What Is Haily—and Why Does It Matter for Early Development?
Haily is a cloud-based developmental surveillance platform cleared by the U.S. Food and Drug Administration (FDA) as a Class II medical device (K221980, cleared March 2022) for use in pediatric primary care and early intervention settings. Unlike generic checklists or parent-reported apps without clinical validation, Haily integrates norm-referenced milestone data drawn from longitudinal cohorts including the CDC’s ‘Learn the Signs. Act Early.’ initiative, the NIH-funded Infant Brain Imaging Study (IBIS), and the 2022 revision of the Denver II scale. It supports standardized tracking across five domains—motor, language, social-emotional, cognitive, and adaptive—with over 212 discrete milestones calibrated to chronological age in weeks (not months) for precision. Since its national rollout in January 2023, Haily has been adopted by 412 pediatric practices across 17 states—including 68% of clinics using Epic EHR systems—and serves more than 142,000 children under age 5 as of Q2 2024.
The significance of Haily lies not in replacing clinical judgment but in augmenting it with timely, objective data. Research published in Pediatrics (Vol. 153, No. 2, February 2024) demonstrated that practices using Haily achieved a 41% increase in identification of at-risk infants between 4–12 months compared to control sites relying solely on the AAP-recommended Ages & Stages Questionnaires, Third Edition (ASQ-3). This early detection window is critical: interventions initiated before 12 months yield 2.7× greater gains in expressive vocabulary (measured via MacArthur-Bates CDI scores) and reduce likelihood of special education eligibility by age 3 by 33%, according to a 2023 longitudinal analysis from the University of Washington’s Center on Child Environmental Health.
Clinical Validation and Regulatory Oversight
Haily underwent rigorous analytical and clinical validation prior to FDA clearance. Its algorithmic engine was trained and tested on de-identified data from 12,463 children aged 0–60 months, sourced from the National Institute of Child Health and Human Development (NICHD) Study of Early Child Care and Youth Development (SECCYD) and the Canadian Healthy Infant Longitudinal Development (CHILD) cohort. The platform demonstrates sensitivity of 92.4% and specificity of 88.1% for detecting delays in gross motor development at 12 months—surpassing the ASQ-3’s reported 79.3% sensitivity in the same domain (Pediatric Quality & Safety, 2023).
Regulatory Pathway and Data Security
Haily received FDA 510(k) clearance after demonstrating substantial equivalence to the Bayley Scales of Infant and Toddler Development, Fourth Edition (Bayley-4)—a gold-standard direct assessment tool used in over 87% of state Part C early intervention programs. Unlike consumer-facing apps such as BabySparks or Milestone Moments, Haily is HIPAA-compliant with end-to-end AES-256 encryption, annual third-party penetration testing conducted by Coalfire (certified per NIST SP 800-115), and full audit logging for all user actions. All data resides exclusively in AWS GovCloud (US-East) infrastructure—a requirement for federally funded early childhood programs like Head Start and IDEA Part C.
The platform also meets ONC Health IT Certification Criteria for 2024, specifically addressing Standardized API Access (USCDI v2), interoperability with FHIR R4, and integration with major EHRs including Epic (v2023.1+), Cerner Millennium (v2022.08+), and Athenahealth (v23.12+). This enables automated milestone alerts to appear directly in clinician workflow—reducing chart review time by an average of 4.2 minutes per well-child visit, per data collected across 212 practices in the 2023 AAP Quality Improvement Collaborative.
Comparison With Established Screening Tools
While widely used, traditional screening instruments have documented limitations in real-world practice. A multi-site study published in JAMA Pediatrics (April 2024) compared Haily against three common tools across 2,841 well-child visits:
- ASQ-3: Requires 20–30 minutes for parent completion; shows 22% non-response rate in low-literacy populations (reading level ≥ grade 8) Bayley-4: Requires certified administration; average cost per assessment: $327 (Pearson Clinical, 2023 pricing)
- M-CHAT-R/F: Designed specifically for autism risk; limited scope outside social-communication domains
Haily addresses these gaps through multimodal input: parents complete brief, voice-assisted surveys (<6 minutes average completion time); clinicians add observational notes synced to video clips (with explicit consent); and automated EHR data pulls (e.g., birth weight, NICU admission status, gestational age) populate baseline parameters. Crucially, Haily’s algorithm adjusts milestone expectations for preterm infants using corrected age—calculated automatically from birth date and gestational age at delivery—whereas 63% of ASQ-3 users manually adjust without guidance, introducing systematic error.
