SenseToKnow Autism Screening Device Validation Study

SenseToKnow STAR Study: A Study of Technologies for Assessing Children's Development

This is a pivotal, prospective, double-blind, study to evaluate the sensitivity and specificity of the SenseToKnow device for the detection of autism spectrum disorder in children 16-36 months of age.

Study Overview

• Status: Recruiting
• Conditions: Autism Spectrum Disorder; Autism

Detailed Description

This is a pivotal, prospective, double-blind, study to evaluate the sensitivity and specificity of the SenseToKnow device for the classification of autism spectrum disorder when administered by parents in a sample of patients 16-36 months of age. The trial design is a non-interventional cross-sectional study comparing the SenseToKnow device classification of autism spectrum disorder ("autism") versus non-autism with the patient's diagnostic status based on expert clinical diagnosis in a population of pediatric patients.

• Study Type: Observational
• Enrollment (Estimated): 350

Contacts and Locations

• Study Contact: Name: Geraldine Dawson, PhD; Phone Number: 9196680070; Email: geraldine.dawson@duke.edu
• Study Contact Backup: Name: Charlotte Stoute, BA; Phone Number: 919-681-9730; Email: charlotte.stoute@duke.edu

Study Locations

• United States
  • North Carolina
    • Durham, North Carolina, United States, 27705
      • Recruiting
      • Duke University
      • Contact: Geraldine Dawson, PhD; Phone Number: 919-668-0070

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child
• Accepts Healthy Volunteers: Yes

• Sampling Method: Non-Probability Sample

Study Population

Participants will be patients 16-36 months of age recruited from > 6 sites comprised of pediatric medical clinics that are part of the broader Duke University Health System in North Carolina.

Description

Inclusion Criteria:

1. Duke Health pediatric patient at enrollment
2. 16-<37 months of age at enrollment
3. Parent/legal guardian speaks English or Spanish
4. Parent/legal guardian understands and voluntarily provides informed consent

Exclusion Criteria:

1. Severe motor impairment that precludes study measure completion
2. Known genetic disorders
3. Severe hearing or visual impairment as determined on physical examination according to parent report
4. Acute illnesses likely to prevent successful or valid data collection
5. Uncontrolled epilepsy or seizure disorder
6. History or presence of a clinically significant medical disease, or a mental state that could confound the study or be detrimental to the subject as determined by the investigator
7. Acute exacerbations of chronic illnesses likely to prevent successful or valid data collection
8. Receiving therapies that affect vision
9. Parent/legal guardian and/or investigator believes that the child will be unable/unwilling to sit in the parent's lap to watch the app videos
10. Parent/legal guardian indicates that they or their child is unwilling or unable to complete the app administration, surveys, or diagnostic assessment
11. Participants who are otherwise judged as unable to comply with the protocol by the investigator
12. Any other factor that the investigator feels would make the study measures invalid

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort

Pediatric patients, 16-36 months of age, recruited through pediatric medical clinics; Consecutive pediatric participants will be recruited and enrolled via >= 6 participating sites comprised of pediatric medical clinics (e.g., primary care and family medicine clinics) that are part of the broader Duke University Health System (DUHS) located in North Carolina. Enrollment will proceed until the targets of N = 150 participants diagnosed with autism spectrum disorder and N = 200 without autism are reached.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Sensitivity of the SenseToKnow screening device based on a machine learning algorithm that combines SenseToKnow digital data with data from the SenseToKnow Caregiver survey for autism detection	Sensitivity = #participants positive for autism on both (1) the SenseToKnow screening device based on a machine learning algorithm that combines SenseToKnow digital data with the SenseToKnow Caregiver Survey data and (2) expert clinical diagnosis / #participants positive for autism on both SenseToKnow and expert clinical diagnosis	Will be calculated based on data from Baseline/Timepoint 1
Specificity of the SenseToKnow screening device based on machine earning algorithm that combines SenseToKnow digital data with data from the SenseToKnow Caregiver survey for autism detection	Specificity = #participants negative for autism on both (1) the SenseToKnow screening device based on a machine learning algorithm that combines SenseToKnow digital data with the SenseToKnow Caregiver Survey data, and (2) expert clinical diagnosis / #participants negative for autism on autism by expert clinical diagnosis	Will be calculated based on data from Baseline/Timepoint 1

