple-choice options were provided (e.g., for race: White [Non-Hispanic], Black/Af rican American, Hispanic, Asian/Pacific Islander, Native American, Multi-Racial/Et hnic, Other). The Pediatric Healthcare Provider Self-Report Questionnaire is presented in Appendix B. The Pediatric Healthcare Pr ovider Self-Report Questionnaire was reviewed by an expert panel to assess its content-relate d validity. The panel comprised group members who developed the ASC trainings ( n = 6), including pediatricians ( n = 3), professors from related fields ( n = 2), and the director of an autism center in the region ( n = 1). No members from the expert pane l participated in this study. Copies of the questionnaire were distributed to the expert panel to obtain feedback on each of the items. The researcher conducted semi-structured intervie ws with members of the expert panel to

PAGE 63

55 review the items on the questi onnaire. The panel believed th at the first draft of the questionnaire was too long with 47 items, and as a result, participants might be less likely to complete the questionnaires. Therefor e, the panel recommended removing 12 items from the questionnaire that we re not as strongly related to the content of the ASC trainings as the remaining items. The panel also suggested reducing the length of the questionnaire so that the 35 items fit onto tw o pages. These changes were made, and the final version of the copy was distributed to the panel and was subsequently approved by all panel members. The questionnaire was completed both pr ior to and approximately two to three months after completion of the training session. This time frame was chosen after discussion with several pediatricians from the expert panel. It was agreed among the panel members that two to three months wa s an appropriate time period to measure change in screening practices for pediatric healthcare providers. A pproximately one week following the initial mailing of the post-test questionnaire, a follow-up postcard was sent to participants to remind them to complete and return the post-test questionnaire. Approximately two weeks following the remi nder postcard, the post-test questionnaire was resent to all participants who had not returned the comp leted post-test questionnaire. Two weeks after the second ma iling of the post-test questionnaire, a final reminder postcard was sent to participants who ha d not returned the completed post-test questionnaire. Procedures The procedures for this study included the researcher identifying the content to be disseminated in the Autism System of Care trainings, developing the Pediatric Healthcare

PAGE 64

56 Provider Self-Report Questionnaire based on the content of the trainings, and obtaining approval from the Institutional Review Bo ard. After the training materials were developed and finalized by the ASC group me mbers, the researcher developed the questionnaire based on the content that was to be included in the Autism System of Care training sessions. The Autism System of Care training wa s developed from a one year extension grant that was funded by the Florida Devel opmental Disabilities Council. The previous grant (i.e., year one) surveyed pediatricians throughout Florida, a nd based on the results of the survey responses, it was determined that a need existed in Florida to provide more information about early screening and referr ing for ASD. The Autism System of Care training sessions were one hour in length, and were struct ured as workshops for the participants. The locations of the training sessions were: (a) Clewiston, FL, (b) Jacksonville, FL, and (c) Tampa, FL. Two pedi atricians involved in the development of the ASC training facilitated the training sessions. One of the presenters is a developmental pediatrician who practices medi cine in Pinellas County, Florida. The other presenter is a general pediatri cian who practices medicine in Duval County, Florida. Both physicians had considerable knowledge rela ted to ASD and the importance of early identification. The goal of the training was to improve pediatric healthcare providers’ screening practices for the early identifica tion of children with ASD. Therefore, the sessions included a brief overview of ASD (e.g., definition, areas of impairment, and etiology), a review of the wa rning signs of ASD, and a review of common screening instruments. In addition, a discussion regard ing the importance of early screening and identification and the barriers to routine scr eening practices occurred. Finally, a model for

PAGE 65

57 improving screening practices was discussed, an d participants were asked to create aim statements for changing their practices and develop the next steps to initiate these changes. The presenters utilized a PowerPoi nt presentation to en sure that the same material was presented to all participants in the three traini ng sessions. Each session included a PowerPoint presenta tion and individual and sma ll group activities related to the material in the presentation. In addition, participants received handouts containing the slides from the presentation, as well as handouts containing additional information regarding screening instruments. To ensure that all learner objectiv es were covered in each of the training sessions, an implementation checklist was developed by the researcher (see Appendix C). The implementa tion checklist containe d 27 items directly from the PowerPoint presentation. The research er indicated the extent to which each item was covered by placing a checkmark in either the “Yes,” “Partially,” or “No” column. In regard to the questionnaires, arbitr ary identification numbers created by the researcher were included on the actual quest ionnaire forms to ensure anonymity. A cover sheet was attached to each questionnaire (i.e ., preand post-test que stionnaires) detailing the purpose of the study and expressing grat itude for participation in the study. In addition, the cover sheet included informati on regarding an incentive for completing and returning the post-test questi onnaire. When a post-test qu estionnaire was returned, the identification number on the questionnaire wa s entered into a draw ing. There were two $25 gift certificates awarded; one to a part icipant who completed the ASC training and one to a participant from the control group. Participants were asked to include their names and email addresses on the bottom of the pre-test cover letter to receive summaries of individual and/or overall resu lts from the study. The email addresses also were used to

PAGE 66

58 contact the two winners from the drawings. Participants were asked to provide their names on the cover sheets for the pre-test questionnaire only. The cover sheets attached to the post-test questionnaires did not requir e participants to provide their names. Each questionnaire cover sheet containe d a unique identification number and was attached to a questionnaire with the co rresponding identification number. The cover sheets attached to the pre-te st questionnaires subsequently was detached by the researcher to separate identifying information (i.e., na mes) from the completed questionnaires. The identification numbers on the pre-test questi onnaire cover sheets were used to match participants’ names to the corresponding pre-test questionnaires. The post-test questionnaire utilized the sa me identification number as the pre-test questionnaire for each individual participant. A sign-in sheet was posted at each trai ning session. When participants arrived for the trainings, they were asked to sign in and provide their current mailing addresses. Immediately prior to the tr aining session, the researcher distributed the pre-test questionnaires to the participants and collected the completed questionnaires and cover sheets before the start of the session. A master list of participants was created to match each participant’s name with his/her unique identification number and corresponding pretest questionnaire. The post -test questionnaires, along with a return addressed, stamped envelope, were mailed to the participants approximately 8 to 12 weeks after completion of the training. These questionnaires contai ned only the unique identification numbers printed on the cover sheets and questionnaires. Therefore, the particip ants were able to return the post-test qu estionnaires without any identifying information.

PAGE 67

59 The researcher matched the post-test que stionnaires’ identification numbers with the corresponding pre-test quest ionnaires’ identification num bers. Once the preand posttest questionnaires were ma tched, all identifying information was removed and discarded. All data obtained from the que stionnaires were accessible only to the researcher and were stored in a locked file cabinet. The cont rol group completed the questionnaire twice, during the same time period as the experi mental groups. That is, the control group completed the first copy of the questionnair e (pre-test) sometime between April and May of 2005, which was the time period for the thr ee ASC training sessions. Participants in the experimental groups and the control group co mpleted the post-test in July or August of 2005. Once all questionnaires were return ed, the researcher conducted analyses, interpreted the results, and presented the findings. Analyses Pre-Test Analyses Once all pre-test questionnaires were completed by the participants and returned to the researcher, the data were analyzed. De scriptive statistics were computed for both preand post-test data pertai ning to all of the scales. In addition, descriptive statistics were computed for the participants’ dem ographic information. Next, an exploratory factor analysis was conducted to examine th e construct-related validity of scale scores, and to reduce the number of items within ea ch scale by grouping items that were moderately to highly correlat ed with one another (Frae nkel & Wallen, 2003). For each scale and subscale that emerged from the fact or analysis, score reliability coefficients were computed using Cronbach’s alpha for each treatment group (i.e., control and experimental groups) and as a whole (Glass & Hopkins, 1996).

PAGE 68

60 To assess regional differences in responses at pre-test, a series of independent samples t -tests was used to compare participants from the rural areas and those from urban areas across the subscales that comp rised the three main sections of the questionnaire (e.g., perceived levels of knowle dge and self-efficacy, screening practices, and perceived barriers). The urban and rura l groups were not differentiated if no statistically significant diffe rences emerged. In addition, to examine pre-existing differences between treatment groups, a series of independent samples t -tests was used to compare participants from the experiment al group and those from the control group across the subscales that comprised the three main sections of the questionnaire. If no statistically significant differe nt emerged between the experimental and control groups, then it was assumed that the experimental groups and control groups did not differ on the outcome variables prior to the intervention. To determine the relationship between pediatric healthcare providers’ perceived barriers to utilizing screening instruments and their actual us e of screening instruments before completion of the ASC training, a Spearman rank correlation coefficient was conducted on the pre-test measures. A 5% le vel of significance was used to test this relationship. An effect size was interprete d if statistical significance was found. Cohen’s (1988) criteria were used to interpret effect sizes (i.e., .1 = small, .3 = medium, .5 = large). Post-Test Analyses The following research questions we re addressed in this study: Research question 1. What is the effect of the Autism System of Care (ASC) training on use of developmental and autism-sp ecific screening inst ruments by pediatric

PAGE 69

61 healthcare providers? To address this question a 2 X 2 repeated-measures ANOVA was conducted. The repeated-measures ANOVA was conducted to determine if there was a statistically significant difference in outcomes between the experimental group and control group. In addition, th e results from the repeat ed-measures ANOVA indicated whether there was a statistically significant difference in outcomes between the two time points (i.e., pre-test and post-te st), as well as if there was a statistically significant interaction between group and time (Maxwe ll & Delaney, 1990). Th e repeated-measures ANOVA, also called a split-plot design, is es pecially valuable for comparing groups across time longitudinally (Maxwell & Delane y, 1990). The between-subjects factor in this analysis was group (i.e., e xperimental group vs. control group) The within-subjects factor was time (i.e., pre-test and post-te st). Because two outcomes (i.e., dependent measures) were of interest, namely, use of developmental screening instruments (Range = 3 to 15), and use of autism-specific screening in struments (Range = 3 to 15) two sets of 2 X 2 repeated-measures ANOVA were conducted. Research question 2. What is the effect of the ASC training on the use of developmental screening instruments in regards to age of patient? To address this question, Wilcoxon’s signed rank test was undert aken. A 5% level of significance (i.e., alpha level of .05) was used (Glass & Hopkins 1996). Bonferroni’s adjustment was used to keep alpha level of significan ce at 5% (Glass & Hopkins, 1996). Research question 3. What is the effect of the Au tism System of Care training on pediatric healthcare providers’ perceived barriers to increasing the use of screening instruments and/or referri ng patients? Two sets of 2 X 2 repeated-measures ANOVA were conducted to address this question. The between-subjects factor in this design was

