Effects of Technology-Supported Social Communication Baby Navigator Intervention to Improve Outcomes of Toddlers With Communication Delays: Preliminary Study
Effects of Technology-Supported Social Communication Baby Navigator Intervention to Improve Outcomes of Toddlers With Communication Delays: Preliminary Study
A confluence of empirical research on poverty, environmental risk factors, and brain development shows that early experiences are uniquely powerful in the first years of life, providing further impetus for early detection and intervention. Measures of early-developing social communication skills offer a viable solution for earlier detection of children with language delays and subsequent educational challenges. The investigators propose testing a new technology-supported platform with three parent-mediated intervention components for babies with early communication delays using mobile technology: 1) the Social Communication Growth Charts, a self-guided app to explore video clips illustrating early milestones and to chart their child's development; 2) Baby Navigator Webinar, webinars open to the public designed as a companion to the Growth Charts; and 3) Mobile Coaching, individual weekly telehealth sessions to coach parents in their everyday activities. The investigators propose using a multiphase optimization strategy (MOST) as a framework for development, optimization, and evaluation of our dynamic web intervention platform. The Investigators will recruit 80 children with communication delays at 12-18 months of age. Families will be invited to the Social Communication Growth Charts bundled with the Baby Navigator Webinar. At 18-21 months, slow responders will be randomized to continue the bundle with or without Mobile Coaching, and responders will continue the bundle. This study will enhance the sustainability, scalability, and lead to transformative changes to efficiently and effectively improve healthcare delivery via the use of innovative technology, an implementation science methodology, and user-friendly tools and web platform.
BACKGROUND AND AIMS:
About one in three children in the U.S. enter kindergarten unprepared to learn and lacking sufficient communication skills. Despite the federal mandate for early intervention (EI), less than 25% of children who will require special education at school age are receiving EI services under three years of age. Lower income, African American, and rural families may receive a diagnosis 1.5 years later. Twenty-five years after Hart and Risley described the 30-million-word gap between children from higher and lower socioeconomic backgrounds, "catching up" earlier is emphasized more than ever, and intervention beginning during infancy or preschool has a greater impact on outcomes for children than providing services at school age. A confluence of empirical research on poverty, environmental risk factors, and brain development shows that early experiences are uniquely powerful in the first years of life with lasting effects on developmental trajectories and significant impact on lifelong well-being, providing further impetus for early detection and intervention.
There is a pressing need to leverage resources by coordinating care provided by the many different agencies serving young children and test new methods to support families as soon as their children begin to show small lags in communication development. The investigators' research has identified robust predictors of later language development that can be identified at 9-18 months of age. Because delays in communication are apparent long before language delays are evident, measures of early-developing social communication skills offer a viable solution for earlier detection of children with language delays and subsequent educational challenges.
The investigators propose studying a new technology-supported intervention for babies with early communication delays-Social Communication "Baby Navigator" Intervention-that builds on the findings of our efficacy trial of the Early Social Interaction (ESI) Model to coach families of toddlers with autism to embed evidence-based intervention strategies in their everyday activities. The content of Baby Navigator is based on responsive parenting research that has a strong evidence-base for children from very low resource settings worldwide and is recommended to improve child health and development by the World Health Organization (WHO). Responsive parenting defined by WHO entails observing the child's cues, such as movements and sounds, interpreting these signals accurately, and acting swiftly, consistently and efficiently to meet the child's needs. Baby Navigator is a web platform with three parent-mediated intervention components using mobile technology: 1) the Social Communication Growth Charts, a self-guided application (app) to explore hundreds of video clips illustrating early milestones with narration on responsive parenting and to chart their child's development; 2) Baby Navigator Class, group weekly online meetings designed as a companion to the Growth Charts to engage parents while learning the milestones and how to support their child's development, and connect with other families; and 3) Mobile Coaching, individual weekly telehealth sessions to coach parents and provide reflective feedback in their everyday activities, adapted from the ESI model.
The investigators will address the following research aims:
Aim 1. Document the feasibility and effects of the Social Communication Growth Charts and the Baby Navigator Webinar bundled from baseline to child and parent outcomes 6-12 months post baseline.
Aim 2. Compare the effects of optimization with or without Mobile Coaching for slow responders in the 2nd half of the study on child and parent outcomes at 6-12 months post baseline.
Aim 3. Explore individual child and family characteristics that may moderate response to the technology-supported intervention components of the new Baby Navigator.
Aim 4. Examine responsive parenting and parent engagement with each technology-supported intervention component and prepare for a full-scale SMART that will control for maturation.
The expected outcomes of this preliminary study will inform immediate and rapid deployment of a full-scale SMART that can be implemented across the US. This study will enhance the sustainability, scalability, and lead to transformative changes to efficiently and effectively improve healthcare delivery via the use of innovative technology, an implementation science methodology, and user-friendly tools and web platform.