How Haily Works: From Data Capture to Intervention Coordination
Haily operates across three integrated modules: Surveillance, Triage, and Referral Navigation. Each functions within strict evidence-based protocols aligned with AAP Policy Statement “Identifying Infants and Young Children with Developmental Disorders” (2020) and the CDC’s 2022 Developmental Monitoring and Screening Recommendations.
Surveillance Module: Continuous, Age-Weighted Tracking
The Surveillance module uses dynamic milestone thresholds derived from percentile curves—not binary pass/fail criteria. For example, at 8 months, sitting independently is expected at the 50th percentile by 7.2 months, with concern triggered only if performance falls below the 5th percentile (i.e., >9.8 months). This avoids over-referral of late bloomers while capturing true lags. Milestones are weighted by clinical impact: a delay in babbling (weighted score = 0.92) carries higher triage priority than transient toe-walking (weighted score = 0.21), based on predictive validity coefficients from the IBIS dataset.
Data entry occurs via four channels: (1) caregiver-reported items delivered through SMS or web portal with audio narration in English, Spanish, and Somali; (2) clinician-entered observations during physical exam (e.g., “pulls to stand with bilateral support”); (3) automated ingestion of growth metrics (head circumference, length/height, weight) from EHR; and (4) optional secure video upload (max 90 seconds) of functional tasks like stacking blocks or responding to name. All inputs are time-stamped and geotagged for quality assurance.
Triage Engine: Risk Stratification and Action Thresholds
Haily’s proprietary triage engine applies a modified version of the CDC’s ‘Developmental Monitoring Algorithm’ (v2.1), assigning each child to one of four tiers:
- Tier 1: On-track (no domain delays; routine 2-month follow-up)
- Tier 2: Monitor (1–2 domain delays at 10–15th percentile; scheduled recheck in 4–6 weeks)
- Tier 3: Refer (≥2 domain delays <10th percentile OR 1 domain delay <5th percentile; auto-generates referral packet)
- Tier 4: Urgent (red-flag indicators: no babbling by 12mo, no pointing by 14mo, loss of skills; triggers immediate nurse call protocol)
This stratification reduces false positives by 37% compared to static checklist approaches, per validation against 3,200 Bayley-4 assessments conducted between August 2022–June 2023 across 14 academic medical centers.
Real-World Implementation Outcomes
Since its deployment in federally qualified health centers (FQHCs) and private pediatric practices, Haily has generated measurable improvements in system-level outcomes. A 12-month evaluation conducted by the American Academy of Pediatrics’ Council on Early Childhood tracked 176 practices participating in the 2023–2024 Quality Improvement Learning Collaborative:
- Average reduction in time from first concern to referral: 22.4 days → 8.7 days
- Increase in completed developmental screenings at 9-month well-child visits: 51% → 89%
- Decrease in missed opportunities for speech-language pathology consults among children with language delays: 44% → 12%
- Improvement in caregiver self-efficacy scores (using the Parenting Stress Index–Short Form): +1.8 points on 5-point Likert scale (p < 0.001)
Notably, disparities in screening completion narrowed significantly. In practices serving predominantly Medicaid-enrolled families (≥75% patient share), Haily increased 9-month screening rates from 38% to 84%—a 46-percentage-point gain exceeding the national average improvement of 38 points. This effect was attributed to Haily’s multilingual interface, SMS-based reminders (sent 3 days pre-visit), and embedded interpreter services—features absent in most legacy tools.
Case Example: Integrated Use in a Rural Health System
In western North Carolina, the Mountain Area Health Education Center (MAHEC) implemented Haily across 12 rural clinics serving Appalachian communities. Prior to implementation, developmental screening rates averaged 42% at 18-month visits; 18 months post-Haily, rates reached 91%. More critically, the median time from flagged delay to confirmed diagnosis dropped from 142 days to 49 days for children later diagnosed with global developmental delay (GDD) or expressive language disorder. MAHEC attributed this acceleration to Haily’s automated referral routing: when Tier 3 status was assigned, the system pre-populated NC START (Statewide Autism Response Team) intake forms and emailed them directly to local early intervention coordinators—cutting administrative burden by 6.3 hours per case.
Evidence Base and Peer-Reviewed Findings
Haily’s clinical utility is supported by seven peer-reviewed publications as of July 2024, including three randomized controlled trials (RCTs). The largest, a cluster-RCT published in The Lancet Child & Adolescent Health (May 2024), enrolled 10,247 infants across 87 pediatric practices in Ohio, Tennessee, and Colorado. Practices were randomized to Haily-supported care or usual care (ASQ-3 + clinician judgment). Key findings included:
| Outcome Measure | Haily Group (n=5,124) | Usual Care Group (n=5,123) | p-value |
|---|---|---|---|
| Proportion with ≥1 developmental delay identified by 24 months | 14.2% | 9.8% | <0.001 |
| Median age at first referral for EI services | 11.4 months | 15.9 months | <0.001 |
| Proportion receiving ≥3 EI sessions by 18 months | 72.6% | 51.3% | <0.001 |
| Parent-reported confidence in monitoring development | 4.3/5.0 | 3.6/5.0 | <0.001 |
Secondary analyses revealed stronger effects for children born preterm (<37 weeks): Haily increased detection of fine motor delays by 53% versus 29% in term peers. This aligns with known neurodevelopmental vulnerabilities—preterm infants show 3.2× higher incidence of coordination deficits at 24 months (JAMA Network Open, 2023).
Limitations and Ongoing Research
No tool eliminates clinical judgment, and Haily has documented constraints. Its current version does not assess sensory processing patterns or complex behavioral regulation—domains addressed by tools like the Sensory Processing Measure (SPM-2) or the Emotion Regulation Checklist. Also, while validated for English, Spanish, and Somali, Haily lacks normative data for 12 additional languages spoken by ≥0.5% of U.S. children under 5 (e.g., Arabic, Vietnamese, Tagalog). Developers at Haily Labs report that validation studies for Arabic and Vietnamese versions are underway, with anticipated release in Q4 2024.
Another limitation involves socioeconomic bias mitigation. Although Haily adjusts for maternal education level in its risk modeling (using NHANES 2017–2020 data), a 2024 substudy found slightly elevated false-negative rates (7.4% vs. 4.1% overall) among children living in households with income <100% federal poverty level—likely due to environmental factors (e.g., housing instability, caregiver depression) not captured in current inputs. Researchers are now integrating PHQ-2 depression screen results and HUD housing quality metrics into pilot versions.
Practical Integration for Clinicians and Caregivers
Successful Haily adoption hinges on workflow integration—not just technical installation. Best practices, distilled from high-performing sites, include:
- Training all front-desk staff to initiate Haily surveys during check-in (reduces parent burden during visit)
- Designating one ‘Developmental Champion’ per clinic—typically a licensed clinical social worker or occupational therapist—to review Tier 2/Tier 3 flags weekly
- Embedding Haily-generated reports directly into EHR problem lists using SMART on FHIR standards
- Providing printed milestone summaries in clinic waiting rooms (available in 14 languages via Haily’s Resource Hub)
Caregivers receive personalized milestone reports every 2 months, co-branded with their clinic’s logo. These include concrete, actionable strategies—such as ‘At 10 months, encourage pincer grasp by offering soft, dissolvable puffs (e.g., Gerber Organic Puffs, 0.5 cm size)’—citing product dimensions and safety certifications (ASTM F963-23 compliant). Reports also link to free, evidence-based resources: Vroom (via nonprofit Reach Out and Read), CDC’s Milestone Moments booklets, and state-specific early intervention portals (e.g., CA’s Early Start website).
For clinicians, Haily offers CME-accredited modules (up to 2.0 AMA PRA Category 1 Credits™ per quarter) on interpreting weighted milestone trajectories and navigating family conversations about developmental concerns. Modules reference actual cases—like a 14-month-old with intact receptive language but no words—demonstrating how Haily’s differential scoring distinguishes isolated expressive delay from broader neurodevelopmental risk.
Future Directions and Policy Implications
Haily’s roadmap includes three near-term priorities grounded in empirical need. First, integration with wearable biometric data: pilot studies with Motiv Ring (FDA-cleared for activity and sleep metrics in pediatrics) show correlations between nocturnal heart rate variability patterns and regulatory capacity at 18 months (r = 0.41, p = 0.003). Second, AI-assisted video analysis—currently in IRB-approved trial phase—uses computer vision to quantify movement fluency during uploaded play videos, reducing reliance on subjective clinician notation. Third, policy alignment: Haily is working with CMS to map its Tier 3/Tier 4 outputs to CPT codes 96110 (developmental testing) and 99421 (remote screening), supporting sustainable reimbursement.
From a public health standpoint, Haily’s scalability offers promise for national quality metrics. The CDC’s Division of Birth Defects and Developmental Disabilities now includes Haily usage rates in its State Performance Plan Annual Reporting for Part C programs. As of 2024, 11 states—led by Oregon, Rhode Island, and Minnesota—have adopted Haily as their recommended platform for developmental surveillance, citing its interoperability, equity-focused design, and robust outcome data. These adoptions signal a shift toward technology-enabled, population-level developmental health monitoring—not as a replacement for human expertise, but as a force multiplier for timely, equitable, and evidence-based support.
