A Multisite Study Evaluating the Benefits of Early Intervention via Telepractice

Abstract

This study sought to determine the effectiveness of telepractice as a method of delivering early intervention services to families of infants and toddlers who are deaf or hard of hearing. A comparison group design was applied to ascertain the child, family, and provider outcomes via telepractice compared with traditional in-person home visits. A total of 48 children and their families, along with 15 providers from 5 early intervention programs, across the country participated. Children in the telepractice group received more intervention, although the number of prescribed sessions was equal across groups. Analyses of covariance demonstrated that children in the telepractice group scored statistically significantly higher than children in the in-person group on the PLS-5 Receptive Language subscale and PLS-5 Total Language standard scores, and the groups scored similarly on other language measures. There were no statistically significant differences between groups in regard to family outcomes of support, knowledge, and community involvement. Analysis of video recordings of telepractice versus in-person home visits resulted in higher scores for provider responsiveness and parent engagement. This study supports the effectiveness of telepractice in delivering early intervention services to families of children who are deaf or hard of hearing. Further research involving randomized trials with larger, more diverse populations is warranted.

HEARING LOSS is the most common congenital condition (^{White, 2003}) with approximately 16,000–18,000 infants and toddlers identified nationally each year (^{Cole & Flexer, 2011}). Prior to 1998, hearing loss typically was not identified until approximately 2.5–3 years of age (^{Harrison, Roush, & Wallace, 2003; Hoffman & Beauchaine, 2007; White, Forsman, Eichwald, & Munoz, 2010}). Now 97% of infants in the United States receive a newborn hearing screening (^{Centers for Disease Control and Prevention, 2010}), lowering the average age of identification to 2 months of age (^{White et al., 2010}). This has resulted in substantially more infants and young children who are deaf/hard-of-hearing (DHH) entering the early intervention (EI) system and needing services from providers with experience serving this population.

Earlier identification combined with advanced hearing technology (such as cochlear implants and digital hearing aids) and EI has resulted in improved outcomes for children who are DHH (^{Bond et al., 2009; Sparreboom et al., 2010; Turchetti, Bellelli, Palla, & Forli, 2011}); however, access to specialized personnel is key to these improved outcomes. In addition to an increased number of children identified as DHH, there is also a trend for more families to choose a listening and spoken language communication modality for their children. In fact, 80%–90% of families of newly identified children who are DHH are now choosing a listening and spoken language communication approach when such therapies are available (^{Alberg, Wilson, & Roush, 2006; Brown, 2006}). Unfortunately, there is a shortage of professionals trained to provide specialized services, such as validation of hearing technology and development of auditory skills, to families who choose this communication approach (^{Muñoz, Bradham, & Nelson, 2011}).

Part C EI services are designed to support the development of infants and toddlers with special needs and to strengthen the capacity of families to foster their development (^{“Workgroup on Principles and Practices,” 2008}). The home is considered the primary location of services, and it is viewed by many programs as important in the effort to provide services in a “natural environment” as deemed necessary under the Individuals with Disabilities Education Improvement Act (IDEA; ^{Campbell & Sawyer, 2007; IDEA, 2004}). IDEA specifies that these services are to be family-centered, with families being actively engaged in setting goals and implementing intervention. The role of the provider as a “coach” to support families of children with disabilities is a critical practice of family-centered services, allowing more opportunities to foster the child's development via the family implementing activities within their daily routines (^{Campbell, 1997; Dunst, 2002; Korfmacher et al., 2008; Rush & Sheldon, 2011}).

In keeping with these principles, telepractice is being used to deliver services by a rapidly growing number of Part C programs, particularly those serving children who are deaf or hard of hearing (^{Houston & Stredler-Brown, 2012}). Telepractice refers to the use of telecommunications technology to the delivery of speech–language pathology and audiology professional services at a distance by connecting the clinician to client for assessment, intervention, and/or consultation (^{American Speech-Language-Hearing Association, 2015}). It has become an increasingly popular service modality to address the shortages of personnel trained to serve children in remote and/or rural areas, and it is used as a strategy to reduce travel costs and increase flexibility among programs serving the broader Part C population (^{Cohn & Cason, 2012; Olsen, Fiechtl, & Rule, 2012}).

Several studies have reported benefits of telepractice, specifically, high provider and family satisfaction (^{Cason, 2009; Crutchley & Campbell, 2010; Heimerl & Rasch; 2009; Kelso, Fiechtl, Olsen, & Rule, 2009; Peters-Lalios, 2012}). These studies reported that telepractice resulted in high overall satisfaction with services in terms of child progress toward speech and language goals, clinician responsiveness, and clinician accessibility. ^{McCarthy, Muñoz, and White (2010)} also reported high provider and family satisfaction, increased participation and use of the parent-coaching model, and decreased cancellations with the use of telepractice. ^{Behl, Houston, Guthrie, and Guthrie (2010)} reported increased access to specialized services (i.e., auditory–verbal therapy) with telepractice in addition to increased flexibility in scheduling and reduced travel for families. ^{Blaiser, Behl, Callow-Heusser, and White (2013)} provided the most relevant evidence to date via the implementation of a randomized study with 27 families of infants and toddlers who are DHH in Utah. The findings showed statistically significant expressive language scores in favor of the Telepractice group using a criterion-referenced tool. In addition, this study reported statistically significant enhanced family engagement in favor of the Telepractice group. However, the sample size of this study was relatively small and the average intensity of intervention was 2 visits per month, which is less than what is typically recommended for this population.

One challenge to conducting comparison design research with this population is the difficulty in obtaining a sufficiently large sample size. Few programs serve the number of families needed to confidently match on key characteristics, such as age, degree of hearing loss, and other disabilities. As a practical solution, a collaborative study was conducted with five EI programs serving infants and toddlers who are DHH. The purpose of this study was to compare the outcomes of telepractice to traditional in-person services for children, families, and providers in Part C EI programs. Specifically, the study sought to answer the following questions: (1) How do families and their children who are DHH who receive services via telepractice compare to those who receive services via in-person visits in regard to child and family outcomes? (2) How does the quality of home visit interaction for those served via telepractice compare to those served via in-person visits, and (3) How does the use of telepractice impact the intensity of services delivered to families of children who are DHH compared to families receiving only in-person services?

METHODS

A study using a comparison group design was conducted through a national policy center. The center recruited five EI programs using telepractice to participate in this study. The administrators of these programs all wanted to conduct research about the efficacy surrounding telepractice, yet each was lacking a sufficient sample size to conduct such a study on their own. All but one site had at least 1–2 years of experience in implementing telepractice to deliver EI services.

Participants

The five participating programs were located in Maine, Missouri, Utah, Washington, and Oregon. Each site was responsible for the recruitment of families and providers, obtaining informed consent, and administration of the research protocols. The participating programs also were responsible for providing their own equipment and ensuring adequate Internet connectivity for their providers and families. Data were sent to the researchers for analysis. The institutional review board approval was obtained through Utah State University.

Children and families

Children were assigned to the Telepractice group or the In-person group after matching on child chronological age, degree of hearing loss, other disabilities/conditions, amount of prescribed EI sessions, and their EI provider. Random assignment was possible with about half of the subjects, while other subjects were assigned on the basis of logistical considerations. For example, some families needed to receive services via telepractice only due to their personal life circumstances; in such cases, comparison group children were then selected after matching the children in each group on the aforementioned key characteristics.

Table 1 compares the characteristics of the active subjects in the in-person group compared with the telepractice group. As can be seen, children in the two groups were well matched on all key characteristics. The children were 19–20 months of age on average at the start of the study. There were no statistically significant differences between groups, and each group had at least four sessions per month. As shown in Table 1, the families in the telepractice group reported slightly more use of computers and videoconferencing prior to the study, and the majority of families had experienced telepractice prior to the study, in comparison to only one of the In-person group subjects. Table 2 presents the degree of hearing loss for participating children; a chi-square test for differences showed that there were no statistically significant differences among groups (p = .71).

Table 1. - Child and Family Characteristics

Characteristic	Telepractice (n = 23)	In-Person (n = 25)	p
Child age (months)	20.2	19.0	.72
Age of HL diagnosis	3.3	5.7	.25
Age at the start of EI	6.7	6.9	.95
% Listening/spoken language	82	92	.44
Primary caregiver's age (years)	32.8	34.5	.42
Primary caregiver's education	14.5	14.7	.86
Income	$60,886	$68,680	.57
Other therapy received (%)	50	40	.69
Computer usea	3.9	3.4	.04
Video conference usea	2.0	1.6	.17
Prior telepractice (%)	56.5	8.0	.01
Prescribed sessions	4.4	4.2	.77

Note. EI = early intervention; HL = hearing loss.

^a1 = never or less than once a month; 2 = more than once a month but not every week; 3 = at least once a week but not every day; 4 = every day.

Table 2. - Degree of Hearing Loss by Group

Degree of Hearing Loss	Telepractice (n = 23)	In-Person (n = 25)
Unilateral, atresia, or bilateral mild	5	5
Bilateral mild/moderate	3	4
Bilateral moderate	2	3
Bilateral moderate/severe	6	4
Bilateral severe	0	2
Bilateral severe/profound	1	0
Bilateral profound	6	6

Service provider characteristics

Fifteen providers across five sites participated in the study; all of the providers served families in both the telepractice and in-person groups. Table 3 shows demographic characteristics for the providers serving children in each group. Averages are weighted on the basis of the number of children served by each provider. For example, if a provider served three children in the telepractice group, the provider was counted three times. Roughly half (eight) of the providers were deaf educators; four identified themselves as speech–language pathologists and three identified themselves as auditory–verbal therapists. All of the providers had at least 3 years of experience working with children who are DHH in EI. The amount of experience delivering services via telepractice was quite variable, ranging from no experience to more than 2 years.

Table 3. - Characteristics of Service Providers

Average provider characteristics
Years in early intervention	11.9
Years serving hearing loss	16.8
Months of experience in telepractice	15.2
Age in years	43.2
Professional role
Number of deaf educators	8
Number of SLPs	4
Number of AVTs	3

Note. AVT = auditory verbal therapists; SLP = speech/language pathologists.

Intervention

The intervention period for this study was 6 months, with the focus of the sessions in both groups being on coaching the caregivers to enhance the child's receptive and expressive language development, such as through turn taking, modeling language, and getting the child's attention. Activities during the sessions were based on the family's daily routines within their natural environments. During this study period, the telepractice group families received 90% of their EI services via telepractice, with only 10% of their visits delivered in-person. Telepractice services were delivered via a variety of technologies. Both families and providers in two of the sites (11 of the 23 children in the Telepractice group) used iPads and FaceTime software; these programs provided the iPads for families. Laptops were employed at the other sites, and the software included systems such as MOVI, Zoom, or Vidyo. These families typically used their own computer hardware. The in-person group received their services via traditional in-person home visits. The number of scheduled EI sessions for each family adhered to what was prescribed in their Individualized Family Service Plan, with an average of 3.8 visits per month for the children in the telepractice group and 4.1 visits for the children in the in-person group at the time of group assignment.

Measures

A variety of assessments were used. The specific assessments to measure service delivery as well as child and, family outcomes are described later.

Service delivery measures

The Monthly Time and Activity Form was a self-report form completed by providers for 5 of the 6 months of the intervention period to document the amount of time required to implement telepractice compared with traditional in-person visits. For the first visit of each month during the intervention period, providers reported the amount of time spent doing specific activities, such as preparation, travel, intervention, and documentation. For cancelled visits, providers were asked to rate the impact of this cancellation on their work ability, with 1 being little impact to 3 being major impact. In addition, the providers reported on this form the number of sessions delivered and cancelled during the previous month; this was used to calculate average sessions provided per month. These forms were developed by the authors and revised on the basis of the forms used in previous telepractice research (^{Blaiser et al., 2013}).

Child listening and language measures

Ultimately, the desired goal of EI is to promote optimal developmental growth in children with disabilities. This is achieved by partnering with families to support their children's development and ensuring that intervention fits with the family's routines, values, and needs. In the case of children who are deaf or hard of hearing, the focus of EI is on increasing the children's communication via increasing the parents' knowledge and skills in fostering their children's communication. Three tools were administered by program providers to assess growth in these areas:

• Preschool Language Scale–5th Edition (PLS-5; ^{Zimmerman, Steiner, & Pond, 2011}), a standardized, norm-referenced assessment that provides standardized scores for receptive and expressive language, along with a total score.
• The Auditory Skills Checklist (ASC; ^{Caleffe-Schenck, 2006}), a criterion-referenced tool, used to measure listening skills, an important area of focus for children who are DHH. A total numerical raw score was calculated in accordance with the publisher's recommendations.
• The MacArthur-Bates Communication Developmental Inventory Words and Sentences (MBCDI; ^{Fenson et al., 1993}), a parent self-report tool that provides raw scores and percentile scores. Parents document the child's production and use of hundreds of words divided into semantic categories. The parent then identifies the words the child understands or uses, providing written examples of the child's three longest sentences.

Family-centered service measures

Two tools were used to measure the impact of mode of service delivery on the extent to which services supported the family's needs and the demonstration of family-centered intervention principles, such as inclusion in the community, active engagement in intervention sessions, and positive parent–child interactions:

• The Family Outcomes Survey (^{The Early Childhood Outcomes Center, 2014}) was used to assess important EI outcomes pertaining to supporting, educating, and increasing community inclusion of families. This is a nationally recognized tool used to assess Part C-related outcomes for families. The first five scales represent the caregiver perceptions of how well they are able to support their child's needs. The last three scales are ratings of the caregivers' perceptions of how helpful EI has been to the family. Caregivers assessed themselves and the program using a Likert-type scale with high scores being better.
• The Home Visit Rating Scales–Adapted and Extended (HOVRS-A+, v2.0; ^{Roggman et al., 2010}) was used to measure the quality of intervention and interpersonal dynamics of the intervention sessions. The HOVRS-A+ was designed to measure excellence based on evidence-based practices for home visits with families of children ages birth to 24 months. The scales measure four indicators of home visit quality based on the home visitor's relationship with family, responsiveness to family, facilitation of the parent–child interaction, and nonintrusiveness and collaboration. Three additional scales provide indicators of home visitor effectiveness during the home visit: parent–child interaction, parent engagement, and child engagement. A 7-point scale is used, with a rating of 1 indicating “needs training,” 3 = adequate, 5 = good, and 7 = excellent. Strong internal consistency is reported (^{Vogel et al., 2011}). One video recording was attempted for each family; the recorded sessions were used for the Telepractice group. Recordings were independently coded by one of the authors of the HOVRS-A+.

Analysis

Pretest scores and the time and activity data were analyzed using analyses of variance to compare the telepractice group with the in-person group. To determine the impact of the 6-month intervention period for the two groups, analyses of covariance (ANCOVAs) were used with child and family outcomes measures. Because of the small sample size and the lack of complete randomization, ANCOVA modeled the effect of the intervention while controlling for the effects of covariates. This statistical technique was chosen to increase statistical power, which otherwise was relatively low due to the sample size, and it helped control for pretest differences among the groups (although there were few documented differences as mentioned previously). Two covariates were selected for each model. These covariates were selected on the basis of their theoretical importance and their relatively high correlations with the outcome, thereby adjusting for these factors that may have influenced the results. Standardized mean difference effect sizes were calculated for variables for which there were statistically significant differences to reflect the size of the differences between groups. For example, an SMDES of 1.0 reflects a difference of 1 standard deviation difference, and 0.10 reflects 1/10th of a standard deviation difference.

RESULTS

The results comparing the telepractice group with the in-person group are provided, beginning with the data pertaining to service delivery. The child and family outcome data are presented next, reporting the pretest scores followed by posttest analysis.

Service delivery outcomes

The average time spent for activities associated with provision of an EI session was analyzed on the basis of provider reports gathered for 5 of the 6 months that comprised the intervention period. Table 4 presents the mean and median number of minutes spent per activity for the telepractice group versus the in-person group, followed by the average number of sessions and minutes delivered per month in Table 5, and reasons for cancellations presented in Table 6.

Table 4. - Provider Report of Time Spent per Activity by Group in Minutes

Activity	Telepractice		In-Person
	Mean	Median	Mean	Median
Preparation time	12	10	13	10
Coordination time	15	15	14	13
Drive time	NA	NA	61	48
Drive mileage	NA	NA	39	34
Technology troubleshooting	11	5	NA	NA
Intervention time	55	60	55	60

Note. NA = not applicable.

Table 5. - Average Number of Reported EI Sessions, Average Number of Session Cancellations, and the Average Minutes in Sessions Summed Over the Month per Child

Month	Telepractice				In-Person
	N	Sessions	Cancellations	Total Minutes in Sessions	N	Sessions	Cancellations	Total Minutes in Sessions
December	21	3.48	0.95	221	23	2.13	1.00	130.9
January	23	3.82	0.86	225.6	24	2.58	0.67	144.4
February	22	3.83	0.59	217.6	25	2.48	0.76	134.2
March	19	3.53	1.79	209.3	20	2.5	0.75	142.5
April	23	3.61	1.22	201.7	25	2.36	3.12	127.8
Totala	23	3.74	1.10	218.7	25	2.43	1.23	137

Note. EI = early intervention.

^aGroup differences tested via t tests across the months: average number of sessions per month (p < .001), total average number of cancellations per month (p = .780), and average number of minutes in sessions per month (p < .001).

Table 6. - Reason for Cancellations of EI Sessions

Type of Cancellation	Telepractice	In-Person
Sick	6	6
Weather	0	2
Transportation	0	0
Family scheduling	7	3
Provider scheduling	1	0
Computer/Internet	3	0
Family “no show”	5	3
Total	22	14
Average impact of cancellationsa	1.24	1.25

^a1 = minor impact on work; 2 = moderate impact; 3 = major impact.

Preparation, coordination, and intervention time

The mean and median scores for these activities were essentially the same for the telepractice group versus the in-person group. The typical session was an hour in duration, with approximately 12–13 min spent in preparation for the session and 14–15 min spent in coordinating with other team members in relation to the EI session.

Drive time and mileage for in-person group

As shown in Table 4, the average drive time for providers to serve families in the in-person group was 61 min per visit. The average mileage for providers driving to the homes of families in the in-person group was 39 miles, with the median being 34 miles.

Time spent setting up and troubleshooting technology

Providers reported the average amount of time spent in setting up the technology needed to conduct a telepractice group session as well as time spent in any trouble-shooting activities when problems arose. The average time spent on technology aspects of telepractice sessions was 11 min, with the median being 5 min.

Number of intervention sessions delivered

Providers were asked to complete monthly logs on the number of sessions completed, the number of sessions cancelled, and the total number of min spent delivering services. As shown in Table 5, the telepractice group received a statistically significantly greater number of visits and more minutes of intervention, with the total number of minutes of intervention also being more for the telepractice group. The number of cancellations was variable across both groups, with the average number being similar across groups.

Reasons for cancellations

Providers reported the reasons for cancelled visits to understand the extent that telepractice reduced cancellations due to weather, transportation problems, or illness. As shown in Table 6, the primary reason for a cancelled visit was illness, and this was equal across both the telepractice and in-person groups. Weather impacted only the in-person group, and computer/Internet issues impacted only the telepractice group. Family scheduling problems and “no shows” were more predominant in the telepractice group. However, based on chi-square analysis, these differences were not statistically significant across groups. In addition, providers were asked to rate the impact of a cancelled visit on their schedule. For example, a provider would record a “1” if the cancellation had minimal impact, (i.e., the provider was able to use the time efficiently), versus “3,” which meant that there were severe constraints on productive use of that time. Cancellations resulted in relatively minimal impact, on average, for providers serving families in both groups.

Child outcomes

A slightly smaller sample size was used in the analyses of the PLS child outcome data due to omissions in the data for the full cohort; thus, only subjects for whom full PLS data at pre- and posttest times are reported. As shown in Table 7, there were no statistically significant differences between groups at pretest time on the scores obtained from the PLS-5, the ASC, or the MBCDI. For the children across both groups, the scores reflect low standard scores compared with the normative population.

Table 7. - Child Pretest Scores by Group

Outcome	Telepractice			In-Person			p
	N	Mean	SD	N	Mean	SD
PLS-5 expressive	20	91.6	15.4	25	94.0	16.3	.61
PLS-5 receptive	20	86.9	19.5	25	90.7	16.2	.49
PLS-5 total	20	88.5	17.1	25	91.9	16.6	.50
MBCDI vocab. count	23	114.4	157.7	25	117.3	179.7	.88
Auditory skills checklist	23	28.3	17.3	25	27.6	20.8	.80

Note. MBCDI = MacArthur-Bates Communicative Development Inventories; PLS-5 = Preschool Language Scale–5.

The PLS-5 was again administered at posttest time. Both the telepractice group and the in-person group made statistically significant gains (p = .001) in age equivalent scores, with an average of 9.6 months for the telepractice group and 6.6 for the in-person group. To assess between-group differences at posttest, ANCOVA was conducted. Covariates consisted of the subjects' pretest scores and the number of sessions received. The results shown in Table 8 reflect statistically significant differences in favor of the telepractice group compared with the in-person group on Receptive Communication (p = .02) and Total Language Scores (p = .05) on the PLS-5, with standardized mean difference effect sizes (SMDES) ranging from 0.26 to 0.30, respectively. There was no statistically significant difference on the PLS-5 Expressive Communication subscale (p = .21). Results of ANCOVA showed no statistically significant differences in the MBCDI vocabulary between these two groups (p = .88); in addition, the ASC posttest scores for the two groups were statistically similar (p = .14).

Table 8. - Child Posttest Scores: Analysis of Covariance (ANCOVA), N, Means, Standard Deviations, and Adjusted Means

Outcome	Telepractice				In-Person				p	Effect Sizea
	N	Mean	SD	Adj. Mean	N	Mean	SD	Adj. Mean
PLS-5 Expressiveb	20	98.4	12.5	98.4	25	93.6	17.1	93.5	.21	0.17
PLS-5 Receptiveb	20	97.6	16.0	98.9	25	89.7	18.1	88.6	.02	0.30
PLS-5 Totalb	20	97.9	14.5	98.8	25	91.1	18.1	90.4	.05	0.26
MBCDI Vocab. Countc	23	217.1	216.5	214.5	25	184.5	240.0	207.0	.88	0.01
Auditory skills checklist	23	39.2	18.2	38.3	25	32.9	20.3	33.7	.14	0.12

Note. Each ANCOVA model included the respective test's pretest score and average number of minutes in sessions per month as covariates.

^aEffect size based on the mean of the two groups' standard deviation.

^bPreschool Language Scale-5.

^cMacArthur-Bates Communicative Development Inventories.

Family-centered service outcomes

As shown in Table 9, there were no statistically significant differences at pretest for any of the subscales on the Family Outcomes Survey. Based on ANCOVA-adjusted means, there were also no statistically significant differences at posttest time (see Table 10). In general, mean scale scores reflect that families viewed themselves as “almost always” having the skills in question, and that the EI services they received were rated as “generally helpful” or above average.

Table 9. - Pretest Family Outcomes Survey Subscale Scores by Group

Family Outcome Survey Subscale Themesa	Telepractice (n = 23)		In Person (n = 25)		p
	Mean	SD	Mean	SD
We understand my child's strengths, needs, abilities (0–16 points)	13.9	2.3	13.3	2.5	.39
We know our rights and advocate for our child (0–20 points)	16.8	2.6	15.5	3.5	.15
We help our child develop and learn (0–16 points)	14.4	1.6	14.3	1.9	.89
We have support systems (0–20 points)	17.9	2. 5	16.9	4.2	.05
We have access to community (0–24 points)	23.1	1.6	21.1	3.6	.01
The program helped me know my rights (0–20 points)	16.1	3.6	15.4	3.2	.46
The program helped me communicate child's needs (0–24 points)	21.4	2.6	20.0	3.3	.11
The program helped me learn how to support my child's development (0–24 points)	20.8	3.3	19.9	3.6	.39

^aHigher scores reflect higher positive ratings.

Table 10. - Family Outcomes Survey by Group ANCOVA results: Unadjusted Means, Standard Deviations, Adjusted Means, P Values, and Effect Sizes^a

Outcome	Telepractice				In-Person				p	Effect Sizea
	N	Mean	SD	Adj. Mean	N	Mean	SD	Adj. Mean
We understand my child's strengths, needs, and abilities	23	14.4	2.02	14.4	25	14.1	1.69	14.2	.91	0.03
We know our rights and advocate for our child	23	17.0	3.24	16.6	25	16.2	2.99	16.6	.95	0.02
We help our child develop and learn	23	14.6	1.95	14.6	25	14.4	1.79	14.4	.71	0.11
We have support systems	23	17.6	2.71	17.2	25	16.4	2.59	16.8	.62	0.14
We have access to community	23	23.1	1.29	22.4	25	21.5	2.77	22.2	.56	0.13
The program helped me know my rights	23	17.1	3.86	17.0	25	16.7	3.46	16.9	.94	0.02
The program helped me communicate child's needs	23	22.5	2.19	22.3	25	22.0	2.51	22.3	.99	0.01
The program helped me learn how to support my child's development	23	22.3	2.05	22.2	25	21.5	2.40	21.6	.33	0.25

Note. Each ANCOVA model included the respective test's pretest score and child age in months at pretest as covariates. ANCOVA = analysis of covariance.

^aEffect size based on the mean of the two groups' standard deviation.

^bHigher scores reflect higher positive ratings.

Table 11 provides the mean scores for the HOVRS-A+ scales. Recordings of sufficient quality for coding were obtained for 17 subjects in the telepractice group and 19 in the in-person group. Overall, ratings reflected above average performance for both groups. Analyses of covariance results showed that the telepractice group scored statistically significantly better in regard to the provider's responsiveness to the caregiver as well as stronger caregiver engagement. There were no statistically significant differences on the other scales, although the mean scores were higher for the telepractice group than for the in-person group for 5 of the 6 other subscales and the average SMDES for all 7 subscales was 0.32.

Table 11. - Home Visit Rating Scale by Group: Unadjusted Means, Standard Deviations, and ANCOVA Adjusted Scores

Home Visit Rating Subscales	Telepractice (n = 17)			In-Person (n = 19)			p	Effect Sizea
	Mean	SD	Adj. Mean	Mean	SD	Adj. Mean
Provider responsiveness to familyb	5.24	0.83	5.35	4.65	0.99	4.48	<.01	0.95
Provider relationship with family	4.53	0.87	4.53	4.80	0.62	4.78	.33	−0.33
Provider facilitation of parent–child interaction	4.94	0.97	4.98	4.80	1.20	4.70	.40	0.26
Provider nonintrusiveness and collaboration	5.06	1.03	5.11	4.70	1.08	4.59	.15	0.49
Parent–child interaction	5.41	0.71	5.41	5.05	1.05	5.00	.21	0.48
Parent engagementb	5.76	0.56	5.77	5.35	0.93	5.32	.04	0.60
Child engagement	5.29	1.05	5.31	5.10	1.37	5.14	.71	0.12

Note. Each ANCOVA model included the primary caregiver's education and the severity of the hearing loss as covariates. ANCOVA = analysis of covariance.

^aEffect size estimated using mean of standard deviation.

^bStatistically significant at an alpha level of .05.

DISCUSSION

While there is increasing interest in and implementation of telepractice for children who are deaf or hard of hearing across the country, there are few studies that systematically examine the differences in outcomes of children served using a telepractice model compared with children who have received services via a traditional in-person model. The multisite study reported here plays an important role in building on the knowledge gained related to the earlier study of the effectiveness of telepractice conducted by ^{Blaiser et al. (2013)}. This study contributes to the research by (1) including more children who are DHH than have been involved in previously reported research studies; (2) using well-established norm-referenced, standardized tools of child language development; and (3) engaging diverse EI programs and providers from across the country.

The results pertaining to the impact of telepractice on service delivery itself are important, as demonstrated by the finding that the number of visits and minutes of intervention received were higher for the telepractice group. This can be viewed as a positive outcome of telepractice, while it also may be viewed as a product that confounds the child and family outcome data. To address this, the authors used the number of sessions received as a covariate to adjust for this factor and measure the impact of the service delivery mode alone. Future research that would systematically control the number of sessions received across groups would be useful. The service delivery data did not reflect dramatic reductions in cancelled visits per se, nor did telepractice appear to reduce negative impact on provider schedules as an often-theorized consequence of in-person visits. This may be due to a likely increased ease of rescheduling visits without necessarily counting the initial scheduled appointment as a cancellation. Because such outcomes have been reported in previous studies, these factors are important to be assessed in future research. Data from this study also reinforce the premise that telepractice can reduce drive time for providers, an aspect that impacts costs to programs. However, there is the need to consider time required for trouble-shooting and addressing technical issues when implementing telepractice. A report by the authors on a cost analysis of telepractice versus in-person visits is anticipated to be reported in the near future, and these nuanced issues will be addressed.

