Daily and weekly rehabilitation delivery for young children with gross motor delay: a randomized clinical trial protocol (The DRIVE study)

Abstract

Purpose:

The proposed project tests the principle that frequency of rehabilitation is an important regulator of therapeutic response in infants.

Methods:

We will randomize 75 infants with cerebral palsy, 6 – 24 months of age and GMFCS III-V (higher severity), to determine the short term and long-term effects of 3 dosing protocols consisting of an identical number of 2 hour sessions of the same motor learning based therapy applied over a different total number of calendar weeks.

Results and Conclusions:

The results will inform clinicians, families, and scientists about dosing and will provide needed recommendations for frequency of rehabilitation in order to optimize motor function and development of young children with cerebral palsy.

INTRODUCTION AND PURPOSE

Cerebral palsy (CP), the most common and serious childhood physical disability, is associated with multiple functional impairments, poor health, and economic burden over the life course for an estimated 15 million people worldwide.¹ Dosing parameters for rehabilitation of participants with CP vary with frequency ranging from a low of 1 hour per month to a high of 6 hours per day for 21 days.^2,3 Infants with more severe functional limitations (classified as higher levels on the Gross Motor Function Classification System (GMFCS)) are often excluded from daily intervention programs due to an unresolved debate about the relative benefits of providing a concentrated or intermediate burst of treatment versus potential risks and fatigue.^4,5 This research protocol is informed by principles of motor learning and infant development^6–9 as well as preliminary results from a case series (n=2) of daily intervention for infants with CP GMFCS V.¹⁰

Determining the appropriate parameters for a dose of treatment is critical for achieving an optimal therapeutic response. In order to capture the inherent plasticity of a developing nervous system, especially in infants with CP or CP risk, we propose that it is especially important to understand the impact of different frequencies of dose on functional outcomes in pediatrics. In the proposed project, defining the dose requires defining the frequency (number of sessions per week and the number of weeks), intensity (how intensely the infant or child works), and time (how many total hours) of rehabilitation treatment.^11–13 A characteristic of infant and child development for children developing typically or atypically is a pattern characterized by spurts in physical growth and in motor and cognitive development.^8,14,15 A concentrated dose of rehabilitation designed to mimic this pattern may be more effective at producing long-term change than a lengthier and slower pattern of treatment. These spurt-like changes also require a careful long-term follow-up to reliably measure developmental changes associated with the treatment dosing.

Determining optimal frequency of treatment has implications for shaping the future of pediatric rehabilitation. There are variations in the number of hours per week of treatment in current rehabilitation programs for infants and children with CP, suggesting clinical uncertainty.^16,17 Usual weekly therapy at 1 – 2 hours per week for 6 months or longer is the most commonly implemented frequency of dose for infants with CP 6 – 24 months of age. However, the decision about frequency is often made based on clinical reasoning and scheduling, not on principles of rehabilitation, child development, or evidence from strongly designed randomized controlled trials (RCTs). The proposed study will fill this gap by directly comparing the effects of 3 frequency levels of therapy – concentrated daily, intermediate, and weekly in children with CP 6 – 24 months of corrected age at the initiation of treatment and following these participants for 2 years.

Concentrated daily programs are typically several hours per day for several days and sequential weeks. Concentrated daily programs have used specific types of treatment, such as constraint induced movement therapy and treadmill training, and have demonstrated rapid improvements in motor development.^3,5,18–22 There are 2 issues of importance: (1) none of the concentrated daily programs were compared to an equal time of another treatment, making it impossible to differentiate which of the following factors produced the therapeutic response: type of treatment, total amount of treatment, or frequency of treatment^23,24 and (2) most of the concentrated daily programs targeted for children with CP have focused on older but not younger children. Our program will target infants with CP in higher GMFCS levels at a developmental time when their rate of gross motor change is the greatest given the natural history of CP.^25–27 Children with CP who do not crawl or walk independently (i.e. higher GMFCS levels) are typically excluded from concentrated daily treatment studies because their motor skills are not adequate for the type of treatment or outcome measures and there is an assumption that they cannot tolerate a concentrated burst of therapy. Our project individualizes therapy so it is appropriate to age and skill level of each child; uses appropriate outcome measures that can be measured in our age ranges and severity levels; and challenges the assumption that children with more severe functional levels of CP cannot tolerate concentrated daily or intermediate therapy.

Cerebral palsy comprises 20 −25% of all childhood motor disability²⁸ and rehabilitation is a commonly prescribed service for children with CP.^17,29,30 A therapist may spend a minimum of 20–25% of her time providing therapy to children with CP. Efficacy has been established for motor learning approaches to rehabilitation.^7,31 The dosing of therapy services for children with CP has been identified as a national priority.^11,23,32 The next question is to determine clear dosing parameters. Specifically, establishing frequency parameters for rehabilitation is critical for informing clinical decision-making, health policy, and guidelines for reimbursement.^10,10,33

Specific aim 1:

To test the hypothesis that infants participating in more concentrated therapy will make greater short-term and long-term gains in motor function. We will compare short-term (6 months following initiation of therapy) and long-term (1, 1.5, and 2 years following the initiation of therapy) outcomes in terms of gross motor function, domains of development, goal attainment scaling (GAS), and spontaneous play. We predict that most infants participating in more concentrated therapy will make greater short-term and long-term gains in motor function, with GMFCS Level 1 greater than Level 2 and greater than Level 3. Our primary outcome measure is the Gross Motor Function Measure (GMFM-88). Our secondary outcomes include: The Bayley Scales of Infant Development 3rd edition (Bayley-III), GAS, Infant-Preschool Play Assessment Scale (I-PAS), and the Family-Professional Partnership Scale survey to measure parent satisfaction.

Specific aim 2:

To test the hypothesis that factors including age and severity will be correlated with the quality of outcome for children with CP or risk for CP, and that these factors will have similar effects across the 3 dosing protocols. We expect that children who are younger, have higher levels of severity, and have higher levels of parent engagement will make greater relative gains, especially for measures of motor function and for the development scales.

METHODS

Design

The study design is a prospective randomized controlled trial (RCT) in which 6 – 24 month-old infants with CP or risk of CP (n = 75) who meet enrollment criteria will be randomly assigned to 1 of 3 dosing groups. Infants in all 3 groups will receive the same total hours of outpatient therapy (40 hours) with systematic documentation of therapy. Key evidence-based elements of this therapy for all 3 groups include goal-directed activities, functional training, shaping, and massed practice based on principles of motor learning, and use of play and everyday activities. All 3 groups include monthly monitoring and a home program by a therapist masked to dosing group. The 3 therapy groups differ systematically, however, in their frequency of outpatient therapy delivery: Level 1: High - uses a concentrated daily therapy schedule for 2 hours per weekday for 4 weeks; Level 2: Intermediate - uses an intermediate therapy schedule for 2 hours on 3 weekdays for 6.6 weeks; and Level 3: Low - uses a usual weekly therapy schedule for 2 hours, 1 day per week, for 20 weeks. This project began enrollment April 25, 2016 and will end April 2021.

Dose: Frequency, Intensity and Timing

The treatment phase of this study is 5 months for a total of 40 hours of one-on-one therapy for each group. Level 1 daily therapy is 2 hours of therapy per day for 20 consecutive weekdays. Level 2 intermediate therapy is 2 hours of therapy per day 3 days per week for 6.6 weeks. Level 3 usual weekly therapy is 2 hours of therapy one day per week for 20 weeks. All therapy sessions will occur at a large comprehensive pediatric medical center with satellite sites.

Frequency:

The number of sessions per week of therapy is different for each group. Level 1 (high) will receive concentrated daily therapy, 5 days per week for 4 weeks; Level 2 (intermediate) will receive intermediate therapy 3 days per week for 6.6 weeks; and Level 3 (low) will receive usual weekly therapy, 1 day per week for 5 months.