RECRUITMENT OF PARENT/CHILD DYADS:
Children and their families will be recruited from the FIRST WORDS Project online Social Communication Check Up (SoCo CheckUp) screening portal available at https://babynavigator.com. Based on our preliminary studies, the investigators anticipate up to a 20% positive screen rate in a primary care setting on the SoCo CheckUp indicating risk for communication delay, and a 5% positive screen rate for autism. Families of children with a positive screen will be invited to complete the CSBS Caregiver Questionnaire online and upload a home observation video, which will be rated for social communication delays and early signs of autism using the SORF. The investigators will be recruiting families across the US, and will make an effort to have a stratified sample drawn equally from 4 regions of the country: West, Midwest, Northeast, and Southeast.
DATA AND POWER ANALYSIS:
Data and Power Analysis for Aims 1 and 2. The investigators will use a series of quantitative analyses to assess outcomes for Aims 1-2. The investigators will document feasibility and effects of the Social Communication Growth Charts and the Baby Navigator Webinar bundled from baseline to child and parent outcomes at 6 and 12 months post baseline for the entire sample using paired t-tests (Aim 1). The initial sample size will be 80 children. Allowing for 15% attrition, the investigators plan to have 68 children complete the MOST. The investigators will also test for differences over time using pre-post assessments to compare the effects of optimization with or without Individual-ESI for slow responders from 6-12 months post baseline on child and parent outcomes at 6-12 months post baseline using ANOVAs (Aim 2). The investigators anticipate that 44 children will meet criteria for slow responders based on previous research and be randomly assigned to Baby Navigator Bundle + Individual-ESI or Baby Navigator Bundle Only. The remaining 24 children (responders) will continue with the Baby Navigator Bundle.
Power calculations were computed a priori using G*Power 3.1 with 0.05 significance and power of .80 using Cohen's d to estimate the sample size required for a two-tailed paired t-test, repeated measures ANOVA, and one-way ANOVA. Results indicated that the study would need 67 children to detect a medium effect size of .35 for the t-test, 30 children for the MAETS and CQ and 36 for the CBCL to detect a medium effect for the RM-ANOVA, and 66 children to detect a large effect for the one-way ANOVA.
Data and Power Analysis for Aims 3 and 4. The investigators will use a combination of quantitative analyses to assess outcomes and measures for Aims 3-4. The study will explore individual child (e.g., communication and language, sex) and family characteristics (e.g., parent education) that may moderate response to our technology-supported intervention components using multiple linear regression (Aim 3). The investigators will also examine responsive parenting using the MAETS and parent engagement with the technology (Aim 4). It is predicted that parental online tool usage will be a mediator variable while moderating variables will be captured in the family measures (e.g., responsive parenting, demographic information, hours in other interventions). In previous studies of children recruited by the FIRST WORDS Project, typically developing females were observed to achieve significantly better receptive language skills on the Mullen Scales of Early Learning and use of words for communication on the CSBS when compared to males. These findings are consistent with existing literature that documents a slight female advantage in the development of early language and communication, and highlight the need to examine sex as a covariate. If the child's outcomes for the dependent variables can be explained by the addition of the mediator or moderator variables, it will provide evidence that these variables impact the children's communication development.
A priori power analyses suggested that the number of predictors that the investigators could examine in a multiple linear regression is five total and two at a time assuming a sample size of 68, an R2 of .5. The number of predictors is dependent upon the amount of variance explained. If the R2 is higher, the investigators will be able to examine more predictors in the model. The quantitative analyses outlined are sufficiently powered, will provide important information regarding the feasibility and effects of the intervention, and are appropriate given the goals of this study.
SOURCES OF MATERIALS AND DATA:
The sources of research material obtained from the children and families participating in this study will be in the form of written records, online screeners and measures, database records for assessment measures, and audio/video recordings documenting assessments and intervention sessions. It will include video files for the home observations, forms and questionnaires filled out by the parents, data from the language evaluation, intervention session audio-video recordings and notes, and coach notations on the ESI coaching videos. This information is being obtained for research purposes; however, it may also serve a clinical purpose for the family. The information obtained will help the families determine whether their child may qualify for publicly funded early intervention services.
DATA AND VIDEO STORAGE AND SECURITY:
All data, documents, records, video recordings (whether digital or hardcopy), consent forms will be stored indefinitely. Long-term storage depends on the media being stored.
Hardcopy video recordings are stored in a locked room at the FSU Autism Institute. The room is only accessible by the PI and the digital media specialists. The digital media specialists are also the only people who have the equipment needed to work with the recordings. A backup copy of all hardcopy recordings is stored in locked filing cabinets in the FSU COM Department of Clinical Sciences.
Electronic video recordings and all electronic documents are stored on FSU Autism Institute computer servers. The Autism Institute has four computers with the processing and memory hardware system requirements to form a virtual dedicated server with the capability for flash video streaming to host project websites. This provides redundancy to prevent system failure and is located in the FSU COM. The Autism Institute also has space on an EMC VNX model storage device to use as a networked secure centralized storage source. The Content Storage Server provides continuous data protection for digital video files that need to be accessed by users. This system offers the storage capacity and input/output processing speed needed to access the videos efficiently by many users simultaneously. Further, the Content Storage Server has built-in redundancy to provide protection against disk failure in order to better protect files and ensure continuous access to users.
Paper files are stored in locked filing cabinets or locked rooms within the FSU Autism Institute and are accessible to the research coordinator and PI. The entire Autism Institute is a Sonitrol locked facility.
Data will be stored in the HexDB and the RexDB-which function as a data repository. Data will also be stored in HIPAA-compliant Smartsheet.
RexDB. Prometheus Research has created a database for the FSU Autism Institute for ongoing large longitudinal and multisite research projects that will be available for this project. This database is only for participants who complete an in-person diagnostic/developmental evaluation and are therefore located only in the United States. This database maintains the central data core for Institute projects, as well as all electronic data capture (EDC) interfaces, using its secure, Web-based platform, RexDB™. The RexDB has a specialized secure database with tables linked by an assigned identification code to enter and store data.
HexDB. The Autism Institute has also developed the FIRST WORDS Project Research Database-called HexDB-as an in-house alternative to the RexDB. The HexDB does not currently store diagnostic/developmental evaluation data. Data from the SoCo CheckUp and Baby Navigator tools and resources utilization data, is automatically stored on the HexDB.
Both the RexDB and the HexDB are housed on web servers with encrypting and password protection in compliance with HIPAA regulations. In addition, HexDB is GDPR-compliant and can be used for participants located in the European Union and other countries. It is necessary for the research databases to have identifying information, since clinical reports are generated from the project and because the investigators will keep the data indefinitely and may contact research participants in the future for additional research opportunities. Because the databases are stored on HIPAA-compliant servers accessed over the internet, if the actual project computers were stolen, the user could not access the databases, thus protecting the security and confidentiality of the records.
Access. HexDB and RexDB employ role-based access to role relevant information. System administrators assign researchers and data entry staff access to only the specific sets of data needed based on their relation to the study in which the user is participating. These access levels are verified on a quarterly basis. The Principal Investigator has view-level access to the databases. The Director of IT & Data Management-functioning as an Honest Broker of Identifiable Information-is the only research staff who can transfer/export data from the databases with encrypting and password protection. Queries are designed so that only de-identified data using the participant's identification code will be exported and used for data analysis.
QUALITY CONTROL OF DATA:
Data within RexDB and HexDB will ideally possess the following qualities:
Accuracy: accurate data is "correct" or valid. It represents exactly what was intended to be collected. Example: reconciliation of double entered measures data in RexDB helps ensure its accuracy (i.e., an ADOS with a correctly entered score at the item level).
Reliability: reliable data is collected in a consistent way at the interface level and is consistently reproducible at the database level. Example: a field label remains the same and consistent over time and is not altered during the collection period. Questions do not change in their intent, meaning, or sets of possible values. Running a report on that data produces the same possible values over time.
Precision: precise data is captured at the correct level of detail. Example: collecting the number of weeks early to calculate adjusted age at evaluation to the appropriate decimal value (typically thousandths) for a greater precision level during statistical analysis of the data.
Security: secure data is collected, transmitted, and stored in a HIPPA compliant way. Example: data moving via API among systems should be encrypted using industry standard best practice protocols.
AUDIT PROCEDURES:
Duplicate Data Identification and Handling. Duplicate data is a common problem in any system, especially in a system that allows its users to create data (e.g., adding multiple instances of the same child accidentally or intentionally). Removal of duplicate data should be handled on an ongoing basis (at minimum, quarterly) and double-checked before any major data pulls for reports or publications. Duplicate management also prevent emails, texts, etc. from going out to parents for a duplicate child and prevents duplicate data from being pulled into reports.
Fictitious Data. Although test data should never intentionally be inserted into a production database as a general best practice; it occasionally is necessary. HexDB includes a boolean field called "fictitious" that will allow a DBA to mark data as being fictional. Records marked as fictitious are excluded from reports.
Identification. Fictitious data can be harder to detect than duplicate data, although there are some tricks to finding it. Looking for name data that includes "test", "delete", etc. will periodically turn up data that has yet to be marked as fictitious. Running a simple query where substituting firstName and lastName with different keywords can be helpful in identifying test data. Staff that encounter fictitious data notify the DBA of said data, and the DBA then flags the data manually.
Handling procedures. Once identified, the fictitious field on a fictitious record is set to 1. This data can be allowed to persist, as deleting test records often only encourages the creation of more test records. The primary goal is to exclude this data from any analysis by flagging it appropriately and ensuring that any reports or queries on the database include the flag to exclude fake data. Alternately, the data can be deleted, but the data from all associated systems will be removed prior to deletion from HexDB.
Inclusion Criteria for parent/child dyads:
Exclusion Criteria: