University of Southern California Dissertations and Theses

The impacts of the COVID-19 pandemic on therapy utilization among racially/ethnically and socio-economically diverse children with autism spectrum disorder

(USC Thesis Other)

The impacts of the COVID-19 pandemic on therapy utilization among racially/ethnically and socio-economically diverse children with autism spectrum disorder

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content The Impacts of the COVID-19 Pandemic on Therapy Utilization Among Racially/Ethnically and
Socio-Economically Diverse Children with Autism Spectrum Disorder
by
Cassin Wolff Gonzales
A Thesis Presented to the
FACULTY OF THE USC OF DORNSIFE COLLEGE OF LETTERS, ARTS AND SCIENCES
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
MASTER OF ARTS
PSYCHOLOGY
May 2022

=

ii

TABLE OF CONTENTS
List of Tables ................................................................................................................................ iii
List of Figures.................................................................................................................................iv
Abstract .......................................................................................................................................... v
Introduction..................................................................................................................................... 1
Differential Access and Utilization of ASD Related Services............................................ 4
Extended Donabedian Model of Healthcare Quality.......................................................... 6
Core Structure – Process- Outcome Factors........................................................... 7
Extended Contextual Factors.................................................................................. 8
Current Study.................................................................................................................... 10

Method.......................................................................................................................................... 13
Overview........................................................................................................................... 13
Participants........................................................................................................................ 13
Measures........................................................................................................................... 14
Hours of ABA Therapy…………………............................................................. 14
Race/Ethnicity………………............................................................................... 15
Primary Payer………………............................................................................... 16
Procedure.......................................................................................................................... 17

Results........................................................................................................................................... 19
Overview........................................................................................................................... 19
Missing Data..................................................................................................................... 20
Descriptive Statistics......................................................................................................... 20
Main Effects of Time on Hours of ABA Therapy…………………................................ 21
Interaction and Moderator Analyses…………………..................................................... 23
Results Summary………………….…………………..................................................... 27

Discussion..................................................................................................................................... 28
Change in Hours of ABA Therapy Over Time................................................................. 28
Differential Impact of COVID19 on Hours of ABA Therapy by Child’s Primary Payer 30
Differential Impact of COVID-19 on ABA Therapy By Child’s Race/Ethnicity……… 34
Study Limitations.............................................................................................................. 34
Future Direction................................................................................................................ 36
Conclusion........................................................................................................................ 37

References..................................................................................................................................... 38

=

iii

LIST OF TABLES

1. Means and Standard Deviations of Total Hours of ABA Therapy by Time Period,
Race/Ethnicity Group, and Primary Payer Group......................................................................... 21

2. One-Way Repeated Measures Analysis of Variance of Hours of ABA therapy by Time........ 23

3. Three-Way Mixed Analysis of Variance of Hours of ABA Therapy by Time, Primary Payer,
and Race/Ethnicity. ...................................................................................................................... 25

=

iv

LIST OF FIGURES
1. Extended Structure-Process-Outcome Model for Applied Behavior Analysis Therapy…….. 12
2. Visual Representation of the Current Study’s Hypotheses....................................................... 13
3. Line Graph Representation of Mean Hours of ABA Therapy at Times One, Two, and Three 23
4. Three-Way Interaction Between Time, Race/Ethnicity, and Primary Payer............................ 26
5. Simple Two-Way Interaction Between Time and Primary Payer on Hours of ABA Therapy. 27
6. Partial Eta Squared Findings Represented on the Extended SPO Model for the Current
Study……………………………………………………………………………………………. 27

=

v

Abstract
Early research on the impacts of the COVID-19 pandemic on children with ASD provides
preliminary evidence of ASD related service disruption and worsening behavioral and affective
outcomes. The current study aims to add to this literature by describing change in hours of ABA
therapy before and through the COVID-19 pandemic and by evaluating if there is differential
disruption based on the race/ethnicity and socio-economic status of the child. Additionally, the
current study utilizes an extended structure-process-outcome model to interpret the implication
of the findings for the quality of the ABA therapy care system as a whole and in the context of
patient variability and environmental stress. Retrospective clinical data on children’s
demographics and client ABA therapy utilization were collected from one ABA agency with 5
branches throughout California (N=203). Using repeated measures ANOVA, we evaluated
change in therapy hours through time and the moderating effects of child’s race/ethnicity and
child’s primary therapy funder. We found that there was a significant effect of time on hours of
ABA therapy so that there was a reduction in hours between pre COVID-19 and the beginning of
COVID-19 with no significant changes in hours of ABA therapy between the beginning of
COVID-19 and 6 months into the pandemic. Analysis of moderators revealed no significant
effect of race or race x payer on the relationship between time and hours of ABA therapy. There
was a significant moderating effect of primary payer on the degree to which time predicted hours
of ABA therapy so that children who receive primary funding from school districts had a more
severe drop in ABA therapy hours during the pandemic while children with other funding
sources did not. These findings indicate that ABA therapy hours may have been disrupted for
longer periods than anticipated and implications for service access to ASD therapies for children
during historical moments of health care disruption are discussed
1
Introduction
Global restrictions to in-person gathering brought on in March 2020 by the COVID-19
pandemic have significantly impacted children’s access to educational and healthcare services.
The impact of COVID-19 restrictions is of particular concern for children and adolescents with
autism spectrum disorder (ASD) who are recommended to receive time-intensive and face-to-
face therapeutic services through early and middle childhood (Elder, 2017; Helt, 2008; Magiata,
2012). Further, the body of research on inequitable access to, and utilization of, ASD-related
services indicates that special attention should be given to the differential impacts of restrictions
brought on by the COVID-19 pandemic on children with ASD from different racial, ethnic, and
socio-economic backgrounds.
On March 11
th
, 2020, the World Health Organization declared COVID-19 (Sars-Cov-2) a
pandemic and by March 19
th
, 2020 California became the first US state to issue a stay at home
order (Exec. Order No. N-33-20, 2020) This order mandated that all residents must stay at home
except to go to an essential job or access essential needs (e.g. medical care, grocery shopping).
Around the same time, the Center for Medicaid and Medicare Services approved funding for
health services delivered through tele-health and several health systems began utilization of tele-
health over in-person health visits (Center for Medicare and Medicaid Services, 2020). By Fall
2020, many health systems, including therapy services, had developed tele-health systems as the
COVID-19 restrictions to in-person gathering remained prominent (Reay, Looi & Keightley,
2020)
Initial research on the impacts of the COVID-19 restrictions on the ASD community
paints a picture of increased need and disrupted services across all families, regardless of
racial/ethnic or socioeconomic background. Survey studies that examined the experience of

=

2

families with a child with ASD found that a majority (93.9%) of parents reported that the
COVID-19 pandemic posed difficulties for their family (e.g., example of difficulties from
survey; Colizzi et al, 2020) and between 41.5-45% of parents reported that, in comparison to the
time prior to the COVID-19 pandemic, their child had more intense and frequent episodes of
challenging behavior (Colizzi, 2020; Nuñez, 2021). The COVID-19 pandemic has also been
shown to have both emotional and behavior impacts on children with ASD. For example,
students with ASD experienced higher levels of stress due to disruption to in-person instruction
compared to their typically developing counterparts (Hosokawa et al, 2021). The same study
demonstrated that students with ASD engaged in a greater amount of restricted repetitive
behaviors after school closures took effect than they experienced prior to the pandemic
(Hosokawa et al, 2021). These findings suggest that children with ASD may actually need more
behavioral support during the COVID-19 restrictions, rather than the reductions in support that
appear to be occurring on a large scale (Jeste et. al, 2020; White, 2021).
In addition to increased behavioral and emotional challenges among children with ASD,
research on the impacts of the COVID-19 pandemic have also shown increased stress at the
family level. Parents of children with ASD experienced higher levels of stress in the time period
following initial COVID-19 restrictions compared to parents of typically developing children
(Corbett, 2021). Among parents of children with ASD, higher stress levels were associated with
younger age of child, child’s higher symptom severity, and higher ASD service utilization prior
to the pandemic (Manning, 2020; Neece, 2020). In addition, parents of children with ASD
reported substantial concerns about the impacts that lost therapeutic, educational, and social
opportunities may have on their child’s long-term development (Neece, 2020). Given the
increased need for support in the autism community, researchers have called for the development

=

3

of more accessible and flexible healthcare services for people with ASD and their families during
the COVID-19 pandemic and beyond (Eshraghi et al, 2020).
One such essential service for families of individuals with ASD is Applied Behavior
Analysis (ABA). ABA is an evidence-based body of behavioral interventions and is a core,
recommended part of treatment for individuals with ASD (Hyman, Levy, & Myers, 2020).
Children with ASD are recommended to receive 20-40 hours per week of behavioral intervention
prior to beginning kindergarten, at which point behavioral services are generally gradually
transitioned to the classroom setting (Behavior Analyst Certification Board, 2014).
During the COVID-19 pandemic, ABA services were deemed an essential service by the
Centers for Disease Control (CDC), prompting researchers to quickly develop guidelines for safe
in-person service delivery, telehealth services, and specific interventions that teach COVID-19
safety behaviors (i.e., masking and social distancing; Ellison, 2021; Halbur et al, 2021; Kornack,
2020). Despite these efforts, the nature of ABA for children with ASD makes successful
administration of intervention difficult in the context of COVID-19. The telehealth setting is not
an ideal replacement for in-person services because children with ASD benefit from active
therapy where they can interact with objects and move around the space with the therapist
(Baweja, 2020). Additionally, masking and hygiene requirements for in-person therapy impact
the ability to use shared toys between clients and utilize face-to-face interactions to teach non-
verbal communication skills (Baweja, 2020). Finally, it is possible that factors related to the
COVID-19 pandemic led to a reduced therapist workforce that, in turn, would reduce the number
of therapists available to see clients. ABA therapist workforce changes have not been explicitly
evaluated in COVID-19; however, reductions in therapist personnel have been reported in other

