Effectiveness of school-based physiotherapy intervention for children

Abstract

Purpose

To evaluate the effectiveness of school-based physiotherapy interventions for improving students’ participation in school settings.

Method

A systematic review was reported using PRISMA guidelines. Four databases were searched for studies investigating physiotherapy outcomes of school-based physiotherapy interventions in children. Studies were categorised by intervention type and evaluated based on evidence level and conduct.

Results

Fifteen intervention types (23 studies) met criteria. Strong positive evidence supported treadmill training without bodyweight support (n = 1), and upper limb interventions (n = 2). Moderate positive evidence supported robotic-assisted gait training (n = 1), Gross Motor Activity Training with Multimodal Education-Based Therapy (GMAT + MET) (n = 2), neurodevelopmental treatment (n = 2), and rock climbing (n = 1). Weak positive evidence supported environmental modifications (n = 1), Ergonomic Health Literacy (n = 3), GMAT (n = 1), GMAT with progressive resistance exercise (GMAT-PRE) (n = 1), hippotherapy (n = 1), MET alone (n = 7), overground gait training (n = 2), treadmill training with partial body-weight support (n = 1), and non-immersive virtual reality (n = 3).

Conclusion

There is preliminary supporting evidence for a variety of school-based physiotherapy interventions which is consistent with evidence for interventions with established efficacy in other contexts. The evidence for interventions in school contexts alone is insufficient to guide current practice. Future research should specifically evaluate the effectiveness of physiotherapy approaches in school settings.

IMPLICATIONS FOR REHABILITATION

• Preliminary supporting evidence exists for a variety of school-based physiotherapy interventions, primarily those with established efficacy in other contexts.

• Successful interventions were designed with a direct focus on assisting participants to improve their ability to engage in school activities.

• Relevant participation-focussed outcome measures should be used to evaluate the effectiveness of intervention provided in a school context.

• Interventions with proven effectiveness for specific population groups in other contexts are likely to be effective in schools, however the impact on participation at school is yet to be determined.

Keywords:

• Physiotherapy
• school
• children
• intervention
• review

Background

School-based physiotherapy interventions are delivered in school contexts with the aim of improving students’ access to learning environments, and their involvement in classroom, physical education or social or recreational activities at school [1–3]. Interventions are provided in educational settings such as mainstream schools, special schools, and kindergartens, by physiotherapists who are employed by individual schools, school districts, or external providers [3,4]. Physiotherapists work collaboratively with students, their families, and teachers, to empower students to increase their independence while accessing and engaging in school activities by delivering interventions that identify and minimise the effects of impairments. Interventions may be delivered directly, whereby the student receives therapy separately from classroom activities or integrated within the classroom, or in a consultative manner on the student’s behalf, where a physiotherapist would provide support to the teacher or other staff [4].

Internationally, it is accepted that school-based physiotherapy must be educationally focused and relevant to students’ learning activities, rather than targeting general performance. This is facilitated differently depending on countries’ education and health systems. For example, in the United States, the Individuals with Disabilities Education Act specifies that physiotherapy is delivered as a “related service” within the school system [5]. In the United Kingdom, physiotherapy is a special education provision, delivered in collaboration with a Special Educational Needs Coordinator under an Education, Health, and Care plan [6]. In Australia, school-based physiotherapy is governed by state education departments, but commonality exists between states in that the purpose of interventions must be to assist students to participate in school activities, thereby enhancing learning outcomes [2,7]. Participation is defined as attendance and involvement in a life situation (schooling) in the context of fulfilling a social role (being a student) and should be measured using instruments that consider the person’s preferences in types and levels of participation [8,9].

Several student populations receive and may benefit from school-based physiotherapy [4]. Students with disabilities may receive school-based physiotherapy in educational settings such as mainstream schools, special education programs in mainstream schools, or special schools. Seven million American students with disabilities are enrolled in public schools, and 89% of Australian students with disabilities attend mainstream school in regular (71%) or special (18%) classes [10,11]. Schools must provide the necessary supports to accommodate students’ needs, minimising the effect of functional impairments or activity limitations on educational participation. A shift towards increased social and educational inclusion in recent years has facilitated the involvement of physiotherapists in schools to support children to improve their participation in learning and recreational activities, and therefore educational outcomes [12,13]. Physiotherapists may also assist typically developing students whose teachers have identified difficulties with school-specific activities which may benefit from therapeutic input. In this case, interventions typically address difficulties with school-specific activities, such as handwriting or the gross motor skills required in physical education classes [2]. Physiotherapists may also deliver interventions to large groups of students, targeting issues that have been shown to impact the broader student population, such as childhood obesity or back pain [14,15].

Recent studies examining school-based physiotherapy in the United States have established an understanding of common interventions [16] and the influence of participation-based goals on service delivery and outcomes [17]. It was reported that neuromuscular, mobility, and musculoskeletal interventions were delivered most frequently, with most time spent on mobility, transitions, and physical education or recreation activities [16]. However, specific intervention efficacy was not reported, nor was the distribution of intervention types for children with varying conditions [16]. When considering goal setting, it was found that students with participation-based goals received a greater proportion of services than students with non-participation-based goals [17]. In addition, correlation was identified between the focus of goals and students’ level of gross motor function. Students with lower gross motor function tended to have participation-based goals and received interventions in the classroom, whereas students with higher level gross motor skills received interventions with a focus on dynamic gross motor activities or quality of skills performance, separate from the classroom [17].

Although recent literature has identified common interventions, several aspects of school-based physiotherapy are currently not well reported. Many studies report the effectiveness of interventions delivered to school-aged children in community settings, but it is uncertain whether this evidence is transferable to school contexts. Lack of consolidated evidence specific to a school setting limits therapists’ ability to optimise treatment planning. In addition, the aims of school-based physiotherapy are specific to the school context, and therefore differ from interventions delivered in the community, resulting in a lack of clarity as to if interventions are effective for addressing educationally focused goals.

This review aims to identify the interventions delivered by physiotherapists in school settings and to evaluate the effectiveness of these interventions. Identification and evaluation of the type, dose, and format of these interventions will facilitate recommendation of appropriate physiotherapy management for children in schools.

Method

This systematic review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [18]. The review was registered prospectively with the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42022318751).

Search strategy

A systematic search of the literature was conducted in March 2022, with a follow up search conducted in June 2023, using four electronic databases: PubMed, EMBASE, CINAHL (EBSCOhost), and ERIC. Search terms were selected to identify articles pertaining to the population of interest (“child” OR “children”), location of intervention (“school”), and the intervention (“physiotherapy” OR “physical therapy”) AND (“intervention” OR “treatment”). Secondary search strategies included citation tracking and reference list checking of all included articles. Titles and abstracts of all identified articles were independently examined by two authors (KA and LJ or GC). Full-text articles that satisfied inclusion criteria were then retrieved, in addition to articles requiring further clarification as to fulfilment of inclusion criteria. The third author (LJ or GC) was consulted if agreement on inclusion was not reached following review by two of the authors. Full text articles were examined if they met inclusion criteria or required further clarification as to relevance for inclusion.

Inclusion and exclusion criteria

Articles were included if (1) they were full-text articles in peer-reviewed journals, published in English, after 2010; (2) study participants were children attending a formal education setting; (3) physiotherapy intervention was prescribed and provided by or under the direction of a physiotherapist; (4) intervention was relevant to the child’s education; and (5) at least one outcome measure reported was performed by or under the supervision of a physiotherapist. Articles were excluded if they did not report original data (e.g., systematic reviews, protocols).

Data extraction and quality appraisal

Data from each article was extracted by two authors, and included details of participants, methods, interventions, outcomes, and results. All missing data was specified in the results table. Quality of each article was examined by two authors, with level of evidence and conduct evaluated. Levels of evidence were assigned to group research designs using the classification proposed by Sackett [19]. Methodological quality of studies classified as Level I-III evidence was evaluated by the Cochrane Risk of Bias tool (RoB-2) [20]. The Risk Of Bias In Non-randomised Studies—of Interventions (ROBINS-I) [21] was used to evaluate quality of non-randomised studies that compared the effects of two or more interventions, and the National Heart, Lung, and Blood Institute (NHLBI) quality assessments were used to evaluate quality of case series and pre-post study designs [22]. The RoB-2 uses a five-domain questionnaire to assess the presence of bias. Studies are rated as having “low risk of bias” if low risk across all domains, “some concerns” if concerns are raised in at least one domain but in the absence of high-risk judgements, or “high risk of bias” if there is high risk in at least one domain or some concerns across multiple domains [20]. The ROBINS-I uses a seven-domain questionnaire to assess the presence of bias, with a similar rating scale to the RoB-2 [21]. The NHLBI quality assessments are specific to certain study designs, and are designed to identify potential flaws in study methods or implementation, and classify studies as “good,” “fair,” or “poor” [22]. Analysis of study conduct for studies of Level I-III Evidence was conducted per the American Academy for Cerebral Palsy and Developmental Medicine (AACPDM) guidelines [23]. The quality of evidence was summarised for each intervention type using the Grading of Recommendation, Assessment, Development and Evaluation (GRADE) framework, which applies ratings of high, moderate, low, or very low, based on the level of evidence, and additional factors which may increase or decrease the certainty of the evidence [24]. The International Classification of Functioning, Disability and Health (ICF) framework and its coding system [25] were used to classify each outcome measure, to determine whether the intervention demonstrated improvements in body functions and structures, activity limitations, participation restrictions, or environmental factors. Studies were categorised by the content of interventions using categories established in previous literature, such as Gross Motor Activity Training (GMAT) [26] or in the case of interventions that have not previously been defined, were given a relevant categorisation, for example Multimodal Education-based Therapy (MET).