PAGE 70

62 group (i.e., experimental group vs control group) and perceive d levels of barriers were the dependent variables, whose preand pos t-measures served as the within-subjects factor. Research question 4. What is the effect of the Au tism System of Care training on pediatric healthcare providers’ perceived levels of knowledge related to Autism Spectrum Disorders? A 2 X 2 repeated-measures ANOVA was conducted to address this research question. The between-subjects factor was gr oup (i.e., experimental group vs. control group) and perceived levels of knowledge were the dependent variab les, whose preand post-measures served as the within-subjects factor. Research question 5. What is the effect of the Au tism System of Care training on the self-efficacy of pediatric healthcare providers regardi ng the ability accurately to screen and refer a child suspected of having an Autism Spectrum Disorder? Two separate 2 X 2 repeated-measures ANOVAs were conduc ted to address the fourth research question. The between-subjects factor was gr oup (i.e., experimental group vs. control group) and perceived levels of self-efficacy of participants were the dependent variables, whose preand post-measures served as the within-subjects factor. Research question 6. What is the relationship between pediatric healthcare providers’ perceived barriers to utilizing sc reening instruments and their actual use of developmental and autism-specific screening instruments before and after completion of the training? To address this research que stion, two Spearman rank (order) correlation coefficients were conducted (i.e ., for preand post-test scores ) to determine the degree of relationship between each pair of scores (Fraenkel & Wallen, 2003). The independent variables were level of perceived barriers, and the dependent variab les were (a) use of

PAGE 71

63 developmental screening instruments (Range = 3 to 15), and (b) use of autism-specific screening instruments (Range = 3 to 15). Research question 7. What is the relationship be tween perceived barriers to utilizing screening instruments and the use of developmental scr eening instruments in regard to age of patients befo re and after completion of th e training? To address this research question, a series of Spearman rank correlation coefficients was computed for both preand post-test scores. The independent variable was level of perceived barriers and the (seven) dependent va riables were the frequency of use of developmental screening instruments in re gard to seven age ranges of children. The Bonferroni adjustment was used to control for the in flation of Type I error. Specifically, a Bonferroni-adjusted alpha value of .007 (i.e., .05 /7) was used to reflec t the fact that seven correlations were computed.

PAGE 72

64 Chapter 4 Results Treatment of the Data The data were entered into an Excel spreadsheet by the researcher and another school psychology graduate stude nt following the completion of both the pre-test and post-test questionnaires. Each score was ente red for every participant on each individual item. Missing data were coded as a blank sp ace in the Excel document. The researcher and another school psychology gr aduate student checked the data by randomly selecting participants’ identification numbers and matching the data in the database to the entrees completed by hand by the randomly selected pa rticipant. Additionally, extreme values were checked across each participant for each it em to ensure that th e data were entered correctly. Inter-rater reliability was 100%. Missing Data Analysis At pre-test a total of 49 pediatric heal thcare providers participated, comprising 25 in the experimental group and 24 in the contro l group. At post-test a total of 26 pediatric healthcare providers participated, consisting of 13 in the experimental group and 13 in the control group. This represente d an overall completion rate of 53.1%, a completion rate of 52% for the experimental group a nd 54.2% for the control group. Of those who completed the study, only one item was not completed by one person in the experimental group (i.e., item asking to indicate the age at which they believe they are able to screen a child suspected of having an ASD). For the control

PAGE 73

65 group, every item was completed by all participan ts. For the repeated measures analysis of variance (ANOVA), a sample size of 40 was n eeded to detect a statistically significant difference between the experimental and cont rol group with a power coefficient of .80 at the 5% level of significance. Because the pos t-test sample size was 26 the statistical power for the repeated measures ANOVA was below the desired level. The power also was low for the correlation analyses (i.e., Spearman rank). Pre-Test Analyses After all data were transferred from th e Excel spreadsheet into an SPSS data editor file, the data were analyzed. Descriptiv e statistics were comput ed for both preand post-test data (see Table 1). In addition, de scriptive statistics were computed for the participants’ demographic information (see Tabl e 2). Next, a series of exploratory factor analyses was conducted to examine the underlyi ng structure of the items within each section and to reduce the dimensionality of the it ems within each section by grouping items that were moderately to highly corre lated with one anothe r (Fraenkel & Wallen, 2003).

PAGE 74

66 Table 1 Descriptive Statistics for Pre-Test and Post-Test Data Variable Pre-Test Only ( n = 23) (%) Pre-Test and Post-Test ( n = 26) (%) Chi Square (df) Age 3.78 (4) <30 21.7 23.1 31-40 26.1 26.9 41-50 26.1 30.8 51-60 13.0 19.2 >60 13.0 0.0 Gender 2.37 (1) Male 39.1 19.2 Female 60.9 80.8 Race 3.04 (5) White (NonHispanic) 69.6 80.8 Black/African American 8.7 3.8 Hispanic 8.7 11.5 Asian/Pacific Islander 4.3 3.8 Native American 0.0 0.0 Multi-Racial/Ethnic 4.3 0.0 Other 4.3 0.0 Location of Practice 2.92 (2) Rural 34.8 19.2 Suburban 34.8 26.9 Urban 30.4 53.8 Other 0.0 0.0 Profession 6.10 (4) Pediatrician 69.6 73.1 Family Practice 4.3 3.8 Registered Nurse 4.3 0.0 Nurse Practitioner 8.7 23.1 Other 13.0 0.0 Setting of Practice 5.77 (4) Hospital 21.7 19.2 Clinic 26.1 34.6 Private Practice 17.4 15.4 UniversityAffiliated Center 17.4 30.8 Other 17.4 0.0 p < .05

PAGE 75

67 Table 1 (Continued) Descriptive Statistics for Pre-Test and Post-Test Data Variable Pre-Test Only (%) ( n = 23) Pre-Test and PostTest (%) ( n = 26) Chi Square (df) Years in Practice 19.72 (20) 1-10 56.3 61.4 11-20 17.4 22.9 21-30 17.3 11.4 31-40 8.6 3.8 ASD Trainings 7.22 (6) 0 45.5 64.0 1-3 36.3 36.0 4-5 13.6 0.0 Change in Practice Trainings 2.28 (4) 0 73.7 75.0 1-3 21.1 25.0 4-7 5.3 0.0 p < .05

PAGE 76

68 Table 2 Demographic Characteristics of Experime ntal and Control Groups at Pre-Test Variable Experimental (%) (n = 25) Control (%) (n = 24) Total (%) (n = 49) Age <30 4.0 41.7 22.4 31-40 20.0 33.3 26.5 41-50 36.0 20.8 28.6 51-60 28.0 4.2 16.3 >60 12.0 0.0 6.1 Gender Male 32.0 25.0 28.6 Female 68.0 75.0 71.4 Race White (Non-Hispanic) 64.0 87.5 75.5 Black/African American 8.0 4.2 6.1 Hispanic 16.0 4.2 10.2 Asian/Pacific Islander 4.0 4.2 4.1 Native American 0.0 0.0 0.0 Multi-Racial/Ethnic 4.0 0.0 2.0 Other 4.0 0.0 2.0 Location of Practice Rural 44.0 8.3 26.5 Suburban 12.0 50.0 30.6 Urban 44.0 41.7 42.9 Other 0.0 0.0 0.0 Profession Pediatrician 72.0 70.8 71.4 Family Practice 8.0 0.0 4.1 Registered Nurse 4.0 0.0 2.0 Nurse Practitioner 4.0 29.2 16.3 Other 12.0 0.0 6.1 Setting of Practice Hospital 4.0 37.5 20.4 Clinic 52.0 8.3 30.6 Private Practice 20.0 12.5 16.3 University-Affiliated Center 8.0 41.7 24.5 Other 16.0 0.0 8.2

PAGE 77

69 Table 2 (Continued) Demographics Characteristics of Sample at Pre-Test as a Function of Treatment Group Variable Experimental (%) (n = 25) Control (%) (n = 24) Total (%) (n = 49) Years in Practice 1-10 36.0 83.3 59.2 11-20 36.0 4.2 20.4 21-30 20.0 8.3 14.3 31-40 8.0 4.2 6.1 ASD Trainings 0 32.0 75.0 53.1 1-3 48.0 20.8 34.7 4-5 8.0 4.2 6.1 Change in Practice Trainings 0 40.0 91.7 65.3 1-3 32.0 8.4 20.3 4-7 4.0 0.0 2.3 Exploratory Factor Analyses The first exploratory factor analysis was conducted on the items pertaining to the General Knowledge section (see Table 3). Us ing the eigenvalue-gre ater-than-one rule, one factor was extracted, explaining 62.19% of the variance (Kaiser, 1958).

PAGE 78

70 Table 3 Exploratory Factor Analysis for Gene ral Knowledge Scale Items (Pre-Test) ________________________________________________________________________ Item Factor Communality Coefficient 1 ________________________________________________________________________ Knowledge Area Autism Spectrum Disorders .999 .999 Early Warning Signs of ASD .821 .673 Developmental Screeners .413 .170 Autism-Specific Screeners .545 .297 ________________________________________________________________________ Trace 2.487 2.14 % of variance explained 62.19 62.19 ________________________________________________________________________ n = 49 Note: All bolded coefficients within this f actor had effect sizes greater than the cut-off value of 0.3 recommended by Lambert and Durand (1975). The second exploratory fact or analysis was conducted on the Screening Patterns section. Table 4 presents the re sults from this analysis. The factor analysis revealed two factors. The three items in Subsection A repr esented one factor, and the three items in subsection B represented anothe r factor. The first factor was named Screening Patterns: Developmental scale, and this factor explained 49.19% of th e variance. The second factor was named Screening Patterns: Autism-Specifi c scale, and this f actor explained 24.44% of the variance. The total scale explained 73.63% of the variance.

PAGE 79

71 Table 4 Exploratory Factor Analysis for Screening Scale Items (Pre-Test) ________________________________________________________________________ Item Factor Communality Coefficient 1 2 ________________________________________________________________________ Ages & Stages* .367 .483 .367 PEDS* .088 .833 .701 CSBS DP* .136 .823 .695 CHAT** .968 .089 .944 M-CHAT** .983 .145 .986 PDDST** .454 .208 .249 ________________________________________________________________________ Trace 2.948 1.467 3.943 % of variance explained 49.14 24.44 73.582 ________________________________________________________________________ n = 49 Note: All bolded coefficients within this f actor had effect sizes greater than the cut-off value of 0.3 recommended by Lambert and Du rand (1975), and had larger coefficients than in the other factor. represent developmental screening instruments ** represent autism-specific screening instruments The final exploratory factor analysis was conducted on the Potential Barriers section. This factor analysis revealed two factors. The firs t two items (i.e., Insufficient time and Lack of staff to assist with screenings) in the Potential Barriers section represented one factor and, therefor e, these items were labeled as Potential Barriers: Time and Personnel Assistance scale This factor explaine d 41.54% of the variance (Kaiser, 1958). The next four items in the se ction (i.e., Insufficient information regarding referral resources, Cost of screening inst ruments, Inadequate reimbursement, and Concern regarding emotional impact on the fa mily) represented anot her factor, and these

PAGE 80

72 items were labeled as Potential Barriers: Financia l and Emotional Costs scale (see Table 5). This factor explained 18.42% of the vari ance. The total scale explained 59.96% of the variance. The last item in the section (i.e., Be lief that clinical impression is sufficient) was not highly correlated with either factor ; therefore, this item was removed from the scale. Table 5 Exploratory Factor Analysis for Po tential Barriers Items (Pre-Test) ________________________________________________________________________ Item Factor Communality Coefficient 1 2 ________________________________________________________________________ Insufficient Time .989 .143 .999 Lack of Staff .733 .151 .560 Insufficient Info. (referral resources) .344 .510 .378 Cost of Instruments -.015 .651 .424 Inadequate Reimbursement .249 .840 .767 Concern regarding emotional impact .106 .440 .204 Belief that clinical impre ssion sufficient .234 .283 .135 ________________________________________________________________________ Trace 2.908 1.290 3.469 % of variance explained 41.54 18.420 59.958 ________________________________________________________________________ n = 49 Note: All bolded coefficients within this f actor had effect sizes greater than the cut-off value of 0.3 recommended by Lambert and Du rand (1975), and had larger coefficients than in the other factor. Score Reliability of Pre-Test Measures For each scale and subscale, score reliabi lity coefficients were computed using Cronbach’s alpha for each treatment group (i.e., contro l and experimental groups) and as a whole. For the General Knowledge scale, Cronbach’s alpha was computed for (a) Pre-

PAGE 81

73 test scores from participants in the experi mental group and the control group combined, (b) Pre-test scores from participants in the experimental group only, and (c) Pre-test scores from participants in the control group only. The same procedures were carried out for the remaining scales on the questionnaire. Overall, the Cronbach’s alpha was high for scores pertaining to all measur es except for the Screening Patterns: Developmental scale for the control group and the Potential Barrie rs: Financial and Emotional Costs scale for the control group (see Table 6). Table 6 Score Reliabilities (Cronbach’s Alpha) for a ll Measures by Treatment Group: Pre-Test ________________________________________________________________________ Scale Experimental Control All ________________________________________________________________________ General Knowledge .70 .84 .78 Screening Patterns: Developmental .83 .49 .74 Screening Patterns: Autism-Specific .92 .72 .84 Screening Patterns: Age of Patient .99 .99 .99 Potential Barriers: Time & Personnel .84 .88 .86 Potential Barriers: Fin. & Emot. Costs .82 .43 .72 ________________________________________________________________________ n = 49 Assessing Group Equivalence Urban versus rural. To determine whether there was a difference in participants’ scores based on geographic region (i.e., ur ban versus rural), independent samples t -tests were conducted. Prior to conducting these analyses, the normality assumption was evaluated. Tables 7 and 8 present the skewne ss and kurtosis coefficients for each of the pre-test scales for the urban and rural samp les, respectively. According to Onwuegbuzie and Daniel (2002), (a) standardized skewne ss (i.e., skewness divi ded by its standard error) and kurtosis coefficients (i.e., kurtosis divided by its sta ndard error) that lie within 2 suggest no serious departures from normality; (b) coefficients outside this range but

PAGE 82

74 within the 3 boundary signify slight depart ures from normality; and (c) standardized coefficients outside the 3 range indicat e important departures from normality. For the urban group, the sta ndardized skewness and kurtosi s coefficients indicated that the following five scales did not depa rt from normality: (a) General Knowledge, (b) Screening Patterns: Developmental, (c) Screening Patterns: Age of Patient, (d) Potential Barriers: Time & Personnel Assistance, and (e ) Potential Barriers: Financial & Emotional Costs (see Table 7). In contrast, the Screening Patterns: Autism-specific scale scores were both extremely positively skewed and indica ted a leptokurtic dist ribution (i.e., more peaked than the normal distribution). This finding was confirmed by the histograms (not presented). For the rural group, both the Screening Patterns: Devel opmental Scale scores and Screening Patterns: Autism-Specific scale sc ores deviated from normality; both were extremely positively skewed and indicated leptokurtic distributions (see Table 8). Because the assumption of normality was violat ed for at least one group with respect to the Screening Patterns: Developmental and Screening Patterns: Autism-Specific scales, a nonparametric independent samples t -test (i.e., Mann-Whitney) was used for these two scales. The Bonferroni adjustment was used to control for the inflat ion of Type I error. Specifically, a Bonferroni-adjusted alpha valu e of .008 (i.e., .05/6) was used. The results of these t -tests are presented in Table 9. From th ese results it can be seen that for the General Knowledge scale, a statistically si gnificant difference was found between urban and rural participants. For the remaining scales no statistically signi ficant difference was found between urban and rural participants.

PAGE 83

75 Table 7 Skewness and Kurtosis Coefficients for Pre-Test Scales: Urban Group ( n = 8) General Knowledge Screening Patterns: Developmental Screening Patterns: AutismSpecific Screening Patterns: Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs Skewness -.37 1.17 3.02 .73 -1.09 .33 Std. Error of Skewness .52 .52 .52 .54 .52 .52 Standardized Skewness -.71 2.25 5.81 1.35 -2.10 .63 Kurtosis .37 .77 8.45 -.82 .47 -1.49 Std. Error of Kurtosis 1.01 1.01 1.01 1.04 1.01 1.01 Standardized Kurtosis .37 .76 8.37 -.79 .47 -1.48 Table 8 Skewness and Kurtosis Coefficients for Pre-Test Scales: Rural Group ( n = 15) General Knowledge Screening Patterns: Developmental Screening Patterns: AutismSpecific Screening Patterns: Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs Skewness .96 2.35 2.48 -.10 -.43 -.48 Std. Error of Skewness .43 .43 .43 .43 .43 .43 Standardized Skewness 2.23 5.47 5.77 -.23 -1.00 -1.12 Kurtosis .82 6.18 5.69 -1.46 -.78 -.44 Std. Error of Kurtosis .83 .83 .83 .83 .83 .83 Standardized Kurtosis .99 7.45 6.86 -1.76 -.94 -.53

PAGE 84

76 Table 9 T-Tests Comparing Participants’ Scores Based on Geographic Region: Pre-Test ( n = 49) ________________________________________________________________________ Urban Rural Scale M SD M SD t -value U p -value ________________________________________________________________________ General Knowledge 9.00 2.08 7.27 2.03 2.88 .006 Screening (Developmental) 4.68 2.08 4.07 2.10 221.50 .131 Screening (Autism-specific) 3.53 1.61 3.43 1.01 263.50 .493 Screening (Age of Patient) 16.89 10.52 19.93 10.06 -.998 .324 Barriers (Time & Personnel) 6.53 1.81 5.77 1.89 1.395 .170 Barriers (Cost) 9.95 4.21 10.27 2.43 -.337 .738 ________________________________________________________________________ U denotes Mann-Whitney’s test statistic. Experimental versus control. To determine whether there was a difference in the selected outcomes between the experimental an d control groups, a se ries of independent samples t -tests was conducted. Prior to conduc ting this analysis, the normality assumption was evaluated. Tables 10 and 11 present the skewness and kurtosis coefficients for each of the pre-test scales for the experimental and control samples, respectively. For the experimental group, the standardized skewness and kurtosis coefficients indicated that the following four scales did not depart from normality: (a) General Knowledge, (b) Screening Patterns: Ag e of Patient, (c) Potential Barriers: Time & Personnel Assistance, and (d) Potential Barriers: Financia l & Emotional Costs Scale. In contrast, the Screening Patterns: Devel opmental scale was positively skewed and the Screening Patterns: Autism-specific scale scores were both extremely positively skewed and had a leptokurtic distributi on (i.e., more peaked than th e normal distribution). This finding was confirmed by the histograms. For the control group, only the Screening Patterns: Autism-Specific scale scores devi ated from normality. Because the assumption of normality was violated for at least one gr oup with respect to the Screening Patterns:

PAGE 85

77 Developmental scale and the Screening Patte rns: Autism-specific scale, a nonparametric independent samples t -test (i.e., Mann-Whitney) was used for these two scales. The results of these t -tests are presented in Table 12. Thes e data indicate that no statistically significant differences emerged. Table 10 Skewness and Kurtosis Coefficients fo r Pre-Test Scales: Experimental Group ( n = 25) General Knowledge Screening Patterns: Developmental Screening Patterns: AutismSpecific Screening Patterns: Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs Skewness -.12 1.69 3.54 .60 -.95 .14 Std. Error of Skewness .46 .46 .46 .47 .46 .46 Standardized Skewness -.26 3.64 7.62 1.28 -2.05 .30 Kurtosis -.48 2.47 12.00 -1.02 -.03 -1.54 Std. Error of Kurtosis .90 .90 .90 .92 .90 .90 Standardized Kurtosis -.54 2.74 13.30 -1.11 -.04 -1.71

PAGE 86

78 Table 11 Skewness and Kurtosis Coefficients for Pre-Test Scales: Control Group ( n = 24) General Knowledge Screening Patterns: Developmental Screening Patterns: AutismSpecific Screening Patterns: Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs Skewness 1.02 1.32 2.12 -.18 -.37 -.50 Std. Error of Skewness .46 .46 .46 .47 .46 .46 Standardized Skewness 2.22 2.87 4.61 -.38 -.80 -1.09 Kurtosis 1.25 .32 3.84 -1.37 -.80 .18 Std. Error of Kurtosis .92 .92 .92 .92 .92 .92 Standardized Kurtosis 1.36 .35 4.17 -1.49 -.87 .20 Table 12 T-Tests Comparing Participants’ Scores Based on Treatment Group: Pre-Test ( n = 49) ________________________________________________________________________ Experimental Control Scale M SD M SD t -value U p -value ________________________________________________________________________ General Knowledge 8.60 2.18 7.25 2.05 2.23 .030 Screening (Developmental) 4.64 2.46 3.96 1.60 257.50 .325 Screening (Autism-specific) 3.40 1.41 3.54 1.10 254.50 .157 Screening (Age of Patient) 16.88 10.19 20.71 10.12 -1.31 .198 Barriers (Time & Personnel) 6.32 1.93 5.79 1.82 .986 .329 Barriers (Cost) 8.32 3.41 8.29 2.05 .035 .972 ________________________________________________________________________ U denotes Mann-Whitney’s test statistic. Check of Normality Assumptions for Post-Test Scores Urban and rural. The skewness and kurtosis coeffici ents were computed for each of the post-test scales for the urban and rura l samples, respectivel y. For the urban group, the standardized skewness and kurtosis coeffi cients indicated that the following four

PAGE 87

79 scales did not depart from normality: (a) Screening Patterns: Developmental, (b) Screening Patterns: Age of Patient, (c) Pote ntial Barriers: Time & Personnel Assistance, and (d) Potential Barriers: Financial & Emotional Costs. In contrast, the General Knowledge scale scores had a le ptokurtic distributi on (i.e., more peaked than the normal distribution) and the Screening Patterns: Autism-specific scale scores were both positively skewed and had a leptokurtic distribution. This finding was confirmed by the histograms (not presented). For the rural group, none of the six scales departed from normality. Experimental and control. Tables 13 and 14 present the skewness and kurtosis coefficients for each of the post-test scal es for the experimental and control groups, respectively. For the experimental group, the standardized skewness and kurtosis coefficients indicated that the following five scales did not depart from normality: (a) General Knowledge, (b) Screening Patterns: De velopmental, (c) Screening Patterns: Age of Patient, (d) Potential Barriers: Time & Personnel Assistance, and (e) Potential Barriers: Financial & Emotional Costs. In contrast, the Screening Patterns: Autismspecific scale scores were positively sk ewed. This finding was confirmed by the histograms (not presented). For the control group, none of the six scales departed from normality.

PAGE 88

80 Table 13 Skewness and Kurtosis Coefficients fo r Post-Test Scales: Experimental Group ( n = 25) General Knowledge Screening Patterns: Developmental Screening Patterns: AutismSpecific Screening Patterns: Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs Skewness -.03 -.11 1.98 .53 -1.24 .22 Std. Error of Skewness .62 .62 .62 .62 .62 .62 Standardized Skewness -.05 -.17 3.20 .85 -2.02 .36 Kurtosis 1.55 -1.57 2.98 -1.07 2.01 -1.11 Std. Error of Kurtosis 1.19 1.19 1.19 1.19 1.19 1.19 Standardized Kurtosis 1.30 -1.31 2.51 -.90 1.69 -.93 Table 14 Skewness and Kurtosis Coefficients for Post-Test Scales: Control Group ( n = 24) General Knowledge Screening Patterns: Developmental Screening Patterns: AutismSpecific Screening Patterns: Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs Skewness .05 .80 1.34 -.14 -1.52 -.30 Std. Error of Skewness .62 .62 .62 .62 .62 .62 Standardized Skewness .07 1.30 2.18 -.22 -2.46 -.49 Kurtosis -.98 -.62 1.42 -1.00 1.70 -1.00 Std. Error of Kurtosis 1.19 1.19 1.19 1.19 1.19 1.19 Standardized Kurtosis -.82 -.52 1.19 -.84 1.42 -.84

PAGE 89

81 Post-Test Analyses Score Reliability of Measures For each scale and subscale, score reliabi lity coefficients were computed using Cronbach’s alpha for each treatment group (i.e., contro l and experimental groups) and as a whole. For the General Knowledge scale, Cronbach’s alpha was computed for (a) Posttest scores from participants in the experi mental group and the control group combined, (b) Post-test scores from pa rticipants in the experiment al group only, and (c) Post-test scores from participants in the control group only. The same procedures were carried out for the remaining scales on the questionnaire. Overall, the Cronbach’s alpha was high for all measures except for the General Knowledge scale for the experimental group and the Screening Patterns: Developmental for both groups and the groups combined (see Table 15). Table 15 Score Reliabilities (Cronbach’s Alpha) for a ll Measures by Treatment Group: Post-Test ________________________________________________________________________ Measure Experimental Control All ________________________________________________________________________ General Knowledge .41 .82 .77 Screening Patterns: Developmental .08 .63 .41 Screening Scale (subsection B) .76 .89 .84 Screening Patterns: Age of Patient .99 .99 .99 Potential Barriers: Time & Personnel .94 .83 .88 Potential Barriers: Fin. & Emot. Costs .82 .73 .77 ________________________________________________________________________ n = 26 The first research question in this study was: What is the effect of the Autism System of Care (ASC) trai ning on use of developmental and autism-specific screening

PAGE 90

82 instruments by pediatric healthcar e providers? To answer the fi rst part of this research question (i.e., effect of ASC training on use of developmenta l screening instruments), a 2X2 repeated measures ANOVA was conducted. Box’s M test indicated that the assumption of equality of the covariance matrices was not violated ( M = 5.24, p > .05). The repeated measures ANOVA revealed no st atistically significant interaction between time and treatment group ( F [1, 24] = 1.17, p > .05). Further, no stat istically significant main effect due to time was found ( F [1, 24] = 0.13, p > .05). Finally, no statistically significant difference was found (i.e., the be tween-subjects main effect) between the treatment and control group ( F [1, 24] = 4.92, p > .05). The experimental group’s scores decreased (i.e., indicating a decline in the us e of developmental screening instruments) whereas the control group’s scores increased—not a desira ble outcome. However, these changes were not statistically significant. Additionally, while the mean score for the experimental group decreased from 5.15 ( SD = 2.64) at pre-test to 4.85 ( SD = 1.41) at post-test, the effect size associated with th is decline was .15 which may be considered small and representing chance. To answer the second part of this resear ch question (i.e., effect of ASC training on use of autism-specific screening instrume nts), a 2X2 repeated measures ANOVA was computed. Box’s M test indicated that the assumpti on of equality of the covariance matrices appeared to be violated ( M = 11.95, p < .05). However, the fact that the sample sizes were equal did not give cause for conc ern (Stevens, 2002). The repeated measures ANOVA revealed no statistically significant interaction between time and treatment group ( F [1, 24] = 1.07, p > .05). Further, no statistically significant main effect due to time was found ( F [1, 24] = 4.76, p > .05). Finally, no statisti cally significant difference

PAGE 91

83 was found (i.e., the between-subjects main e ffect) between the treatment and control group ( F [1, 24] = 0.71, p > .05). Both the experimental and control groups’ scores increased in a similar fashion; therefore, th e change was not statis tically significant. The second research question from this st udy was: What is the effect of the ASC training on the use of developmental screening in struments in regard to age of patient? To address this question, a Wilcoxon’s signed rank test was computed. The results of this test are presented in Table 16. No statistically significant effect was found for any of the age ranges. Table 16 Wilcoxon Test for Screening Patterns: Age of Patients Scale Scores: Preand Post-Test ________________________________________________________________________ Age Positive Negative Ties Z p -value 0-6 months 5 1 7 -1.63 .10 7-12 months 6 1 6 -1.89 .06 13-18 months 5 1 7 -1.63 .10 19-24 months 6 1 6 -1.89 .06 25-36 months 6 1 6 -1.89 .06 37-48 months 4 2 7 -.33 .74 Older than 48 months 3 2 8 .00 1.00 ( n = 49) The third research question from this st udy was: What is the effect of the Autism System of Care training on pediatric hea lthcare providers’ perceived barriers to increasing the use of screening instruments a nd/or referring patient s? To address this question, two 2 X 2 repeated-measures ANOVAs were computed. Box’s M test indicated that the assumption of equality of the covariance matrices was not violated ( M = 4.26, p > .05). For the Potential Barriers: Time & Pe rsonnel Assistance Scale, the repeated measures ANOVA revealed no statistically significant interaction between time and

PAGE 92

84 treatment group ( F [1, 24] = 1.71, p > .05). Further, no statistica lly significant main effect due to time was found ( F [1, 24] = 0.00, p > .05). Finally, no stat istically significant difference was found (i.e., the between-subjects main effect) between the treatment and control group ( F [1, 24] = 1.69, p > .05). For the Potential Barriers: Financial & Emotional Costs Scale, Box’s M test indicated that the assumption of equality of the covariance matrices was not violated ( M = 6.97, p > .05). The repeated measures ANOVA re vealed no statis tically significant interaction between time and treatment group ( F [1, 24] = 3.21, p > .05). Further, no statistically significant main ef fect due to time was found ( F [1, 24] = 1.37, p > .05). Finally, no statistically signifi cant difference was found (i.e., the between-subjects main effect) between the treatment and control group ( F [1, 24] = 0.31, p > .05). For both scales, the experimental group scores decreased (i.e., indica ting that the set of barriers decreased their likelihood to impede the use of screeni ng instruments) whereas the control group scores increase d—a desirable outcome. However, these changes in scores were not statistically significant. The fourth research question was: What is the effect of the Autism System of Care training on pediatric healthcare providers ’ perceived levels of knowledge related to Autism Spectrum Disorders? A 2 X 2 re peated-measures ANOV A was conducted to address this question. Box’s M test indicated that the a ssumption of equality of the covariance matrices was not violated ( M = 0.79, p > .05). The repeated measures ANOVA revealed no statistically significant interaction between time and treatment group ( F [1, 24] = 1.13, p > .05). Both the experimental and control groups’ scores increased slightly from preto post-test; however, these increases in scores were not

PAGE 93

85 statistically significantly different from each other. However, a statistically significant main effect due to time was found ( F [1, 24] = 6.67, p < .05; 2 = .18). The effect size associated with the time main effect was small to moderate. Specifically, across groups, general knowledge of ASD was higher after the intervention ( M = 8.42, SD = 2.08) than before ( M = 7.77, SD = 2.20). Also, a statistically signi ficant difference was found (i.e., the between-subjects main effect) betw een the treatment and control group ( F [1, 24] = 11.39, p < .05; 2 = .29). The effect size associated with this difference was moderate. Specifically, across the pre-te st and post-test, the expe rimental group (M = 9.23, SE = .48) rated their general knowle dge of ASD to be higher than did the control group (M = 6.96, SE = .48). The fifth research question addressed in th is study was: What is the effect of the Autism System of Care training on the self-e fficacy of pediatric healthcare providers regarding their ability to screen accurately and refer a child suspected of having an Autism Spectrum Disorder? To answer this research question, two separate 2 X 2 repeated-measures ANOVAs were conducted. Box’s M test indicated that the assumption of equality of the covariance matrices was not violated ( M = 1.98, p > .05). With regard to the item, “Please indicate the age at which you believe it is possible to accurately screen and refer a child suspected of having an Autism Spectrum Disorder,” the repeated measures ANOVA revealed no statistically significant interaction between time and treatment group ( F [1, 23] = 0.64, p > .05). Further, no statistica lly significant main effect due to time was found ( F [1, 23] = 0.77, p > .05). Finally, no stat istically significant difference was found (i.e., the between-subjects main effect) between the treatment and control group ( F [1, 23] = 3.09, p > .05).