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Positive Predictive Value of SenseToKnow screening device (based on a machine learning algorithm using the SenseToKnow digital data, combined with the SenseToKnow Caregiver Survey data) for autism detection in comparison to expert clinical diagnosis	The likelihood that a participant with a positive test result (based on a machine learning algorithm using the SenseToKnow digital data, combined with the SenseToKnow Caregiver Survey data) has a diagnosis of autism (based on expert clinical diagnosis). Positive Predictive Value will be calculated with and without adjustment for population prevalence.	Will be calculated based on data from Baseline/Timepoint 1
Negative Predictive Value of SenseToKnow screening device (based on a machine learning algorithm using the SenseToKnow digital data, combined with the SenseToKnow Caregiver Survey data) for autism detection in comparison to expert clinical diagnosis	The likelihood that a participant with a negative test result (based on a machine learning algorithm using the SenseToKnow digital data, combined with the SenseToKnow Caregiver Survey data) does not have a diagnosis of autism (based on expert clinical diagnosis). Negative Predictive Value will be calculated with and without adjustment for population prevalence.	Will be calculated based on data from Baseline/Timepoint 1
Receiver Operating Characteristic Curve and Area Under the Curve with respect to the accuracy of the SenseToKnow screening device (using the SenseToKnow digital data and SenseToKnow Caregiver survey data) for autism versus non-autism classification	Receiver Operating Characteristic Curve (ROC) is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied. Area Under the Curve (AUC) measures the area underneath the entire ROC curve. Accuracy of test is based on a machine learning algorithm using the SenseToKnow digital data, combined with the SenseToKnow Caregiver Survey data, in comparison to expert clinical diagnosis.	Will be calculated based on data from Baseline/Timepoint 1
Sensitivity of SenseToKnow screening device based on a machine learning algorithm using only the SenseToKnow digital data for autism detection	Sensitivity = #participants positive for autism on both (1) the SenseToKnow screening device based on a machine learning algorithm using only the SenseToKnow digital data and (2) expert clinical diagnosis / # participants positive for autism on expert clinical diagnosis	Will be calculated based on data from Baseline/Timepoint 1
Specificity of SenseToKnow screening device based on a machine learning algorithm using only the SenseToKnow digital data for autism detection	Specificity = #participants negative for autism on both (1) the SenseToKnow screening device based on a machine learning algorithm using only the SenseToKnow digital data and (2) expert clinical diagnosis / #participants negative for autism on expert clinical diagnosis.	Will be calculated based on data from Baseline/Timepoint 1
Positive Predictive Value of SenseToKnow screening device based on a machine learning algorithm using only the SenseToKnow digital data for autism detection in comparison to expert clinical diagnosis	The likelihood that a participant with a positive test result has a diagnosis of autism (based on expert clinical diagnosis). Positive Predictive Value will be calculated with and without adjustment for population prevalence.	Will be calculated based on data from Baseline/Timepoint 1
Negative Predictive Value of SenseToKnow screening device based on a machine learning algorithm using only the SenseToKnow digital data for autism detection in comparison to expert clinical diagnosis	The likelihood that a participant with a negative test result does not have a diagnosis of autism (based on expert clinical diagnosis). Negative Predictive Value will be calculated with and without adjustment for population prevalence.	Will be calculated based on data from Baseline/Timepoint 1
Receiver Operating Characteristic Curve and Area Under the Curve with respect to the accuracy of the SenseToKnow device using only the SenseToKnow digital data for autism versus non-autism classification	Receiver Operating Characteristic Curve (ROC) is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied. Area Under the Curve (AUC) measures the area underneath the entire ROC curve. Accuracy of test is based on a machine learning algorithm using only the SenseToKnow digital data, in comparison to expert clinical diagnosis.	Will be calculated based on data from Baseline/Timepoint 1

Collaborators and Investigators

• Sponsor: Duke University
• Collaborators: Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD)
• Investigators: Principal Investigator: Geraldine Dawson, PhD, Duke University

Study record dates

Study Major Dates

• Study Start (Actual): July 7, 2023
• Primary Completion (Estimated): December 1, 2027
• Study Completion (Estimated): December 1, 2027

Study Registration Dates

• First Submitted: May 12, 2023
• First Submitted That Met QC Criteria: May 12, 2023
• First Posted (Actual): May 25, 2023

Study Record Updates

• Last Update Posted (Actual): March 6, 2026
• Last Update Submitted That Met QC Criteria: March 4, 2026
• Last Verified: February 1, 2026

More Information

Terms related to this study

• Keywords: Diagnosis; Digital device
• Additional Relevant MeSH Terms: Mental Disorders; Pathologic Processes; Neurodevelopmental Disorders; Child Development Disorders, Pervasive; Pathological Conditions, Signs and Symptoms; Autism Spectrum Disorder; Disease; Autistic Disorder
• Other Study ID Numbers: Pro00111371; 2P50HD093074 (U.S. NIH Grant/Contract)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: YES
• IPD Plan Description: All individual-level data that meets PHI and IRB confidentiality requirements will be submitted to the NIH/NIMH Data Repository by the end of the grant period.
• IPD Sharing Time Frame: We will submit an electronic version of the final, peer-reviewed work, including the statistical analysis code, to the National Library of Medicine PubMed Central, to be made publicly available no later than 12 months after the official date of publication.
• IPD Sharing Access Criteria: Publically available via PubMed Central
• IPD Sharing Supporting Information Type: ANALYTIC_CODE

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No