Results

Initial searches identified 1205 unique articles. Titles and abstracts of these articles were then screened, leaving 92 articles for full-text review. Of these, 23 studies (27 articles) met the final inclusion criteria. The flow of studies and reasons for exclusion are summarised in a PRISMA diagram (Figure 1). Details of all included studies are presented in Table 1. The 23 included studies, ranging from Evidence Levels II-V, included 15 types of school-based physiotherapy: environmental modifications (n = 1, Evidence Level IV); ergonomic health literacy (n = 3, Evidence Level II-IV); gait training—overground (n = 2, Evidence Level IV); gait training—treadmill (no body-weight support) (n = 1, Evidence Level II); gait training – treadmill (partial body-weight support) (n = 1, Evidence Level II); gait training – robotic-assisted (n = 1, Evidence Level IV); Gross Motor Activity Training (GMAT) (n = 1, Evidence Level IV); GMAT with Multimodal Education-based Therapy (GMAT-MET) (n = 2, Evidence Level II); GMAT with Progressive Resistance Exercise (GMAT-PRE) (n = 1, Evidence Level V); hippotherapy (n = 1, Evidence Level V); MET (n = 7, Evidence Level IV); Neurodevelopmental Treatment (NDT) (n = 2, Evidence Level II-V); non-immersive virtual reality (n = 2, Evidence Level II-IV); rock climbing (n = 1, Evidence Level II); and upper limb intervention (n = 2, Evidence Level II).

Figure 1. PRISMA flow diagram [19].

Table 1. Summary of school-based physiotherapy interventions in children – interventions and results.

Reference	Design, evidence level, conduct rating (level I-III) and GRADE rating	Participants	Setting	Intervention and dosage	Outcomes measured and [ICF coding]
Environmental modifications
[27]	Pre-post design Level IV Low quality/weak recommendation	Diagnosis: Developmental Coordination Disorder; typically-developing children n = 41 Age: mean age 7.85 years 23 F, 18 M	Country: South Africa School type: Mainstream primary school	Content: Health Promotion Program with aims of encouraging physical activity, providing a safe and healthy environment with appropriate equipment, collaborating with teachers to provide skills-based health education, and communicating with families to provide advice for accessing further support. Dose: nine week “whole school” initiative; physiotherapy students present at the school for 26 h per week. Delivered by: 4 physiotherapy students and their clinical educator. Target: Students, teachers, families Group 1: Participants with DCD (n = 22) Group 2: Typically-developing participants (n = 19) Both groups received the same intervention.	Within group analyses Group 1 (pre-post) Improved gross motor motor skills [d4] on Total Standard Score of MABC-2 (p < 0.0001). Improved balance [b755, b760] on balance component of MABC-2 (p < 0.0001). Improved anaerobic power [b730] on Muscle Power Sprint Test (p not reported). Group 2 (pre-post) No change in gross motor skills [d4] on Total Standard Score of MABC-2 (p = 0.69) No change in balance [b755, b760] on balance component of MABC-2 (p = 0.49). Improved anaerobic power [b730] on Muscle Power Sprint Test (p not reported). Between group analyses Greater improvement in balance (balance subscale of the MABC-2) in group 1 (p < 0.0001) compared to group 2 (p = 0.49). Greater improvement in anaerobic fitness in group 2 (p = 0.008).
Ergonomic health literacy
[28]	Individually randomised parallel-group trial Level II Moderate quality/moderate recommendation	Diagnosis: none; school children n = 194 Age: 9–11 years 110 F, 84 M	Country: Belgium School type: Mainstream primary schools	Group 1: Back education program (n = 96) Content: Education based on ten guidelines for correct posture and spinal health, delivered to one class group at a time. Dose: 6 h (6 weeks × 1 session per week × 1 h) Delivered by: Physiotherapist Target: Students Group 2: No intervention (n = 98)	Within group analyses Group 1 (pre-post) Improved back-care knowledge (p ≤ 0.011) (d = 0.84) except specific back care knowledge at the 8-year follow-up (p = 0.112) (d = 0.28). Higher pain prevalence [b280] at all test moments (p < 0.001). Higher fear avoidance beliefs for the item ‘lifting a heavy object’ (p = 0.025). Group 2 (pre-post) Improved back-care knowledge over time (p not reported). Improved spinal care behaviour [d4105, d430], most evident at 1-year follow-up (p not reported). Between group analyses Lower scores for spinal care behaviour (p ≤ 0.014) and self-efficacy (p ≤ 0.003) at the 8-year follow-up compared to previous test moments in both groups.
[29]	Case series Level IV Low quality/weak recommendation	Diagnosis: none; school children n = 206 Age 12–19 years 97 F, 109 M	Country: Portugal School type: Mainstream primary and high schools	Content: Theoretical-practical session targeting body mechanics and posture (static and dynamic) during daily activities. Dose: Single occasion, 1.5 h session Delivered by: Physiotherapist Target: Students	Improved performance of practical tasks [d4105, d430] – increase in 53.75% of students able to perform all tasks correctly (p not reported). Improved ergonomic knowledge 1 month after the intervention (p not reported). Decreased percentage of students in 8th grade with an overweight backpack [d430] (p ≤ 0.001). Increased percentage of students in 10th grade with an overweight backpack [d430] (p ≤ 0.001).
[30]	Case series Level IV Low quality/weak recommendation	Diagnosis: none; school children n = 84 Age: 10–13 years 42 F, 42 M	Country: Portugal School type: Mainstream primary school	Content: Theoretical-practical session targeting body mechanics and posture (static and dynamic) during daily activities. Dose: 1.5 h (2 weeks × 1 session per week × 45 min) Delivered by: Physiotherapist Target: Students	Improved ergonomic knowledge – improvement in theoretical test scores (mean scores improved from 7.7 to 10.83) (p ≤ 0.001). Improved performance of practical tasks [d4105, d430] – practical test scores (mean scores improved from 4.14 to 11.8) (p ≤ 0.001). No change in backpack weight [d430] following intervention (p = 0.285).
Gait training – overground
[31]	Case report Level IV Low quality/weak recommendation	Diagnosis: Cerebral palsy (diplegic) n = 1 Age: 15 years 1 F	Country: United States School type: Mainstream high school	Content: Intervention focused on progressive gait training, per the student’s IEP goals. Dose: 60 min/week of direct PT, and 30 min/month of consultative services over 3.5 years from 9th to 12th grade. Delivered by: Physiotherapist Target: Student	Improved gross motor function [d4] − 12.4% increase in GMFM score (p not reported). Improved walking tolerance and functional capacity [d450-d469] – final 6MWT distance of 152.4 m with KAFOs and 2 axillary crutches (initially able to mobilise 45.7 m with crutches; time not specified) (p not reported). Improved mobility with crutches and light assistance [d450-d469] – able to ascend and descend 26 steps with a landing; progressed to stand-by assistance (p not reported). Improved independence with personal cares and dressing [d540; d5] – including donning, doffing, and locking KAFOs. No rating scale reported (p not reported).
[32]	Case report Level IV Low quality/weak recommendation	Diagnosis: Developmental disability n = 1 Age: 8 years 1 M	Country: United States School type: Mainstream primary school with special education class	Content: Use of forward chaining procedure to teach the student to use his walker when verbally prompted. Dose: 100 min (5 days × 20 min per day); one additional day to assess maintenance. Delivered by: Physiotherapist Target: Student	Improved mobility using walker [d450–469] – able to perform sit-to-stand and walk 30 feet with only verbal or visual cues by conclusion of intervention (previously unable to stand or walk short distances without assistance from teachers).
Gait training – treadmill (no body-weight support)
[33]	RCT Level II Strong High quality/strong recommendation	Diagnosis: Cerebral palsy (spastic tetraplegic or diplegia), GMFCS I – III n = 22 Age: 13–19 years 9 F, 13 M	Country: Greece School type: Special high school	Group 1: Treadmill training (no body weight support) (n = 11) Content: Static stretching warm-up, 10 min; individualised treadmill training without body weight support, at a comfortable speed increased gradually, for a maximum of 30 min; cool-down stretches, 5 min. Dose: 27 h (12 weeks × 3 sessions per week × 45 min) Delivered by: Physiotherapists Target: Students Group 2: Usual PT (n = 11) Content: Mat activities, balance, gait training, functional gross motor tasks. Dose: 27 h (12 weeks × 3 sessions per week × 45 min) Delivered by: Physiotherapists Target: Students	Within group analyses Group 1 (pre-post) Improved gross motor function [d4] − 3.87% change in mean GMFM score (p not reported). Improved self-selected walking speed [d450] − 8.06 m/min change in mean self-selected walking speed (p not reported). Increase in mean treadmill speed [d450] from 1.76 km/hr to 3.00 km/hr (p = 0.00). Increase in mean time spent training from 12–20 mins to 23.9 − 29.8 mins (p not reported). Group 2 (pre-post) Improved gross motor function [d4] − 0.69% change in mean GMFM score (p not reported). Improved self-selected walking speed [d450] − 1.31 m/min change in mean self-selected walking speed (p not reported). Between group analyses Greater improvements in gross motor function [d4] on GMFM (p = 0.007) (d = 0.38) in group 1 [d450 – d469]. Greater improvements in self-selected walking speed [d450] (p = 0.000) (d = 1.13) in group 1 [d450 – d469]. No significant differences in spasticity between groups (p > 0.05) [b735].
Gait training – treadmill (partial body-weight support)
[34]	RCT Level II Strong Moderate quality/moderate recommendation	Diagnosis: Cerebral palsy, GMFCS II-III n = 30 Age: 6–18 years 10 F, 20 M	Country: Australia School type: Special school (primary and high school)	Group 1: Treadmill training (partial body weight support) (n = 15) Content: Graded progression with reduced body weight support, increased speed, and increased duration as appropriate. Walking harness attached to ceiling hoist. Dose: 8 h (8 weeks × 2 sessions per week × 30 min) Delivered by: Physiotherapists Target: Students Group 2: Over ground walking training (n = 15) Content: Participants used usual assistive devices and practiced walking in school corridors. Dose: 8 h (8 weeks × 2 sessions per week × 30 min) Delivered by: Physiotherapists Target: Students	Within group analyses Group 1 (pre-post) Improved walking speed [d450] − 0.051 m/s improvement in 10MWT between weeks 4 and 8 (p not reported). Improved functional capacity and walking tolerance [d450-d469] − 13.67 m improvement in 6MWT between weeks 4 and 8 (p not reported). Improved gross motor function [d4] − 6.26% improvement in GMFM (D) score and 7.4% improvement in GMFM (E) score between weeks 4 and 8 (p not reported). Group 2 (pre-post) Improved walking speed [d450] − 0.125 m/s improvement in 10MWT between weeks 4 and 8 (p not reported). Improved functional capacity and walking tolerance [d450-d469] − 15.54 m improvement in 6MWT between weeks 4 and 8 (p not reported). Improved gross motor function [d4] − 7.86% improvement in GMFM (D) score and 8.93% improvement in GMFM (E) score between weeks 4 and 8 (p not reported). Between group analyses Greater minimum distance covered in group 2 (p < 0.001) (d = 1.71). Greater average (p = 0.017) (d = 0.46) and maximum (p < 0.001) (d = 2.65) time spent walking in group 1. Greater minimum time spent walking in group 2 (p < 0.001) (d = 2.71). No significant difference between groups for change in scores in any outcome measure (GMFM, 10MWT, 6MWT) (p = 0.244–0.967).
Gait training – robotic-assisted
[35]	Pre-post design Level IV Low quality/weak recommendation	Diagnosis: Cerebral palsy (GMFCS IV/V) n = 27 Age: 5–8 years 10 F, 17 M	Country: United Kingdom School type: Special school (primary and high school)	Content: Robotic gait retraining using the Innowalk Pro – a robotic rehabilitation trainer that allows children with significant physical disabilities to mobilise in a weight-bearing position by providing assisted, guided and repetitive gait-like movements. Delivered in combination with usual physiotherapy care. Dose: 12 h (6 weeks × 4 sessions per week × 30 min) Delivered by: Physiotherapists, physiotherapy assistants, teachers and teaching assistants. Target: Students	Improved health-related quality of life [b, d, e, s] – on CPChild (p = 0.004) Improved individual goal attainment [d4] – on Goal Attainment Scale goals (p < 0.001) Improved lower limb range of motion [b710, b735] – on goniometric assessment (p = 0.004 − 0.17) Reduced lower limb spasticity [b735] – on the Modified Tardieu Scale (p = 0.02 − 0.27)
Gross motor activity training (GMAT)
[36]	Pre-post design Level IV Low quality/weak recommendation	Diagnosis: Developmental coordination disorder n = 46 Age: 6–10 years 22 F, 24 M	Country: South Africa School type: Mainstream primary schools	Group 1: Neuromotor task training (n = 27) Content: Part-task practice of components of soccer, netball, variations of tagging games, and other playground games. Dose: 13.5–18 h (9 weeks × 2 sessions per week × 45–60 min) Delivered by: Physiotherapists. Target: Students Group 2: Wii Fit (n = 19) Content: Games designed to mimic cycling, soccer, skateboarding or skiing, by encouraging weight shift in lateral and anterior directions, and stepping or humping over virtual obstacles. Games incorporating arm movements using the hand controller were also used. Dose: 9 h (6 weeks × 3 sessions per week × 30 min) Delivered by: Physiotherapists. Target: Students	Within group analyses Group 1 (pre-post) Improved gross motor skills [d4] on Total Standard Score of MABC-2 (p < 0.0001). Improved functional strength [b730] on Functional Strength Measure (p < 0.01). Improved anaerobic power [b730] on Muscle Power Sprint Test (p < 0.01). Group 2 (pre-post) No change in gross motor skills [d4] on Total Standard Score of MABC-2 (p = 0.26). No change in functional strength [b730] on Functional Strength Measure (p = 0.89). Improved anaerobic power [b730] on Muscle Power Sprint Test (p = 0.01). Between group analyses No significant difference in motor performance between groups (TSS of MABC-2) (p = 0.12). Greater improvement in functional strength (Functional Strength Measure) (p < 0.01) in group 1. No significant difference in upper limb muscle strength between groups (isometric contraction measured with dynamometer) (p not reported). No significant difference anaerobic power between groups (Muscle Power Sprint Test) (p not reported).
Gross motor activity training + multimodal education-based therapy (GMAT-MET)
[37]	RCT Level II Moderate Moderate quality/moderate recommendation	Diagnosis: Developmental Coordination Disorder n = 93 Age: 5–9 years 27 F, 66 M	Country: Australia School type: Mainstream primary schools	Group 1: Participant’s school with a school assistant (n = 37) Content: Motor retraining program, comprising fine motor (e.g., cutting and pasting), body awareness, gross motor circuit or skills practice (task-specific training of fundamental skills). Intervention delivered in small groups of 4–6 children, staff:children 1:3. Dose: 13 h (13 weeks × 1 session per week × 1 h) Delivered by: School assistant Target: Students Group 2: Participant’s school with a PT (n = 30) Content: As above Dose: 13 h (13 weeks × 1 session per week × 1 h) Delivered by: Physiotherapist Target: Students Group 3: Community clinic with a PT (n = 26) Content: As above Dose: 13 h (13 weeks × 1 session per week × 1 h) Delivered by: Physiotherapist Target: Students	Within group analyses Group 1 (pre-post) Improved motor skills [d410-d429; d450-d469; b755, b760] – on MABC and TGMD-2 (p < 0.05). Improved perceived physical competence [d4] – on PSPCSA from pre-intervention to 6 months post (p < 0.01). Group 2 (pre-post) Improved motor skills [d410-d429; d450-d469; b755, b760] – on MABC and TGMD-2 (2 (p < 0.05). Improved perceived physical competence [d4] – on PSPCSA from post-intervention to 6 months post (p = 0.01) and pre-intervention to 6 months post (p < 0.01). Group 3 (pre-post) Improved motor skills [d410-d429; d450-d469; b755, b760] – on MABC and TGMD-2 (2 (p < 0.05). Improved perceived physical competence [d4] – on PSPCSA immediately post-intervention (p = 0.01) and pre-intervention to 6 months post (p < 0.01). Between group analyses Significant improvement in motor skills in all groups on MABC (p = 0.00) (d = 0.98) & TGMD-2 (locomotor (p = 0.00) and object control (p = 0.00)) (d = 1.37) [d410-d429; d450-d469; b755, b760]. No group effect between modes of intervention (p = 0.09) for improvement in motor skills on MABC and TGMD-2. No group effect between modes of intervention (p = 0.62) for improved in perceived physical competence over time (p = 0.01) (d = 0.73) [d4].
[38]	RCT Level II Moderate Moderate quality/moderate recommendation	Diagnosis: Profound hearing impairment n = 21 Age: 6–11 years 7 F, 14 M	Country: India School type: School for the deaf	Group 1: Vestibular-specific neuromuscular training (n = 11) Content: Eye movement exercises, balance training (static stance in varied positions & walking with narrow base of support), and fundamental motor skills practice. Dose: 13.5 h (6 weeks × 3 sessions per week × 45 min) Delivered by: Physiotherapists Target: Students Group 2: Regular school activities (n = 12) No details of content or dose reported.	Within group analyses Group 1 (pre-post) Improved postural control [b755, b760] – Paediatric Reach Test (p = 0.001), Single Leg Stance test (p = 0.03), anteroposterior sway (eyes open p = 0.007, eyes closed p = 0.03), mediolateral sway eyes open p = 0.014, eyes closed p = 0.017). Improved gross motor skills [d4] – TGMD (p = 0.02), throw for distance (p = 0.042), jump for distance (p = 0.001), 15-yard dash (p = 0.001). Improved health-related quality of life on PedsQL (p = 0.01). Group 2 (pre-post) No changes in postural control [b755, b760] (p values not reported). No changes in gross motor skills [d4] (p values not reported). No changes in health-related quality of life (p values not reported). Between group analyses Greater improvement in postural control [b755, b760] (p < 0.05) in group 1. Greater improvement in gross motor skills [d4] (p < 0.05) in group 1. Greater improvement in health-related quality of life (p < 0.05) in group 1.
GMAT + progressive resistance exercise (GMAT-PRE)
[39]	Case series Level V Low quality/weak recommendation	Intellectual disability n = 2, 1 F, 1 M 18–19 years	Country: United States School type: Mainstream high school with special education program	Participant 1 Content: Therapeutic exercise, functional activities, participation in APE program, and a home program. Dose: 9 h (18 weeks × 1 session per week × 30 mins) Delivered by: Physiotherapist, in collaboration with school teachers. Target: Student Participant 2 Content: Therapeutic exercise, functional activities, instruction to school team regarding proper squat technique and teaching strategies, a home program, and continued participation in APE. Dose: 6 h (12 weeks × 1 session per week × 30 mins) Delivered by: Physiotherapist, in collaboration with school teachers. Target: Student	Participant 1 Improved lifting performance [d430] – able to perform 10–25 reps of lifting activities (previously performed 9–10 reps) (p not reported) Improved squat performance [d4101] – able to move from squat to stand and stand to squat (previously unable); able to hold squat position for 60 s (previously unable) (p not reported) Improved body mechanics for squatting [d4101] – scoring 8/8 on mechanics checklist (previously 1/8) (p not reported) Participant 2 Improved lifting performance [d430] – able to perform 15–20 reps of lifting activities (previously performed 9–15 reps) (p not reported) Improved squat performance [d4101] – able to move from squat to stand and stand to squat (previously unable); able to hold squat position for 32 s (previously unable) (p not reported) Improved body mechanics for squatting [d4101] – scoring 6/8 on mechanics checklist (previously 1/8) (p not reported)
Hippotherapy
[40]	Case report Level V Low quality/weak recommendation	Diagnosis: Cerebral palsy with mild cognitive impairment n = 1 Age: 14 years 1 F	Country: Slovenia School type: Mainstream primary school with special education program	Intervention Type 1 Content: Hippotherapy exercises on unsaddled horseback to improve motor function through muscle relaxation, balance responses and postural control. Dose: Single occasion, 45 min. Delivered by: Physiotherapist Target: Student Intervention Type 2 Content: NDT (before class) with a focus on inhibition of pathological patterns and facilitation of normal movement to induce correct motor responses. Dose: Single occasion, 45 min Delivered by: Physiotherapist Target: Student Intervention Type 3 Content: NDT (after class) with a focus on inhibition of pathological patterns and facilitation of normal movement to induce correct motor responses. Dose: Single occasion, 45 min. Delivered by: Physiotherapist Target: Student	Within group analyses Intervention type 1 (pre-post) No change in reading speed [d166] – no change in time to complete a set reading task (p not reported). Improved writing speed [d170] – faster time to complete a set writing task (p not reported). Improved accuracy of writing [d170] – fewer mistakes made in exact joining of letters (contact lineation) (p not reported). Between group analyses: No difference between groups for contact lineation (p not reported) or reading speed (p not reported).
Multimodal education-based therapy (MET)
PT COUNTS (single study comprised of the following five articles) [41] [42] [43] [44] [45]	Case series Level IV Low quality/weak recommendation	Diagnoses: Cerebral palsy (n = 103); down syndrome (n = 46); other genetic syndromes (n = 40); global developmental delay (n = 32); other (n = 75) n = 296 Age: 5–12 years 130 F, 166 M	Country: United States School type: Mainstream primary schools	Content: School physiotherapy intervention based on students’ IEP goals. Activities included practice of sitting, standing, and transitions, mobility, and physical education/ recreation. More time spent on behalf of student (consultation/collaboration, documentation, in-service, or curriculum development) as opposed to delivering direct therapy. Dose: Average of 534.4 min of PT services delivered per child in 6 months (duration of each service occasion not reported); average of 86.9 min of group service delivery, and 262.7 min of services on behalf of the student. Delivered by: Physiotherapists Target: Students, teachers	Goal Attainment Scale: GAS goals achieved: 75.68% of students achieved their primary GAS goal. [41] Improved GAS (posture/mobility): [d410-d429; d450-d469] 41% exceeded goals for posture/mobility. More minutes of service on behalf of student and self-care activities associated with exceeding expectation (p < 0.05). [43] Improved GAS (recreation/fitness): [d920; d570]. 40% exceeded goals for recreation/fitness. Greater use of functional strength, mobility for playground access, cognitive and behavioural interventions was associated with exceeding expectation (p < 0.05). [43] Improved GAS (self-care): [d5] 40% exceeded goals for self-care. Students received less PT time with no other students present (p = 0.05), less documentation time (p < 0.02) and more neuromuscular interventions (p < 0.04) than students who did not exceed expectation. [43] Improved GAS (academic): [d820] 27% exceeded academic goals. Students received more mobility interventions than students who did not exceed expectation (p < 0.03). [43] School Function Assessment: Improved SFA (Participation): [d] 51% improved on the Participation domain of the SFA (p not reported). Improved SFA (Task Support Adaptations): [e] 46% improved on the Task Support Adaptations domain of the SFA (p not reported). Improved SFA (Maintaining and Changing Position): [d410–429] 46% improved on the Maintaining and Changing Position domain of the SFA (p not reported). Improved SFA (Manipulation with Movement): [d430–449] 44% improved on the Manipulation with Movement domain of the SFA (p not reported). Mobility, sensory, motor learning, aerobic/ conditioning, functional strengthening, and playground access intervention services correlated positively with SFA outcomes, as well as higher student participation during therapy (p not reported). Services with negative correlation to SFA outcomes included group delivery, within school activities, with students not in special education, during recreation activities, and with positioning, hands-on facilitation, sensory integration, orthoses, and equipment interventions (p not reported). [45]
[46]	Case series Level IV Low quality/weak recommendation	Diagnosis: Down syndrome n = 46 Age: 5–11 years 25 F, 20 M	Country: United States School type: Mainstream primary school with special education classes	Intervention Group Content: School physiotherapy intervention based on individual goals from the students’ IEP, e.g., transfers, mobility, physical education participation, lunchtime recreation participation. Dose: Average of 11 h 50 min of services per student (20 weeks × 35.6 min per week); 8 h direct therapy (20 weeks × 24 min per week) plus 3 h and 50 min of services on behalf of the student (20 weeks × 11.6 min per week). Delivered by: Physiotherapists Target: Students, teachers	Intervention Group (pre-post) Improved GAS (posture/mobility): [d410-d429; d450-d469] 66.7% met/exceeded goals for posture/mobility (p not reported). Improved GAS (recreation): [d920] 73.7% met/exceeded goals for recreation (p not reported). Improved SFA: [d820] mean subtest scores increased by 2.2–3.7 points (p not reported).
[47]	Case report Level IV Low quality/weak recommendation	Diagnosis: Cerebral palsy (spastic hemiplegia), GMFCS I n = 1 Age: 5 years 1 M	Country: United States School type: Kindergarten at mainstream school	Content: Mobility and stability tasks relevant to goal of ascending ladder to playground. Combination of strength and endurance training, ROM and body awareness activities, to address identified constraints to function. Dose: 19 sessions (unspecified duration) Delivered by: Physiotherapist Target: Student	Improved ability to ascend ladder to access slide [d4551] (previously unable). No rating scale reported (p not reported).
Neurodevelopmental treatment (NDT)
[40]	Case report Level V Moderate quality/moderate recommendation	Diagnosis: Cerebral palsy with mild cognitive impairment n = 1 Age: 14 years 1 F	Country: Slovenia School type: Mainstream primary school with special education program	Intervention Type 1 Content: Hippotherapy exercises on unsaddled horseback to improve motor function through muscle relaxation, balance responses and postural control. Dose: Single occasion, 45 min. Delivered by: Physiotherapist. Target: Student Intervention Type 2 Content: NDT (before class) with a focus on inhibition of pathological patterns and facilitation of normal movement to induce correct motor responses. Dose: Single occasion, 45 min Delivered by: Physiotherapist. Target: Student Intervention Type 3 Content: NDT (after class) with a focus on inhibition of pathological patterns and facilitation of normal movement to induce correct motor responses. Dose: Single occasion, 45 min. Delivered by: Physiotherapist. Target: Student	Within group analyses: Intervention type 2 (pre-post) Improved reading speed [d166] – faster time to complete a set reading task (p not reported). Improved writing speed [d170] – faster time to complete a set writing task (p not reported). Improved accuracy of writing [d170] – fewer mistakes made in exact joining of letters (contact lineation) (p not reported). Intervention type 3 (pre-post) No change in reading speed [d166] – no change in time to complete a set reading task (p not reported). Improved writing speed [d170] – faster time to complete a set writing task (p not reported). Improved accuracy of writing [d170] – fewer mistakes made in exact joining of letters (contact lineation) (p not reported). Between group analyses: Greater improvement in writing speed after NDT (before class) compared to other interventions (χ^2 = 44.68; χ^2 0.05 (3) = 7.81). No difference between groups for contact lineation (p not reported) or reading speed (p not reported).
[48]	Crossover RCT Level II Strong Moderate quality/moderate recommendation	Diagnosis: Cerebral palsy (spastic; bilateral), GMFCS I – III; age-matched typically developing children n = 8 Age: 7–18 years 2 F, 6 M	Country: Germany School type: Special school (primary and high school)	Intervention group: Climbing therapy (then usual PT) (n = 4) Content: Indoor rock climbing at an orthopaedic hospital for children. Dose: 18 h (6 weeks × 2 sessions per week × 1.5 h); 4 weeks wash-out prior to study and between interventions. Delivered by: Climbing therapists. Target: Students Comparator group: Usual PT (then climbing therapy) (n = 4) Content: NDT and functional tasks Dose: 18 h (6 weeks × 2 sessions per week × 1.5 h); 4 weeks wash-out prior to study and between interventions. Delivered by: Physiotherapists Target: Students	Within group analyses – Comparator group Group 2 (pre-post) Improved self-selected walking speed, step length, and step time [d450-d469] in participants with cerebral palsy (p not reported). Improved gait kinematics on Gait Profile Score [b770] (–1.4°) (p not reported). Improved knee flexion (–4.5°) and ankle dorsiflexion (1.1°) [b770] (p not reported). Deterioration in the hip peak internal rotation [b770] (p not reported). Between group analyses No significant difference in spatiotemporal parameters of gait between groups [d450–469] (p not reported). Greater improvement in knee flexion at initial contact in the comparator group (p = 0.039) [b710].
Non-immersive virtual reality (NVR)
[36]	Pre-post design Level IV Low quality/ weak recommendation	Diagnosis: Developmental coordination disorder n = 46 Age: 6–10 years 22 F, 24 M	Country: South Africa School type: Mainstream primary schools	Group 1: Neuromotor task training (n = 27) Content: Part-task practice of components of soccer, netball, variations of tagging games, and other playground games. Dose: 13.5–18 h (9 weeks × 2 sessions per week × 45–60 min) Delivered by: Physiotherapists. Target: Students Group 2: Wii Fit (n = 19) Content: Games designed to mimic cycling, soccer, skateboarding or skiing, by encouraging weight shift in lateral and anterior directions, and stepping or humping over virtual obstacles. Games incorporating arm movements using the hand controller were also used. Dose: 9 h (6 weeks × 3 sessions per week × 30 min) Delivered by: Physiotherapists. Target: Students	Within group analyses Group 1 (pre-post) Improved gross motor skills [d4] on Total Standard Score of MABC-2 (p < 0.0001). Improved functional strength [b730] on Functional Strength Measure (p < 0.01). Improved anaerobic power [b730] on Muscle Power Sprint Test (p < 0.01). Group 2 (pre-post) No change in gross motor skills [d4] on Total Standard Score of MABC-2 (p = 0.26). No change in functional strength [b730] on Functional Strength Measure (p = 0.89). Improved anaerobic power [b730] on Muscle Power Sprint Test (p = 0.01). Between group analyses No significant difference in motor performance between groups (TSS of MABC-2) (p = 0.12). Greater improvement in functional strength (Functional Strength Measure) (p < 0.01) in group 1. No significant difference in upper limb muscle strength between groups (isometric contraction measured with dynamometer) (p not reported). No significant difference anaerobic power between groups (Muscle Power Sprint Test) (p not reported).
[49]	RCT Level II Strong Moderate quality/moderate recommendation	Diagnosis: Cerebral palsy, GMFCS III – IV n = 18 Age: 6–14 years 7 F, 11 M	Country: Hong Kong School type: Special school	Group 1: Non-immersive VR and usual PT (n = 9) Content: Computer games using a force plate, designed to target postural control. Usual care interventions not specified. Dose: 8 h (6 weeks × 4 sessions per week × 20 min) Delivered by: Physiotherapists Target: Students Group 2: Usual PT alone (n = 9) Content: Typical care (interventions not specified). Dose: Not specified. Delivered by: Physiotherapists Target: Students	Within group analyses Group 1 (pre-post) Improved dynamic balance [b755, b760] − 3.28 cm improvement in mean forward reach distance on Paediatric Reach Test (p not reported). Improved gross motor function [d4] − 0.51 increase in mean GMFM score (p not reported). Improved walking ability and functional capacity [d450-d469] − 1.43 m increase in mean walking distance on 2MWT (p not reported). Group 2 (pre-post) Improved dynamic balance [b755, b760] − 5.22 cm improvement in mean forward reach distance on Paediatric Reach Test (p not reported). Improved gross motor function [d4] − 2.04 increase in mean GMFM score (p not reported). Improved walking ability and functional capacity [d450-d469] − 10.4 m increase in mean walking distance on 2MWT (p not reported). Between group analyses No significant difference between groups on measures of dynamic balance (Paediatric Reach Test), gross motor function (GMFM), or walking ability (2MWT), at any testing moment (p > 0.001).
[50]	Case series Level IV Low quality/weak recommendation	Diagnosis: Cerebral palsy (mild – moderate hemiplegic and diplegic), GMFCS I – II n = 11 Age: 4–11 years 6 F, 5 M	Country: Spain School type: Mainstream primary school	Content: Non-immersive VR using Kinect Xbox 360. Games used to target balance, trunk movements, visual-manual coordination, and limb movements. Dose: 8 h (2 months × 2 sessions per weeks × 30 min) Delivered by: Physiotherapists Target: Students	Improved gross motor function [d4] on GMFM (p = .001). Improved quality of performance of activities of daily living [d230] on the Assessment of Motor and Process Skills (motor (p = 0.001) and process (p = 0.01)). Improved dynamic balance [b755, b760] on Paediatric Reach Test (p = 0.005). Improved walking speed [d450] on 10MWT (p = 0.029). Improved hand function [d440, d445] on most domains of Jebsen-Taylor Hand Function Test; differences between basal and follow-up mostly not significant (p not reported).
Rock climbing
[48]	Crossover RCT Level II Strong Moderate quality/moderate recommendation	Diagnosis: Cerebral palsy (spastic; bilateral), GMFCS I – III; age-matched typically developing children n = 8 Age: 7–18 years 2 F, 6 M	Country: Germany School type: Special school (primary and high school)	Intervention group: Climbing therapy (then usual PT) (n = 4) Content: Indoor rock climbing at an orthopaedic hospital for children. Dose: 18 h (6 weeks × 2 sessions per week × 1.5 h); 4 weeks wash-out prior to study and between interventions. Delivered by: Climbing therapists. Target: Students Comparator group: Usual PT (then climbing therapy) (n = 4) Content: NDT and functional tasks Dose: 18 h (6 weeks × 2 sessions per week × 1.5 h); 4 weeks wash-out prior to study and between interventions. Delivered by: Physiotherapists Target: Students	Within group analyses – Intervention group Group 1 (pre-post) Improved self-selected walking speed, step length, and step time [d450-d469] in participants with cerebral palsy (p not reported). Deterioration in gait kinematics on Gait Profile Score [b770] (p not reported). Improved hip peak internal rotation [b770] (0.1°) (p not reported). Deterioration in knee flexion and ankle dorsiflexion [b770] (p not reported). Between group analyses No significant difference in spatiotemporal parameters of gait between groups [d450–469] (p not reported).
Upper limb
[51]	RCT Level II Strong High quality/strong recommendation	Diagnosis: Cerebral palsy (spastic hemiplegia), GMFCS I – III n = 32 Age: mean age 10.43 years 15 F, 17 M	Country: Turkey School type: Mainstream primary schools (School-based intervention (1x 1h session/week) or group sessions at a rehabilitation centre (2x 1h session/week), combined with home practice)	Group 1: mCIMT (n = 16) Content: Gross motor and fine motor activities, manipulative games, puzzles, arts and crafts, board and card games, and functional activities, with use of the less affected hand restricted. Dose: 75 h face-to-face therapy (10 weeks × 3 sessions per week × 2.5 h) and home practice (1 h per day during intervention period and 2 h per day during follow-up period). Delivered by: Physiotherapists Target: Students Group 2: Bimanual Training (n = 16) Content: Fine and manipulative gross motor activities requiring the use of both hands, particularly tasks which were skilled, repetitive, and structured in nature. Dose: 75 h face-to-face therapy (10 weeks × 3 sessions per week × 2.5 h) and home practice (1 h per day during intervention period and 2 h per day during follow-up period). Delivered by: Physiotherapists Target: Students	Within group analyses: Group 1 (pre-post) Improved grip strength [b730] of more affected hand (p < 0.001) (d = 0.53), retained during follow-up (p < 0.01) (d = 0.4). Improved quality of bimanual hand use [d445] (d = 0.91–0.96). Improved quantity of use of more affected limb [d445] (p < 0.001). Improved global participation (d = 1.08) and participation in school activities (d = 1.1) (p < 0.001) [d820] on the Child and Adolescent Scale of Participation. Group 2 (pre-post) No improvement in grip strength [b730] (p > 0.05). Improved quality of bimanual hand use [d445] (d = 0.63–0.75). Improved quantity of bimanual hand use [d445] (p < 0.001). Improved global participation (d = 0.67) and participation in school activities (d = 0.69) [d820] on the Child and Adolescent Scale of Participation. Between group analyses: Greater improvement in unimanual capacity (QUEST) [d440, d445] after mCIMT (p < 0.001). Greater improvement in manual ability in everyday activities (ABILHAND-Kids) [d440, d445] after mCIMT (p < 0.001) (d1=3.75, d2=1.12).
[52]	RCT Level II Moderate High quality/strong recommendation	Diagnosis: Handwriting problem diagnosed by HPSQ n = 60 Age: 5–7 years 19 F, 41 M	Country: India School type: Mainstream primary schools	Group 1: Physiotherapeutic exercises targeting handwriting (n = 30) Content: Exercises to improve upper limb stability, visual-motor development, and fine motor skills for handwriting. Dose: 48 h (3 months × 4 sessions per week × 1 h) Delivered by: Physiotherapists Target: Students Group 2: Ergonomic advice and postural correction (n = 30) Content: Ergonomic advice on handwriting, taught appropriate writing posture Dose: 1 session (duration not reported Delivered by: Physiotherapist, parents, teachers Target: Students	Within group analyses Group 1 (pre-post) Improved handwriting proficiency [d170] on HPSQ (t = 8.413), MHA total (t = 42.17), and MHA subscale scores following intervention. Group 2 (pre-post) Improved handwriting proficiency [d170] on HPSQ (t = 6.71), MHA total (t = 34.00), and MHA subscale scores following intervention. Between group analyses Greater improvement in handwriting proficiency (HPSQ) scores in group 1 1 (p < 0.05) (d = 26.72).

GRADE: Grading of Recommendation, Assessment, Development, and Evaluation; DCD: developmental coordination disorder; MABC-2: Movement Assessment Battery for Children 2nd Edition; IEP: Individualised Education Program; PT: physiotherapy/physiotherapist; GMFM: Gross Motor Function Measure; 6MWT: 6 min Walk Test; KAFOs: knee ankle foot orthoses; RCT: randomised controlled trial; GMFCS: Gross Motor Function Classification Scale; 10MWT: 10 Metre Walk Test; CPChild: Caregiver Priorities and Child Health Index of Life with Disabilities; MABC: Movement Assessment Battery for Children; TGMD-2: Test of Gross Motor Development 2nd Edition; PSPCSA: Pictorial Scale of Perceived Competence and Social Acceptance; TGMD: Test of Gross Motor Development; PedsQL: Paediatric Quality of Life Inventory; APE: Adapted Physical Education; NDT: neurodevelopmental treatment; GAS: Goal Attainment Scale; SFA: School Function Assessment; ROM: range of motion; VR: virtual reality; 2MWT: 2 Metre Walk Test; mCIMT: modified Constraint Induced Movement Therapy; QUEST: Quality of Upper Extremity Skills Test; MHA: Minnesota Handwriting Assessment; HPSQ: Handwriting Proficiency Screening Questionnaire.

Interventions were delivered to school children with cerebral palsy (n = 11), developmental coordination disorder (n = 3), down syndrome (n = 2), intellectual disability (n = 1), developmental disability (n = 1), profound hearing impairment (n = 1), and other presentations (n = 5). Studies were conducted in both mainstream (n = 15) and special education settings (n = 6), including kindergartens (n = 1), primary schools (n = 17), and high schools (n = 8). When considering ICF coding of the 43 unique reported physiotherapy outcome measures, 30 related to activity, nine measures to body structures and functions, and only four measures to participation.

Details of study design, evidence level, conduct rating (for Level I-III studies), participants, location and setting, intervention type, results, outcome measures, and ICF coding [25] are presented for each study in Table 1. Analysis of study conduct, per the AACPDM guidelines [23] is presented in Table 2, and risk of bias per the RoB-2 [20], ROBINS-I [21], and NHLBI quality assessments [22] is presented in Tables 3–6, respectively. Quality of evidence, per the GRADE framework [24], is presented in Table 7. In the following sections, each intervention type is described in alphabetical order, with discussion of the overall strength of evidence, context, intervention dose, and efficacy.

Table 2. AACPDM conduct of studies of school-based physiotherapy interventions – group design studies – Level of Evidence I-III.

Study	Evidence level/quality	1 Inclusion criteria	2 Int/con participants described	3 Measures appropriate	4 Blinded assessor	5 Statistical calculations	6 Dropout reporting	7 Controlling confounders
Ergonomic health literacy
[28]	II (4/7) moderate	No	Yes/yes	Yes	Yes	No	No	Yes
Gait training – treadmill (partial body-weight support)
[34]	II (7/7) strong	Yes	Yes/yes	Yes	Yes	Yes	Yes	Yes
Gross motor activity training + multimodal education-based therapy (GMAT-MEBT)
[37]	II (6/7) moderate	Yes	Yes/yes	Yes	Yes	No	Yes	Yes
[38]	II (4/7) moderate	Yes	Yes/no	Yes	No	No	Yes	Yes
Rock climbing /NDT
[48]	II (6/7) strong	Yes	Yes/yes	Yes	Yes	No	Yes	Yes
Non-immersive virtual reality
[49]	II (5/7) strong	Yes	Yes/no	Yes	Yes	No	Yes	Yes
Upper limb
[51]	II (7/7) strong	Yes	Yes/yes	Yes	Yes	Yes	Yes	Yes
[52]	II (4/7) moderate	Yes	Yes/yes	Yes	No	No	No	Yes

1. Were inclusion and exclusion criteria of the study population well described and followed? 2. Was the intervention well described and was there adherence to the intervention assignment? (for 2-group designs, was the control exposure also well described?) Both parts of the question need to be met to score ‘yes.’ 3. Were the measures used clearly described, valid and reliable for measuring the outcomes of interest? 4. Was the outcome assessor unaware of the intervention status of the participants (i.e., were the assessors masked)? 5. Did the authors conduct and report appropriate statistical evaluation including power calculations? Both parts of the question need to be met to score ‘yes.’ 6. Were drop out/loss to follow-up reported and less than 20%? For 2-group designs, was drop out balanced? 7. Considering the potential within the study design, were appropriate methods for controlling confounding variables and limiting potential biases used? [25].

Table 3. Cochrane risk-of-Bias tool for randomised trials (RoB-2) – conduct of studies of school-based physiotherapy interventions – group design studies – Level II Evidence.

Study	1 Risk of bias arising from the randomisation process	2 Risk of bias due to deviations from the intended interventions (Effect of assignment to intervention)	2 Risk of bias due to deviations from the intended interventions (Effect of adhering to intervention	3 Missing outcome data	4 Risk of bias in measurement of the outcome	5 Risk of bias in selection of the reported result	Overall risk of bias
[28]	Low	Low	Low	Low	Low	Low	Low
[33]	Low	Low	Low	Low	Low	Low	Low
[34]	Low	Low	Low	Low	Low	Low	Low
[37]	Low	Low	Low	Low	Low	Low	Low
[38]	Low	Low	Low	Low	Low	Low	Low
[48]	Low	Low	Low	Low	Low	Low	Low
[49]	Low	Low	Low	Low	Low	Low	Low
[51]	Low	Low	Low	Low	Low	Low	Low
[52]	Low	Low	Low	Low	Low	Low	Low

Table 4. Risk of Bias in non-randomised studies – of interventions (ROBINS-I) – conduct of studies of school-based physiotherapy interventions – non-randomised group design studies – Level IV Evidence.

Study	1 Bias due to confounding	2 Bias in selection of participants into the study	3 Bias in classification of interventions	4 Bias due to deviations from intended interventions	5 Bias due to missing data	6 Bias in measurement of outcomes	7 Bias in selection of the reported result	Overall risk of bias
[27]	Moderate	Low	Low	Low	Low	Low	Low	Low
[36]	Moderate	Low	Low	Low	Low	Low	Low	Moderate

Table 5. NHLBI quality assessment – conduct of studies of school-based physiotherapy interventions – pre-post design studies – Level IV Evidence.

Study	1 Was the study question or objective clearly stated?	2 Were eligibility/ selection criteria for the study population prespecified and clearly described?	3 Were the participants in the study representative of those who would be eligible for the test/ service/ intervention in the general or clinical population of interest?	4 Were all eligible participants that met the pre-specified entry criteria enrolled?	5 Was the sample size sufficiently large to provide confidence in the findings?	6 Was the test/ service/ intervention clearly described and delivered consistently across the study population?	7 Were the outcome measures prespecified, clearly defined, valid, reliable, and assessed consistently across all study participants?	8 Were the people assessing the outcomes blinded to the participants’ exposures/ interventions?	9 Was the loss to follow-up after baseline 20% or less? Were those lost to follow-up accounted for in the analysis?	10 Did the statistical methods examine changes in outcome measures from before to after the intervention? Were statistical tests done that provided p values for the pre-to-post changes?	11 Were outcome measures of interest taken multiple times before the intervention and multiple times after the intervention (i.e., did they use an interrupted time-series design)?	12 If the intervention was conducted at a group level (e.g., a whole hospital, a community, etc.) did the statistical analysis take into account the use of individual-level data to determine effects at the group level?	Quality rating
[31]	Yes	No	Yes	Cannot determine	No	Yes	Yes	No	NA	No	Yes	NA	Poor
[32]	Yes	No	Yes	Cannot determine	No	Yes	Yes	No	NA	No	Yes	NA	Poor
[35]	Yes	Yes	Yes	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Fait
[47]	Yes	No	Yes	Cannot determine	No	Yes	Yes	No	NA	No	No	NA	Poor

Table 6. NHLBI quality assessment – conduct of studies of school-based physiotherapy interventions – case series studies – Level IV Evidence.

Study	1 Was the study question or objective clearly stated?	2Were eligibility/selection criteria for the study population prespecified and clearly described?	3 Were the cases consecutive?	4 Were the subjects comparable?	5 Was the intervention clearly described?	6 Were the outcome measures clearly defined, valid, reliable, and implemented consistently across all study participants?	7 Was the length of follow-up adequate?	8 Were the statistical methods well-described?	9 Were the results well-described?	Quality rating
[29]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Good
[30]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Good
[39]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	Fair
[41–45]	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Fair
[46]	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Fair
[50]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Good

Table 7. GRADE quality of evidence table.

No.of studies	Design	Limitations (Risk of bias)	Indirectness of patients, intervention, and comparator	Inconsistency	Imprecision	Quality of evidence	Strength of recommendation
Environmental modifications
1	Pre-post (n = 1)	High	Direct	No major inconsistencies	Moderate uncertainty	Low	Weak
Ergonomic health literacy
3	Case series (n = 2) Cohort study (n = 1)	High	Direct	Some inconsistency	Moderate uncertainty	Low	Weak
Gait training – overground
2	Case report (n = 2)	High	Direct	No major inconsistencies	Moderate uncertainty	Low	Weak
Gait training – treadmill (no body-weight support)
1	RCT (n = 1)	Low	Direct	No major inconsistencies	None	High	Strong
Gait training – treadmill (partial body-weight support)
1	RCT (n = 1)	Low	Direct	Some inconsistency	None	Moderate	Moderate
Gait training – robotic-assisted
1	Pre-post (n = 1)	Medium	Direct	No major inconsistencies	Moderate uncertainty	Moderate	Moderate
Gross motor activity training (GMAT)
1	Pre-post (n = 1)	Medium	Direct	Some inconsistency	Moderate uncertainty	Low	Weak
Gross motor activity training + multimodal education-based therapy (GMAT-MET)
2	RCT (n = 2)	Low	Direct	Some inconsistency	None	Moderate	Moderate
GMAT + progressive resistance exercise (GMAT-PRE)
1	Case report (n = 1)	High	Direct	Some inconsistency	Moderate uncertainty	Low	Weak
Hippotherapy
1	Case report (n = 1)	High	Direct	Some inconsistency	Moderate uncertainty	Low	Weak
Multimodal education-based therapy (MET)
3*	Case series (n = 2) Case report (n = 1)	High	Direct	Some inconsistency	Moderate uncertainty	Low	Weak
Neurodevelopmental treatment (NDT)
2	RCT (n = 1) Case report (n = 1)	Medium	Direct	No major inconsistencies	Moderate uncertainty	Moderate	Moderate
Non-immersive virtual reality (NIVR)
3	RCT (n = 1) Case series (n = 1)	Medium	Direct	Moderate inconsistency	Moderate uncertainty	Low	Weak
Rock climbing
1	Crossover RCT (n = 1)	Low	Direct	No major inconsistencies	Moderate uncertainty	Moderate	Moderate
Upper limb
2	RCT (n = 2)	Medium	Direct	No major inconsistencies	None	High	Strong

*Including 1 study comprised of 5 articles. Total of 7 articles.

Environmental modifications (n = 1)

Environmental modification interventions included those which altered the school environment, provided appropriate equipment, or collaborated with teachers to provide appropriate and inclusive education. One study (Evidence Level IV) [27] evaluated a whole-school health-promotion program delivered over 234 h (9 weeks × 26 h per week) and compared effects between students with developmental coordination disorder (DCD) and typically developing peers. Students with DCD demonstrated improvement in balance and gross motor skills, while typically developing students improved their anaerobic power. The available evidence provides a weak recommendation in favour of environmental modification interventions for children with DCD in schools.

Ergonomic health literacy (n = 3)

Ergonomic health literacy interventions involved educating children about body mechanics and postures as they relate to daily activities and protecting physical health. Three studies (Evidence Level II-IV, moderate n = 1) supported the use of ergonomic health literacy for school children. Students improved their ergonomic knowledge [28–30] and spinal care behaviour [29,30]. Intervention dose ranged from 90 min (1–2 sessions) to six hours (6 weeks × 1 session × 1 h). Considering the available evidence, a weak recommendation can be made for ergonomic health literacy interventions in schools.

Gait training (n = 4)

Gait training interventions include the practice of walking. Four approaches were identified in this review, including overground (n = 2), treadmill without body-weight support (BWS) (n = 1), treadmill with partial BWS (n = 1), and robotic-assisted (n = 1). The evidence for each of these was analysed separately due to the reported differences in the effectiveness of these interventions in other contexts [53].

Gait training – overground (n = 2)

Overground gait training involves progressive practice of walking with or without a mobility aid. Two studies (both Evidence Level V) supported overground gait training for school children with disabilities [31,32]. Intervention dose ranged from 100 min (5 days × 20 min) to 60 min weekly (duration unspecified). Students improved their independent mobility; one student with a developmental disability learnt to perform a sit-to-stand and walk 30 feet using a walker [32], and another student with cerebral palsy made gains in walking distance (6MWT [54]) and gross motor function (Gross Motor Function Measure (GMFM) [55]) [31]. Available evidence provides a weak recommendation for overground gait training for children with cerebral palsy and developmental disabilities in schools.

Gait training – treadmill (no body-weight support, n = 1)

Treadmill training without bodyweight support involved repetitive gait practice using treadmill equipment. One study (Evidence Level II, strong) [33] provided positive evidence for treadmill training without BWS for ambulant school children with cerebral palsy. Intervention dose was 27 h (12 weeks × 3 sessions × 45 min). Students improved their walking speed during training, training duration, self-selected walking speed, and gross motor function (GMFM) [55]. For these outcomes, treadmill training was more effective than MET, consisting of balance, gross motor, and overground gait training. A strong recommendation can be made for treadmill training without BWS for children with cerebral palsy (Gross Motor Function Classification System (GMFCS) I-III [56]) in schools.

Gait training – treadmill (partial body-weight support, n = 2)

Treadmill training with partial BWS was performed as above, with the addition of a body-weight support system. One study (Evidence Level II, strong) [34] provided evidence for treadmill training with partial BWS for ambulant school children with cerebral palsy. Intervention dose was 8 h (8 weeks × 2 sessions × 30 min). Students improved their walking speed during training, training duration, self-selected walking speed, and gross motor function (GMFM [55]). For these outcomes, treadmill training with partial BWS was no more effective than overground gait training [34]. Available evidence provides a weak recommendation for treadmill training with partial BWS for children with cerebral palsy (GMFCS II-III) in schools.

Gait training – robotic assisted (n = 1)

Robotic-assisted gait training involved assisted, guided, and repetitive gait-like movements, facilitating mobility in a weight-bearing position. One study (Evidence Level IV) [35] provided evidence for robotic-assisted gait training for school children with cerebral palsy. Intervention dose was 12 h (6 weeks × 4 sessions × 30 min). Students improved their health-related quality of life (CPChild [57]) individual goal attainment, lower limb range of motion, and lower limb tone. Improvements were most notable immediately post-intervention, with outcome measures returning to or worsening beyond pre-intervention measures by three months post-intervention. Available evidence provides a weak recommendation for robotic-assisted gait training for children with cerebral palsy (GMFCS IV-V) in schools.

Gross motor activity training (GMAT) (n = 4)

Gross Motor Activity Training (GMAT) is designed to target gross motor development through practice of functional gross motor activities [26]. This review identified one GMAT intervention delivered in isolation (n = 1), and two GMAT interventions delivered in combination with adjunct physiotherapy: GMAT-MET (n = 2) and GMAT-PRE (n = 1).

GMAT alone (n = 1)

One study (Level IV evidence) [36] evaluated the efficacy of GMAT compared to a non-immersive virtual reality (NVR) intervention, for children with developmental coordination disorder. In this study, GMAT included part-task practice of components of various sports and playground games. Intervention dose ranged from 13.5 to 18 h (9 weeks × 2 sessions × 45–60 min). Students improved their gross motor skills (Movement Assessment Battery for Children – 2nd edition (MABC-2) [58]), functional strength (Functional Strength Measure (FSM) [59]), and anaerobic power (Muscle Power Sprint Test (MPST) [60]). GMAT was superior to NVR for improving functional strength, and similarly as effective for improving gross motor skills, upper limb muscle strength, and anaerobic power. For these outcomes, a weak recommendation can be made for GMAT interventions for children with developmental coordination disorder in schools.

GMAT with multimodal education-based therapy (GMAT-MET) (n = 2)

Two studies (Evidence Level II, strong n = 1, moderate n = 1) provided strong positive evidence for GMAT + MET. MET was defined as individually tailored physiotherapy designed to address students’ individual education-based goals. In these studies, MET included fine motor, body awareness, and skills practice activities for children with developmental coordination disorder [37], and oculomotor exercises and balance training for children with profound hearing impairment [38]. Intervention dose was similar, at 13 h (13 weeks × 1 session × 60 min) and 13.5 h (6 weeks × 3 sessions × 45 min), with some variation in session frequency. Students with developmental coordination disorder demonstrated improvement in gross motor skills (Movement Assessment Battery for Children [58] and Test of Gross Motor Development (TGMD) [59]) and perceived physical competence [37]. Students with profound hearing impairment improved their postural control, gross motor skills (TGMD [61]), and health-related quality of life [38]. As such, a moderate recommendation can be made for GMAT + MET interventions for children with developmental coordination disorder and profound hearing impairment in schools.

GMAT with progressive resistance exercise (GMAT-PRE) (n = 1)

One GMAT-PRE intervention was identified (Evidence Level V) [39], which involved gross motor activities such as sit-to-stands and squats, performed with additional resistance, such as free weights. Intervention dose ranged from 6 h (12 weeks × 1 session × 30 min) to 9 h (18 weeks × 1 session × 30 min). Two students improved their performance of lifting and squatting tasks, with a greater number of repetitions performed and improved body mechanics [39]. The available evidence provides a weak recommendation for GMAT-PRE for children with intellectual disabilities in schools.

Hippotherapy (n = 1)

Hippotherapy involves performing physiotherapist-prescribed exercises on horseback, utilising the natural gait and movement of the horse to provide motor and sensory input to the participant [62]. One study (Evidence Level V) [40] compared hippotherapy to NDT for improving performance in an academic task for one child with cerebral palsy. The student demonstrated improved speed and accuracy of writing following 45 min of hippotherapy (single occasion); however, no change was observed in reading speed. For these outcomes, available evidence provides a weak recommendation for hippotherapy for children with cerebral palsy in schools, in specific circumstances.

Multimodal education-based therapy (MET) (n = 7)

MET consists of individually tailored physiotherapy designed to address students’ individual education-based goals. Goals may be set across domains such as mobility, recreation, education, or self-care, but must be educationally focussed. Three studies (seven articles) (Evidence Level IV-V) [41–47] provided evidence for MET. Intervention dose ranged from 19 sessions (frequency and duration unspecified) to an average of 8 h and 40 min (6 months × 20 min per week). Students improved their performance of education activities (School Function Assessment [63]) [39–42,61,62], their attainment of individual education-based goals (Goal Attainment Scaling [45,64]) [41–46], and ability to ascend a playground ladder [47]. Available evidence provides a weak recommendation for MET in schools, for children with a variety of conditions.

Neurodevelopmental treatment (NDT) (n = 2)

Neurodevelopmental treatment is an approach to therapy intervention involving sensory facilitation, management of motor compensations, and an overall strategy that considers the emotional, social, and functional problems of the individual [65]. Two studies (Evidence Level II-V, strong n = 1) [40,48] evaluated efficacy of NDT for children with cerebral palsy. Intervention dose ranged from 45 min (single occasion) to 18 h (6 weeks × 2 sessions × 90 min). Students improved their self-selected walking speed, step length, step time, and gait kinematics [48], and reading speed, writing speed, and text copying ability [40]. Neurodevelopmental treatment was superior to hippotherapy [40] and equally effective as rock climbing [48]. A moderate recommendation can be made for NDT in schools, for children with cerebral palsy (GMFCS I-III).

Non-immersive virtual reality (NVR) (n = 3)

Non-immersive virtual reality interventions involve interactions with a virtual environment through computer games [49] or gaming consoles [50] and may also incorporate input from external devices such as force plates or kinetic sensors. Three studies (Evidence Level II-IV, strong n = 1) provided evidence for NVR for children with cerebral palsy [49,50] and developmental coordination disorder [36]. Intervention dose ranged from 8 h (6 weeks × 4 sessions × 20 min, or 8 weeks × 2 sessions × 30 min) to 9 h (6 weeks × 3 sessions × 30 min). Students with cerebral palsy demonstrated improvements in gross motor function (GMFM [53]) and balance [49,50], as well as their walking speed, quality of daily activities and tasks, and hand function [49]. Students with developmental coordination disorder improved their anaerobic power (MPST [60]) but did not demonstrate significant improvements in gross motor skills (MABC-2 [58]) or functional strength (FSM [59]) [36]. One study [49] suggested that NVR is inferior to “usual physiotherapy” for children with cerebral palsy (content not specified); however, there was no comparator in the other study of children with cerebral palsy [50]. For children with developmental coordination disorder, NVR was inferior to GMAT for improving functional strength, and similarly as effective for improving gross motor skills, upper limb muscle strength, and anaerobic power [36]. A weak recommendation can be made for NVR in schools, for children with cerebral palsy (GMFCS I-IV) and developmental coordination disorder.

Rock climbing (n = 1)

Rock climbing combines participation in a sports activity with therapeutic strength, flexibility, and coordination training [66]. One study (Evidence Level II, strong) [48] evaluated the efficacy of an indoor rock-climbing program for children with cerebral palsy and age-matched typically developing children, when compared to NDT. Intervention dose was 18 h (6 weeks × 2 sessions × 90 min). Students improved their self-selected walking speed, step length, and step time; however, deterioration was noted in knee flexion and ankle dorsiflexion range during gait [48]. There was no significant difference in spatiotemporal parameters of gait between rock climbing and NDT [48]. A moderate recommendation can be made for rock climbing intervention for children with cerebral palsy (GMFCS I-III) in schools.

Upper limb (n = 2)

Upper limb interventions target the use of one or both hands to improve upper limb function relevant to school activities. One study (Evidence Level II, strong) provided positive evidence for the use of school-based modified Constraint-Induced Movement Therapy (mCIMT) and bimanual therapy for children with cerebral palsy [51]. Another study (Evidence Level II, moderate) supported the use of interventions targeting upper limb stability, visual-motor development, and fine motor skills, for children with poor handwriting [52]. Intervention dose ranged from 48 h (12 weeks × 4 sessions × 60 min) to 75 h (10 weeks × 3 sessions × 2.5 h). Students improved their grip strength, quality and quantity of hand use, and participation in school activities [51], and their handwriting proficiency [52]. Receiving mCIMT was superior to bimanual therapy for improving unimanual capacity (Quality of Upper Extremity Skills Test [67]) and manual ability in everyday activities (ABILHAND-Kids [68]). Physiotherapeutic exercises for handwriting were superior to a single session of ergonomic advice [52]. A strong recommendation can be made for providing upper limb interventions in schools, for children with cerebral palsy (GMFCS I-III) or handwriting difficulties.

Discussion

This systematic review identified 23 studies involving 14 types of school-based physiotherapy interventions delivered to children in their school environment. Each intervention aimed to improve students’ ability to engage in school activities. Strong positive evidence was available for treadmill training without BWS (Evidence Level II) and upper limb (Evidence Level II). Moderate positive evidence was provided for GMAT-MET (Evidence Level II), NDT (Evidence Level II-V), and rock climbing (Evidence Level II). Weak positive evidence was provided for environmental modifications (Evidence Level IV), ergonomic health literacy (Evidence Level II-IV), GMAT (Evidence Level IV), GMAT-PRE (Evidence Level V), hippotherapy (Evidence Level V), MET (Evidence Level IV), NVR (Evidence Level II-IV), overground gait training (Evidence Level IV), robotic-assisted gait training (Evidence Level IV), and treadmill training with partial BWS (Evidence Level II).

Interventions that were shown to be effective in schools were consistent with the effectiveness of those interventions reported in previous literature in non-school contexts. For example, students with hemiplegic cerebral palsy receiving mCIMT or bimanual training [51] improved their quality and quantity of hand use, manual ability in everyday activities, and participation in school activities. This is consistent with previous systematic reviews [69,70] evaluating upper limb interventions for children with hemiplegic cerebral palsy in community contexts, which suggested that CIMT [69,70] improves unimanual capacity of the more affected limb and bimanual training [69] improves bimanual performance.

The strength of evidence for all interventions evaluated in school contexts is lesser than that in other contexts, given the lower quantity and quality of studies. Many included studies provided weak positive evidence for interventions which have strong supporting evidence in other contexts. For example, a previous systematic review [71] of gait training for children with cerebral palsy in the community reported improvements in independent mobility and gross motor skills similar to included studies evaluating overground gait training [31,32]. Despite similar improvements, previous literature, which consisted primarily of randomised controlled trials, reported statistically significant changes, whilst school-based studies in this review did not report statistics and did not have sufficient sample size or conduct to accurately quantity efficacy. Future research for all school-based physiotherapy interventions should aim to use more robust study designs with greater sample sizes, to optimise methodological quality and improve clarity of results.

This review identified weak positive evidence for NVR for children with cerebral palsy in schools [49,50]. This is somewhat inconsistent with a previous review [68], which found moderate evidence in favour of NVR for children with cerebral palsy in community settings. Whilst both included studies reported that students improved on measures of gross motor skill, dynamic balance, and walking ability, one study [49] reported that NVR was inferior to usual physiotherapy (content not specified), and the other study did not use a comparator [50]. One additional study [36] investigated NVR for children with developmental coordination disorder, and reported no change in gross motor skills or functional strength, although did observe improvements in anaerobic power. It is possible for these studies that the sample size and dose of interventions were insufficient to properly measure effects. In addition, the poor description of the usual physiotherapy intervention in one study [49] means it is difficult to determine the relative efficacy of NVR. Further high quality research is necessary to better understand the efficacy of NVR in school contexts, particularly in comparison to other interventions with proven efficacy.

Several interventions evaluated in this review differed from descriptions in previous literature in content, dose, and mode of delivery. Gross motor activity training, first described by Clutterbuck et al. [26] is one such example. In this review, GMAT was identified predominantly in combination with other interventions, including GMAT-MET [37,38] and GMAT-PRE [39]. Although GMAT-PRE has been described previously [26], this is the first time that GMAT-MET has been described in the literature. Students receiving both GMAT-MET and GMAT-PRE demonstrated improvement across a range of outcomes measures, indicating improvement in gross motor skills [37–39], as well improvement in perceived physical competence [37], postural control [38], and body mechanics [39]. Only one study analysed the efficacy of GMAT alone in schools for children with developmental coordination disorder, and provided weak positive evidence for improving gross motor skills, functional strength, and anaerobic power [36]. Analysis of included studies shows that GMAT is a positive component of school-based physiotherapy, however further research is required to determine efficacy of GMAT alone for more diverse student populations.

As dose varied widely within and between intervention types, it was difficult to determine if frequency, duration, and intensity of interventions were associated with effectiveness. Intervention dose ranged from single occasions of 45-min duration [40], to 75 h over 10 weeks [51]. The comparison of dose between studies was made challenging by variation in delivery and reporting of physiotherapy contact time; this was particularly poorly reported in studies evaluating MET [41–47]. Analysis of previous literature [72] shows that dose is largely variable in physiotherapy for children, ranging from single occasions to intensive treatment models, with the most effective dose differing dependent on interventions, conditions, goals, and available funding. Moreover, the reported dose of interventions is significantly greater in community settings such as hospitals and clinics, than was reported in schools. This may be attributed to the requirement of intervention to fit within students’ schedules, as well as time constraints for school physiotherapists, limiting ability to conduct intensive school-based intervention research. Future research must aim to determine the lowest effective dose and preferred setting and mode of delivery for each intervention, to maximise feasibility within schools.

Improved participation is the primary objective of school-based physiotherapy, however many studies included in this review failed to assess the effectiveness of interventions to improve participation with an appropriate outcome measure. Only four studies [35,41–45,51] used outcome measures to assess participation. The Child and Adolescent Scale of Participation [73] was used to evaluate attendance [51], measured as the frequency of attending, or the diversity of activities an individual partakes in [74]. The School Function Assessment [75] was used to evaluate both attendance and involvement [41–45], which considers the experience of the individual whilst attending the life situation [74]. The Caregiver Priorities and Child Health Index of Life with Disabilities (CPChild) [57] and Paediatric Quality of Life Inventory (PedsQL) [76] were used to evaluate health-related quality of life across a range of domains, including the participants’ perceived impact of their disability on their ability to attend and be involved in life situations [35,38]. Although interventions may be delivered with aims of improving participation, their effectiveness in improving attendance or involvement can only be measured objectively with future research that includes appropriate outcome measures.

In the absence of strong evidence in school contexts, the findings of this systematic review suggest that interventions with established efficacy in other contexts, and for specific diagnostic groups, are likely to be effective in school settings. In addition, interventions should focus on assisting participants to improve their ability to engage in school activities.

A strength of this review is that it provides a comprehensive summary and synthesis of the literature on physiotherapy for children in schools, which had not previously been completed. In addition, this review illustrates the diversity of student needs at various ages and levels of ability in the school system, and therefore the requirement for physiotherapists to possess a broad skill set across several clinical areas. This review has also identified the need for development and validation of interventions and outcome measures to address participation in educational settings.

Several issues in the design, reporting, and conduct of included studies were identified in this review, resulting in poor quality of the literature. First, small sample size is a consistent issue among included studies, likely due to smaller populations of children with certain conditions in each school compared to in clinical settings. In combination with poor reporting of power calculations, small sample size creates difficulty in attributing the cause of non-significant findings to inadequate power or inefficacy of interventions. Second, the dose of interventions was likely insufficient to measure the effect of treatment in some studies, particularly Žgur [40] and Minghelli [29] in which the effects of single occasions of the intervention were evaluated. Third, only six of the 24 included studies were classified as high-quality evidence (Level II-III), usually due to the lack of randomisation or blinding within studies. In addition, there was an absence of control interventions in many studies, and in the case that control interventions were reported, dose and content were poorly described, particularly in the context of usual care. Finally, there were a limited number of studies available for each intervention, meaning that conclusions could not be drawn on the effectiveness of interventions for specific sub-populations of children with disabilities. Further research is required to determine the effectiveness of frequently used school-based physiotherapy interventions, especially for specific population groups.

The small number of published studies does not necessarily reflect of the frequency of interventions’ use in school settings. Barriers to publication of high-quality literature within school-based physiotherapy contexts may include lack of funding and time constraints for physiotherapists to conduct research in schools. School-based physiotherapists typically have limited non-clinical time in which to conduct research, as well as a lack of funding from education departments to facilitate research. It is likely that a great quantity of data is being collected clinically that could be evaluated for physiotherapy in schools. As such, the poor quality of the existing literature is likely not representative of actual practice, but rather a weakness of reporting.

Conclusion

School-based physiotherapy is broad in content and varied in participants, dosage, and outcome measures used. There is currently insufficient high-quality evidence to clearly indicate which interventions are most effective for specific populations of children in schools. It is recommended that (1) education jurisdictions conduct, report on, and publish results of audits for physiotherapy services delivered, to facilitate better understanding of referral reasons, students’ goals, services delivered, and subsequent outcomes, and (2) further research be conducted to inform standards for physiotherapy practice in schools. It is critical that future research seeks to deliver higher quality evidence, with (1) larger sample sizes from multi-site studies, (2) more diverse samples that are representative of school populations, (3) sufficient intervention dose for clinical effect, (4) optimal pairing of interventions and outcome measures to measure efficacy, and (5) measurement of participation outcomes to evaluate effectiveness of interventions in schools. Future research could also utilise single-case experimental designs, as opposed to randomised controlled trials, to accommodate diverse student populations and low incidence conditions [77]. Furthermore, it is recommended that school-specific participation outcome measures be developed to more accurately determine the true efficacy of interventions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.