=

4

face-to-face therapy fields (e.g. psychologists, physical therapists; Abrams, 2020; Oster et al,
2020).
Reports from parents substantiate concerns about service disruption during the initial
COVID-19 restrictions in early 2020. A survey of 818 parents of children with ASD from 46
U.S. states and non-U.S. countries found that 74% of parents reported that their child lost access
to at least one therapy or educational service, and 56% of parents reported that their child
received at-least some continued services through telehealth (Jeste et al., 2020). Another survey
of 3,502 parents of children with ASD in the US found the majority of parents reported
disruption to ABA services (77%) and special education (80%) (White et al, 2021). Broken
down by age group, pre-school aged children with ASD experienced significantly more
disruption to services compared to school-aged children and dependent adults with ASD (White
et al, 2021). While these studies provide evidence that services for children with ASD have been
disrupted during COVID-19, the parent-reported format provides limited information on the
degree of service disruption, change in degree of disruption over time, and differential degrees
of disruption between subgroups of children with ASD.
Differential Access and Utilization of ASD Related Services
The existing literature on the impact of COVID-19 on children with ASD demonstrates
disruption across the ASD community, but provides limited information on the differential
impacts on children with ASD of different races/ethnicities and socio-economic statuses. Pre-
existing research on ASD service access and utilization inequities, however, raise questions
about whether these inequities deepened during the pandemic.
Research conducted prior to COVID-19 has shown that children with ASD of different
racial/ethnic groups experienced different levels of service access, services intensity (i.e. hours

=

5

per week), and breadth of services (Smith, 2020). This body of literature suggests that, in
comparison to white children with ASD, Black and Latinx children with ASD have significantly
less access to general healthcare services (i.e., primary care, specialty medicine, social services),
and Latinx children have more unmet needs with regard to psychotherapy, intensive ASD
services (i.e., 20 hours of ABA per week), occupational therapy, and speech language therapy
(Irvin et. al., 2011; Liptak, 2008; Magaña et al., 2013). With regard to school-based services,
Latinx students with ASD were significantly less likely to receive an individualized education
program compared to white students with ASD (Harstad et. al., 2013). These findings suggest
that racial/ethnic disparities in ASD service access are widespread and indicate that racial/ethnic
disparities in healthcare quality should be examined across all facets of the care delivery system.
Socio-economic status has also been shown to predict different levels of ASD service
access (receipt of any quantity of services) and utilization (amount of service use) among
children with ASD (Liptak 2008; Patten, 2013; Siller, 2014; Smith 2020). Higher intensity (i.e.,
greater hours) of ASD-related services is associated with higher level of parental education,
higher parental job prestige, higher annual household income, and family ownership of home
(Siller et.al., 2004). Higher level of parental education is also associated with utilization of more
distinct types of ASD services (e.g., behavioral therapy, occupational therapy, psychotherapy,
etc; Patten, 2013) and greater likelihood of receiving an individualized education program from
the child’s school (Harstad et al., 2013). Finally, studies that examine ASD service utilization
among Medicaid enrollees have found evidence of differential service access compared to
children who get funding from other sources. For example, one study found that Medicaid spent
more money per child with ASD compared to private insurance (Wang et al., 2013), while others

=

6

have found that Medicaid enrollees with ASD faced long wait times to access behavior therapy
services and limited hours of therapy utilization (Ruble et al., 2005; Yingling, 2018).
While there is limited existing research of racial/ethnic and socio-economic disparities in
access to ASD related services during the COVID-19 pandemic, medical and public health
research on differential impacts of COVID-19 between individuals or different backgrounds
indicates that racial/ethnic minorities experienced worse outcomes in several areas of healthcare
(CDC, 2021). Pre-existing inequities in ASD care, in combination with other demonstrated
health inequities during COVID-19, raises the need for further examination of the differential
impacts of COVID-19 on children with ASD from different racial and socio-economic
backgrounds.
Extended Donabedian Model of Healthcare Quality
Specific impacts of the disruptions brought on by COVID-19 may be better understood
and ameliorated when conceptualized as impacts to the quality of the ABA therapy healthcare
system. Donabedian’s quality of care conceptual framework has been the predominant model of
evaluating the quality of healthcare systems used by researchers across disciplines (e.g.,
Ayanian & Markel, 2016; LoPorto, 2020; Sandler, 2020). Donabedian proposed that the quality
of care of any given health delivery system should be evaluated in three domains: structure,
process, and outcome; often referred to as the SPO model (Donabedian, 1980). Donabedian’s
SPO model is commonly used to understand problems in existing healthcare delivery systems
and identify potential changes that would improve the overall quality of care. In more recent
years, researchers have extended the model to include environmental context and patient
characteristics as influential factors into the original conceptual framework (Amir et. al., 2017;
Mahdavi et. al., 2018; Qu et. al., 2010). This extension allows for consideration of how clients in

=

7

various times or places may experience different levels of quality within the same health delivery
system. The extended SPO model places equal importance on structure, process, and outcome in
determining the quality of a health care system, and considers each of the three domains as
interdependent (Donabedian, 1980). Donabedian proposed that high quality structure allows for
high quality process, and high quality process allows for high quality outcome. In the current
study, an extended version of the Donabedian SPO model is used to understand the mechanisms
and vulnerabilities of the ABA therapy care delivery system and identify how environmental
context (i.e., the COVID-19 pandemic) and patient demographic factors (race and socioeconomic
status) influence the quality of behavioral healthcare (Figure 1).
Core Structure – Process- Outcome Factors
Structure refers to the stable context in which care is delivered, including physical spaces,
funding sources, equipment and tools, and organization (Donabedian, 1980). In the context of
ABA therapy, structure refers to physical clinics, telehealth networks, evidence-based resources
and tools for therapists, clinical personnel, and funding sources. The distinct funding options that
cover ABA services include private insurance, public insurance, school districts, and regional
centers. Care may be funded by one or multiple funding pathways depending on access to
personal insurance, public resources, and school-based services.
Process refers to the interactions between healthcare providers and patients in the
healthcare setting (Donabedian, 1980). This includes both provider activities such as diagnosis,
treatment, and recommendations as well as patient activities such as seeking care and treatment
adherence (Donabedian, 1980). In the context of ABA therapy, process includes creation of valid
and appropriate goals, hours of therapy delivered, flexibility of therapy setting, responsiveness to
needs of the client, and therapeutic alliance.

=

8

Outcome refers to changes in the patient’s life brought on by the health system, and is
typically the factor that is evaluated to determine quality of health care (Donabedian, 1980).
Outcomes can refer to patient loss of diagnosis, reduction in symptoms, improved quality of life,
increased knowledge, and satisfaction with the health system (Donabedian, 1980). In ABA
therapy, outcome refers to achievement of client-specific goals (e.g., ability to get ready for
school independently, play with other children), change in adaptive functioning skills, change in
quality of life, and change in stress on the family system or academic system.
Extended Contextual Factors
The extended SPO model includes two contextual factors, environment context and
patient variability, that influence each domain individually and the healthcare model as a whole.
Environmental context refers to any transient influences outside of the immediate model
including policy changes, social perception of the health system, and unexpected crises. For
example, education policy changes (e.g., No Child Left Behind Law, 2002) could influence the
volume of student referrals to ABA clinics and render the previous number of therapists
(structural factor) insufficient. Social perception changes, such as the neurodiversity movement
(Kapp, 2013), could change the process of how clients expect the therapists to create goals.
Unexpected crises, such as the COVID-19 pandemic and the 2008 United States financial crisis,
have the potential to affect funding systems (structure), ability for patients and clients to meet at
the appropriate frequency (process), and raise client daily stress levels, thereby diminishing the
positive effects of treatment (outcomes).
Patient variability refers to any characteristics or beliefs individual to the patient. These
include race, ethnicity, income, education level, employment status, political affiliation, and
culture. Like the environment factor, patient variability can potentially alter the quality of the

=

9

system as a whole and the quality of each domain. Structural elements such as evidence-based
tools may be of lower quality with members of non-westernized cultures with whom the tools
have not been tested (DuBay, Watson, & Zhang, 2018). Process variables such as client initiation
of treatment may be impacted if a client identifies with a racial group that tends to have a
mistrust of medical providers and/or services based on historical racial injustices in medical
research (Armstrong et al., 2007). Even the definition of what is considered to be a successful
outcome or adaptive functioning may vary across cultural identity groups (Matson & Rieske,
2014). For example, high SES families may have different goals for their children with ASD
(e.g., academic achievement) as compared to the goals for children of families from lower SES
populations (e.g., independent daily self-care)
The extended SPO model allows for the reality of environmental context and patient
factors to be accounted for when considering the quality of a healthcare network holistically.
Given the pre-existing evidence that different subgroups of children with ASD receive
differential levels of ASD service access, the extended model is expected to be a useful for
understanding potential inequities in impact of COVID-19 on ASD service utilization. The
interaction between environment, patients, and the health system is complex, and research on
health care quality needs to more clearly account for these factors in order to suggest more
realistic avenues of change.

=

10

Figure 1:
Extended Structure-Process-Outcome Model for Applied Behavior Analysis Therapy

Current Study
The aim of the current study is to utilize the extended SPO model in order to understand
the degree to which the COVID-19 pandemic affected ABA therapy care over time, and to
examine the potential moderating role of child characteristics (e.g., race/ethnicity, insurance
carrier; Figure 2). The use of the extended SPO model allows findings to be interpreted within
the context of the ABA therapy system and, therefore, provides opportunity to identify potential
mechanisms of findings and interpret implications of findings for the ABA health care model as
a whole. Hypotheses are visually summarized in figure 2.
The primary dependent variable of this study is the process variable, hours of ABA
therapy per month. A process variable was chosen for the dependent variable over an outcome
variable for several reasons. First, there is substantial evidence from the ASD treatment literature

=

11

that early intensive ABA intervention leads to a multitude of positive outcomes including
improvements in cognitive abilities, adaptive functioning, language utilization, and quality of life
(Elder, 2012; Magiata, 2012). Therefore, the current model includes hours of ABA therapy as a
useful proxy of positive outcomes based on previously-established evidence. Additionally, hours
of therapy are directly observable and generalizable across clients, while outcomes for children
with ASD vary greatly based on the specific needs of the child. Because of this difficulty, there
is little consensus on how to measure outcomes across children in a standardized or generalizable
way (Matson & Rieske, 2014). Consideration of hours of therapy alone may prevent nuanced
analysis of therapy progress, but the use of hours allows for larger scale comparisons to be made.
The first hypothesis is that hours of ABA therapy will be significantly lower in the time
periods during the COVID-19 pandemic compared to hours of service prior to the COVID-19
pandemic. Specifically, we expect a decrease in hours of therapy between time period one (Sep –
Nov, 2019) and time period two (Mar-May, 2020), followed by an increase in hours of therapy
between time period two and time period three (Sep-Nov, 2020).
The second hypothesis is that, when age and gender are controlled for, the association
between time and hours of ABA therapy will be moderated by child’s primary funder pathway so
that children who are primarily funded by public insurance will have a greater reduction of hours
through time compared to children who are primarily funded through private insurance. The
directionality of hours received by school funded children is not specifically predicted. No
direction is hypothesized because, (a) school and school-based services are accessible to all
children with ASD, regardless of race/ethnicity and socio-economic status, as mandated by the
Individuals with Disabilities in Education Act (IDEA), and (b) ABA therapy support in schools
occurs during the standard school day while private and public funded hours typically occur in

=

12

the child’s home and, therefore, hours of ABA in a school setting likely has less variation
compared to hours delivered in the child’s home. Despite these factors making directionality
difficult to predict, findings on inequities in access to school-based services create precedent for
examining if children who are primarily funded by schools experience a differential impact of
time on hours of therapy compared to children who receive funding from other sources.
The third hypothesis is that, when age and gender are controlled for, the association
between time (Pre COVID-19 and two times during COVID-19) and monthly hours of ABA
therapy will be moderated by the racial/ethnic identity of the child so that ‘White, non-Latinx’
and ‘Asian, non-Latinx’ children will have a lesser reduction in ABA therapy hours compared to
children of other racial ethnic identities, including those who identify as Latinx. Given the
consensus in pre-COVID-19 ASD literature that demonstrates race and ethnicity-based inequities
in ASD service access, it is important to fill the gap in research on differential impacts of the
COVID-19 pandemic on therapy access between children of different racial/ethnic groups
The fourth and final hypothesis posits that there will be a three way interaction between
time, child race/ethnicity, and child’s primary funder when age and gender are controlled. In
other words, the impact of time (pre-COVID-19 and two times during COVID-19) on hours of
ABA therapy per month will be influenced by the unique combination of the child’s primary
funder and the child’s racial/ethnic identity. This is an exploratory hypothesis with no specific
directionality predicted.
In the analyses of these hypotheses, we control for both the child’s gender and the child’s
age. These covariates were included in order to control for potential effects of differential ASD
prevalence between genders (Manner & Shaw, 2020) and for natural variation in ABA therapy
utilization throughout development (Behavior Analyst Certification Board, 2014).

=

13

Figure 2:
Visual Representation of the Current Study’s Hypotheses.

Method
Overview
This study utilized retrospective, de-identified data from one ABA agency that primarily
serves children with ASD. Data consisted of client demographic information and detailed client
ABA therapy logs. Data came from each of the agency’s five branches (Contra Costa County,
East Los Angeles, Hermosa Beach, North Los Angeles, and West Los Angeles), representing a
racially and socioeconomically diverse set of geographical regions in California.
Participants
Participants included in the sample consisted of children who received ABA behavior
therapy from this agency, at any of the branches, as of September 1, 2019. Potential participants
were excluded from the sample if the they had more than six consecutive months of zero hours
of ABA therapy between September 1, 2019 and February 28, 2021 or if the participant had

=

14

exactly six consecutive months of zero hours of ABA therapy with no return to therapy by
February 28, 2021. This exclusion criteria was implemented in order to remove participants from
the sample who naturally transitioned out of ABA therapy due to reduced need/clinical
improvement during the study time period. The six-month time increment was chosen in order to
retain participants in the sample who experienced temporary disruption of ABA therapy services
due to COVID-19. Finally, potential participants were excluded from the sample if they received
zero funding for services from private insurance, public insurance, or school districts. For
example, those who received exclusive funding from regional centers or private pay were
excluded. These funding sources were excluded from the sample because they are typically
transient (i.e. provide funding for six months or less) and the purpose of the study was to
examine changes in long term ABA care.
Out of 224 participant records obtained, 14 were excluded for having more than six
consecutive months of zero hours of ABA therapy and 7 were excluded for having exclusive
funding from regional centers (and therefore, not private insurance, public insurance, or school
district). After exclusion, total sample size consisted of 203 participants records (23.6% Asian,
6.4% Black, 21.2% Latinx ethnicity (of any race), 33.5% White, 1% other race/ethnicity, 14.3%
not reported). The age of participants in the sample ranged from 2-14 years, as of September 1,
2019. The sample contained 44 participants who identified as female, 156 participants who
identified as male, and 1 client who did not report gender.
Measures
Hours of ABA Therapy
The primary independent variable in the current study is time period: time period one
(pre-COVID-19, August – November 2019), time period two (during COVID-19, March-May

=

15

2020), and time period three (during COVID-19, August – November 2020). Two COVID-19
time periods were chosen in order to account for the changing nature of the COVID-19
pandemic. The months immediately following initial restrictions in mid-March 2020 comprised
a rapid adjustment period where delays in treatment as programs adjusted to the restrictions were
somewhat expected. By the beginning of the new school year in August 2020, the ABA therapy
health system had had 6 months to adjust and modify care delivery plans and a perception of a
“new normal” became predominant. In order to assess the quality of the ABA therapy system, it
is important to understand how quality varies in times of stress (COVID-19) both in the short
term and long term.
In addition to covering different meaningful phases of the COVID-19 pandemic, these
time periods were selected in order to best control for expected fluctuation of therapy hours
throughout the calendar year. Under normal circumstances, it is expected that children with ASD
receive different levels of therapy and services over the summer, in the absence of school.
Additionally, some disruption of services is expected in December and January due to winter
break and religious holidays. The fall months (September, October, November) and the spring
months (March, April, May) were chosen for comparison because both time periods consist of
school instruction with minimal breaks (one week off expected for both periods – Thanksgiving
week and spring break).
Race/Ethnicity
The de-identified data that was received from the agency included the child’s parent-
reported racial/ethnic identity in open question format. Based on this information, participants
were categorized as Asian, Non-Latinx; white, Non-Latinx; or other racial/ethnic identity other
than Asian or white, including those who identify as Latinx and any other race. Participants were

=

16

grouped in the Asian, non-Latinx group if they reported their race as “Asian” or if they identified
with a nation/ethnic group associated with continental Asia (e.g. Vietnamese, Chinese, Thai) and
if they did not report Latinx/Hispanic/Chicano ethnic identity. Participants were grouped in the
white, non-Latinx group if they reported their race as “white”, “Caucasian”, or specified a
country within the United Kingdom and if they did not specify Latinx/Hispanic/Chicano ethnic
identity. Participants were grouped in the “racial/ethnic identity other than Asian or white”
group if they did not meet group criteria for the Asian, non-Latinx group or the white, non-
Latinx group. For example, this group included those who identified as Black, African
American, Latinx, Hispanic, Chicano, Native American, “other”, and “mixed”.
These race/ethnicity group divisions are consistent with prior literature demonstrating
that Black and Latinx children receive less ASD related services compared to White children
(Smith et al, 2020). Findings on service access among Asian children with ASD have been
mixed. In 2014, Michael Siller and colleagues found that White and Asian children had higher
intensity of therapy hours compared to Black and Hispanic children while Dwight Irvin and
colleagues found that Asian children received less occupational therapy compared to white
children (2011). The current study groups children by Asian (non-Latinx), white (non-Latinx),
and other racial/ethnic identity in order to expand on the existing literature on racial inequities in
access to ASD services while retaining maximum power for between group comparisons.
Primary Payer
Participants were grouped into one of three categories of primary payer: private insurance
as primary payer, public insurance as primary payer, or school district as primary payer.
Because any one child could receive funding from multiple sources (e.g., both private insurance

=

17

and school), the following system was developed in order to categorize each participant into one
of the three primary payer categories.
The obtained data consisted of minutes of ABA therapy for each session as well as the
funder of that particular session for each child, organized by month. First, each funder that
contributed to a child’s ABA therapy was identified for each participant. Next, each of those
funders were categorized as private insurance (e,g., Aetna, Kaiser, Blue Shield), public insurance
(e.g., LA Care, Medical, Medicaid, Tri-Care), or school district (e.g., Los Angeles Unified
School District). Then, each participant’s total hours of ABA therapy, broken down by payer
type, was calculated using data from all months between September 1, 2019 and February 28,
2021. Participants were grouped into the private insurance group if more than 50% of their hours
were funded by a private insurer; into the public insurance group if more than 50% of their hours
were funded by a public insurer; and into the school district group if more than 50% of their
hours were funded by a school district.
These groups reflect the three primary funding pathways through which children can
receive ABA therapy. While funding source is considered a structural variable, distinct funding
pathways control access to a child’s care, and therefore may potentially predict differential
quality of ABA therapy care. The structural funding source pathway that any given child is on is
primarily determined by patient variables (represented by the dashed green line in figure two).
For example, public insurance in the US (Medicaid) is reserved for certain groups of individuals
including low income families, qualified pregnant women, and recipients of supplementary
security income (Centers for Medicaid and Medicare Services). Therefore, access to the
Medicaid funding pathway is only available to individuals who meet criteria. The employment
status of a child’s parent may influence if a child is able to receive funding through private

=

18

insurance, given that private insurance is often provided to families through employment.
School funding of services is determined by the recommendations made on a child’s
Individualized Education Program. As previously described, white racial identity and higher
parental level of education are both associated with higher likelihood of receiving an IEP. While
these examples are not exhaustive, they demonstrate how a child’s structural funding pathway is
a proxy of patient level factors (e.g. race/ethnicity, income, employment status). Child’s funding
pathway is included in the current study in order to assess if patient level factors, via the child’s
funding pathway, will influences the degree to which COVID-19 impacts their ABA therapy
hours over time.
Procedure
The agency selected a team of internal agency employees to collect and transfer data.
The principal investigator provided the agency team with detailed instruction on what
information should be included in the transferred data set. Deidentified data were included for
demographics (date of birth, reported race/ethnicity, and reported sex/gender) and ABA therapy
delivery information (therapy session duration, therapy session funder, therapy session setting)
for the time period between September 1, 2019 through February 28
th
, 2021. The agency team
obtained information by downloading records from their internal client management system and
organizing information by identification numbers. The key linking identification numbers to
client personal health information was not shared with the investigators of the current study.
Because only de-identified data was collected, consent forms were not collected from clients
included in the sample.

=

19

Data obtained from the ABA agency consisted of duration (minutes) of each session a
participant received organized by month. For each participant, the sum total duration of ABA
therapy sessions was calculated for each month.
Results
Overview
Three statistical models were conducted in order to evaluate all hypotheses. Hypothesis
one, that there would be a significant difference in hours of ABA therapy between the three
timepoints among all children in the sample, was evaluated with a one-way repeated measures
analysis of variance (ANOVA). Next, we ran a three-way mixed ANOVA in order to evaluate if
there was a significant three-way interaction between time periods, primary payer, and
race/ethnicity (hypothesis four). Results of the three-way mixed ANOVA were also examined
for any indication of two-way interactions between each of the predictors (race/ethnicity and
primary payer) and time period. Then, we ran a pair of two-way mixed ANOVAs in order to
evaluate if primary payer (hypothesis two) and race/ethnicity (hypothesis three) significantly
moderated the relationship between time period and hours of ABA therapy. Finally, post-hoc
analyses were conducted for all significant two-way interactions in order evaluate differences
between levels of the predictors.
Characteristics of the sample distribution were examined in order to evaluate any
violations of model assumptions. Both the pre-COVID-19 time period and the post-COVID-19
time period had a single outlier (between 1.5 and 3 SD), as assessed by inspection of boxplot.
Participants with outliers were included in the final analysis. The assumption of normal
distribution was violated as assessed by the Shapiro Wilk test; however, results were still

=

20

interpreted due to the robustness of ANOVA to non-normality. All analyses were conducted in
IBM SPSS Statistics (Version 27, IBM Corp 2017).
Missing Data
Across all three timepoints, there was no missingness within measures for hours of ABA
therapy or primary payer. The race/ethnicity variable had missing data for 14.3% percent of the
sample. The agency that provided the data reported that missing race/ethnicity data was due to
race/ethnicity omission on evaluation reports received from diagnosing clinicians at ABA
therapy intake. A sensitivity analysis was conducted, comparing the study models including and
excluding participants with missing race/ethnicity data. There was no significant difference in
results between these two analyses, therefore clients with missing race/ethnicity data were
included in the final analyses in order to retain maximum power.
Descriptive Statistics
Descriptive statistics were calculated for all three timepoints by race/ethnicity and
primary payer; means and standard deviations are presented in Table 1. All participants had a
significant loss of ABA therapy hours between time periods one and time periods two and
between time periods one and three. There was no difference in likelihood of reduction in
therapy hours over time between members of the different racial/ethnic groups. Participants in
the school district funded group were more likely to receive a greater number of ABA therapy
hours compared to children who received funding from different sources at both time periods one
and two, but not at time period three.

=

21

Table 1:

Means and Standard Deviations of Total Hours of ABA Therapy by Time Period, Race/Ethnicity
Group, and Primary Payer Group.

Variable Race/Ethnicity Primary Payer
Full Sample
(n = 203)
Asian,
Non-
Latinx
(n = 48)
White,
Non-
Latinx
(n = 68)
Other
Racial /
Ethnic
Identity,
including
Latinx
(n = 58)
Private
Insurance
(n = 100)
Public
Insurance
(n = 27 )
School
District
(n = 47 )
M(SD) Min Max M(SD) M(SD) M(SD) M(SD) M(SD) M(SD)
Time 1 249.51
(132.61)
17.67 832.40
216.96
(155.28)
270.16
(120.48)
243.04
(130.07)
201.20
(110.66)
208.44
(130.23)
364.55***
(114.02)
Time 2 178.80***
(136.27)
10.25 567.20
155.85
(139.93)
192.53
(144.29)
179.122
(135.72)
122.13
(105.05)
121.85
(113.22)
328.33***
(106.02)
Time 3 173.63***
(100.82)
3.0 495.58
148.20
(89.4)
197.78
(102.52)
156.22
(92.80)
172.93
(97.64)
133.27
(74.85)
185.80
(106.17)

Note: *p < .05, **p < .01, *** p <.001.

Main Effects of Time on Hours of ABA Therapy
A one-way repeated measure analysis of variance (ANOVA) was conducted in order to
evaluate if there was a significant difference in hours of ABA therapy delivered between the
three time periods (hypothesis one). This test consisted of one within-subject variable, time, with
three levels: time period one (September-November 2019), time period two (March-May 2020),
and time period three (September-November 2020). The dependent variable was the sum total
duration (hours) of ABA therapy. Pairwise comparisons were calculated for significant findings
in order to evaluate differences between each combination of within-subject levels. For this test,
the assumption of sphericity was violated, as assessed by Mauchly's test of sphericity, χ
2
(2) =

=

22

13.11, p = .001. Therefore, a Greenhouse-Geisser correction was applied (ε = 0.949). Results are
summarized in table 2.
Hours of ABA therapy utilization was statistically significantly between the three
timepoints, F(1.88, 380.00) = 52.20, p < .001, partial η
2
= .21 (table 2). Therapy hours decreased
between time period one (M = 249.50, SD = 132.6 hours), time period two (M = 178.79, SD =
136.27 hours), and time period three (M = 173.63, SD = 100.82 hours) (figure 3). Post hoc
analysis with a Bonferroni adjustment revealed that the total number of ABA therapy hours
decreased significantly between time period one (pre COVID-19) and time period two (during
COVID-19) (M = 70.71 hours, 95% CI [53.37, 88.04], p < .001), as well as between time period
one (pre COVID-19) and time period three (during COVID-19) (M = 75.87 therapy hours, 95%
CI [54.59, 97.15], p < .001). There was no significant difference in hours of ABA therapy
between time period two and time period three (M = 5.16 therapy hours/3 month period, 95% CI
[-16.07, 26.39], p = 1).
Table 2:

One-Way Repeated Measures Analysis of Variance of Hours of ABA therapy by Time

Predictor df Mean
Square
F p Partial η
2

Time 1.881 364829.249 52.197 < .001 .205
Error
(Time)
380.003 7430.809 - - -

=

23

Figure 3:
Line Graph Representation of Mean Hours of ABA Therapy at Times One, Two, and Three.

Interaction and Moderator Analyses
The second, third, and fourth hypotheses were investigated through a three-way mixed
analysis of variance (ANOVA). This analysis included one within-subject predictor (time – three
levels) and two between-subject predictors (race/ethnicity – three levels; primary payer – three
levels). This analysis also included two covariates: gender (two levels, nominal) and age in
years (ordinal). There was one outcome variable, hours of ABA therapy (continuous). All
significant two-way interactions found in the three-way mixed ANOVA were decomposed by
running a two-way mixed ANOVA. Pairwise comparisons were interpreted for all significant
two way interactions.
There was homogeneity of variances for hours at time one (p = .697), hours at time two
(p = .998), and hours at time three (p = .45), as assessed by Levene's test for equality of
variances. Mauchly's test of sphericity indicated that the assumption of sphericity had been

=

24

violated, χ
2
(2) = 6.90, p = .032. Therefore, a Greenhouse-Geisser correction was applied (ε =
0.96). Results are summarized in Table 3.
The three-way interaction between time, race, and primary payer was not statistically
significant, F(7.68, 312.96) = 1.58, p=.13 (Figure 4). There was a statistically significant two-
way interaction between time and primary payer, F(3.84, 312.96) = 27.90, p < .001. All other
two-way interactions, including covariates, were not statistically significant (p > .05). Statistical
significance of a simple main effect was accepted at a Bonferroni-adjusted alpha level of .025.
There was a statistically significant simple main effect of primary payer at time one, F(2, 312.96)
= 30.52, p < .001 (partial η
2
= .27) and at time two, F(2, 312.96) = 48.01, p < .001 (partial η
2
=
.37), but not at time three, F(2, 312.96) = 1.26, p = .29.
All pairwise comparisons were performed for statistically significant simple main effects
(times one and two but not time three) (figure 5). Bonferroni corrections were made with
comparisons within each simple main effect. Adjusted p-values are reported. At time one,
participants in the school district funded group received more ABA therapy hours compared to
participants in both the private insurance funded group (mean difference of 179.94, 95% CI
[123.99, 235.89], p < .001), and in the public insurance funded group (mean difference of
153.79, 95% CI [78.74, 228.83], p < .001). There was no statistically significant difference in
number of ABA therapy hours between participants in the private insurance funded and public
insurance funded groups (mean difference of 26.157, 95% CI [-37.57, 89.88], p = .967) at time
one. At time two, participants in the school district funded group received more hours of therapy
compared to participants in both the private insurance funded group (mean difference of 210.96,
95% CI [157.76, 164.19], p < .001), and public insurance funded group (mean difference of
208.05, 95% CI [136.67, 279.45], p <.01). There was no statistically significant difference in

=

25

number of ABA therapy hours between participants in the private insurance funded group and
public insurance funded group (mean difference of 2.90, 95% CI [-57.72, 63.53], p = 1) at time
two.
Table 3:

Three-Way Mixed Analysis of Variance of Hours of ABA Therapy by Time, Primary Payer, and
Race/Ethnicity.

Predictor
df
Mean
Square
F p Partial η
2

Time 1.92 60055.75 11.3 < .001 .065
Time x Age 1.92 7819.99 1.48 .231 .009
Time x Gender 1.92 862.73 0.16 .850 .001
Time x Race/Ethnicity 3.84 8137.02 1.54 .194 .018
Time x Primary Payer 3.84 147848.66 27.90 < .001 .255
Time x Race/Ethnicity
x Primary Payer
7.68 8366.80 1.58 .134 .037
Error (Time) 312.96 5299.28 - - -

=

26

Figure 4:
Three-Way Interaction Between Time, Race/Ethnicity, and Primary Payer
.

=

27

Figure 5:
Simple Two-Way Interaction Between Time and Primary Payer on Hours of ABA Therapy

Results Summary
Figure 6 depicts a summary of significant findings and effect sizes (partial eta squared)
with reference to each hypothesis in the context of the extended SPO model.
Figure 6
Partial Eta Squared Findings Represented on the Extended SPO Model for the Current Study

Note: Numbers in the figure denote partial eta square, a metric of effect size.
* = p < .05 ;** = p < .01 *** = p < .001

=

28

Discussion

The purpose of the current study was to examine (a) the relationship between time period
during the COVID-19 pandemic and hours of ABA therapy delivered to children with ASD, and
(b) if that relationship was moderated by child’s racial/ethnic identity and/or child’s primary
funding source. Recent research on parent-reported impacts of the COVID-19 pandemic on the
ASD community have found that, during the pandemic, there were more instances of challenging
behaviors and stress among children with ASD, higher levels of stress among families with a
child with ASD, and perceptions of disruptions to necessary ASD related service for children
with ASD (Colizzi et. al., 2020; Corbett 2020; Hosokawa, 2021; Jeste, 2020; Manning, 2020;
Neece, 2020; Nuñez, 2021, White, 2021). While the existing literature on COVID-19 and ASD
provides limited information on differential impacts based on child characteristics, pre-existing
literature consistently shows differences in access to, and utilization of, ASD-related services
among children of different race/ethnicities and socioeconomic statuses (Irvin et. al., 2011;
Liptak 2008; Magaña et al., 2013; Patten, 2013; Siller, 2014; Smith 2020).
The current study contributes to the existing literature in two primary ways, (1) by
describing objectively measured changes in service utilization through different times before and
during the COVID-19 pandemic, and (2) by examining differential impacts of COVID-19 on
children with ASD from different racial/ethnic and socioeconomic backgrounds. The current
study’s use of the extended SPO model puts specific findings into context of a larger health
delivery system framework and, therefore, allows for consideration of the implications on the
ABA care system as a whole.
Change in Hours of ABA Therapy Over Time
We first examined changes in hours of ABA therapy through time. We found that there
was a significant reduction in ABA therapy hours delivered between the pre-COVID-19 time

=

29

period (time one) and the “during” COVID-19 time periods (times two and three). This aligns
with existing literature on parent-reported impacts of COVID-19 that found perception of ASD
service disruption at the start of the COVID-19 pandemic (Jeste, 2020).
Analysis of changes between the three time periods yielded findings that partially
supported to our hypothesis. Aligned with our hypothesis, we found that there was a significant
reduction in hours between the pre COVID-19 time period and each of the during COVID-19
time periods, but no difference between the two COVID-19 time periods (March-May 2020 and
September through November 2020). Contrary to our hypothesis, six months after the start of the
COVID-19 pandemic there was no significant increase in hours of ABA therapy delivered to
children with ASD. This lack of increase in ABA therapy hours was found in spite of
developments in adaption of services during that six-month period. In the context of the extended
SPO model, this finding indicates that environmental stress does negatively impact the process of
ABA therapy delivery. In combination with the literature that demonstrates the association
between more hours of ABA therapy with positive outcomes, this finding suggests that
environment stress, via the process of ABA therapy delivery, leads to worse outcome for
children with ASD. Additionally, the finding that there was no change in hours of ABA between
time periods two and three (6 months apart) suggests that the process of ABA therapy delivery
did not improve when the environmental stressor is sustained over a long period of time.
Without parent and clinician reports of why the continued disruption occurred, it is
difficult to identify the exact cause of sustained disruption. However, the extended form of the
SPO model of ABA therapy care provides some insight on potential mechanisms. First, based on
employment information from other therapy fields (Abrams, 2020; Oster et al, 2020), it is
possible that there was a reduction in therapist personnel brought on by the COVID-19 pandemic

=

30

that may have impacted quality at the structural level. There may have been a reduction in
therapist personnel due to therapists limiting hours to avoid COVID-19 exposure, taking time off
because of infection with COVID-19, and moving away from ABA clinics in order to save
money by co-living with family members in other locations. Additionally, financial strain across
funding sources at the structure level may have led to a reduction in the number of hours children
with ASD were approved to receive. Finally, COVID-19 may have impacted other process
variables such as mode of therapy delivery (in-person versus telehealth) which may have
diminished positive outcomes of ABA therapy. Diminished positive outcomes, in turn, may have
led to parents of children with ASD to reduce the number of hours of therapy received by the
child.
Regardless of mechanism, the findings remain that hours of ABA therapy were reduced
at the beginning of COVID-19 and did not increase after 6 months. A disruption of ABA therapy
for this length of time could negatively impact short and long term outcomes for children with
ASD. Previous literature on the effects of delaying the start of early intensive ABA intervention
found that children with ASD who begin treatment earlier in life experienced more positive
outcomes including higher IQ gain and higher likelihood of placement in a general education
classroom compared to children who started treatment later in life (Tarbox, Persicke, & Kenzer,
2013). This indicates that children with ASD who experienced long term disruption to ABA
therapy may experience worse outcomes due to this disruption. Special attention should be paid
to help children with ASD make up for gains lost during COVID-19 service disruption in order
to minimize the negative effects of this disruption.
Differential Impact of COVID-19 on Hours of ABA Therapy by Child’s Primary Payer

=

31

We found that primary payer significantly moderated the relationship between time and
hours of ABA therapy so that children who were primarily funded by school districts received
more hours of ABA therapy at times one and two compared to children who received funding
from other sources. There was no significant difference in change in hours between children who
were primarily funded by private insurance and children who were primarily funded by public
insurance. The lack of difference between private insurance and public insurance funding, in this
sample, may suggest similar quality of ABA therapy care between these two funding pathways
during times of stress, however more precise evidence on this relationship would be needed to
substantiate and generalize this claim.
Participants funded by school districts, however, had differential impact of time on hours
of ABA therapy compared to children with different primary funding sources. While the
deviation between participants in the school district funded group and participants in the other
groups occurred at time periods one and two, it is the lack of deviation at time period three that,
perhaps, carries the more notable implications. Because time period one takes place prior to the
COVID-19 pandemic, hours of ABA therapy during this time reflect how the ABA therapy
system works when not under stress. Based on distributions at time period one, children who
receive primary funding from school districts receive more hours of ABA therapy compared to
children who receive funding from other sources under normal circumstances. This discrepancy
is likely due to the context in which school district funded therapy takes place (during the typical
7-hour school day) compared to the context in which insurance funded hours take place (in the
child’s home or at the agency). If a student receives consistent support in the classroom setting,
they would get support for most of the school day whereas ABA therapy outside of the school

=

32

setting is typically delivered in smaller doses (e.g. 1-3 hours/day, Behavior Analyst Certification
Board, 2014).
Under this interpretation, the finding that participants in the school district group had
significantly more hours of ABA therapy compared to participants in other funding groups at
time two suggests that COVID-19 restrictions did not change the typical relationship between
funding group and hours of ABA therapy in the early months of the pandemic. The lack of
significant difference in hours of ABA therapy between funding groups at time three reflects a
stark reduction of hours between the early months of the pandemic and six months into the
pandemic for children who are primarily funded by school districts. In other words, the hours of
ABA therapy received by participants in the school district funded group reduced to such a
degree between time periods two and three that their hours were similar to participants in other
groups who typically receive less hours of therapy, under normal circumstances. In sum, this
indicates that, while school district-funded children received the same amount of ABA therapy
hours before and at the beginning of COVID-19, they received a significantly lower amount of
hours when the new school year started in Fall 2020.
The extended SPO model may provide insight as to why school district funding,
specifically, was associated with reduction in hours of ABA services in the new school year. In
addition to the factors previously described (changes in funding, reduction of therapist personnel,
and parent adjustment of services), the school-funded therapy setting poses specific challenges
that may contribute disruption of therapy hours. For example, school district funded hours are
typically delivered in the classroom setting so that ABA therapists can provide the child with
therapeutic support in their learning environment (Disabilities Rights California, 2021). While
this modality has been shown to be efficient in the in-person classroom (Stahmer, 2007), remote

=

33

learning may have posed specific challenges. The United States Health Portability and
Accountability Act prohibits ABA therapists from disclosing many of their client’s identities to
other students in the classroom (HIPAA, 1996). For these students, during in-person instruction,
the therapist is able to subtly support their client without disclosing the identity of the treatment
recipient. However, in remote learning, any verbal cues or guidance given to the client during
virtual class would disrupt the entire class and identify the client. This obstacle limits the
interaction between therapists and clients and likely diminishes the benefits to the child of having
a therapist present in class. This diminished benefit of the online platform may have led parents,
teachers, and therapists to decide to reduce hours of school district-funded ABA therapy hours in
the 2020/2021 school year. In sum, the process-level change of therapy setting brought on by the
COVID-19 pandemic (in-person school to virtual school) may have led to diminished quality of
ABA therapy followed by the decision to reduce hours.
Regardless of the validity of the reasoning for reduction in hours of ABA therapy, the
current study’s findings show that children who primarily received ABA therapy funding
through school districts were not accessing the same quantity of therapy in Fall 2020 that they
had been receiving prior to the COVID-19 pandemic. The primary clinically-indicated reason for
a reduction in hours of this magnitude is that the child no longer needs support in order succeed
in school, a phenomenon unlikely to take place in the virtual learning environment. Given
findings of worse student performance in remote learning compared to in-person learning
(Becker, 2020; Kuhfield, 2020), this scenario appears unlikely. Therefore, the finding that
school-funded ABA therapy hours decreased from Spring 2020 to Fall 2020 suggests that
children with ASD may have experienced worse school-based outcomes (e.g., academic skill

=

34

attainment, adaptive functioning in a classroom setting, and social skill attainment), specifically,
during the COVID-19 pandemic.
Differential Impact of COVID-19 on ABA Therapy By Child’s Race/Ethnicity
The third hypothesis, that race would moderate the relationship between time and hours
of ABA, was found to be insignificant. This means that the change in hours of ABA therapy
across time points did not significantly vary between participants of different racial/ethnic
groups. The fourth hypothesis, that ABA therapy hours would be effected by a three way
interaction between time, child’s race/ethnicity, and child’s primary payer, was also not
supported. There was no difference in ABA therapy hours through time between children of the
same race/ethnicity who had different primary payers, nor between children with the same
primary payer of different races/ethnicities.
These results indicate that racial/ethnic identity did not have a significant impact on hours
of ABA therapy received during the COVID-19 pandemic. However, and assuming no type II
error (limitations of interpretation outlines below), these findings speak to only one aspect of
ASD-related care, hours of ABA therapy among children who are already enrolled in ABA
therapy. It is possible that group differences would be more apparent if we examined differential
rates of enrollment or differential outcomes, for example. Additionally, while no significant
results were found, it is possible that racial/ethnic identity influenced the type of primary payer a
child utilized and, therefore, indirectly influences the moderating effects of primary payer on
hours of ABA therapy through time.
Study Limitations
The primary limitation of the current study is that the sample is restricted to children with
ASD who receive ABA therapy services from one agency in California. The five clinics whose

=

35

data were analyzed for the study serve a socially and geographically diverse collection of
California, the most populous state in the US, containing 17% of the US population. Still, it is
entirely possible that unknown variables systematically bias the families who were able to access
services through the one agency and therefore participate in the current study sample. Despite
this limitation, the sample sizes provides sufficient power for analyses and significant findings
were attained.
Second, the current study does not explicitly examine quality of ABA therapy outcomes.
Implications regarding the quality of ABA therapy outcomes made in this study rely on the
evidence-based assumption that more hours of ABA therapy are associated with better outcomes
among children with ASD. While this association is heavily supported in the literature (Elder,
2012; Magiata, 2012), the connection between hours of ABA therapy and positive outcomes
could have been fortified if the study included parent, child, therapist, and teacher perceptions of
ABA therapy quality throughout the COVID-19 pandemic , and/or other measures of learning
(e.g., standardized assessments of social skill development, ASD symptoms, etc.).
Finally, the reality of ABA therapy hour fluctuation throughout the calendar year posed
limitations to data analysis. Hours of ABA therapy for children with ASD are expected to
naturally fluctuate throughout the calendar with school breaks and transitions into the new school
year. In order to understand the immediate and delayed impacts of COVID-19 on ABA therapy
hours, the current study utilized six month time increments with similar child and service
administration expectations (i.e., school year was well under way and similar disruption due to
school closures). While uniformity of expected ABA hours across months would add further
controls to the study, any attempts to manipulate hours to attain uniformity would come at the
cost of ecological validity. Put simply, the purpose of the study was to evaluate the disruption in

=

36

treatment that the pandemic caused, as it actually occurred, not to select an artificial overly-
controlled sample of reality to experimentally control variability.
Future Directions
Several potential research questions stem from the findings of the current study. First,
future studies should include participants who receive services from a variety of ABA agencies
across states in order to generalize results more broadly. In addition, future research should
evaluate the effects of pandemics and other crises on the delivery of care to people with ASD in
other countries with different levels of wealth and health care systems.
In addition to expanding the sample make up, future studies could extend on the present
study by evaluating more areas of the extended SPO model. For example, were the rates of
enrollment in ABA therapy different between children of different race/ethnicities of socio-
economic statuses during the COVID-19 pandemic? It is possible that racial/ethnic group
differences would be seen when considering enrollment as opposed to continued utilization
among children who are already enrolled. Future studies could also examine outcomes, did
parents and children with ASD experience diminished benefits of therapy during the COVID-19
pandemic even when hours remained stable? The current study provides an important
foundational step in understanding how the ABA therapy system was impacted by COVID-19
restrictions, but there is likely more to be learned about the nuance of these impacts.
Finally, future studies may also provide further insight by expanding the study,
longitudinally. Examination of ABA therapy hours over the following 12 months (December
2020 – December 2021) may provide insight on how hours of service changed with the
prominence of the COVID-19 vaccine and the Fall 2021 return to in-person learning. These
studies may also consider controlling for expected fluctuations in service across the academic

=

37

year by comparing the same month of the year across multiple years. Perhaps of greatest import
would be to attempt to identify longitudinal effects of pandemics and other crises on individuals
with ASD’s long-term development and habilitation. For example, are children whose childhood
therapy was disrupted less likely to achieve outcomes such as learning functional vocal
communication, attending college, or obtaining paid employment?
Conclusion
This study expands upon previous literature by examining the ways in which the COVID-
19 pandemic and associated restrictions impacted therapeutic services among children with ASD.
These findings carry crucial implications for children with ASD who live through unexpected
world events and significant disruptions to school and health care systems. These findings should
be used to inform the extent to which children whose therapy was disrupted during the COVID-
19 pandemic should receive additional support to make up for therapy time lost. Additionally,
findings should be used to inform potential areas of improvement to the ABA care system so that
future world events result in less negative impact on children with ASD. While we cannot
change how the ABA care system reacted during COVID-19, we can learn from these events to
increase its resilience and improve care for children with ASD in the future.

=

38

References

Abrams, Z. (2020). Economic uncertainty: How COVID-19 has hit psychologists. Monitor on
Psychology, 51(7). http://www.apa.org/monitor/2020/10/economic-uncertainty

Amir, Y., Tan, F. E. S., Halfens, R., Lohrmann, C., & Schols, J. (2017). Pressure Ulcer
Prevalence and Care in Indonesian Hospitals: A Multicenter, Cross-sectional Evaluation
Using an Extended Donabedian Model. Ostomy/wound Management, 63(2), 8–23.

Armstrong, K., Ravenell, K. L., McMurphy, S., & Putt, M. (2007). Racial/Ethnic Differences in
Physician Distrust in the United States. American Journal of Public Health (1971), 97(7),
1283–1289. https://doi.org/10.2105/AJPH.2005.080762

Ayanian, J. Z., & Markel, H. (2016). Donabedian’s Lasting Framework for Health Care Quality.
The New England Journal of Medicine, 375(3), 205–207.
https://doi.org/10.1056/NEJMp1605101

Baweja, R., Brown, S. L., Edwards, E. M., & Murray, M. J. (2021). Covid-19 pandemic and
impact on patients with autism spectrum disorder. Journal of Autism and Developmental
Disorders, doi:http://dx.doi.org.libproxy1.usc.edu/10.1007/s10803-021-04950-9

Becker, S. P., Breaux, R., Cusick, C. N., Dvorsky, M. R., Marsh, N. P., Sciberras, E., et al.
(2020). Remote learning during COVID-19: Examining school practices, service
continuation, and difficulties for adolescents with and without attention-
deficit/hyperactivity disorder. Journal of Adolescent Health, 67(6), 769-777.
doi:http://dx.doi.org.libproxy1.usc.edu/10.1016/j.jadohealth.2020.09.002

Board, B. A. C. (2014). Professional and ethical compliance code for behavior analysts.
Littleton, CO: Behavior Analyst Certification Board
Center for Medicare and Medicaid Services (2020). COVID-19 Frequently Asked Questions
(FAQs) on Medicare Fee-for-Service (FFS) Billing.
Centers for Disease Control and Prevention (2021). Health Equity Considerations and Racial and
Ethnic Minority Groups
Colizzi M, Sironi E, Antonini F, Ciceri ML, Bovo C, Zoccante L. Psychosocial and Behavioral
Impact of COVID-19 in Autism Spectrum Disorder: An Online Parent Survey. Brain
sciences. 2020;10(6):341-. doi:10.3390/brainsci10060341
Corbett, B. A., Muscatello, R. A., Klemencic, M. E., & Schwartzman, J. M. (2021). The impact
of COVID‐19 on stress, anxiety, and coping in youth with and without autism and their
parents. Autism Research, 14(7), 1496-1511. doi: 10.1002/aur.2521

Disabilities Rigths California (2021). Access to ABA Therapy. Pub. #7167.01.
https://www.disabilityrightsca.org/publications/access-to-aba-therapy

=

39

Donabedian, A. (1980). The definition of quality and approaches to its assessment . Health
Administration Press.

DuBay, M., Watson, L. R., & Zhang, W. (2018). In Search of Culturally Appropriate Autism
Interventions: Perspectives of Latino Caregivers. Journal of Autism and Developmental
Disorders, 48(5), 1623–1639. https://doi.org/10.1007/s10803-017-3394-8

Elder Jh, Kreider CM, Brasher Sn, Ansell M. Clinical impact of early diagnosis of autism on the
prognosis and parent-child relationships. Psychology Research and Behavior
Management. 2017;10:283-292.

Ellison, K. S., Guidry, J., Picou, P., Adenuga, P., & Davis, T. E. (2021). Telehealth and autism
prior to and in the age of covid-19: A systematic and critical review of the last decade.
Clinical Child and Family Psychology Review,
doi:http://dx.doi.org.libproxy1.usc.edu/10.1007/s10567-021-00358-0

Eshraghi, et al (2020). COVID-19: Overcoming the challenges faced by individuals with autism
and their families. The Lancet Psychiatry, 7(6), 481-483.

Exec. Order No. N-33-20 (2020), California.

Helt M, Kelley E, Kinsbourne M, et al. Can Children with Autism Recover? If So,
How? Neuropsychology Review. 2008;18(4):339-366. doi:10.1007/s11065-008-9075-

Halbur, M., Kodak, T., McKee, M., Carroll, R., Preas, E., Reidy, J., & Cordeiro, M. C. (2021).
Tolerance of face coverings for children with autism spectrum disorder. Journal of
Applied Behavior Analysis, 54(2), 600-617.

Harstad E, Huntington N, Bacic J, Barbaresi W. Disparity of Care for Children With Parent-
Reported Autism Spectrum Disorders. Academic Pediatrics. 2013;13(4):334-339.
doi:10.1016/j.acap.2013.03.010

Helt M, Kelley E, Kinsbourne M, et al. Can Children with Autism Recover? If So,
How? Neuropsychology Review. 2008;18(4):339-366. doi:10.1007/s11065-008-9075-

Hosokawa, R., Kawabe, K., Nakachi, K., Yoshino, A., Horiuchi, F., & Ueno, S. (2021).
Behavioral affect in children with autism spectrum disorder during school closures due to
the covid-19 pandemic in japan: A case-controlled study. Developmental
Neuropsychology,
doi:http://dx.doi.org.libproxy1.usc.edu/10.1080/87565641.2021.1939350

Hyman, S. L., Levy, S. E., & Myers, S. M. (2020). Identification, Evaluation, and Management
of Children With Autism Spectrum Disorder. Pediatrics (Evanston), 145(1), e20193447–.
https://doi.org/10.1542/peds.2019-3447

=

40

IBM Corp. (2017). IBM SPSS Statistics for Windows. Armonk, NY: IBM Corp. Retrieved from
https://hadoop.apache.org

Irvin DW, McBee M, Boyd BA, Hume K, & Odom SL. (2012) Child and family factors
associated with the use of services for preschoolers with autism spectrum disorder. Res
Autism Spectr Disord. 2012;6(1):565–572

Jeste, S., Hyde, C., Distefano, C., Halladay, A., Ray, S., Porath, M., et al. (2020). Changes in
access to educational and healthcare services for individuals with intellectual and
developmental disabilities during covid‐19 restrictions. Journal of Intellectual Disability
Research,

Kapp, S. K., Gillespie-Lynch, K., Sherman, L. E., & Hutman, T. (2013). Deficit, Difference, or
Both? Autism and Neurodiversity. Developmental Psychology, 49(1), 59–71.
https://doi.org/10.1037/a0028353

Kuhfeld, M., Soland, J., Tarasawa, B., Johnson, A., Ruzek, E., & Liu, J. (2020). Projecting the
Potential Impact of COVID-19 School Closures on Academic Achievement. Educational
Researcher, 49(8), 549–565. https://doi.org/10.3102/0013189X20965918

Kornack, J., Williams, A. L., Johnson, K. A., & Mendes, E. M. (2020). Reopening the Doors to
Center-Based ABA Services: Clinical and Safety Protocols During COVID-19. Behavior
Analysis in Practice, 13(3), 543–549. https://doi.org/10.1007/s40617-020-00462-7

Liptak S, Benzoni B, Mruzek W, et al. Disparities in Diagnosis and Access to Health Services
for Children with Autism: Data from the National Survey of Childrenʼs Health. Journal
of Developmental & Behavioral Pediatrics. 2008;29(3):152-160.
doi:10.1097/DBP.0b013e318165c7a0

LoPorto, J. (2020). Application of the Donabedian Quality-of-Care Model to New York State
Direct Support Professional Core Competencies: How Structure, Process, and Outcomes
Impacts Disability Services. Journal of Social Change, 12(1).
https://doi.org/10.5590/JOSC.2020.12.1.05

Maenner MJ, Shaw KA, Baio J, et al. Prevalence of Autism Spectrum Disorder Among Children
Aged 8 Years — Autism and Developmental Disabilities Monitoring Network, 11 Sites,
United States, 2016. MMWR Surveill Summ 2020;69(No. SS-4):1–12.
DOI: http://dx.doi.org/10.15585/mmwr.ss6904a1external icon

Magaña, S., Lopez, K., Aguinaga, A., & Morton, H. (2013). Access to diagnosis and treatment
services among latino children with autism spectrum disorders. Intellectual and
Developmental Disabilities, 51(3), 141-153.
doi:http://dx.doi.org.libproxy1.usc.edu/10.1352/1934-9556-51.3.141

=

41

Magiati I. Early comprehensive behaviorally based interventions for children with autism
spectrum disorders: a summary of findings from recent reviews and meta-
analyses. Neuropsychiatry. 2012;2(6):543-570. doi:10.2217/npy.12.59

Mahdavi, M., Vissers, J., Elkhuizen, S., Van Dijk, M., Vanhala, A., Karampli, E., Faubel, R.,
Forte, P., Coroian, E., & Klundert, J. (2018). The relationship between context, structure,
and processes with outcomes of 6 regional diabetes networks in Europe. PloS One, 13(2),
e0192599–e0192599. https://doi.org/10.1371/journal.pone.0192599

Manning, B. (2020). Perceptions of Families of Individuals with Autism Spectrum Disorder
during the COVID-19 Crisis. Journal of Autism and Developmental Disorders, 1–9.
https://doi.org/10.1007/s10803-020-04760-5

Matson, J., Rieske, R. (2014). Are outcome measures for early intensive treatment of autism
improving? Research in Autism Spectrum Disorders, 8(3), 178 – 185.
https://doi.org/10.1016/j.rasd.2013.11.006

Neece, M. (2020). Examining the impact of COVID‐19 in ethnically diverse families with young
children with intellectual and developmental disabilities. Journal of Intellectual
Disability Research, 64(10), 739–749. https://doi.org/10.1111/jir.12769

Nuñez, A., Le Roy, C., Coelho-Medeiros, M. E., & López-Espejo, M. (2021). Factors affecting
the behavior of children with ASD during the first outbreak of the COVID-19 pandemic.
Neurological Sciences, 42(5), 1675–1678. https://doi.org/10.1007/s10072-021-05147-9

Patten E, Baranek GT, Watson LR, Schultz B. Child and Family Characteristics Influencing
Intervention Choices in Autism Spectrum Disorders. Focus on Autism and Other
Developmental Disabilities. 2013;28(3):138-146. doi:10.1177/1088357612468028

Qu, H., Shewchuk, R. M., Chen, Y.-Y., & Richards, J. S. (2010). Evaluating the quality of acute
rehabilitation care for patients with spinal cord injury: An extended donabedian
model. Quality Management in Health Care, 19(1), 47–61.

Oster, N., Skillman, S., Stubbs, B., Dahal, A., Guenther, G., Frogner, B. (2020) The Physical
Therapist Workforce in the U.S.: Supply, Distribution, Education Pathways, and State
Responses to the COVID-19 Emergency. Center for Health Workforce
Studies/University of Washington. https://familymedicine.uw.edu/chws/wp-
content/uploads/sites/5/2020/11/PT_PB_Nov_23_2020.pdf

Qu, H., Shewchuk, R. M., Chen, Y.-Y., & Richards, J. S. (2010). Evaluating the quality of acute
rehabilitation care for patients with spinal cord injury: An extended donabedian
model. Quality Management in Health Care, 19(1), 47–61.

Reay, R. E., Looi, J. C., & Keightley, P. (2020). Telehealth mental health services during
COVID-19: summary of evidence and clinical practice. Australasian Psychiatry : Bulletin

=

42

of the Royal Australian and New Zealand College of Psychiatrists, 28(5), 514–516.
https://doi.org/10.1177/1039856220943032

Ruble, L., Heflinger, C., Renfrew, J., & Saunders, R. (2005). Access and Service Use by
Children with Autism Spectrum Disorders in Medicaid Managed Care. Journal of Autism
and Developmental Disorders, 35(1), 3–13. https://doi.org/10.1007/s10803-004-1026-6

Sandler AD. Value-Driven Care in Developmental-Behavioral Pediatrics, Part 2: Measuring
Quality and Meeting the Challenge. J Dev Behav Pediatr. 2019 Jul/Aug;40(6):479-488.
doi: 10.1097/DBP.0000000000000686. PMID: 31107770.

Siller M, Reyes N, Hotez E, Hutman T, Sigman M. Longitudinal change in the use of services in
autism spectrum disorder: Understanding the role of child characteristics, family
demographics, and parent cognitions. Autism. 2014;18(4):433-446.
doi:10.1177/1362361313476766

Smith, K. A., Gehricke, J.-G., Iadarola, S., Wolfe, A., & Kuhlthau, K. A. (2020). Disparities in
service use among children with autism: A systematic review. Pediatrics
(Evanston), 145(Suppl 1), S35–S46. https://doi.org/10.1542/peds.2019-1895G

Stahmer, A. C. (2007). The Basic Structure of Community Early Intervention Programs for
Children with Autism: Provider Descriptions. Journal of Autism and Developmental
Disorders, 37(7), 1344–1354. https://doi.org/10.1007/s10803-006-0284-x

Tarbox, J., Persicke, A., Kenzer, A. (2013) Home-Based Services. In D.D. Reed, F.D.
DiGennaro Reed, J.K. Luis (Eds.) Handbook of Crisis Intervention and Developmental
Disabilities (Vol. 38). Springer New York.

Wang, L., Mandell, D. S., Lawer, L., Cidav, Z., & Leslie, D. L. (2013). Healthcare Service Use
and Costs for Autism Spectrum Disorder: A Comparison Between Medicaid and Private
Insurance. Journal of Autism and Developmental Disorders, 43(5), 1057–1064.
https://doi.org/10.1007/s10803-012-1649-y

White, L. C., Law, J. K., Daniels, A. M., Toroney, J., Vernoia, B., Xiao, S., et al. (2021). Brief
report: Impact of covid-19 on individuals with asd and their caregivers: A perspective
from the spark cohort. Journal of Autism and Developmental Disorders,
doi:http://dx.doi.org.libproxy1.usc.edu/10.1007/s10803-020-04816-6

Yingling, Y. (2018). Time-Lag Between Diagnosis of Autism Spectrum Disorder and Onset of
Publicly-Funded Early Intensive Behavioral Intervention: Do Race–Ethnicity and
Neighborhood Matter? Journal of Autism and Developmental Disorders, 48(2), 561–571.
https://doi.org/10.1007/s10803-017-3354-3

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Effects of preferred physical activity on stereotypical behaviors in children with autism spectrum disorder: adapting from in-person to telehealth

PDF

A preliminary evaluation of a telehealth approach to acceptance and commitment training (ACT) for enhancing behavioral parent training (BPT) for Chinese parents

PDF

Use of a clinical decision model to support the transition of services for individuals with autism spectrum disorder

PDF

Let’s do something different: evaluation of lag schedules of reinforcement to increase flexibility and promote generalization in children with autism

PDF

Pregnancy in the time of COVID-19: effects on perinatal mental health, birth, and infant development

PDF

Mixed-method analysis for school-aged children with access and functional needs during the Covid-19 pandemic

PDF

Using Acceptance and Commitment Training and a virtual patient model for supporting families with children with autism spectrum disorder

PDF

The impact of the COVID-19 pandemic on cancer care delivery

PDF

The impact of the COVID-19 pandemic on K–12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

The impact of the COVID-19 pandemic on K–12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

The impact of the COVID-19 pandemic on K-12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

Evaluating the effects of virtual reality and acceptance and commitment therapy on music performance anxiety

PDF

The impact of the COVID-19 pandemic on K–12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

Social norms intervention to promote help-seeking with depressed Asian and European Americans: A pilot study

PDF

The impact of the COVID-19 pandemic on K–12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

The impact of the COVID-19 pandemic on K-12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

The impact of the COVID-19 pandemic on K–12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

Developmental trajectories of sensory patterns in young children with and without autism spectrum disorder: a longitudinal population-based study from infancy to school age

PDF

The impact of the COVID-19 pandemic on K–12 public school districts in southern California: responses of superintendents, assistant superintendents, and principals

PDF

Effects of online choir and mindfulness interventions on auditory perception and well-being in older adults during the COVID-19 pandemic

Asset Metadata

Creator Gonzales, Cassin Wolff (author)

Core Title The impacts of the COVID-19 pandemic on therapy utilization among racially/ethnically and socio-economically diverse children with autism spectrum disorder

School College of Letters, Arts and Sciences

Degree Master of Arts

Degree Program Psychology

Degree Conferral Date 2022-05

Publication Date 02/25/2022

Defense Date 10/15/2021

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag access,Applied Behavior Analysis,autism spectrum disorder,COVID-19,health insurance,health systems,OAI-PMH Harvest,Race

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Tarbox, Jonathan (committee chair), Huey, Stanley (committee member), Lopez, Steven (committee member), Schwartz, David (committee member)

Creator Email cassiegonzales11@gmaill.com,cwgonzal@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC110765146

Unique identifier UC110765146

Legacy Identifier etd-GonzalesCa-10403

Document Type Thesis

Format application/pdf (imt)

Rights Gonzales, Cassin Wolff

Type texts

Source 20220301-usctheses-batch-914 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu