The SIT-PT Trial Protocol: A Dose-Matched Randomized Clinical Trial Comparing 2 Physical Therapist Interventions for Infants and Toddlers With Cerebral Palsy

Abstract

Objective

Although early intervention for infants at risk for cerebral palsy is routinely recommended, the content of intervention is poorly described, varies widely, and has mixed supporting evidence. The purpose of this study was to compare efficacy of 2 interventions grounded in differing domains of the International Classification of Functioning, Disability and Health on developmental outcomes of infants with or at high risk of cerebral palsy.

Methods

Infants who meet inclusion criteria will be randomized into either Sitting Together and Reaching To Play or Movement, Orientation, Repetition, Exercise Physical Therapy groups. Both groups will receive intervention twice weekly for 3 months and follow-up at 3, 6, 9, and 12 months from baseline. The primary objectives compare changes over time and between groups in sitting, gross motor, and cognitive development. The setting is the infant’s home unless the caregiver requests otherwise. One hundred and fifty infants between 8 and 24 months of age will be enrolled in 3 geographically, racially, and ethnically diverse sites: Los Angeles, California; Omaha, Nebraska; and Seattle, Washington. Enrolled infants will demonstrate motor delays, emerging sitting skills, and signs of neurologic impairment. Sitting Together and Reaching To Play targets activities including sitting, reaching, and motor-based problem solving to improve global development. In contrast, Movement, Orientation, Repetition, Exercise Physical Therapy focuses on strengthening and musculoskeletal alignment while encouraging repeated movement practice. Outcome measures include the Gross Motor Function Measure, Bayley Scales of Infant Development-IV, Assessment of Problem Solving in Play, and a Parent Child Interaction assessment. Enrolled children will maintain usual intervention services due to ethical concerns with intervention withdrawal.

Impact

This will be the first study, to our knowledge, comparing efficacy of early physical therapy with dose-matched interventions and well-defined key principles. The outcomes will inform selection of key principle of intervention in this population.

Introduction

Cerebral palsy is the most common cause of infant-onset physical disability, affecting 17 million people worldwide.^1,2 Although methods of early detection are reducing the age of diagnosis, the dearth of high-quality intervention evidence for infants <2 years of age is staggering and impedes our ability to make evidence-based recommendations.^3–6 Physical therapy for infants often focuses on developmental progression of motor skills, underlying impairments, and prevention of secondary conditions. As such, intervention is grounded heavily within the body/structure/function domain of the International Classification of Function (ICF).^5,7 Recent research from developmental science,^8–10 our team, and others stress the importance of variable, self-directed motor and cognitive exploration, which focuses primarily on the activities/participation domain and environmental factors of the ICF.^11–18

Interventions focused on strength training, range of motion, postural control, and motor repetition have been used to advance motor skills for decades.^19–22 Neurodevelopmental approaches to date rely heavily on physical guidance, facilitation of normal movement patterns/posture, and developmental positioning to promote progression of motor control.³ In a recent study using fidelity measures to describe components of early intervention,²³ the usual care group or community-based services received a higher percentage of their intervention time focused on providing motor assistance to achieve correct alignment of movement patterns and activities addressing the body structure and function domain of the ICF. These aspects of usual care are incorporated into 1 of the study arms in the current study as the Movement, Orientation, Repetition, Exercise-Physical Therapy (MORE-PT).

The development and evaluation of interventions that target motor-based problem solving as a way of advancing both motor and cognitive skills is recent.^12,13,24 The focus in this strategy is on the activities/participation and environmental domains of the ICF. Using a top-down approach that embeds cognitive and social/motivation aspects of the environment, this approach differs from the MORE-PT approach in key ingredients of the intervention.¹⁶ Thus, the START-Play arm of this study uses an intervention that includes variable and self-initiated movement, environmental adaptation, and problem-solving of motor-based activities within a participatory mileau.¹⁶

Few studies have directly compared the efficacy of 2 pediatric physical therapist dose-matched interventions, each based on differing aspects of the ICF model, while monitoring fidelity of the interventions.^3,5,25 Importantly, this study clearly differentiates and defines both the theoretical basis and key components of each intervention approach, which allows use of fidelity measures as an objective indicator of both adherence to and differentiation of the approaches. Such rigor avoids type III error (misleading results because intended intervention was not actually delivered) in a multi-site study comparing interventions.

Direct comparisons of differing physical therapist interventions for infants with cerebral palsy are critical to determine which approaches are more effective overall or in specific populations, at specific ages, or for infants with specific early assessment findings.^5,26,27 To our knowledge, no study has directly compared 2 dose-matched physical therapist interventions grounded in differing theoretical principles for infants and toddlers with cerebral palsy.

The primary purpose of this project is to compare the efficacy of 2 physical therapist interventions based on different theoretical models and domains of the ICF (Tab. 1).²³

Table 1.

The Primary and Secondary Hypothesis^a

Primary Hypothesis

Compared with dose-matched control group (MORE-PT), START-Play group will demonstrate greater improvements in sitting skill as measured by GMFM Sitting Scale at end of intervention and 3, 6, and 9 mo later (at 3, 6, 9, 12 mo post baseline)

Secondary hypotheses

Compared with dose-matched control group (MORE-PT), START-Play group will demonstrate greater improvements in overall gross motor skills as measured by Bayley-IV at end of intervention and at 3, 6, and 9 mo later (at 3, 6, 9, 12 mo post baseline)

Compared with dose-matched control group (MORE-PT), START-Play group will demonstrate greater improvements in cognitive skills on Bayley-IV at end of intervention and 3, 6, and 9 mo later (at 3, 6, 9, 12 mo post baseline)

^a Bayley-IV = Bayley Scales of Infant and Toddler Development-IV; GMFM = Gross Motor Function Measure-88; MORE-PT = Movement, Orientation, Repetition, Exercise Physical Therapy; START-Play = Sitting Together and Reaching To Play.

Methods

This 2-arm randomized clinical superiority trial will include 3 enrollment sites in Los Angeles, California; Omaha, Nebraska; and Seattle, Washington; a data storage and analytic site in Lincoln, Nebraska; and an intervention oversight site in Pennsylvania. The study design is consistent with the CONSORT and SPIRIT statements for clinical trials. All assessment and intervention visits are conducted in the family’s choice of setting, home, or clinic by researchers or clinical therapists trained to participate in this study. Outcomes will be assessed at 5 time points: baseline, end of intervention (3 months post baseline), and 3 follow-up visits at 6, 9, and 12 months post baseline (3, 6, and 9 months post intervention).

Participants

We aim to recruit 150 infants 8 to 24 months of age (corrected for prematurity as applicable) at high risk of cerebral palsy. To identify a homogenous population of infants with a diagnosis of or at high-risk for cerebral palsy, the inclusion criteria are based on the early detection of cerebral palsy and clinical factors used by physicians (Tab. 2).^2,4 A combination of at least 2 of the following was considered a high risk of cerebral palsy: an atypical neurological exam, a motor delay of more than 1.5 SD below the mean, or neuroimaging associated with cerebral palsy. In addition to the inclusion criteria, enrollment will occur at a common motor skill level—emerging sitting (Tab. 2)—to ensure that participating children are working on similar motor skills at enrollment. Exclusion criteria were selected to exclude participation of children whose diagnosis or clinical presentation is not consistent with cerebral palsy and factors that may limit participation in intervention (Tab. 3).

Table 2.

Participant Inclusion and Exclusion Criteria^a

Category	Description
Inclusion criteria
Age at enrollment	Participants must be between ages 8–24 mo (corrected for prematurity as applicable) at study entry
Family language	At least 1 guardian must be able to provide consent in English or Spanish to ensure adequate team support to meet human participants guidelines
Gross motor delay	Bayley-IV gross motor subscale score >1.5 SD below mean
Consistent motor skill level: demonstrates readiness to sit and interact with objects	Ability to hold head upright at least 3 s with trunk supported at axilla Ability to visually focus on toy or person’s face at least 3 s
	Demonstration of spontaneous movement of arms
	Ability to sit using arms for support for 3 s
	Inability to get in and out of sitting position independently
High risk of cerebral palsy, as defined by medical record	2 of following documented in medical record:  • Abnormal Prechtl General Movements (absent, fidgety) at 12–14 wk of age  • HINE score <63 at 6 mo corrected age or older  • MRI abnormalities including periventricular leukomalacia/periventricular hemorrhagic infarcts involving the posterior limb of the internal capsule/cerebellum, cortical/deep gray matter lesions (basal ganglia or thalamus lesions, watershed injury, multicystic encephalomalacia), or stroke.
Baseline screening	Baseline screening (in addition to the motor delay criteria, at least 1 of the following):
	• HINE score < 63
Exclusion criteria
Medical complications that severely limit participation in intervention and assessments	• Severe visual impairment: unable to track a toy • Congenital/orthopedic anomalies that limit sitting or reaching
	• Uncontrolled seizures at any time throughout the study: seizures continue for >2 mo despite medical management
Family plans	• Family plans to move from local area before 1 y from enrollment
	• Child has major surgery planned that may affect physical performance
Primary diagnosis known to have medical or developmental course different from cerebral palsy	• Achondroplasia • Autism spectrum disorder as primary diagnosis
	• Down syndrome as primary diagnosis
	• Fetal alcohol spectrum disorder
	• Fragile X
	• Prader Willi
	• Progressive disorder such as muscular dystrophy or leukodystrophy

^a HINE = Hammersmith Infant Neurological Exam; MRI = Magnetic Resonance Imaging.

Table 3.

Intervention Key Principles^a

Key Principles
Intervention Area	START-Play	MORE-PT
Focus of activities	Cognitive constructs embedded in motor problem-solving	Strengthening, stretching, advancement of motor skills
Practice	Variable exploration with self-initiated movement	Practice within normal movement patterns, with physical assistance as needed
Information sharing with caregiver	Explanation of cognitive and motor interaction	Explanation of exercise principles
Activities for child and caregiver	Brainstorm ideas for next level of challenge in motor blended with cognitive	Prescribe specific exercises to advance motor skill
Parent participation	Invites parent to assist, create, or lead activities	Invites parent to repeat demonstrated activities

^a Trained interventionists were required to include these key principles in all intervention sessions and videotaped sessions were scored for adherence. MORE-PT = Movement, Orientation, Repetition, Exercise Physical Therapy; START-Play = Sitting Together and Reaching To Play.

The research team will complete a neurological exam and motor assessment that can be used to meet the inclusion criteria. The Hammersmith Infant Neurological Exam (HINE) is a standardized and scoreable neurological exam that is highly recommended as part of a comprehensive clinical profile used to diagnose cerebral palsy.^4,28,29 The HINE will be completed with all infants at baseline and will be used for inclusion determination as well as stratification. Given the frequent delay in early diagnosis of cerebral palsy, it is likely that some infants enrolled in the study will not be clinically diagnosed during their 12 months of participation in the study. The HINE will be completed and videotaped a second time for all infants at the visit closest but prior to 2 years of age. Infants who score <63 will be considered as having an abnormal HINE score. Using a combination of medical records, HINE score, and/or a physician with expertise in neurodevelopmental disabilities review of the HINE video recording, all infants will be classified as having cerebral palsy or not by 2 years of age, or by the end of their study participation.²⁹ This will allow for post-hoc analysis of the efficacy of the intervention between those who have a diagnosis of cerebral palsy at 2 years and those who were enrolled as high risk but do not meet the criteria for a cerebral palsy diagnosis at age 2 years.

