506 - Feasibility and engagement-related differences in infant sleep outcomes in a Medicaid ACO pilot of a digital sleep app

Publication Number: 4497.506

Nisha Fahey, University of Massachusetts Medical School, Worcester, MA, United States; Alonzo M. Aguilar, University of Massachusetts Medical School, Worcester, MA, United States; Julia Ferranto, University of Massachusetts Medical School, Somerville, MA, United States; Madeleine Leake, UMass Memorial Children's Medical Center, Worcester, MA, United States; Xiang Li, University of Massachusetts Medical School, Worcester, MA, United States; Jessica Toh, Huckleberry Labs, Inc., Irvine, CA, United States; John Tan, Huckleberry Labs, Menifee, CA, United States; Apurv Soni, University of Massachusetts Medical School, Worcester, MA, United States

Background: Reduced sleep duration and quality in early infancy is common and can have negative effects. Poor sleep has been linked to long-term consequences for parental wellbeing and child development. Traditional behavioral sleep interventions often require in-person instruction, which can be resource-intensive and inaccessible for many families.

Objective: To investigate feasibility of Huckleberry, an infant-sleep tracking and guidance app, and explore associations between app engagement and infant sleep behaviors and parental wellbeing among a Medicaid Accountable Care Organization patient population.

Design/Methods: 80 parent–infant dyads (infants < 12 months old) enrolled in a 3-month single-arm pilot. Participants received access to the Huckleberry app and completed surveys at baseline and 3-month follow-up. App use was passively recorded and dichotomized into high-engagement and low-engagement based on frequency of sleep logging. Linear mixed-effects models included Group, Time, and Group×Time interaction, adjusting for demographic covariates. Data analysis was performed using R.

Results: 74 participants (92.5%) completed the study. Most participants (87.8%) self-reported using the app. Participants logged sleep a median of 15 days (IQR: 9–21), and about a quarter received a personalized sleep plan. Overall, infants woke less often overnight and slept longer, while parents reported modest gains in sleep duration and reductions in depressive and stress symptoms (p < 0.05 for time effects in mixed models) over the course of the study period. There was equal distribution by engagement category, n=37 in each group. High engagement participants logged sleep more often (mean: 28.7days, SD: 26.8) than low engagement participants (mean: 1.3, SD:1.1; p< 0.01) and were more likely to obtain a personalized sleep plan (45.9% vs 0%, p< 0.01). The overall improvement in infant and parental sleep metrics, depressive, and stress symptoms observed did not vary significantly by engagement. However, infant longest overnight sleep was one hour longer in the high-engagement group (β = +1.08 h; p = 0.12), translating to an estimated 70% greater gain in consolidated sleep.

Conclusion(s): Delivering an infant-sleep guidance app in a Medicaid ACO was feasible, with high retention and measurable engagement. In our cohort, the modal baseline longest stretches of sleep clustered around 4–6 hours, so an added hour in the high-engagement group represents roughly a 15–25% increase in consolidated nighttime sleep, an increment that plausibly supports more stable sleep–wake patterns and daytime self regulation during early infancy.

Figure 1. Infant nightly sleep duration over time by engagement group
Huckleberry Figure 1.pdf