PAGE 94

86 With regard to the item “Please i ndicate the age at which you believe you are able to accurately screen and refer a child suspected of having an Autism Spectrum Disorder,” another repeated measures ANOVA was conducted. Box’s M test indicated that the assumption of equality of the covariance matrices was not violated ( M = 3.75, p > .05). The repeated measures ANOVA revealed no st atistically significant interaction between time and treatment group ( F [1, 22] = 0.60, p > .05). Further, no stat istically significant main effect due to time was found ( F [1, 22] = 0.48, p > .05). However, a statistically significant difference was found (i.e., the be tween-subjects main effect) between the treatment and control group ( F [1, 22] = 7.13, p < .05; 2 = .20). The effect size associated with this difference was moderate Specifically, across the pre-test and posttest, the experimental group ( M = 17.34, SE = 2.06) believed that they could accurately screen a child suspected of having an ASD at a lower age than did the control group ( M = 24.81, SE = 1.89). The sixth research question was: What is the relationship between pediatric healthcare providers’ perceived barriers to uti lizing screening instruments and their actual use of developmental and autism-specific screening instruments before and after completion of the training? To address this research question, two Spearman rank (order) correlation coefficients were c onducted (i.e., for preand posttest scores). For pre-test scores, no statistically signifi cant relationship was found betwee n barriers associated with time and personnel assistance and (a) use of developmental screening instruments ( rs = .10, p > .05), and (b) use of autism-specific screening instruments ( rs = .11, p > .05). Further, no statistically significant relations hip was found between barriers associated with financial and emotional costs and (a) the use of developmental screening

PAGE 95

87 instruments ( rs = .13, p > .05), and (b) the use of autismspecific screening instruments ( rs = .05, p > .05). For post-test scores, no statistically significant relationship was found between barriers associated with time and personne l assistance and (a) use of developmental screening instruments ( rs = -.26, p > .05), and (b) use of autism-specific screening instruments ( rs = -.12, p > .05). Further, no statistically significant relationship was found between barriers associated with financia l and emotional costs and (a) the use of developmental screening instruments ( rs = -.12, p > .05), and (b) the use of autismspecific screening instruments ( rs = -.31, p > .05). The seventh and final research questi on from this study was: What is the relationship between perceived barriers to ut ilizing screening instru ments and the use of developmental screening instruments in rega rd to age of patients before and after completion of the training? To address this fi nal research question, a series of Spearman rank correlation coefficients was comput ed for both preand post-test scores. Specifically, a Bonferroni-adjusted alpha value of .007 (i.e., .05/7) was used to reflect the fact that seven correlations were computed. No statistica lly significant relationship was found between barriers associated with time and personnel assi stance and barriers associated with financial and emotional cost s and the age of patients screened for both preand post-test scores. Tables 17 and 18 present the Spearman rank correlation coefficients for pre-test and post-test scores, respectively.

PAGE 96

88 Table 17 Spearman Rank Correlation Coefficients: Pre-Test ( n = 49) Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs 0-6 months -.13 .09 7-12 months -.12 .10 13-18 months -.10 .06 19-24 months -.11 .03 25-36 months -.15 -.03 37-48 months -.13 -.02 Older than 48 months -.15 -.08 p < .05 Table 18 Spearman Rank Correlation Coefficients: Post-Test ( n = 26) Age of Patient Potential Barriers: Time & Personnel Assistance Potential Barriers: Financial & Emotional Costs 0-6 months -.13 -.09 7-12 months -.13 -.10 13-18 months -.15 -.07 19-24 months -.15 -.07 25-36 months -.16 -.11 37-48 months -.21 -.14 Older than 48 months -.29 -.21 p < .05

PAGE 97

89 Chapter 5 Discussion Summary of Study The present study was conducted to expl ore the effectivenes s of the Autism System of Care (ASC) trainings by measuri ng change in pediatric healthcare providers’ methods of identifying young ch ildren at-risk for autism spect rum disorders (ASD). The principal investigator develope d preand post-test questionna ires to measure change in participants’ screening practices in relation to ASD. This study was novel because it was a pilot study whereby the rese archer developed a questionnai re based specifically on the ASC training to determine the training’s eff ect on participants’ methods of identifying children with ASD. A pretest-posttest nonequi valent-groups design was used. This design was chosen because the experimental and control groups were naturally assembled groups, and therefore, such a design was the mo st practical one for this study. In addition, this design enabled the researcher to gain insight on potential changes over time among the participants in the experimental group and the control group. In this chapter, a summary of results is presented, implications of the results are discussed, limitations are examined, and suggestions for future research are provided. Summary of Results Prior to examining the highlights from Ch apter 4, it is important for the reader to put the findings from this study in the prope r context. Based on the limited number of participants who completed the ASC trainings and subsequently, completed both the pre-

PAGE 98

90 and post-test questionnaires, it is difficult to de termine if the relative lack of change from preto post-test was due to an ineffectual trai ning, or if the lack of change is due mainly to the relatively small sample size in the study. Unfortunately no statistically significant differences were found when examining the preand post-test scores of the participants. Therefore, the hypotheses presented in the intr oduction of this documen t that stated that the ASC trainings would impact participants ’ practices in relation to ASD and early screening were not confirmed. Notable findings from the measures. The results from the measures administered in this study produced some interesting fi ndings. Although these findings were not statistically significant, it is believed that the findings are clinica lly significant. For the experimental group and the control group, scor es on the General Knowledge scale were higher at post-test compared to pre-test scores. In other wo rds, participants from the experimental group and the control group te nded to rate their knowledge of autism spectrum disorders higher on the post-test than their ratings on the pre-test. This was an encouraging finding as a primary goal of the study was to increase participants’ knowledge of autism spectrum disorders. As ci ted in the section of this document that reviews the related literature there are significant impli cations for children with ASD who are identified at young ages. The outcome for these children is significantly improved when they are identified and rece ive supports and servic es early in their development (Dawson et al., 2000; Dawson & Osterling, 1997; Filipe k et al., 2000; Oser & Shaw, 2001; Wetherby et al., 2004). Other notable findings from the study reve aled that across the preand post-test scores, the experimental group rated their ge neral knowledge of ASD to be higher than

PAGE 99

91 did the control group, and they believed they could accurately screen a child suspected of having an ASD at a lower age than did the control group. However, this finding is likely due to the differences in experience between the two groups. That is, the average number of years in practice for the pa rticipants in the experimental group was 17.6 years, whereas the average number of years in practice for the control group was only 6.4 years. Although it was hoped that the experimental group and the control group would be similar in all demographic variables for the purposes of this study, it is encouraging to find that in this study practitioners’ knowle dge of ASD was associated with increased with experience. This finding indicates a need for more instruction early on in physicians’ training regarding autism spectrum disorders and the importance of early identification and intervention. Material on ASD and vari ous developmental and autism-specific screening measures could be incorporated in didactic trainings th roughout medical school and/or residency training programs. Potent ial barriers to implementing these “best practices” also could be addressed early in phys icians’ training to increase their ability to overcome these frequent barri ers to early screening. The pre-test examination of score reliabilities for each scale exhibited important results. The Cronbach’s alpha was high for sc ores on all scales except for the Screening Patterns: Developmental scale and the Potent ial Barriers: Financial and Emotional Costs scale for the control group. The Screeni ng Patterns: Developmental scale assessed participants’ frequency of use of developmen tal screening instruments, and the Potential Barriers: Financial and Emotional Costs scale assessed the potential financial barriers and emotional barriers (e.g., impact of diagnosis on family) to utilizing sc reening instruments. However at post-test, the Cr onbach’s alpha was moderate for the Developmental scale

PAGE 100

92 and high for the Potential Barriers: Financial and Emotional Costs scale for the control group. Therefore, the scores on these s cales were included in this study. Sixth question. For post-test scores, no statistically signifi cant relationship was found between barriers associated with tim e and personnel assistan ce and (a) use of developmental screening instruments, a nd (b) use of autism-specific screening instruments. Further, no statistically signi ficant relationship was found between barriers associated with financial and emotional cost s and (a) the use of de velopmental screening instruments, and (b) the use of autism-specific screening instruments. Although none of these relationships were statis tically significant, it should be noted that two of these correlations (i.e., the use of developmental sc reening instruments and time and personnel assistance barriers, and the use of autism-sp ecific screening instruments and the financial and emotional costs barriers) appear to repr esent a non-trivial (i.e., moderate) association. Seventh question. At pre-test, the relati onship between Poten tial Barriers: Time and Personnel Assistance and the use of deve lopmental screening instruments at each age level of patients (e.g., 0-6 mont hs, 7-12 months) was both sta tistically non-significant and small. However, the fact that all seven corre lations were negative is notable because it suggests a consistent inverse relationship betw een these two variable s regardless of the age of the patient. Furthermore, although at post-test the relationship between both Potential Barriers scales and the use of deve lopmental screening in struments at each age level of patients was non-significant and small, it is notable that all 14 correlations were negative. Implementation integrity. The design of this study was se lected with the intention that the implementation would be consistent and stable across training sessions. Fortunately,

PAGE 101

93 no deviations from the original plan occu rred, as 100% of the components were fully covered across the three training sessions. Therefore, the implementation integrity in this study was strong as each part icipant did receive the same type of training. Effectiveness of the intervention. Overall, the results of this study revealed no statistically significant effects from the ASC trainings ba sed on participants’ responses to the Pediatric Healthcare Provi der Self-Report Questionnaire. However, because of the small sample size, it cannot be definitively known whether these resu lts were due to the lack of effectiveness of the trainings. Directions for future interventions and research will be discussed later in this chapter. Implications of the Results Although this study did not produce statisti cally significant findings regarding the effect of the ASC trainings on pediatric hea lthcare providers’ practice, some implications from this study have emerged. For example, it is important to point out that there were positive changes in participants’ scores from preto post-test. For example, on average, participants from the experimental group and the control group rate d their use of autismspecific screeners higher at post-test as compared to pre-test. Furthermore, when examining use of developmen tal screening instruments regarding age of patient, on average, participants in the experimental group screened patients in all age categories more frequently than did participants from the control group. This trend might suggest that the ASC training had an impact on the practitioners’ screening practices. After completion of the training whereby participan ts learned about the importance of early identification and use of screen ing instruments with all young patients, these practitioners increased their frequency of use of screen ing instruments with patients of all ages.

PAGE 102

94 However, there were differences betw een experimental group and control group regarding frequency of use of screening inst ruments at pre-test. This finding may be partly explained by the difference in years in practice between the experimental group and the control group as participants in th e experimental group had significantly more years in practice compared to the control group. Through e xperience practicing medicine, practitioners may become more aware of the need for sc reening their young patients. Given the substantial literature that supports the positive effects of ea rly identification of children with ASD, it is encouraging that ex perienced practitioners are screening their patients. However, if practiti oners understand the importance of early identification early in their own training, more practi tioners will be able to iden tify more children earlier in their development. When examining perceived barriers to screening children, participants in the experimental group had a slight change in pr eto post-test scores which indicated a decreased likelihood of the barriers preventing the utilization of sc reening instruments. Specifically, on the Potential Barriers: Ti me & Personnel Assistance scale, scores decreased from 6.77 to 6.38 for the experime ntal group. On the Potential Barriers: Financial & Emotional Costs scale, scor es decreased from 8.62 to 7.15 for the experimental group. For the control group, sc ores on both of these scales actually increased slightly from preto post-tes t (i.e., 5.62 to 6.00, and 8.23 to 8.54, respectively). This is a positive trend because it suggests th at the ASC training may have had an effect on participants’ perceived barriers to utilizi ng screening instruments with their patients. In other words, after completion of the traini ng, participants did not rate the barriers as high in terms of hindering screening practices.

PAGE 103

95 Participants from the experimental gr oup and the control group had changes in scores on the General Knowledge scale, and although the changes were not statistically significant, there was a positive trend. For example, experimental group scores increased from 8.77 to 9.69 from preto post-test, indicating an increase in their perceived knowledge of ASD. The control group scor es increased, from 6.77 to 7.15, indicating a slight increase in their perceived knowle dge of ASD as well. Because there was an increase on the General Knowledge scale fo r the experimental group and the control group, this suggests that partic ipants from both groups had an increase in their knowledge related to ASD. This may indicate that practitioners are gaining knowledge regarding ASD from multiple sources. It would be adva ntageous to ascertain in what capacity practitioners are acquiring this knowledge so that these efforts can be strengthened and continued. Limitations Although the initial intent of the Autism System of Care grant was to increase awareness of ASD and autism-specific screeners, the objective of the grant changed during the course of the fundi ng year. A Health Care Task Force was responsible for the funding and was the main driver of changing th e focus of the grant. Therefore, the focus switched from autism-specific information and screening instruments to general developmental screening instruments and ch anging practice. Therefore, the training sessions ultimately focused less on autism-sp ecific instruments, and more on general developmental screeners and changing screen ing practice. This change in focus may explain in part why statistic ally significant changes were not found through this study.

PAGE 104

96 The original goal of the study was to secure a minimu m of 100 professionals to participate in the ASC trainings, and subsequen tly in this study. It was anticipated that virtually all participants in the trainings would complete the pre-test questionnaire. However, it was expected that approximatel y 50 participants would complete both the preand post-questionnaire becau se research demonstrates th e average response rate to mail surveys for physicians is 54% (Asch et al., 1997). Unfortunately, although the grant for the ASC trainings specified that a mini mum of 100 practitioners would be trained, this number was not obtained. Therefore, the number of potential participants available for this study was decreased significantl y. Out of the 36 tota l practitioners who participated in the ASC trainings, 25 comp leted the pre-test questionnaire and 13 completed both the preand post-test questi onnaires. Although this study acquired a very similar response rate compared to the averag e response rate for physicians (52% versus 54%), the small overall sample size was a signi ficant limitation of this study. If a larger initial number of practitioners had completed the ASC training, this would have been an adequate response rate and might have pr oduced some more stat istically significant results. Although Halfon et al. (2000) found that 65% AAP members surveyed reported less than adequate training, oftentimes it is difficult to attract physicians to trainings for a multitude of reasons. One primary reason is lack of time to fit a training session into an already busy work schedule. The literature in dicates that less than half of physicians agree that there is adequate time to perf orm developmental screenings during typical well-child visits (Sices et al ., 2003). Another difficulty may stem from the belief that one

PAGE 105

97 can identify children with developmental delays without the use of screening instruments and therefore, they do not need addi tional training (Sices et al., 2003). Several methods were implemented in an attempt to attract participants. To address the issue of time constraints, the tr ainings were scheduled at times that were thought to be most convenient for practitioners such as in the evening after their work day was complete, or during their lunch hour To address the pot ential belief that practitioners already are know ledgeable about ASD and early screening instruments, pediatricians who took part in the AS C workgroup attempted to communicate the importance of the ASC trainings to colleagues. Furthermore, in an attempt to acquire more participants, physicians were awarde d one Continuing Medi cal Education (CME) credit for completing the ASC training. Regi stered nurses and nurse practitioners who completed the training were awarded one C ontinuing Education Unit (CEU). In addition, a complimentary dinner was provided at each training. However, based on the small number of pediatric healthcare providers who participated in the Autism System of Care training, it is apparent that more must be unde rtaken to attract provide rs to trainings that will enable them to improve their current pract ices. The small sample size was a threat to external validity because the findings cannot be generalized to the population. Also, the small sample size limited statistical power and therefore made it difficult to find statistically significant results. Another limitation of this study was the method by which participants were placed in groups (i.e., experimental and cont rol). Random selection and assignment were not possible in the current st udy because participants who re gistered for the ASC training were automatically placed in the experimental group, wherea s the control group consisted

PAGE 106

98 of those practitioners who we re not invited to participat e in the ASC training for the singular reason that the trainings were not held near their geographic region. Therefore, the study cannot claim a true experimental design; thus the study utilized a quasiexperimental design. Differential selection of participants, also known as selection bias, presented a threat to internal validity becau se there was an important distinction between participants from the experimental group and participants from th e control group (Best & Kahn, 2003). That is, there was a statistically significant difference in both age and number of years in practice be tween the two groups, with part icipants in the experimental group being older and possessing more years in practice compared to the control group. This difference between groups could have impact ed the results of this study; therefore, it cannot be determined definitively whether the outcome data obtained from the groups were due to the ASC training, selection bi as, or a combination of the two factors. Furthermore, it is important to point out some discrepancies in the demographics of the participants from this study. Significan tly more women particip ated in the training compared to men (71.4% vs. 28.6%). In addi tion, the participants were predominately White (Non-Hispanic) and de fined themselves as pedi atricians (75.5% and 71.4%, respectively). Because of the skewed demogr aphics, the results of this study may not generalize to other types of pediatric healthcare providers (e.g., males, non-Whites, and nurse practitioners). It is important to mention that this was a pilot of the Autism System of Care training. Therefore, it is hypothe sized that issues may arise with the suitability of the material to be presented in the training. Specifically, it may be beneficial to obtain feedback from participants to determine whic h material was least effective and could be

PAGE 107

99 removed from the training, and what was most effective and could th erefore be expanded. Also, it would be advantageous to obtain feedback regarding the group and individual activities. Then, as needed, the ASC traini ng could be reworked so that it is most effective in supporting practitioners in cha nging their practices. Finally, instrumentation may have been a threat to the internal vali dity of this study. Because the questionnaire was designed to obtain information from partic ipants via self-reports, the accuracy of the data cannot be verified or known with co mplete confidence. However, the score reliability of each scale was assessed via Cronbach’s alpha. Also, construct-related validity was examined via factor analysis. Considerations for Future Research First and foremost, to determine whether the Autism System of Care training is truly effective in facilitati ng pediatric healthcare providers to change their practice regarding early screenin g, the trainings must be carried out with a larger number of participants. It also could be beneficial to have the partic ipants complete the post-test questionnaire at an additional time severa l months following the original post-test administration. As the literature has shown, ch ange is a slow process; therefore, it is likely that more significant changes woul d occur over a longer duration of time. Furthermore, in addition to the use of a preand post-test questionna ire, qualitative data could be collected through semi-structured in terviews with participants to ascertain supplementary information regarding their knowledge of ASD, self-efficacy, use of screening instruments, and the potential barriers to the routine use of these instruments. Given the importance of early identifi cation and intervention for children with developmental delays such as autism spectrum disorders, it is critic al that professionals

PAGE 108

100 who are in the position to identify these children obtain the k nowledge and resources necessary to provide children with the supports and services needed as early in their development as possible. Based on the results of this study, it is unclear whether or not the ASC trainings would have been effectiv e had there been a larger sample size. However, based on previous literature on th e effectiveness of trainings in changing practice, it may be beneficial to consid er other methods of changing practice. An effectual start may be to incorpor ate training regarding ASD, early warning signs, and the importance of early, routine sc reening into medical school training as well as residency training programs. This effort could involve the combination of didactic training as well as supervisors and chief re sidents modeling these “best practices” with the use of routine, early scr eenings of all children. Furthermore, information regarding the importance of early identification of ASD could be disseminated during Grand Rounds, journal clubs, and other types of mee tings with medical students and residents. This information also could be disseminated and discussed at na tional conferences and through newsletters, brochures, or handouts. Final Thoughts The evaluation of the Autism System of Care training provided the principal investigator with a novel opportunity to unders tand the effect that the intervention had on pediatric healthcare providers’ practice. While the results of the study did not suggest that the ASC trainings had any major impact on pr actitioners’ screening practices with young children, encouraging trends we re found through the study. It is critical that pediatric healthcare providers are armed with the know ledge and resources necessary to identify children with developmental delays early in their development. An abundance of

PAGE 109

101 literature supports the positive impact of early identification a nd intervention on young children’s developmental outcomes. Alt hough it is recommended by the American Academy of Pediatrics (AAP) th at all infants and children are screened for developmental delays or disabilities (AAP, 2001), the litera ture points out a numb er of barriers that prevent practitioners from carryi ng out these “best practices.” It is quite clear that changes must o ccur for practitioners to use developmental screeners routinely with thei r young patients. To assist prac titioners with this change, knowledge and practical support must be provide d. Therefore, it may first be necessary to identify and overcome the barriers to acqui ring this knowledge and support, whether in the form of trainings or otherwise. Then, the barriers to utilizing screening instruments to identify children with ASD and other devel opmental disabilities can be dismantled, and change in practice can truly take place.

PAGE 110

102 References American Academy of Pediatrics. (2000). Primary care service areas (PCSAs) for Florida and surrounding areas Retrieved January 6, 2005, from http://www.aap.org/mapping/print/FL/FL.pdf. American Academy of Pediatrics. (2001). De velopmental surveillanc e and screening of infants and young children. Pediatrics, 108, 192-195. American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental Disorders (4th ed., text revision). Washington, DC: American Psychiatric Association. Asch, D. A., Jedrziewski, M. K., & Christ akis, N. A. (1997). Response rates to mail surveys published in medical journals. Journal of Clinical Epidemiology, 50, 1129-1136. Baird, G., Charman, T., Baron-Cohen, S., Cox, A ., Swettenham, J., Wheel wright, S. et al. (2000). A screening instrument for autism at 18 months of age: A six-year followup study. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 694-702. Bandura, A. (1994). Self-efficacy. In V. S. Ramachaudran (Ed.), Encyclopedia of human behavior (Vol. 4, pp. 71-81). New York: Academic Press. (Reprinted in H. Friedman [Ed.], Encyclopedia of mental health San Diego: Academic Press, 1998).

PAGE 111

103 Baron-Cohen, S., Allen, J., & Gillberg, C. ( 1992). Can autism be detected at 18 months? The needle, the haystack, and the CHAT. British Journal of Psychiatry, 161, 839843. Baron-Cohen, S., Cox, A., Baird, G., Swettenham, J., Nightingale, N., Morgan, K., et al. (1996). Psychological markers in the dete ction of autism in infancy in a large population. The British Journal of Psychiatry, 168 158-163. Bertrand, J., Mars, A., Boyle, C., Bove, F., Yeargin-Allsopp, M., & Decoufle, P. (2001). Prevalence of Autism in a United States Population: The Brick Township, New Jersey, Investigation. Pediatrics, 108, 1155-1161. Best, J. W., & Kahn, J. V. (2003). Research in education Boston: Pearson Education, Inc. Bettelheim, B. (1967). The empty fortress: Infantile autism and the birth of self. New York: Free Press. Bricker, D., & Squires, J. (1999). Ages and stages questionnair es: A parent-completed, child-monitoring system (2nd ed.) Baltimore, MD: Paul H. Brookes. Bristol-Power, M. M., & Spinella, G. (1999) Research on screening and diagnosis in autism: A work in progress. Journal of Autism and Developmental Disorders, 29, 435-438. Chakrabarti, S., & Fombonne, E. (2001). Pervas ive developmental disorders in preschool children. Journal of the American Medical Association, 285, 3093-3099. Charman, T., Baron-Cohen, S., Baird, G., Cox, A., Wheelwright, S., Swettenham, J., et al. (2001). Commentary: The modified checklist for autism in toddlers. Journal of Autism and Developmental Disorders, 31, 145-151.