Intensity:

The time of each session of therapy for all groups is 2 hours and infers equal intensities among groups because they have an equal amount of time to practice motor skill acquisition during each session of therapy. Since intensity is a measure of child effort during therapy, we will measure intensity by a behavioral check list.

Heart Rate:

A measure of cardiovascular fitness, heart rate measurements, are used in infant populations to determine learning, autonomic nervous system function, and stress in a population that is preverbal and has limited voluntary movement control for communication.^34–36 In this study, heart rate will be monitored with a pulse oximeter before, during, and after treatment during every fifth treatment session as a measure of fatigue and tolerance of activity since higher dosing protocols may have potential risks and produce fatigue in infants.

Behaviors Check List:

A description of negative and positive behaviors and affect will be collected via survey (the child behaviors check list)³⁷ from the parents at the initial, midterm, and final treatment session by therapists. This survey collects information on child behaviors that are indicative of fatigue and poor actively tolerance in infants.

Time:

The total amount of time of therapy for all groups is 40 hours. The total time of treatment was carefully chosen. Previous work using daily therapy to train upper extremity skills for adults and children with stroke have demonstrated efficacy from 30 – 90 total hours, with 90 hours providing more benefits and no harm.^21,38–41 We use the lower end of time in these studies as a guide for the minimum and range of time in the proposed project because of the differences in ability between upper extremity training that uses only small muscle groups and gross motor skill training for infants that uses primarily large muscle groups. It is typical for children in this age range to receive continual therapy for 6 months to 1 year.^17,29

Study population

Inclusion and exclusion criteria

Inclusion criteria: 1) age 6 – 24 months at initiation of treatment, 2) a diagnosis or risk of CP at GMFCS levels III, IV and V, and 3) the ability to tolerate a 2-hour therapy session based on parent report and evaluating therapist. The age will be corrected for eligible children born preterm until they are 2 years of age, as is standard clinical and research practice.⁴³ Participants will be excluded for any of the following criteria: 1) uncontrollable seizures or a co-morbid condition that prevents full participation during treatment, 2) participation in another daily treatment program in the last 6 months, 3) auditory, or visual conditions that prevent full participation during treatment, or 4) progressive neurological disorders with no potential for improvement.

Participants who meet the inclusion and exclusion criteria will be recruited by study personnel or by research coordinator. The initial evaluation by a treating therapist will dually serve as the screen to determine if the participant can tolerate 2 hours therapy sessions and establish the family goal for the outcome of goal attainment scaling (GAS). Participants who complete and pass the screen will undergo baseline outcomes evaluation (OE) and simultaneously be randomized to a treatment group and stratified to assure group balance for gender and GMFCS level. Because withholding therapies during follow-up may be harmful to an infant developing atypically, additional rehabilitation services will be noted and recorded by parent survey, but not withheld. It is likely that all infants with CP GMFCS III, IV, and V (because of their age and severity level) will receive a usual and customary care.

Recruitment and enrollment

Participants are recruited through a comprehensive CP follow up clinic, therapy clinics, CP database, and word of mouth. The IRB approval: This study has been approved by the Institutional Review Board at Children’s Hospital (IRB15–01025). It was assigned a risk level 1—no greater than minimal risk on the Code of Federal Regulations (CFR), (45 CFR 46.404; 21 CFR 50.51).

Intervention

The type of treatment for all groups is functional and goal-directed training using principles of motor learning. Therapy sessions are one-on-one, meaning there is one therapist providing the therapy, and one participant. Successful treatments used for pediatric rehabilitation in research laboratories and clinical programs including functional training,^42–51 goal-directed training,^7,42,52–54 and promoting independence and active practice of motor skills have demonstrated improvements in strength, gross motor skills, functional range of movement, and motor function in children with CP or risk for CP. Their common framework is a treatment based on principles of motor learning and a family-centered approach to treatment. Our project uses this type of treatment because it has known efficacy and is the most common treatment type used clinically for children with CP.⁷ In the proposed project, principles of motor learning including repetition, task-specificity, active practice, generalization of skills, errors, structured practice, and developmentally appropriate feedback with sufficient time to practice are central to the treatment type we will provide.⁵⁵ Standardization of treatment will be assured by formal training of therapists and treatment fidelity checks.

Treatment Fidelity

Treatment fidelity is a key methodological component of rehabilitation research and refers to the extent to which the core components of the intervention are delivered in a reliable manner that is consistent with the goals underlying the research.^56–58 Treatment fidelity is important to increase the internal reliability and replicability of a study. Fidelity also strengthens the internal validity of a study to support the results as well as the external validity of a study to generalize findings to applied settings.^58–60 Key components of intervention fidelity include a well-designed framework, therapist training, monitoring of treatment delivery, and monitoring of treatment receipt (by participants).^56–58,61 Our study will include application of a treatment fidelity framework to strengthen reliability and validity to support study goals and outcomes. Treatment fidelity will specifically include a well-defined and systematic process, specific therapist training for motor learning principles and techniques as well as administration of outcome measures, establishment of inter-rater reliability with outcome measures, treatment session observation, critique of participant goals, and implementation of a therapist feedback process.

Physical Therapy Consultation (PTC)

All participants enrolled in the proposed project will have 1 hour per month of PTC for the treatment period. PTC will focus on home exercise programs (HEP) with parent coaching and will be completed by research physical therapists who are masked to group assignment. The monthly PTC allow for developmentally appropriate and regular monitoring of motor skill development and retention of participants.

Outcome Evaluations (OE)

OE will be performed by masked evaluators at screen, baseline (before treatment) 3 and 6 months to determine short term response, and at 1, 1.5, and 2 years to determine long term response. During baseline, 3, 6 month and 1year OE and PTC sessions will occur during the same visit. The OE is expected to take 2 – 2.5 hours inclusive of PTC. PTC and OE will occur in a pediatric laboratory or in the participant’s home. It is anticipated that half of the families enrolled in the proposed project will choose to come to the laboratory, and half of the families will choose to have us come to their home. All OEs will be done in the same location for an individual child, with the same transportable equipment, insuring consistency of testing. Physically separating the location of treatment/therapy and OE sites was purposeful allowing for unbiased evaluations, strengthening our study design; and allowing for more flexibility in the PTC/OE location for families, eliminating potential travel burden. All masked evaluators will receive training and reliability testing on each outcome measure. Masked evaluators will need to score at least 85% on inter and intra rater reliability measures as is standard for behavioral video coding. OE involves validity and reliability measures with sensitivity to detect change at 3 – 6 month intervals and include: Gross Motor Function Measure (GMFM-88), Goal Attainment Scaling, Bayley Scales of Infant Development-III, and the Multidimensional Assessment of Parental Satisfaction for Children with Special Needs, and the Infant-Preschool Play Assessment (Table 1).

Table1.

Outcome Evaluations

Outcome	Domain/Subtest	Measure	Age
Gross Motor Function Measure (GMFM −88) 26,65,70,71	Lying and rolling Sitting Crawling and kneeling Standing Walking, running and jumping	Total percentile, 0 – 100	5 months to 16 years
Bayley Scales of Infant Development-III (Bayley III)72,73	Cognitive Expressive language Receptive language Fine motor Gross motor Social-emotional Adaptive behavior	Index scores and subtest scaled scores	1 month to 3.5 years
Goal Attainment Scaling (GAS) 66,67,69,74–77	Motor skills and participation in community, school, and family activities	−2 to +2 (5 points)	Any age
Pediatric Evaluation and Disability Inventory (PEDI)70–74	Self-care Mobility Social function	Standard and scales scores	6 months – 7.5 years
Family-Professional Partnership Scale78	Child-focused relationships Family-focused relationships	0 – 5 point scale (average)	Families of children with disabilities

Gross Motor Function Measure (GMFM)-88 is the primary outcome measure. GMFM is an instrument to “evaluate change in gross motor function over time or with intervention” in children with CP from 5 months to 16 years.⁶² It has been used often in the field to determine functional motor change following intervention. The GMFM has 2 versions; GMFM-88 has 88 items and the GMFM-66 has −66 items which can be derived from the GMFM-88. There is evidence that the GMFM-66 may underestimate change in younger children and children classified at higher GMFCS levels (lower functioning) because the majority of the eliminated items come from the section on lying and rolling.^63–65 Thus, we will use the GMFM-88 as the primary outcomes in this study to detect larger changes for these young children with lower motor skill development.

Goal Attainment Scaling (GAS)⁶⁶ creates participant, family, and clinical anchors as the external criterion for improvement by establishing activity or participation goals that reflect what an individual, family, and clinician consider meaningful or relevant. GAS is an established clinical method for quantifying the achievement of goals and offers a forward thinking approach for calculation of meaningful change in participant specific activity or participation goals for pediatric participants.^66,67 We will use GAS to measure effectiveness of frequency in each group in the proposed study and as a comparison among groups. The use of GAS uniquely allows us to consider activity and participation levels of the ICF as important outcomes of improvements.⁶⁶ The GAS method allows for goals to be defined at different levels of mastery and assigned numerical values for score calculation, similar to a Likert scale. The scale will have 5 points representing different levels of mastery of the individual participant’s goal. A score of −2 represents baseline, −1 less change than expected, 0 for the expected level of change, and +1 and +2 for achievement of more change than expected. Each level on the scale will be described and will reflect a single dimension of change that is measurable, achievable, and relevant.^66–69

Data and Safety Monitoring Plan

The Data and Safety Monitoring Committee (DSMC) will include 4 members, 2 members are not associated with the study. The volunteers will have experience in child development and cerebral palsy. This board will convene at the start of the study to be orientated to the study and approve the IRB protocols. The DSMC will review the outcome evaluations collected at each time points for all 3 groups. Reviews will occur early in the grant period to assess whether there are concerns with low, intermediate, or high frequency therapy. The DSMC will conduct the first review of data after 8 participants have completed the 3-month outcomes evaluation and then conduct biannual reviews until all participants have completed treatment. The DSMC will be asked to provide a list of other types of data they would like to review each year at their annual in-person meeting and at what intervals they would like data reports of short-term and long-term outcomes. We will also review the findings when we are twenty-five percent through the trial to see if there is a subgroup of children who are not responding. We feel this is unlikely, but if it is found, we could modify our protocol and refer the subgroup of children to an appropriate health care provider.

The PI will be responsible for data and safety monitoring in accordance with IRB policies. The study procedures and risk assessment are reviewed on an annual basis by the IRB. Adverse events will be reported to the IRB, DSMC, and sponsor per guidelines. If changes to the research or consent process are proposed as a result of the event, an amendment through the IRB will be requested and the PI will be responsible for communicating with the sponsor should protocol changes be required. Informed consent or assent from potential trial participants or authorized surrogates will obtain by approved study staff.

RESULTS

Determining optimal frequency of treatment has implications for shaping the future of pediatric rehabilitation. There are variations in the number of hours per week of treatment in current outpatient rehabilitation programs for children with CP, suggesting clinical uncertainty. Usual weekly therapy at 1 – 2 hours per week for 6 months or longer is the most commonly implemented frequency of dose for children with CP 6 –24 months of age. However, the decision about frequency is often made based on clinical reasoning and scheduling, not on principles of rehabilitation, child development, or evidence from valid RCTs. The proposed study will fill this gap by directly comparing the effects of 3 frequency levels of therapy –concentrated daily, intermediate, and weekly in children with CP 6 –24 months of age at the initiation of treatment and following these participants for 2 years.

Research Electronic Data Capture (REDCap), a secure, web-based application designed to support data capture for research studies, will be used for data management and quality assurance. Participant medical records will be sourced through 1 all-inclusive electronic medical record system (EPIC) for birth history and comorbidities. All study investigators have access to final trial data through REDCap.

For this randomized trial, 75 participants will be randomized into a concentrate daily, intermediate, or usual weekly therapy group, stratified based on gender (rates of CP are 1:1.4 for girls: boys) and the severity of the disease at the baseline (level III, IV, V). The primary outcome endpoint for this study is the GMFM-88 (rated by masked evaluator), secondary outcomes include Bayley III, parental satisfaction, spontaneously play, GAS, and other development measures across domains (Table 1).

All baseline data will be summarized using descriptive statistics such as means, standard deviations, medians, and ranges for continuous variables and frequencies and percentages for categorical variables for the overall sample and by the study groups. Baseline characteristics will be compared among groups and tested using ANOVA for continuous variables and Chi-square test for categorical variables. Nonparametric tests will be used if the assumptions of the parametric tests are not satisfied. Final analyses will compare the study groups on all outcome measures at the end of the study period.

Most scales (such as GMFM-88) can be treated as continuous measures in this study. For all continuous outcomes collected over time, plots over time will be generated for each individual to identify potential outliers visually. Before data analysis, each outlier will be investigated to determine if it is from data input error or not. Linear mixed models for repeated measures will be used to model the changes of each group from baseline to the time points of interest after adjusting for the effects of stratification factors. The baseline score will be considered as covariate in the model. The difference among groups will be compared based on the fixed effect of group estimated for the selected model and tested using their corresponding contrasts. An unstructured covariance (UN) matrix will be assumed for the repeated measures within each subject, alternative covariance structure will be used if its model performance is better than the UN structure based on model fitness test. For categorical data, we will summarize it using corresponding tables and use Chi-square test for hypothesis testing.

For Aim 1, to tightly control for type I error, we will use combined gatekeeping procedures: first test if the changes are different among 3 groups in 2-year GMFM-88 at significance level of 0.05. If the null hypothesis is rejected, we will further test the group-wise difference, and then the difference among groups for the shorter-term effect, and similar approach will be used for other secondary endpoints. If the long term effect is not significant, the secondary analysis of the short term effect and secondary outcomes will be conducted as exploratory analysis. For Aim 2, we will develop a prediction model using multivariate linear regression analysis to predict the short /long term effect based on participants’ baseline characteristics (such as baseline GMFCS (severity) level, age, gender, race), parent engagement, and the treatment assignment. Interaction of the treatment with gender and disease severity will also be included in the model. This analysis also includes a cross-validation procedure: randomly select 70% of the participants from each treatment group to develop the models and the remaining 30% to validate the models. This procedure will be repeated for 10 times as a sensitivity analysis.

Sample Size and Power Calculation from our preliminary data from a daily (high) and weekly (low) therapy program, the GMFM-88 increased 15 and 8 respectively; assuming a conservative standard deviation of 5, the effect size is approxiamtely 1.4. We conservatively estimate that the expected effect size for our study is at least 1 standard deviation difference between two groups. Sample size of 21 per group provides at least 80% power to detect an effect size of 1 standard deviation difference between two groups at significance level of 0.05/2=0.025 based on a two-sided test. Bonferroni multiple comparisons adjustment for type I error were used for two simultaneous comparisons between weekly vs daily, and intermediate vs weekly. Considering up to 15% attrition during the 2-year study period, we will recruit of 75 subjects (~25 per group) for this study.

Our primary analysis will be conducted using the intention to treat (ITT) principle, use SAS v 9.3 (Sas Inc. NC). While we will make every effort to minimize the missing data for this study, missing data are expected due to various reasons. The pattern of data missing will be investigated. While our primary analysis will only use these non-missing data without data imputation for the mixed models, which assumes Missing At Random (MAR) to deal with the missing data problem, sensitivity data analysis will be conducted based on the completed data set (only individuals with no missing data), and multiple imputation data sets simulated based on the identified missing data pattern.