The Efficacy of Telehealth-Delivered Speech and Language Intervention for Primary School-Age Children: A Systematic Review

The purpose of this article is to determine if telehealth-delivered speech-language pathology interventions are as effective as traditional in-person delivery for primary school-age children with speech and/or language difficulties. A systematic review was conducted (in accordance with PRISMA guidelines) using five databases, two journals and reference lists. Titles and abstracts were screened for inclusion, with relevant studies reviewed in full-text. Initial searches identified 132 articles. Following exclusion of non-relevant studies, seven articles remained for inclusion. Results revealed both telehealth and in-person participants made significant and similar improvements when treatment effects were measured through five of the six outcome measures. Findings showed there is limited but promising evidence to support telehealth for delivering speech-language pathology intervention services to school-age children. Whilst this is encouraging, particularly for rural children where in-person services are limited, more rigorous study designs are required to support the efficacy of telehealth for this population.

telecommunications technology to deliver clinical services at a distance by linking clinician to client, caregiver, or any person(s) responsible for delivering care to the client, for the purposes of assessment, intervention, consultation and/or supervision" (SPA, 2014, p. 4). The governing body of the SLP profession, Speech Pathology Australia (SPA), supports and encourages the use of telehealth but recommends that the services provided through this service delivery model be "equivalent to standard clinical care" (SPA, 2014).
Telehealth-delivered SLP services have previously been investigated in reviews regarding a number of practice areas and populations. Mashima and Doarn (2008) conducted an extensive literature review on the application of telehealth in SLP with adults and a small number of studies with children. They reviewed 40 studies investigating disorders relating to adult neurogenic communication, fluency, voice, dysphagia (n=35), and childhood speech and language (n=5). This review suggested that telehealth is a feasible and effective method for providing SLP services at a distance. However, the authors noted that the reviewed literature consisted primarily of pilot studies and anecdotal accounts of telehealth applications rather than large, wellcontrolled, randomised clinical trials (Mashima & Doarn, 2008). Reynolds, Vick, and Haak (2009) conducted a narrative review of 29 studies which were analysed using a quality assessment checklist. These 29 articles focused on assessment and intervention with the adult (n=19) and paediatric (n=7) population as well as an unspecified population (n=3). The authors concluded that the results achieved through the telehealth and in-person service delivery models were equivalent; however, many of the studies noted that telehealth was not a complete replacement for in-person services but may be appropriate for combined practices. These findings were consistent with the review conducted by Theodoros (2012), which investigated 19 studies regarding adult neurogenic communication, voice, stuttering, dysphagia and laryngectomy follow-up and four studies regarding paediatric speech, language and literacy disorders. Edwards, Stredler-Brown, and Houston (2012) conducted a further review investigating 39 studies in the fields of audiology and SLP. The majority of these studies were conducted on adult populations (n=27) with neurogenic communication, voice, dysphagia and fluency disorders. The review was further expanded to include a small number of studies (n=12) focusing on early intervention services. This review by Edwards et al. (2012) suggested that telehealth is an effective way to diagnose and treat both adults and children in the areas investigated, as services provided through telehealth or by conventional in-person means resulted in similar outcomes.
Although these previously conducted reviews included studies targeting the paediatric population, the number of studies investigated was minimal and the focus was primarily on the application of telehealth in SLP with the adult population. These reviews suggest positive results. However, service delivery models and intervention techniques used with children typically differ from those used with adults, as the focus with children tends to be on achieving developmental milestones, as opposed to a rehabilitative approach that is commonly used with adults (Edwards et al., 2012). It can therefore be difficult to apply previous findings that were obtained from primarily adultfocussed studies to the paediatric population.
The potentially detrimental effects of communication difficulties on a child's education and social participation increase the importance of alleviating these where possible, regardless of where the child resides. It is therefore important to focus on this specific population to determine whether telehealth service delivery may be a viable alternative to in-person intervention in locations where this service is not readily available. However, no review to date has focused specifically on evaluating the telehealth studies undertaken with children. Thus, this systematic review evaluated the present literature to determine if telehealthdelivered SLP interventions are as effective as traditional inperson delivery for primary school-age children with speech and/or language difficulties.