This study demonstrated the positive outcomes of telepractice in response to the primary research question, “How do families and their children who are DHH who receive services via telepractice compare to those who receive services via in-person visits in regard to child, family, and service delivery outcomes?” These results support the conclusion that telepractice is a valuable method for providing services to children who are DHH and their families. On average, families and children receiving services via telepractice had at least the same if not better language outcomes and auditory skills than children who received services solely through traditional in-person visits. According to the family self-report measures, families who received telepractice services felt equally supported, knowledgeable, and confident in fostering their children's development as families who received in-person services, demonstrating that telepractice supports the Part C mission of enhancing the capacity of families to meet their children's needs (^{IDEA, Part C §631, 2004}).

One particularly relevant outcome of this study pertains to the extent to which telepractice can promote the recommended practice of coaching families within the context of natural environments. In spite of this being widely accepted as best practice, past research has reported that the use of coaching has been a challenge for Part C providers during traditional home visits (^{Campbell & Sawyer, 2007; Colyvas, Sawyer, & Campbell, 2010; Peterson, Luze, Eshbaugh, Jeon, & Kantz, 2007}). Unfortunately, many providers still demonstrate child-directed intervention while families observe rather than supporting the family as the child's natural teacher. Recent research found that the amount of time spent on parent–child interactions is associated with higher quality visits (^{Aikens et al., 2015}) and that both the parenting environment and child language development are predicted by home visiting quality, particularly parent engagement (^{Roggman et al., 2016}). The analysis of data obtained from the HOVRS-A+ in this study demonstrated that the families who received services via telepractice were more engaged in the intervention than the families in the in-person group, and providers were more responsive to the families when providing services via telepractice compared with those they served in the in-person group. Again, these findings reinforce those found in previous studies (^{Blaiser et al., 2013}). When using telepractice, it is likely that the provider's reliance on communicating directly with the caregiver versus the child naturally fosters the increased use of coaching within the natural environment, thus increasing the use of this Part C-recommended principle.

The limitations of this study are important to keep in mind when interpreting the results and in guiding future research. First, this was an applied research design, which was vulnerable to the constraints of real-world EI programs. Random assignment was possible across some but not all of the sites, so every effort was made to match subjects on relevant characteristics and prescribed service. True random assignment would increase the scientific rigor. Second, although this study had a relatively large sample size in relation to the existing literature pertaining to telepractice with infants and toddlers, a larger sample size would increase statistical power and it would enhance the generalization of these results. Third, the child developmental measures were administered by the family's primary EI provider; ideally, objective testers blind to group assignment would have been used. Fourth, replication with a more culturally and linguistically diverse population of families is needed, given the homogeneity of the subjects in this study. Finally, additional observational data regarding the quality of the intervention sessions would be valuable beyond the single video recording collected in this study. Data pertaining to the technical aspects of the telepractice sessions, such as the number of video and/or audio interruptions, were not collected because of budgetary constraints. Further research that can address these limitations would be an important contribution to the efficacy of telepractice for this population.

This study verifies that telepractice can support the development of infants and toddlers who are deaf or hard of hearing, and it reinforces the use of telepractice to serve families of infants and toddlers with other special needs. However, these results must not be construed to mean that telepractice should take the place of in-person home visits. Rather, these findings reinforce the stance that telepractice is a valuable tool in the overall goal of EI, which is to deliver comprehensive, family-centered services. Ultimately, all aspects of intervention, including the format for service delivery, should be based on family needs, priorities, and cultural considerations. From a practical standpoint, telepractice can help reduce costs associated with provider travel and provider shortages, which are primary deterrents for programs serving this population. Therefore, these findings can inform policy makers, program directors, providers, and families of the value of incorporating telepractice into EI programs that serve not only those whose children are DHH but those serving the broader Part C population.