Recruitment

In a subsample of infants with signs of cerebral palsy from the prior START-Play trial, effect sizes (Hedges g) for primary outcome measures ranged from 0.41 to 0.69 for sitting (Gross Motor Function Measure [GMFM]), from 0.36 to 0.79 for cognition (Bayley Scales of Infant Development-IV [Bayley-IV]), and 0.25 to 0.61 for gross motor (Bayley-IV). Post-intervention attrition in the previous trial was 7% of enrolled infants. A Monte Carlo simulation-based power analysis with 500 replications determined that a sample size of 150, accounting for 7% drop-out, would provide >80% power for all primary and secondary aim comparisons except the 3-month comparison for Bayley-IV gross motor.

Each of the 3 enrollment sites will recruit n = 50 infants for a total of 150. Infants will be identified through medical record searches at local hospitals and through community partners who will share study information with families. Recruitment and consenting will be completed by the study coordinator or site investigator using materials in English and/or Spanish.

Following baseline assessment, infants will be randomized to receive the START-Play or MORE-PT intervention using a blocked randomization design with randomly selected block sizes of 4 and 6. To ensure that the intervention groups are balanced, randomization will be stratified by enrollment site and baseline neurological impairment (HINE ≥40 vs <40). The study statistician created the randomization matrix prior to any enrollment. The randomization module for the study REDCap database allows each site to randomize the participant after entering the baseline HINE score. Parents, interventionists, and study coordinators will be aware of the group assignment to implement the study intervention.

Description of Intervention

Both interventions will be provided by licensed pediatric physical therapists (training described below) in the home or a clinic at the parent’s discretion. Sessions are scheduled twice per week for 24 sessions of 45 to 60 minutes each. Sessions missed due to scheduling or illness will be rescheduled as possible. Parents are provided a notebook at the start of the study that can be used to write notes on the therapy session, suggested activities to work on between sessions, or questions that arise between sessions. However, this is at the parents’ and therapists’ discretion. No other handouts or web-based materials are specifically provided to the family participating. Each interventionist has a set of toys that they can bring to the family home and lend the family to allow for practice of activities between sessions. The set of toys includes a shape sorter, musical instruments, blocks, nesting balls, a small folding bench, and a few foam floor mats.

Although all infants in this study will receive study intervention, it is short term and experimental, and thus it was deemed unethical to withhold usual care intervention. Therefore, all groups will receive a study intervention as well as usual care interventions.

Description of MORE-PT

The MORE-PT intervention was defined based on direct observation and parent report of usual care intervention sessions in a previous multi-site trial.^16,23 The intervention focus is on advancing motor skills through repetition of movement patterns or exercises that increase strength, encourage practice with postural orientation, and prevent impairments such as range of motion limitations, all consistent with the body structure and function domain of the ICF. Consistent with our preliminary data, MORE-PT will train the infant, using physical assistance, to accomplish and practice functional tasks of transitioning between positions using “normal” movement patterns to prevent the development of atypical movement strategies or secondary impairments. Passive range of motion will be used to stretch tight muscles and prepare the child for functional movements. Toys will be used to motivate the child to move in optimal biomechanical alignment and encourage repetition for strengthening and postural control. The parent will be encouraged to be present and interact with the child and therapist. Discussions of the parents’ goals, any general developmental questions, and how to incorporate handling to support “normal” movement will be included. Parents will be coached to practice these motor activities daily with existing equipment accessible to the family.

Description of START-Play

The START-Play intervention focuses on motor based problem-solving through interaction with objects, early sitting, and reaching skills embedded within play.^16,30 Although sitting and reaching are the primary motor tasks encouraged, the experimental intervention emphasizes self-initiated, goal-directed movements to bolster orienting and attending to objects and environmental supports, consistent with the activities/participation and environment components of the ICF.^5,8 In addition, cognitive principles, specifically object affordances, means end reasoning, object permanence, and joint attention, are scaffolded through manipulation of objects and social interaction during play as foundational aspects of participation. The resulting key components of the intervention are described in Table 2. The specific intervention depends on the skill of each child. The focus is not on training the child to practice “normal” movement patterns. Therapists work with parents to learn, discover, and create problem-solving motor challenges for their child, to link small motor changes to cognitive skills, and to increase opportunities for the child to blend motor and cognitive constructs. Leading with a cognitive task increases a child’s drive to initiate movement, which the parents can build on to advance cognitive and motor skills while promoting greater participation with the environment. A full intervention and training manual already exists from previous research.^16,30

Training and Fidelity

All interventionists are provided with an intervention manual and literature relevant to the theoretical background of the intervention and participate in small group trainings for the intervention they are assigned to provide. Initial didactic training was completed virtually due to the ongoing COVID-19 pandemic limiting travel. In addition, an onsite investigator who was part of the prior START-Play study and familiar with the development of the MORE-PT intervention was available at all 3 enrollment sites and coordinated in-person practice sessions, with the lead trainer initially via video conferencing or watching videotaped sessions. Additional therapists will be added using a combination of virtual and in-person training based on travel restrictions and public health concerns. Regardless of the initial training format, all interventionists are required to demonstrate the ability to meet the required fidelity standards (described below) during a practice session before providing independent intervention. In addition, yearly in-person or video conference refresher trainings will be completed for all interventionists. These refresher sessions also serve as opportunities for the interventionists across sites to brainstorm about treatment strategies that support the key principles of their assigned intervention. Only the site PIs were permitted to provide intervention in both groups, and only if they demonstrated fidelity in both interventions. This strict standard is in place to prevent cross-over between interventions.

Our fidelity approach will determine the degree to which all elements of the START-Play and MORE-PT interventions are delivered across the 2 treatment conditions.^23,31,32 As in our previous work, the fidelity measure will be used throughout the study to assess adherence to each intervention (see Table 3 for specific adherence items). Interventionists will complete an adherence check list at the end of each session, and 2 intervention sessions will be videotaped and behaviorally coded for each child participant to (1) determine if the key principles of the intervention were the focus of each session, and (2) quantify the duration of time focused on either the START-Play or MORE-PT intervention. Feedback is provided to each interventionist on their videotaped session to ensure ongoing high adherence to the interventions. In addition to measuring adherence, the coded video tapes of the intervention sessions will be used to ensure intervention activities differ between the intervention groups (program differentiation). If the therapist does not meet the adherence thresholds, they will be asked to meet with their site PI to review the session and record an additional video of intervention to determine if they have improved their adherence. Refresher trainings will be provided as needed for therapists who have more than 2 videos in which their adherence does not meet the threshold. Descriptive statistics will be calculated to evaluate adherence levels relative to a priori established criteria. Statistical significance of intervention group differences will be tested to evaluate program differentiation.

The parent’s adherence will be assessed through a monthly survey sent to the parent throughout the study asking the parent to indicate which intervention-related activities they did with their child in the last 24 hours. These data will be used to describe parent-reported utilization of the intervention principles between sessions and following the 3-month intervention period, and to compare between groups the types of activities that parents complete with their infants.

Outcome Measures

Gross Motor Function Measure

The GMFM is the international standard for quantifying change in gross motor function of children with neuromotor disorders (inter-rater reliability, r = 0.97).^33,34 The GMFM assesses the child’s motor control and ability to complete items that are often difficult for children with cerebral palsy in 5 dimensions: (1) lying and rolling; (2) sitting; (3) crawling and kneeling; (4) standing; and (5) walking, running, and jumping. Given the focus of this intervention on advancing sitting, the GMFM-88 sitting scale will be completed to provide a single score to specifically evaluate the change in sitting skill over time within and between groups. The GMFM-66-IS (Item Sets)³⁴ is a validated modification of the GMFM that reduces the number of items administered based on the child’s ability to sit, stand, or walk.³⁴ The GMFM-66 item will be competed to fully describe global change in motor skill.

Bayley Scales of Infant and Toddler Development, 4th Edition

The Bayley-IV is a widely used, norm-referenced test designed to assess multiple developmental domains in infants 16 days to 42 months of age.³⁵ Because interventions may have global effects on all areas of development, all scales—including the cognitive, expressive, and receptive communication—and gross and fine motor subtests will be administered.³⁶

Assessment of Problem-Solving in Play

The Assessment of Problem-Solving in Play is a play-based observational measure adapted from the Early Problem-Solving Indicator³⁷ and validated to evaluate problem-solving skills in children with motor impairment.³⁸ It measures infants’ problem-solving behaviors while interacting with 3 specific toys each for 2 minutes while in a sitting position. The assessor provides the minimum amount of support the child needs to sit upright and freely use his/her arms. The frequency of key elements of problem-solving skills (look, simple explore, complex explore, function, and solution) are coded for each toy from a video of the assessment. Frequencies are weighted to create a total score considering the element difficulty and frequency.³⁸

Parent-Child Interaction (PCI)

Intervention provided with parent engagement is assumed to change parent behavior over time, but this is rarely measured. PCI will be evaluated under 2 conditions to evaluate the efficacy of the intervention to change PCIs during play or the ability of the parent to spontaneously apply the intervention principles. Custom behavioral coding of a 5-minute free play session is used to assess how the families interact with their child during a typical play session.^18,39 A standard set of toys or the child’s own toys can be used. A second play assessment focuses on ways a parent teaches their child to play with a novel toy that the parent reports the child has not played with previously. The parent will be asked to “do whatever you would normally do when introducing a new toy to your baby.” The degree to which the family uses intervention principles embedded in their play will be coded and used to assess the effect of the intervention on PCIs.

Demographic, Medical, and Environmental Factors

We will collect demographic data, including sex, race, ethnicity, parent education, and income, at the baseline and final study visit to fully describe the sample and identify potential correlates of attrition. Variables that are correlated with dropout will be included in the outcome analyses as auxiliary variables to help meet the assumption that data are missing at random. In addition, we will examine whether the efficacy of the intervention varies as a function of these characteristics (see Data Analysis section).

Parents will be asked to complete questionnaires on changes in medical conditions, hospitalizations, onset of seizures, as well as the child’s current community-based therapy services and access to toys and play space at each assessment.⁴⁰ Parent response to these questions will allow us to confirm ongoing study eligibility, track adverse events as required by the Data Safety Monitoring Board, and complete post-hoc analysis on the access to therapy services for children enrolled in this study.

Data Collection and Participant Retention

Outcomes will be measured in the home (unless the parent requests the clinic) by an assessor blinded to group assignment. Parents are reminded by a coordinator and again by the assessor to not disclose who their therapist is or what intervention group they are assigned to. If the assessor is unblinded, they will complete the assessments and report the unblinding to the site PI. A different assessor will watch the videotaped assessment and score it independently to ensure unbiased scoring. Assessments are typically completed within 1 home visit lasting 2 to 3 hours, with parents present throughout. If necessary, a second visit will be scheduled to complete assessments. If completed within 1 week of the first visit, the data are merged to represent a single visit. If more than 1 week and less than 2 weeks elapses between visits, all assessments are completed at the second visit. If any component of an assessment visit is not rescheduled within 2 weeks, it is recorded as missed.

To enhance study retention, parent-preferred assessment and intervention locations, flexible scheduling, and participant incentives are used. Assessment visits are also scheduled every 3 months with at least 1 contact by phone, text, or email between visits to improve communication with families and study retention.

Data Management

Videos will be stored in a secure cloud-based server that will only be available to research team members with appropriate human participant clearance. At least 20% of the data will be double scored to ensure continued reliability (demonstrating at least 85% agreement or intra class correlation coefficients >0.80). Reliability will be measured on an ongoing basis to allow for retraining as needed. If multiple visits by the same assessor are determined to be below the reliability standard, the video of each of that assessor’s visits will be rescored by an assessor with high reliability. The data management site will be responsible for processing, cleaning, and filing raw data from video coding in addition to data from the REDCap electronic database.

Data Analysis

Consistent with the CONSORT statement and SPIRIT guidelines, the analysis plan for the primary outcomes (efficacy of the intervention) is described here along with several planned subgroup and adjusted analysis. Three general linear piecewise mixed models will be estimated to test intervention effects on children’s sitting (GMFM-88 sitting subscale; primary hypothesis) and gross motor and cognitive skills (Bayley-IV; secondary hypotheses). Each model will include fixed effects of time (with separate slopes for the intervention and post-intervention phases), intervention group, time by intervention group, randomization stratum, and baseline corrected age, as well as random residual, intercept, and slope effects. Time points will be treated as individually varying to account for variation in measurement intervals. The target effects will be estimated as the intervention group difference in the change in outcomes from baseline to 3, 6, 9, and 12 months post baseline. Statistical significance will be assessed using α = .05 (2-tailed), and clinical significance will be assessed via Hedges g.^41,42 Analyses will adhere to an intent-to-treat perspective such that group comparisons will be based on infants’ initial intervention assignment. Full information maximum likelihood estimation will ensure that all eligible infants are included in the analyses, regardless of dropout.

Exploratory analyses will be performed to determine whether sex, PCI, and severity of motor impairment analyses will follow from the models used to test the primary and secondary hypotheses but with an additional main effect (the effect of the exploratory variable) and additional 2-way (time by exploratory variable, intervention group by exploratory variable) and 3-way (time by intervention group by exploratory variable) interactions. The 3-way interaction tests whether intervention efficacy varies as a function of the exploratory variable. Simple 2-way interactions test intervention efficacy within each subgroup represented by a given exploratory variables (eg, intervention efficacy for boys). In addition, in recognition of the impacts of social determinants of health (including systemic racism and access to family resources) on child development, using the same analytic approach we will explore associations between response to intervention and demographic variables, including home environment characteristics, socioeconomic status, race, and ethnicity. The sites included in this study have a track record of enrolling diverse samples, and the inclusion of Spanish-speaking assessors, interventionists, and coordinators on the study teams will increase the likelihood of having a diverse sample. Although the study is not powered to evaluate the impact of social determinants of health on the outcomes, the planned comparisons will ensure these factors are considered and controlled for if needed in the primary analysis.

Monitoring and Auditing

A Data Safety Monitoring Board (DSMB) consisting of 3 scientists (2 physical therapists and 1 physician) who are not otherwise related to the study will monitor safety and adverse event reporting as mandated for all multi-center clinical trials funded by the National Institutes of Health. The frequency of adverse events, reported in the parent questionnaires at each assessment visit, will be compared between groups every 6 months in preparation for the bi-annual DSMB meeting. Although there are no anticipated severe adverse events, a formal tracking system is in place in which any team member can report an adverse event regardless of whether it is related to the study. For example, hospitalization for a respiratory virus 1 month after a child’s last interaction with a research team member is unlikely to be related to the research. However, this is tracked as an adverse event to allow for analysis of the rate of adverse events between groups. A preliminary analysis of the outcomes of the study will be completed after 76 infants are enrolled and have completed the first 3 study assessment visits. The sole purpose of this mid-enrollment analysis is to ensure if it is ethical to continue the study; that is, if one intervention is statistically far superior to the other, it would be unethical to continue to randomize to the less effective intervention. We do not anticipate any adjustments to the study design based on this analysis, but it is required by the funding agency to monitor these metrics.

Ethics

Approval and Consent

This study has been reviewed and approved by the Human Subjects Research board at the University of Southern California. A reliance agreement is in place so all procedures, recruitment materials, and consent forms are the same except for a difference in contact information among sites.

Role of the Funding Source

The funding source has no influence on the design or consult of this study other than the requirement to have a DSMB, which is not related to the study team or funding agency. We anticipate the funding agency will not have a role in results, discussion, or dissemination.

Dissemination

We plan to disseminate the findings in a combination of pediatric and rehabilitation journals, at conferences, and to stakeholders, including parent participants and providers. Significant protocol modifications will be submitted as an erratum to this protocol manuscript and will be included in the outcome papers. The full dataset will be deidentified and available on request.

Discussion

James Heckman, Nobel Laureate in Economics, said, “A comprehensive approach to early childhood development can deliver a 13% per year return on investment.”⁴³ For years, pediatric therapists and early intervention providers have struggled with understanding the necessary components of this comprehensive approach. Interventions including neurodevelopmental treatment, and craniosacral therapy, myofascial release, compression suits, Vojta, and Cuevas Medek Exercise continue to be available despite a lack of theoretical support or empirical evidence on efficacy.^5,44 It is crucial to the field of physical therapy that we evaluate the components of interventions that are effective and may lead to improved clinical decision-making as well as understanding of the mechanisms of action.

This clinical trial has several potential limitations. First, it is possible that the intervention fidelity will be limited, and the 2 interventions will become more similar over time. To address this, fidelity will be measured on an ongoing basis with feedback provided to interventionists, and all interventionists will be required to participate in yearly retraining. Second, this study is designed to evaluate efficacy of intervention for infants with cerebral palsy; however, there is considerable heterogeneity within groups because infants with cerebral palsy are variable in their original brain injury and in their developmental progress. Accordingly, infants will be stratified by their baseline HINE score to reduce within-group variability, which will increase precision of the intervention effect. Third, the interventions in this study are based on principles that could be broadly applied to multiple skills levels. However, to keep the intervention more homogeneous, we opted to enroll children who were all working on sitting skill development. Thus, the results of the study may not be generalizable to all skill levels. Future research will be needed to confirm the potential limitation. Finally, the GMFM-IS, a secondary measure, was initially validated on a sample of children with a mean age of 7 years, with a range from 1 year 4 months to 13 years 8 months. Thus, it is possible that the GMFM-IS will not be sensitive to change in this sample who is enrolled between 8 and 24 months and complete the study between 20 and 36 months. The use of multiple measures, GMFM-88 sitting, GMFM-IS, and Bayley Gross Motor scales will allow for alternative measures to be used in post-hoc analysis.

The innovative approach used in this study to manipulate the key ingredients of the intervention, while matching for dose, will allow for direct comparison of the efficacy of intervention based on the body structure and function domain versus the activity and participation domains of the ICF. Secondary outcomes, which explore parental engagement, type of infant brain injury, and other environmental issues, will provide additional, important insights regarding the effect of different ingredients of therapy on multiple facets of the ICF, which may assist early interventionists with clinical decision-making. Finally, it is only with comparison of interventions that we will truly be able to determine if the content of an intervention impacts efficacy on overall development and which aspects of development may be amenable to change.

Author Contributions

Concept/idea/research design: S.C. Dusing, R.T. Harbourne, N.A. Koziol, S. Jensen-Willett, S. Westcott-McCoy

Writing: S.C. Dusing, R.T. Harbourne, L-Y. Hsu, N.A. Koziol, K. Kretch, B. Sargent, S. Jensen-Willett, S. Westcott-McCoy, D.L. Vanderbilt

Data collection: S.C. Dusing, L-Y. Hsu, K. Kretch, B. Sargent, S. Jensen-Willett

Data analysis: S.C. Dusing, L-Y. Hsu, N.A. Koziol

Project management: S.C. Dusing, R.T. Harbourne, N.A. Koziol, K. Kretch, S. Jensen-Willett

Fund procurement: S.C. Dusing, R.T. Harbourne

Providing participants: S.C. Dusing, S. Jensen-Willett, D.L. Vanderbilt

Providing facilities/equipment: S.C. Dusing, S. Jensen-Willett

Providing institutional liaisons: S.C. Dusing, S. Jensen-Willett

Consultation (including review of manuscript before submitting): L-Y. Hsu, N.A. Koziol, S. Jensen-Willett, S. Westcott-McCoy, D.L. Vanderbilt

Funding

This study was funded by a grant from the National Institutes of Health, Eunice Kennedy Shriver Institute of Child Health and Human Development (5R01HD101900–02).

Ethics Approval

This study has been reviewed and approved by the Human Subjects Research Board at the University of Southern California.

Clinical Trial Registration

This protocol study was registered at ClinicalTrials.gov (NCT04230278).

Disclosures

The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.