PAGE 112

104 Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, N.J.: Erlbaum. Dawson, G., Ashman, S., & Carver (2000). The role of early experience in shaping behavioral and brain development and its implications for social policy. Development and Psychopathology, 12, 695-712. Dawson, G., & Osterling, J. (1997). Early inte rvention in autism. In M. J. Guralnick (Ed.), The effectiveness of early intervention (pp. 307-326). Baltimore: Brookes. Dobos, Jr., A. E., Dworkin, P. H., & Bernstein, B. A. (1994). Pediatricians' approaches to developmental problems: Has the gap been narrowed? Journal of Developmental and Behavioral Pediatrics, 15, 34-38. Dworkin, P. H. (1989). British and American recommendations for developmental monitoring: The role of surveillance. Pediatrics, 84, 1000-1010. Filipek, P. A., Accardo, P. J., Baranek, G. T., Cook, Jr., E. H., Dawson, G., Gordon, B., et al. (1999). The screening and diagnos is of autistic sp ectrum disorders. Journal of Autism and Developmental Disorders, 29 439-484. Filipek, P. A., Accardo, P. J., Ashwal, S., Ba ranek, G. T., Cook, Jr., E. H., Dawson, G., et al. (2000). Practice Parameter: screening and diagnosis of autism. Report of the quality standards subco mmittee of the American Academy of Neurology and the Child Neurology Society. Neurology, 55, 468-479. Fombonne, E. (2003). Epidemiological surveys of autism and other pervasive developmental disorders: An update. Journal of Autism and Developmental Disorders, 33, 365-382.

PAGE 113

105 Fraenkel, J. R., & Wallen, N. E. (2003). How to design and evaluate research in education. New York: McGraw-Hill. Gillberg, C., & Wing, L. (1999). Autism: Not an extremely rare disorder. Acta Psychiatrica Scandinavica, 99, 399-406. Glascoe, F.P. (1998). Collaborating with parents: Using parents’ evaluation of developmental status to detect and address developmental and behavioral problems Nashville, TN: Ellsworth & Vandermeer Glass, G. V., & Hopkins, K. D. (1996). Statistical methods in education and psychology. Needham Heights, MA: Allyn & Bacon. Harris, S. L., & Handleman, J. S. (2000). Age a nd IQ at intake as predictors of placement for young children with autism: A fourto six-year follow-up. Journal of Autism and Developmental Disorders, 30, 137-142. Halfon, N., Hochstein, M., Sareen, H., O'Conno r, K., Inkelas, M., & Olson, L. (2001). Barriers to the provision of developmen tal assessments during pediatric health supervision: American Academy of Pediatrics Periodic Survey of Fellows. Howlin, P., & Moore, A. (1997). Diagnosis of autism: A survey of over 1200 patients in the UK. Autism, 1, 135-162. Kaye, J. A., Melero-Montes, M. M., & Jic k, H. (2001). Mumps, measles, and rubella vaccine and the incidence of autism r ecorded by general practitioners: A time trend analysis. British Medical Journal, 322, 460-463. Klinger, L. G., Dawson, G., & Renner, P. ( 2003). Autistic Disorder. In E. J. Mash, & R. A. Barkley (Eds.). Child Psychopathology (2nd ed.; pp. 409-454). New York: Guilford Press.

PAGE 114

106 Lord, C. (1995). Follow-up of two-year-olds referred for possible autism. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 36 1365-1382. Lotter, V. (1966). Epidemiology of autistic cond itions in young children: I. Prevalence. Social Psychiatry, 1, 124-137. Maxwell. S. E., & Delaney, H. D. (1990). Designing experiments and analyzing data: A model comparison perspective Belmont, CA: Wadsworth. Mundy, P., Sigman, M., & Kasari, C. (1990). A longitudinal study of joint attention and language development in autistic children. Journal of Autism and Developmental Disorders, 20, 115-128. National Research Council (2001). Educating children with autism Committee on Educational Interventions for Children w ith Autism. Division of Behavioral and Social Sciences and Education. Wa shington, DC: National Academy Press. Nicolson, R., & Szatmari, P. (2003). Genetic and neurodevelopmental influences in autistic disorder. Canadian Journal of Psychiatry, 48 526-537. Onwuegbuzie, A. J., & Daniel, L. G. (2002) Uses and misuses of the correlation coefficient. Research in the Schools 9 (1), 73-90. Oser, C., & Shaw, E. (2001). Early identific ation of young children with autism spectrum disorders: Improving practices. Unpublis hed manuscript, University of North Carolina, Chapel Hill, NC. Powell, J. E., Edwards, A., Edwards, M., Pandit, B. S., Sungum -Paliwal, S. R., & Whitehouse, W. (2000). Changes in the incidence of childhood autism and other autistic spectrum disorders in preschool children from two areas in the West Midlands, UK. D evelopmental Medicine and Child Neurology, 42, 624-628.

PAGE 115

107 Prater, C. D, & Zylstra, R. G. (2002). Autism: A medical primer. American Family Physician, 66, 1667-1674. Robins, D. L., Fein, D., & Barton, M. L. (1999). The Modified Checklist for Autism in Toddlers (M-CHAT) Storrs, CT: University of Connecticut. Robins, D. L., Fein, D., Barton, M. L., & Green J. A. (2001). The modified checklist for autism in toddlers: An initial study inve stigating the early de tection of autism and pervasive developmental disorders. Journal of Autism and Developmental Disorders, 31 131-144. Rutter, M. (2000). Genetic studies of autis m: From the 1970’s into the millennium. Journal of Abnormal Child Psychology, 28, 3-14. Scambler, D., Rogers, S. J., & Wehner, E. A. (2001). Can the checklist for autism in toddlers differentiate young children with autism from those with developmental delays? Journal of the American Academy of Child and Adolescent Psychiatry, 40, 1457-1463. Shonkoff, J. E., Dworkin, P. H., Leviton, A ., & Levine, M. D. (1979). Primary care approaches to developmental disabilities. Pediatrics, 64, 506-514. Sices, L., Feudtner, C., McLaughlin, J., Dr otar, D., & Williams, M. (2003). How do primary care physicians identify young ch ildren with developmental delays? A national survey. Journal of Developmental & Behavioral Pediatrics, 24 409417.

PAGE 116

108 Siegel, B. (1998, June). Early screening and diagnosis in autism spectrum disorders: The Pervasive Developmental Disorders Screening Test (PDDST) Paper presented at the National Institute of Health State of the Science in Autism, Screening, and Diagnosis Working Conference, Bethesda, MD. Sigman, M., Ruskin, E., Arbeile, S., Corona, R., Dissanayake, C., Espinosa, M., et al. (1999). Continuity and change in the soci al competence of children with autism, down syndrome, and developmental delays. Monographs of the Society for Research in Child Development, 64(1), 1-114. Sontag, J. C. (1996). Toward a comprehens ive theoretical framework for disability research: Bronfenbrenner revisited. Journal of Special Education, 30, 319-344. Stone, W., Evon, B. L., Ashford, L., Brissie, J., Hepburn, S. L., & Coonrod, E. E. (1999). Can autism be diagnosed accurately in children under 3 years? Journal of Child Psychology and Psychiatry, 40, 219-226. Stone, W., Ousley, O., Yoder, P., H ogan, K., & Hepburn, S. (1997). Nonverbal communication in 2and 3-y ear old children with autism. Journal of Autism and Developmental Disorders, 27, 677-696. Strock, M. (2004). Autism Spectrum Disorders (Pervasive Developmental Disorders) NIH Publication No. NIH-04-5511, National Institute of Mental Health, National Institutes of Health, U.S. Department of Health and Human Services, Bethesda, MD. Retrieved December 9, 2004, from http://www.nimh.nih.gov/publicat/ autism.cfm.

PAGE 117

109 Szatmari, P., Jones, M. B., Zwaigenbaum, L ., MacLean, J. E. (1998). Genetics of autism: Overview and new directions. Journal of Autism and Developmental Disorders, 28, 5168. Teitelbaum, P., Teitelbaum, O ., Nye, J., Fryman, J., & Maur er, R. G. (1998). Movement analysis in infancy may be usef ul for early diagnosis of autism. Proceedings of the National Academy of Sciences of the United States of America, 95, 1398213987. Turner, M. (1999). Annotation: Repetitive beha vior in autism: A review of psychological research. Journal of Child Psychology and Psychiatry, 40, 839-849. U.S. Department of Education. (2001). Twenty-third annual report to Congress on the implementation of the Individuals with Disabilities Education Act. Washington, D.C.: Author. Wetherby, A., Allen, L., Cleary, J., Kublin, K., & Goldstein, H. (2002). Validity and reliability of the Communication and Sy mbolic Behavior Scales Developmental Profile with very young children. Journal of Speech, Language, & Hearing Research, 45, 1202-1218. Wetherby, A., & Prizant, B. (1993). Communication and symbolic behavior scales – normed edition Baltimore, MD: Paul H. Brookes. Wetherby, A., & Prizant, B. (2002). Communication and symbolic behavior scales developmental profile – first normed edition Baltimore, MD: Paul H. Brookes. Wetherby, A., Prizant, B., & Hutchinson, T. (1998). Communicative, social-affective, and symbolic profiles of young children w ith autism and pervasive developmental disorder. American Journal of Speech-Language Pathology, 7, 79-91.

PAGE 118

110 Wetherby, A., Woods, J., Allen, L., Cleary, J ., Dickinson, H., & Lord, C. (2004). Early indicators of autism spectrum diso rders in the second year of life. Journal of Autism and Developmental Disorders, 34, 473-493. Wimpory, D. C., Hobson, R. P., Williams, J. M. G., & Nash, S. (2000). Are infants with autism socially engaged? A study of recent retrospective parental reports. Journal of Autism and Developmental Disorders, 30, 525–536. Wing, L., & Gould, J. (1979). Severe impairment s of social interac tion and associated abnormalities in children: Epidemiology and classification. Journal of Autism and Developmental Disorders, 9, 11-29. Yang, M., & Goldstein, H. (1999). The use of data for school improvement purposes. Oxford Review of Education, 25 469-483. Yeargin-Allsopp, M., Rice, C., Karapurkar, T ., Doernberg, N., Boyle, C., & Murphy, C. (2003). Prevalence of autism in a U.S. metropolitan area. JAMA: The Journal of the American Medical Association, 289, 49-55.

PAGE 119

111 Appendices

PAGE 120

112 Appendix A: Autism System of Care Flyers Statewide Autism System of Care This publication was commissioned, funded by the Florida Developmental Disabilities Council, Inc., and produced through funding provided by the United States Department of Health and Human Services, Administration for Developmental Disabilities Screening and Surveillance of A utism and Related Disabilities How to Change One’s Clinical Practice Presenter Quimby E. McCaskill, M.D., MPH, FAAP Assistant Professor of Pediatrics Associate Director of the Community Pediatrics Training Initiative at the University of Florida 6 p.m. to 7 p.m., Wednesday, May 4, 2005 University of South FloridaFlorida Me ntal Health Institute Westside A & B 13301 Bruce B. Downs Blvd. Tampa Dinner will be provided Learner Objectives Discuss why early screening and surveillance are important. Define red flags of auti sm spectrum disorders. Review developmental screening tools. List barriers preventing change in practice. Describe model for improving screening practices. Create aim statement fo r changing practice. Develop next steps to initiate practice change. This activity has been planned and implemented in acco rdance with the Essential Areas and Policies of the Accreditation Council for Continuing Medical Education (A CCME) through the joint sponsorship of the University of South Florida College of Medicine and USF Florida Mental Health Insitute. The University of South Florida College of Medicine is accredited by the ACCME to provide continuing me dical education for physicians. The University of South Florida College of Medicine desi gnates this educational activity for a maximum of 1.0 category 1 credits towards the AMA physician’s Recognition Award. Each physician should claim only those credits that he/she actually spent in the activity This activity for 1 contact hour is provided by the University of South Florida Colle ge of Nursing, which is accredited as a provider of continuing education in nursi ng by the American Nurses Credentialing Center’s Commission on Accreditation. Each nurse should claim only those hours of credit that he/she actually spent in the educational activity. RSVP by 4-27-05 to Craig Silverstein at 813-974-6464 or csilverstein@fmhi.usf.edu Sponsored by College of Medicine and College of Nursing

PAGE 121

113 Appendix A: (Continued) Statewide Autism System of Care This publication was commissioned, funded by the Florida Developmental Disabilities Council, Inc., and produced through funding provided by the United States Department of Health and Human Services, Administration for Developmental Disabilities Screening and Surveillance of Au tism and Related Disabilities How to Change One’s Clinical Practice Presenter Flora Robinson, M.D. Developmental Pediatrician University of South Florida-All Children’s Hospital 3:00 p.m. 4:00 p.m., Wednesday, May 18, 2005 Duval County Health Department 515 W. 6th Street, Jacksonville (2nd Floor, Conference Room A & B) Learner Objectives Discuss why early screening and surveillance are important. Define red flags of autism spectrum disorders. Review developmental screening tools. List barriers preventi ng change in practice. Describe model for improving screening practices. Create aim statement for changing practice. Develop next steps to initiate practice change. This activity has been planned and implemented in accordance with the Essential Areas and Policies of the Accreditation Council for Continuing Medical E ducation (ACCME) through the joint sponsorship of the University of South Florida College of Medi cine and USF Florida Mental Health Insitute. The University of South Florida College of Medicine is accredited by the ACCME to provide continuing medical education for physicians. The University of South Florida College of Medicine designates this educational activity for a maximum of 1.0 category 1 credits towards the AMA physician’s Recognition Award. Each physician should claim only those credits that he/she actually spent in the activity This activity for 1 contact hour is provided by the University of Sout h Florida College of Nursing, which is accredited as a provider of continuing education in nursing by the American Nurses Credentialing Center’s Commission on Accreditation. Each nur se should claim only those hours of cr edit that he/she actually spent in the educational activity. RSVP by 4-28-05 to Craig Silverstein at 813-974-6464 or csilverstein@fmhi.usf.edu Sponsored by College of Medicine and College of Nursing

PAGE 122

114 Appendix A: (Continued) Statewide Autism System of Care This publication was commissioned, funded by the Florida Developmental Disabilities Council, Inc., and produced through funding provided by the United States Department of Health and Human Services, Administration for Developmental Disabilities Screening and Surveillance of Au tism and Related Disabilities How to Change One’s Clinical Practice Presenter Quimby E. McCaskill, M.D., MPH, FAAP Assistant Professor of Pediatrics Associate Director of the Community Pediatrics Training Initiative at the University of Florida Noon to 1 p.m., Wednes day, June 1, 2005 Hendry Regional Medical Center Conference Room 500 West Sugarland Highway (State Road 27) Lunch will be provided Learner Objectives Discuss why early screening and surveillance are important. Define red flags of auti sm spectrum disorders. Review developmental screening tools. List barriers preventing change in practice. Describe model for improving screening practices. Create aim statement fo r changing practice. Develop next steps to initiate practice change. This activity has been planned and implemented in acco rdance with the Essential Areas and Policies of the Accreditation Council for Continuing Medical Education (A CCME) through the joint sponsorship of the University of South Florida College of Medicine and USF Florida Mental Health Insitute. The University of South Florida College of Medicine is accredited by the ACCME to provide continuing me dical education for physicians. The University of South Florida College of Medicine desi gnates this educational activity for a maximum of 1.0 category 1 credits towards the AMA physician’s Recognition Award. Each physician should claim only those credits that he/she actually spent in the activity This activity for 1 contact hour is provided by the University of South Florida College of Nursing, which is accredited as a provider of continuing educati on in nursing by the American Nurses Credentialing Center’s Commission on Accreditation. E ach nurse should claim only those hours of credit that he/she actually spent in the educational activity. Presented in cooperation with Hendry Regional Me dical Center and Hendry County Health Department RSVP on or before Friday, May 13 to Sue Reese (863) 674-4056, ext. 157 or Suzette_Reese@doh.state.fl.us College of Medicine and College of Nursing

PAGE 123

115 Appendix B: Pediatric Healthcare Provider Self-Report Questionnaire Pediatric Healthcare Provide r Self-Report Questionnaire Purpose : As part of a multidisciplinary co llaborative through the University of South Florida, this instrument was designed to obtain information regarding screening and referral patterns of hea lthcare providers as well as potential barriers to early screening and referral fo r Autism Spectrum Disorders (ASD). ASD includes : autistic disorder, Asperger syndrome, pervasive de velopmental disordernot otherwise specified (PDDNOS), Rett syndrome, and childhood disintegrative disorder. General Information Please rate your overall knowledge of: Poor Fair Good Excellent 1. Autism Spectrum Disorders (ASD) 1 2 3 4 2. Early warning signs of ASD 1 2 3 4 3. Developmental screening inst ruments* (e.g., ASQ, PEDS) 1 2 3 4 4. Autism-specific screening inst ruments* (e.g., CHAT, M-CHAT) 1 2 3 4 *Note. This refers to use of the entire instrument and does not include us e of only a few items from the instrument. 5. Please indicate the age at which you believe it is possible to accurately screen and refer a child suspected of having an Autism Spectrum Disorder: ______ months 6. Please indicate the age at which you believe you are able to accurately screen and refer a child suspected of having an Autism Spectrum Disorder: ______ months II. Screening Patterns A) How often do you use the following developmental screening instruments: Never (0%) Rarely (1-19%) Sometimes (20-49%) Usually (50-99%) Always (100%) 1. Ages & Stages Questionna ires (Bricker & Squires) 1 2 3 4 5 2. Parents’ Evaluations of Developmental St atus (PEDS; Glascoe) 1 2 3 4 5 3. Communication and Symbolic Behavior Scales Developmental Profile In fant-Toddler Checklist (CSBS DP; Wetherby & Prizant) 1 2 3 4 5 B) How often do you use the following autism-specific screening instruments: Never (0%) Rarely (1-19%) Sometimes (20-49%) Usually (5099%) Always (100%) 1. Checklist for Autism in Toddlers (CHAT; BaronCohen, Allen, & Gillberg) 1 2 3 4 5 2. Modified Checklist for Autism in Toddlers (M-CHAT; Robins, Fein, & Barton) 1 2 3 4 5 3. Pervasive Developmental Disorder Screening Test (PDDST; Siegel) 1 2 3 4 5 C) How often do you use developmental screening instruments with patients in the following age ranges: Never (0%) Rarely (1-19%) Sometimes (20-49%) Usually (5099%) Always (100%) 0-6 months 1 2 3 4 5 7-12 months 1 2 3 4 5 13-18 months 1 2 3 4 5 19-24 months 1 2 3 4 5 25-36 months 1 2 3 4 5 37-48 months 1 2 3 4 5 Older than 48 months 1 2 3 4 5

PAGE 124

116 Appendix B: (Continued) III. Potential Barriers to Utilizing Screening Instruments Please indicate the extent to which each item is likely or unlikely to impede your use of screening instruments (e.g., PEDS). Unlikely Somewhat Unlikely Somewhat Likely Very Likely 1. Insufficient time 1 2 3 4 2. Lack of staff to assist with screenings 1 2 3 4 3. Insufficient information rega rding referral resources 1 2 3 4 4. Cost of screening instruments 1 2 3 4 5. Inadequate reimbursement 1 2 3 4 6. Concern regarding emotional impact on the family 1 2 3 4 7. Belief that clinical impr ession is sufficient 1 2 3 4 IV. Demographics Age: <30 31-40 41-50 51-60 >60 Gender: Male Female Race: White (Non-Hispanic) Black/African Amer ican Hispanic Asian/Pacific Islander Native American Multi-Racial/Ethnic Other Location of practice: Rural Suburban Urban Other Profession: Pediatrician Family Practice Registered Nurse Nurse Practitioner Other Subspecialty (if applicable): ____________________ Setting of practice: Hospital Clinic Privat e Practice University-Affiliated Center Other Years in practice: ______ Number of trainings completed related to: Autism Spectrum Disorders: ______ Changing practice/service delivery: ______ Thank you for completing this questionnaire

PAGE 125

117 Appendix C: Implementation Checklist Autism System of Care Trai ningImplementation Checklist Trainer: ____________________ Date: ______________ Reviewer: __________________ ASC Training Content: Yes Partially No Reminded participants about importance of completing questionnaire and turning it in Reviewed 7 Learning Objectives Discussed “triad of impairments” in autism Discussed all 8 “red flags” of autism Reviewed indicators for immediate evaluation Reviewed average age of diagnosis & recommended age Discussed importance of using screening instruments Discussed importance of ear ly screening/intervention Reviewed 5 recommended general screening instruments Reviewed 3 autism-specific screening instruments Detailed review of PEDS Detailed review of Ages and Stages Questionnaire (ASQ) Reviewed AAP’s screening recommendations Reviewed perceived and conc rete barriers to screening Reviewed Model for Improvement (Aim, Measures, Ideas) Discussed the 5 criterion of effective aim statements

PAGE 126

118 Appendix C: (Continued) Reviewed “Plan, Do, Study, Act” Cycles Discussed measurement and data collection Discussed 5 steps in approaching barriers Reviewed example of change in practice to increase early screening Participants completed I ndividual 10-minute activity Reviewed “Tips for Success” Reviewed resources (e.g., websites, articles, books) Addressed 7 Learning Objectives Reminded participants of follow-up questionnaire and importance of completing and returning Obtain presenter’s comments (e.g., How do you feel the traini ng went? Is there any part of the training where you feel more time should be spent, Is there any material that should be added? Removed?) ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________