METHODS
To address this study's aim, a systematic review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (Liberati et al., 2009). The PRISMA flow chart is detailed in Figure 1 (adapted from Liberati et al. (2009)). The current systematic review was registered with the PROSPERO registry: CRD42016052187.

SEARCH STRATEGY
A systematic literature search was undertaken using the PubMed, CINAHL, Scopus, ERIC and SpeechBITE databases. Additional manual searches in two highly relevant journals, the International Journal of Speech-Language Pathology and the International Journal of Telerehabilitation, were also conducted, in order to locate more recent versions of journals that may not yet have been transferred into the databases. Systematic search strategies were adhered to using the following search string: (telehealth OR telepractice OR telerehabilitation OR teletherapy) AND (speech pathology OR speech-language pathology OR speech therapy) AND (child OR paediatric). In addition, citations and references within identified articles were searched for further studies relevant to the review. The authors corresponded with experts in the field to ensure all relevant studies were included within the review.

STUDY SELECTION
The studies identified through the systematic searches were included in the review if they reported studies of speech and language intervention delivered through telehealth to primary school-age children (4-12 years) across various settings (e.g., schools, private practice), provided treatment outcome data on intervention effectiveness and did not describe special client populations (e.g., autism spectrum disorder, childhood apraxia of speech). The year of publication was not restricted, ensuring all available evidence was identified, but the search was limited to articles written in English. Papers were included on speech intervention (speech sound production and intelligibility) and language intervention (receptive and expressive language). Articles describing voice, fluency, pragmatics, literacy or special client populations were excluded to focus on primary speech and language disorders.

DATA EXTRACTION
All articles identified from the initial searches were reviewed and duplicates were removed. The title and abstracts of the articles were screened for inclusion by all authors, with the remaining articles reviewed in full text and the exclusion criteria applied. In the case of disparities between the authors' judgments regarding suitability, they consulted to achieve agreement. Data from the included studies were extracted using a standard table developed specifically for this review (refer to Appendix A). The articles were summarised in terms of intervention type and participants, study aim and design, equipment, methods and main study results.

RESULTS
The initial database and reference list searches conducted during November and December 2016 yielded a total of 120 unique articles. During the initial screening, 68 articles were excluded on title and another 33 articles were eliminated on abstract. The remaining 19 articles were reviewed in full-text. The full-length review excluded a further 12 articles, because they: (1) did not describe speech and language intervention via telehealth with the majority of participants between 4 and 12 years of age, and/or (2) did not provide outcome data on intervention effectiveness. From this selection process, seven articles were retained for the final systematic review. The review process is detailed in the flow chart in Figure 1.

STUDY DESIGN
The seven included studies focused on telehealthdelivered speech and language intervention with primary school-age children. Two of the included studies were randomised controlled trials (level of evidence II) (Grogan- Johnson, Alvares, Rowan, & Creaghead, 2010;Grogan-Johnson et al., 2013) and another two studies were method comparison studies (level of evidence IIIa) (Gabel, Grogan-Johnson, Alvares, Bechstein, & Taylor, 2013;Grogan-Johnson et al., 2011) which were investigating the validity of telehealth-delivered intervention by comparing it with inperson results. A further three studies used a pre versus post study design (level of evidence IV) to determine if telehealth-delivered intervention facilitated improvement in the participants' communication skills, with no comparison group (Fairweather, Lincoln, & Ramsden, 2016;Isaki & Farrell, 2015;Jessiman, 2003). Four of these papers included a participant satisfaction survey (Fairweather et al., 2016;Grogan-Johnson et al., 2010;Isaki & Farrell, 2015;Jessiman, 2003). Only one of the seven studies was conducted in Australia (Fairweather et al., 2016). The intervention services provided within the included studies were undertaken within a structured school/university clinic (n=6) or community health clinic (n=1) environment. The studies varied according to the intervention focus and outcome measures used.

PARTICIPANTS
The majority of the studies (71%) focused only on primary school-age children between the ages of 4 and 12 years (Fairweather et al., 2016;Gabel et al., 2013;Grogan-Johnson et al., 2010;Grogan-Johnson et al., 2011;Grogan-Johnson et al., 2013;Isaki & Farrell, 2015;Jessiman, 2003). However, one study included a very small number of participants from 3 years of age (exact number of participants not specified) (Fairweather et al., 2016) and another included one participant aged 15 years (Gabel et al., 2013). Although these few participants were aged outside of the set criteria, the majority of the participants in the studies were aged between 4 and 12 years, allowing the results to be suitably applied to the primary school-age population. Other studies however were excluded during the initial study selection process due to the majority of participants being aged outside of the set criteria. Five of the seven studies (Fairweather et al., 2016;Grogan-Johnson et al., 2011;Grogan-Johnson et al., 2013;Isaki & Farrell, 2015;Jessiman, 2003)

TELEHEALTH EQUIPMENT
Three papers reported the use of commercial videoconferencing systems (Gabel et al., 2013;Grogan-Johnson et al., 2011;Grogan-Johnson et al., 2013) designed for use with low-speed connections (using a 128 kbit/s internet link). In contrast, three studies reported the use of web-based videoconferencing platforms (Fairweather et al., 2016;Grogan-Johnson et al., 2010;Isaki & Farrell, 2015) and the final study used a custom telehealth videoconferencing system (Jessiman, 2003). Two studies complemented their telehealth equipment with document cameras (Grogan- Johnson et al., 2010;Jessiman, 2003). The seven reviewed studies used real-time videoconferencing.

INTERVENTION TYPE, INTENSITY AND TARGETS
Five of the seven included studies investigated the application of both speech sound and language intervention through telehealth (Fairweather et al., 2016;Gabel et al., 2013;Grogan-Johnson et al., 2010;Isaki & Farrell, 2015;Jessiman, 2003). The remaining two studies focused primarily on the investigation of speech sound intervention (Grogan-Johnson et al., 2011;Grogan-Johnson et al., 2013). Notably, no studies included in the review solely examined the provision of language intervention through telehealth.

SPEECH SOUND INTERVENTION
Two studies conducted only traditional speech sound intervention (Van Riper approach to articulation intervention) through both telehealth and in-person delivery models. The participants in the study conducted by Grogan- Johnson et al. (2011) received 20 minutes of therapy each week between fall (baseline) and spring (post-intervention), whereas Grogan-  provided intervention for 30 minutes twice per week for a five week period. Both studies followed the same session format, however, only one study required the participants to reach a set number of productions prior to progressing through the intervention levels (Grogan- . The intervention targets in both studies were selected based on the participant's current Individualised Education Plan (IEP) goals, with the Grogan-  study also selecting additional targets based on the results of pretesting on the Goldman Fristoe Test of Articulationsecond edition (GFTA-2) (Goldman & Fristoe, 2002).

COMBINED SPEECH AND LANGUAGE INTERVENTION
An examination of a combination of speech sound and language interventions was conducted in five of the seven studies. The duration of intervention varied between studies. Participants in the study by Fairweather et al. (2016) received six 30 minute sessions on a fortnightly basis over a 12 week period, whilst Jessiman (2003) provided hourly treatment sessions twice a week for two months and Isaki and Farrell (2015) provided weekly therapy for two blocks of 15 weeks. Grogan- Johnson et al. (2010) provided one group of participants with telehealth treatment for four months followed by in-person intervention for another four months, while the second group received in-person intervention for four months and then subsequently telehealth-delivered intervention for four months. Further detail regarding the number and frequency of sessions in this study was not provided. Gabel et al. (2013) provided intervention to the telehealth group for 20 minutes per week for one academic year.
Further differences between the studies focussing on both speech and language intervention related to whether or not the treatment sessions were provided on an individual basis or in a group setting. An individual format was adopted in three of the studies (Fairweather et al., 2016;Isaki & Farrell, 2015;Jessiman, 2003), however, in the remaining two studies, the participants in the telehealth groups received mainly individual therapy sessions with some small group sessions also conducted (Gabel et al., 2013;Grogan-Johnson et al., 2010). The in-person participants in these two studies received primarily group sessions with 2-4 students, with some students alternatively receiving an individual pull-out model of intervention (Gabel et al., 2013;Grogan-Johnson et al., 2010).
The intervention provided varied depending on the selected targets. Two studies selected intervention targets based on the participant's IEP goals and objectives (Gabel et al., 2013;Grogan-Johnson et al., 2010), whereas another two studies established therapy goals based on recent assessment results (Isaki & Farrell, 2015;Jessiman, 2003). The fifth study developed goals in collaboration with adults familiar with each participant (Fairweather et al., 2016).

OUTCOME MEASURES
The included studies examined the efficacy of telehealth intervention using various outcome measures. Six different outcome measures were investigated: the Goldman Fristoe Test of Articulationsecond edition (GFTA-2); Functional Communication Measures (FCMs); goal achievement; informal probes; comparison of preintervention baselines with post-intervention production levels; and change reported on quarterly progress reports.

EFFICACY OF THERAPY GOLDMAN FRISTOE TEST OF ARTICULATION -SECOND EDITION (GFTA-2)
Three studies utilised pre-and post-intervention testing with the GFTA-2 to compare telehealth to in-person delivered intervention (Grogan- Johnson et al., 2010;Grogan-Johnson et al., 2011;Grogan-Johnson et al., 2013). Each of these studies revealed no significant difference between the two treatment modalities, with the first study reporting across three measurement points (pre-test p=0.16; post-first treatment period p=0.06; post-second treatment period p=0.21) and the second and third study reporting across two measurement points each (pre-test p=0.805; post-test p=0.805; and pre-test p=0.706; post-test p=0.644, respectively). Using a repeated measure ANOVA, Grogan-   found no significant difference between the two groups on post-intervention GFTA-2 testing (p=0.415); however, a statistically significant change in test scores was evident from pre-to post-intervention for both groups (p=0.020), indicating that both groups made significant and similar progress during intervention. Grogan- Johnson et al. (2011) identified a similar result with both groups making significant improvement in performance (p=0.014) but neither group was found to improve more than the other.

FUNCTIONAL COMMUNICATION MEASURES (FCMS)
Two studies measured outcomes through Functional Communication Measures (FCMs), which are used as a measure of progress in the ASHA K-12 Schools National Outcomes Measurement System (NOMS) database (American Speech-Language-Hearing Association, 2003; Gabel et al., 2013;Grogan-Johnson et al., 2010). This database reports descriptive information on students receiving in-person speech-language intervention in the school system. Gabel et al. (2013) compared their results for the telehealth condition with the subjects reported in the NOMS database (in-person participants). This study revealed similarities between the changes in FCM level for the telehealth group and also the in-person participants for disorders related to intelligibility (66.7% improved at least one level in telehealth and 62.3% in-person) and speech sound production (84.6% in telehealth and 78.4% in-person) (Gabel et al., 2013). For spoken language production, this study revealed a sizable difference between telehealth and in-person results, with 55.6% and 71.1% improving at least one level respectively. Gabel et al. (2013)'s results for spoken language comprehension were varied, with a higher percentage of telehealth participants improving by one level (47.1% vs. 38.2%) and a lower percentage improving by multiple levels in comparison to the in-person group (11.8% vs. 27.8%). The results reported by Gabel et al. (2013) were in contradiction to the results identified by Grogan-Johnson et al. (2010). As part of this research, the FCMs were used to compare progress between two groups of students, one group that received telehealth-delivered intervention and the other via in-person. This study found that a slightly lower percentage of participants in the telehealth group improved at least one level compared to the in-person group for disorders related to intelligibility (63% vs. 70%, respectively). This was similar for the speech sound production measure, with less participants in the telehealth group improving at least one level (71% telehealth vs. 79% in-person). However, for disorders related to spoken language production, a higher percentage of telehealth participants improved by a minimum of one level in comparison to the inperson group (72% vs. 62%, respectively).
The results of these two studies demonstrate conflicting findings; however, neither of the studies conducted statistical analyses of the results and thus the significance of the percentage differences between the two intervention conditions is unknown. The limitations evident in both studies could also likely have introduced confounding factors, which may have affected the results. For instance, one study had a considerable difference in the sample size for the two conditions and did not randomly allocate participants, but instead selected the telehealth participants from a pilot project already being conducted (Gabel et al., 2013). The selected participants were allocated to the telehealth condition and their results were compared with data already stored in the NOMS database, therefore introducing potential bias. Neither of the studies controlled for the type of service utilised (e.g., individual or group therapy) or the methods of treatment provided (Gabel et al., 2013;Grogan-Johnson et al., 2010).

GOAL ACHIEVEMENT
Two studies (Fairweather et al., 2016;Isaki & Farrell, 2015) used goal achievement to determine outcomes, with one study using Goal Attainment Scaling (GAS), a criterionreferenced measure of change rated on a five-point scale, to evaluate the telehealth program (Fairweather et al., 2016). This study revealed that 68.9% of the established goals were achieved at either an expected or greater than expected level. From the 19 participants, 15 (78.9%) achieved at least one goal at or above the expected level and eight participants (42.1%) achieved all their goals. The GAS scores were converted to t-scores to reflect performance above or below the expected level (e.g., achieving the set goal). This analysis revealed that 73.68% of the participants achieved or exceeded their set goal following six telehealth sessions.
The second study evaluated goal completion against a set criterion (Isaki & Farrell, 2015), with the results indicating that for the speech goals targeted, three of the five participants achieved 100%, one achieved 50% and the other achieved 33%. The three participants with language goals all achieved 100%. These results related only to telehealth and did not provide a comparison to in-person treatment.

INFORMAL PROBES
Examination of progress using informal probes was conducted in one study (Jessiman, 2003). The participants' goals or number of goals were not detailed, however, based on informal probes completed after therapy and by parent report it was suggested that the participants made progress in their speech and language goals across the 12 sessions. One participant was reported to have made "substantial" progress while the other participant's progress was "less substantial, but still appeared promising" (Jessiman, 2003, p.48-49). Jessiman (2003) quantified the participant's progress using these terms by determining the number of speech and language skills mastered or progressing within the treatment period.

COMPARISON OF PRE-INTERVENTION BASELINES WITH POST-INTERVENTION PRODUCTION LEVELS
Two studies used the comparison of pre-intervention baselines with post-intervention production levels as an outcome measure (Grogan- Johnson et al., 2011;Grogan-Johnson et al., 2013). The analysis of this outcome measure differed between the two studies, however both studies indicated that progress was achieved regardless of the treatment modality. Grogan- Johnson et al. (2011) measured the change in speech sound production from baseline to the completion of intervention, with the results suggesting that both the telehealth (n=55) and in-person (n=8) groups made similar amounts of progress. The results were comparable for the percentage of improved baselines, with 98% (n=54) in the telehealth and 95% (n=6) in the in-person group. However, the in-person group had a higher percentage of unchanged baselines (2% [n=1] for telehealth and 12.5% [n=1] for in-person) and decreased baselines (0% [n=0] for telehealth and 12.5% [n=1] in-person) (Grogan- Johnson et al., 2011). The varying number of baselines targeted in the intervention may explain the difference in the results for unchanged and decreased baselines between the telehealth and in-person methods in this study. There were 55 baselines targeted for participants in the telehealth group and only eight collected for the in-person participants. Both groups only had one unchanged baseline however, due to the high variance in total baselines targeted, a considerable difference in percentage was indicated.
Grogan-  used listener judgments to compare pre-and post-intervention productions and these results were examined using a repeated measures ANOVA. The listener judgments revealed a statistically significant difference across time for both groups (p=0.007), but no significant difference between the two groups in regard to the amount of change across time (p=0.434). Thus, both groups were deemed to receive benefit from the intervention regardless of the service delivery model.

QUARTERLY PROGRESS REPORTS
The results reported on participants' quarterly progress reports were used as outcome measures for two studies (Grogan- Johnson et al., 2010;Grogan-Johnson et al., 2011). In the study by Grogan-Johnson et al. (2010), quarterly student progress reports after the first treatment period identified that adequate progress or mastery was achieved for 75% (n=58 for telepractice and n=34 for in-person) of objectives in both conditions. A significant difference (p<0.05) between the two intervention conditions was indicated following the second treatment period, with mastery or adequate progress achieved for 88% (n=42) of objectives in the telehealth model and 84% (n=56) of objectives in the in-person model. A similar result was found in Grogan- Johnson et al. (2011)'s study, as more participants in the telehealth group (100%, n=25) mastered or made adequate progress on their IEP goals in comparison to the 87% (n=13) of participants in the inperson group.
The difference in the results for the number of IEP goals achieved between the intervention conditions in these studies can be explained by a disproportionate number of IEP objectives being targeted in the two intervention conditions and across the first (telehealth n=77, in-person n=45) and second treatment period (telehealth n=48, inperson n=67). A larger number of total IEP objectives were targeted in the telehealth group across the two treatment periods.

PARTICIPANT SATISFACTION
Four studies reported satisfaction data through the provision of surveys (Fairweather et al., 2016;Grogan-Johnson et al., 2010;Isaki & Farrell, 2015;Jessiman, 2003). High levels of satisfaction with telehealth-delivered intervention and the progress achieved were found in all studies. Two studies reported that concerns were identified regarding the child's reduced attention in telehealth sessions (Isaki & Farrell, 2015) as well as the need to improve internet connectivity, audio output and communication with stakeholders (Fairweather et al., 2016).

DISCUSSION
The present review investigated the efficacy of telehealth-delivered SLP services when compared to traditional in-person delivery for primary school-age children with speech and/or language difficulties. Evidence was collated through a systematic review of the available telehealth literature. Overall, the findings of the review showed that there is some evidence to support the use of telehealth when delivering SLP intervention services to school-age children. However, it also demonstrated that the amount of research into speech and language intervention for children via the telehealth service delivery model is limited and of variable quality, as the included studies span across the levels of evidence according to the National Health and Medical Research Council (NHMRC) (Australian Government: National Health and Medical Research Council, 2009). A total of six different types of outcome measures were used to investigate the efficacy of telehealth intervention, therefore creating difficulty in directly comparing the studies. The design of the three studies using the goal achievement and informal probe outcome measures did not allow direct comparison between the telehealth and in-person intervention conditions, as the study designs only evaluated the telehealth-delivered intervention, without comparing it to the traditional in-person model (Fairweather et al., 2016;Isaki & Farrell, 2015;Jessiman, 2003). However, these measures demonstrated considerable progress based on the targeted goals during the telehealth intervention.
The remaining reviewed studies directly compared the telehealth and in-person intervention conditions using four different outcome measures. There was convincing evidence in the literature suggesting that speech sound intervention delivered through telehealth to primary schoolage children was just as effective as in-person intervention when measured through the GFTA-2 (Grogan- Grogan-Johnson et al., 2011;. Participants in both intervention conditions made significant improvements in performance and equal gains were demonstrated on the post-intervention testing. Positive results were also identified in the studies that used the comparison of pre-intervention baselines and postintervention production levels to measure outcomes, with both studies indicating that progress was achieved regardless of the treatment modality (Grogan- Johnson et al., 2011;Grogan-Johnson et al., 2013).
Despite these positive results, the two studies that examined telehealth-delivered speech and language intervention and used the FCMs as their outcome measure identified contradictory results. For the intelligibility and speech sound production measures, Gabel et al. (2013) found that more participants improved in the telehealth condition whereas, in the study by Grogan-Johnson et al. (2010), a lower percentage of participants improved in the telehealth condition when compared to the in-person condition, using the same measures. For the spoken language production measure, Gabel et al. (2010) found that a much lower percentage of participants improved in the telehealth condition compared to the in-person condition; however, Grogan-Johnson et al. (2010) found that more participants improved in the telehealth condition, again using the same measure. The authors did not conduct statistical analyses of these results, making it difficult to determine if the difference between the results is significant.
Both speech sound and language interventions were implemented as part of the seven studies included in the review. However, of these reviewed studies, there appeared to be a stronger focus on speech sound intervention, with two studies primarily aiming to assess this range of practice area through telehealth. The remaining five studies investigated the application of both speech sound and language intervention through telehealth, however two focused more heavily on speech than on language, as a greater number of speech goals were targeted in one study (Isaki & Farrell, 2015) and more speech-based FCMs were used as an outcome measure in another study (Grogan-Johnson et al., 2010). Overall, whilst the studies revealed that intervention delivered through telehealth is as effective as in-person intervention, this result seemed to be found more consistently with the provision of speech sound intervention than with language intervention. Although this suggests that speech sound intervention may be more suited to a telehealth approach, this finding is likely to be skewed by the more predominant focus on this range of practice area in the reviewed studies. Another possible explanation for this result is the difficulty in identifying comprehensive measures of language to be used when conducting research relating to telehealth-delivered services, as language is such a broad and highly variable range of practice area.
The uptake of the use of telehealth by speech-language pathologists has been influenced by the need to address the inequity of access to services experienced by Australia's rural population (SPA, 2014). Telehealth allows services to be delivered to clients, including children, within their home and with the assistance of parents/carers, regardless of their location. The majority of the studies included in this review were however undertaken within a structured school or clinic environment, with little or no parent involvement. This results in difficulty drawing conclusions about the effectiveness of telehealth when implemented in the home setting, where the environment is likely to be less structured and full parent involvement is required.
Interestingly, all of the reviewed studies utilised realtime videoconferencing facilities, allowing the clinician and client to visualise each other. This finding is consistent with results from previous reviews (Mashima & Doarn, 2008;Reynolds et al., 2009;Taylor, Armfield, Dodrill, & Smith, 2014), indicating that real-time interactions support the delivery of services and strongly influence the clinical outcomes achieved through telehealth. Delivering speech and language intervention services through real-time videoconferencing facilities is an effective method of service delivery as this medium most closely resembles in-person interactions through the transmission of auditory and visual signals at a distance (Mashima & Doarn, 2008;Reynolds et al., 2009). SLP practice primarily consists of auditory, verbal and visual interactions, therefore allowing services to be easily translated into technology-based environments (Theodoros, 2012). This level of connection enhances the sense of clinician presence and facilitates the development of rapport between clinicians, clients and their families, provided that the necessary bandwidth is available to support the process (Mashima & Doarn, 2008;Reynolds et al., 2009).
Reports that videoconferencing facilities do effectively support real-time interactions between clinicians, clients and families, are consistent with parent, student, and staff satisfaction data that was collected as part of four of the seven studies included in this review (Fairweather et al., 2016;Grogan-Johnson et al., 2010;Isaki & Farrell, 2015;Jessiman, 2003). Satisfaction ratings were high across all stakeholders surveyed despite the concerns raised regarding slightly reduced attention by children in telehealth sessions and difficulties with technology. Similar findings with stakeholder satisfaction have been reported in various studies and reviews (Constantinescu et al., 2014;Crutchley & Campbell, 2010;Lincoln, Hines, Fairweather, Ramsden, & Martinovich, 2015;Mashima & Doarn, 2008;Sicotte, Lehoux, Fortier-Blanc, & Leblanc, 2003), indicating that relevant stakeholders deem telehealth as an effective method of delivering speech and language intervention to children. It is however, important to note that satisfaction ratings related to telehealth are likely to be particularly high in rural areas where in-person SLP services are not typically available.
Overall, the findings from the seven reviewed studies revealed that telehealth is a promising method for treating children with speech and/or language difficulties. However, in spite of this interesting finding, a number of methodological issues limit the quality of the results. The conclusions found in the literature on the effectiveness of telehealth-delivered intervention are dependent on the selected outcome measure. Outcomes for telehealth were more consistently positive when standardised assessments, such as the GFTA-2, were used for the pre-and postintervention testing. The literature also revealed considerable variation in the intensity of therapy, with some studies claiming significant improvement after only a small number of sessions (6) were delivered fortnightly (Fairweather et al., 2016), whereas others reported on a larger number of sessions (10-12) that were delivered twice weekly (Grogan- Jessiman, 2003), making the intervention format more intense. Additionally, the majority of the studies reviewed were based on a small and unequal sample size, resulting in difficulty generalising the results. Furthermore, of the four studies comparing the service delivery models, two studies did not randomly allocate participants to the intervention conditions, therefore introducing potential intervention condition bias. These differences in the studies made direct comparison difficult and therefore, may limit the weight of the findings. Thus, to provide further evidence regarding the effectiveness of telehealth-delivered intervention, studies that use more rigorous methods, such as randomisation of participants and power calculations, need to be performed to ensure that potential key findings can be accurately identified.
The current systematic review also has some limitations that require consideration. Firstly, although two studies included a very small number of participants outside the set age criteria, the primary school-age population was the focus of the review. Therefore, studies which included a large number of children outside this age range were excluded due to the differences in attention span and behaviour between age groups (Owens, 2012). Whilst this allows the results of the review to be appropriately applied to the primary school-age population, without the data being skewed from a mix of different populations, further research in this area is required to confirm if telehealth is as effective as in-person intervention when delivered in an early intervention format or to adolescents. Furthermore, the majority of the studies included in the review were undertaken in the USA, thus generalisability of the results to rural and remote communities within Australia is limited, due to factors such as the frequent lack of adequate and reliable internet connectivity (Australian Bureau of Statistics, 2016;Erdiaw-Kwasie & Alam, 2016;Park et al., 2015). Therefore, rural Australian communities may have difficulty supporting the telehealth service delivery model, an issue that may not have been adequately captured in this review.

CONCLUSIONS
The current review aimed to determine if telehealthdelivered SLP interventions are as effective as traditional inperson delivery for primary school-age children with speech and language difficulties. The reviewed research was limited and of variable quality, however, the evidence presented showed that telehealth is a promising service delivery method for delivering speech and language intervention services to this population. This alternative service delivery model has the potential to improve access to SLP services for children living in geographically remote areas, reducing travel time and alleviating the detrimental effects of communication difficulties on education, social participation and employment. Although some initial positive findings have been published, there is a need for further research using more rigorous study designs to further investigate the efficacy of telehealth-delivered speech and language intervention. Procedure: Both groups received traditional speech sound intervention for 20 minutes weekly. Multiple measures of progress assessed: 1) Pre-and posttesting using GFTA-2; 2) comparison of preintervention baselines with  No significant difference between the TH and inperson groups on the pre-(p = 0.805) and post-tests (p = 0.805).  Both groups had a significant improvement in performance (p = 0.14).  Children in both SDMs improved significantly in their speech production with the telehealth students demonstrating greater IEP goal mastery.

ACKNOWLEDGMENTS
production levels postintervention; and, 3) comparison of quarterly progress reports.
Grogan-Johnson et al.
Telepractice group: N=7, Avg age=8.4yrs, range= 6.4-9.9yrs Side-by-side group: N=7, avg age=9yrs, range= 7.9-10yrs Two SLPs To investigate telehealthdelivered intervention services by comparing speech sound intervention delivered to children in either a telepractice or inperson delivery model in an intervention program.
Study Design: Method comparison study, reporting statistical difference between TH and in-person conditions. Equipment: Laptop, webcamera with microphone and headset. Polycom VC system with 128kbit/s internet link.
Procedure: Both groups received traditional speech sound intervention for 30 minutes twice per week for a 5-week period. Participants were randomly assigned to either the in-person or TH condition. Multiple measures of progress assessed: 1) preand post-intervention testing conducted using subtests of GFTA-2; and 2) pre-and postrecording of single word identification task.
 No significant difference found between two groups on postintervention GFTA-2 through repeated measures ANOVA (p=0.415).  No statistically significant difference between the mean listener judgements for the two groups on the pre-test (p=0.160) but a statistically significant difference in mean listener judgements across time for both groups (p=0.007). Thus, both groups benefitted from intervention and that benefit was the same regardless of intervention condition.

Speech and/or language intervention
Child participants -N=5. Mean age 7.1yrs (range 4.5-9.8 yrs) Adult participants -(not reported in review) To evaluate the effectiveness of Apple iPads to deliver telepractice speech and/or language services.