Cincinnati Children’s Hospital Medical Center and the Pediatric Heart Network have announced the BOND trial, a multicenter adaptive platform study testing HeartGPS psychological support and a direct breastfeeding clinical practice guideline for families facing critical congenital heart disease. Announced on July 14, 2026, the National Heart, Lung, and Blood Institute-funded program plans to recruit 350 families through 17 United States heart centers over three years and follow them from prenatal or neonatal diagnosis through the infant’s first year.
The trial addresses a part of congenital heart disease care that technological progress alone has not resolved. Advances in prenatal diagnosis, surgery and cardiac intensive care have improved survival, but families can still encounter anxiety, traumatic stress, disrupted bonding, feeding difficulties and substantial practical strain. The BOND trial will test whether structured support can be incorporated earlier and delivered more consistently.
This remains a planned clinical investigation rather than evidence that either intervention improves family or infant outcomes. The registered HeartGPS and breastfeeding domains were listed as not yet recruiting around the announcement, and meaningful conclusions will depend on enrolment, adherence, retention, prespecified analysis and several years of follow-up.
How will the BOND adaptive platform test two different gaps in congenital heart disease care?
BOND stands for Building Optimal Strategies to Enhance Parent and Infant Wellbeing in Congenital Heart Disease. Its adaptive platform structure is designed to evaluate multiple supportive-care interventions within a shared research framework, allowing investigators to study different elements of family-centered care more efficiently than a series of disconnected trials.
The first announced domain will evaluate HeartGPS, a virtually delivered psychological intervention beginning after a prenatal congenital heart disease diagnosis. A second domain will test a clinical practice guideline intended to increase direct breastfeeding among infants with critical congenital heart disease during and after intensive care.
The overarching announcement refers to 350 families, while the individual registry records estimate 210 participants for HeartGPS and 266 for the breastfeeding domain. Those figures should not be added together and interpreted as 476 separate families. BOND is a multifactorial platform in which eligibility and participation may overlap across domains, while each intervention retains its own design, comparator and endpoint.
That distinction matters because BOND is not testing a single treatment against a single control. HeartGPS uses an individually randomized, two-arm design comparing usual fetal cardiac care with usual care plus psychological support. The breastfeeding domain uses a cluster design in which participating clinical centers are randomized to a multifaceted guideline or conventional care.
The platform therefore combines two related but analytically separate questions. One asks whether a prenatal psychological intervention can reduce maternal anxiety. The other asks whether changing institutional practices can increase direct breastfeeding among medically complex infants.
Why does HeartGPS begin during pregnancy rather than after cardiac intensive care admission?
HeartGPS is designed for parents who learn between 16 and 30 weeks of pregnancy that their fetus has congenital heart disease expected to require surgery or a transcatheter intervention shortly after birth. The timing is central to the hypothesis because psychological distress can begin at diagnosis, well before the infant enters a cardiac intensive care unit.
The intervention’s name refers to guiding parents through emotions, providing information and support, and strengthening connections. Participants assigned to HeartGPS are expected to receive eight virtual counselling sessions, tailored educational resources and an individualized mental health care plan in addition to usual fetal cardiac care.
Delivering the program virtually may make specialist psychological support available across participating centers without requiring every hospital to employ a large onsite perinatal cardiac psychology team. It may also reduce travel demands during a pregnancy already complicated by specialist appointments and preparations for neonatal intervention.
Telemedicine does not remove every access barrier. Participation still requires reliable connectivity, sufficient privacy for counselling and the ability to attend repeated sessions while managing medical appointments, work and family responsibilities. The registered HeartGPS domain permits assessments in English or Spanish, but that still leaves questions about how the model could eventually serve families speaking other languages.
Eligibility criteria will also affect generalizability. The domain excludes postnatal diagnoses, multiple pregnancies and several maternal or fetal circumstances that could interfere with safe participation or interpretation of infant development. These controls can strengthen the study’s internal validity, but the eventual findings may not apply equally to every family treated in congenital heart disease programs.
What will the HeartGPS primary endpoint establish, and what will it leave unanswered?
The HeartGPS registry identifies maternal anxiety at four months as the primary endpoint, measured using the state component of the State-Trait Anxiety Inventory. This is a more precise objective than the broad idea of helping families thrive. A positive result would indicate that the intervention reduced measured anxiety against usual fetal cardiac care under the conditions of the trial.
The study will also examine depressive symptoms, traumatic stress, parent-infant bonding, maternal stress physiology and infant development. These outcomes could help investigators understand whether any psychological effect extends beyond a questionnaire score and whether the timing of support matters for the wider family experience.
However, the trial should not be interpreted as demonstrating improved infant cardiac outcomes merely because psychological and developmental measures are being collected. Congenital heart disease severity, genetic factors, surgical complexity, intensive care exposures and socioeconomic conditions can all influence infant development. Separating an intervention effect from those overlapping factors will require careful adjustment and transparent reporting.
Usual fetal cardiac care may also differ between participating centers. Some hospitals already integrate psychologists, social workers or peer-support programs, while others provide more limited services. That variability reflects real-world care, but it could narrow or widen the apparent difference between HeartGPS and the control group.
The trial primarily measures outcomes involving mothers and infants. Although the announcement uses the broader language of families, the registered primary population and endpoint may leave less direct evidence about fathers, partners, siblings and other caregivers. Future domains could be required if the Pediatric Heart Network wants the platform to address the entire family system.
Can a cluster-randomized breastfeeding guideline change practice across cardiac intensive care units?
The breastfeeding domain tackles a different problem. Infants with critical congenital heart disease may experience early separation from their mothers, delayed oral feeding, surgery, mechanical ventilation and extended intensive care stays. These factors can make direct breastfeeding difficult even when expressed breast milk remains available.
Participating centers will be randomized to conventional care or a multifaceted clinical practice guideline. The proposed package includes prenatal education, breastfeeding information tailored to congenital heart disease, access to registered lactation consultants, repeated inpatient support, contact after discharge, breast-pump provision and incentives supporting continued pumping or direct breastfeeding.
This is therefore an institutional care pathway rather than one isolated intervention. If direct breastfeeding rates improve, the study may not be able to attribute the effect to counselling, equipment, staff contact or another individual component. Its practical value would instead lie in determining whether the entire implementation package changes behaviour at the hospital level.
The registered objective is to increase the proportion of infants receiving direct breastfeeding at four months. Investigators will continue follow-up through 12 months and evaluate maternal and infant measures, but a higher breastfeeding rate should not automatically be presented as proof of fewer medical complications or improved development. Those outcomes must be analysed independently.
Eligibility also narrows the population. The breastfeeding domain is intended for infants expected to undergo cardiac surgery or transcatheter intervention within 30 days of birth and to enter intensive care during their first week. Registered exclusions include gestational age below 36 weeks, birth weight below the third percentile, certain genetic or anatomical conditions affecting feeding and any medical circumstance that makes direct breastfeeding unsafe.
Those exclusions are clinically understandable, but they mean that some of the infants experiencing the greatest feeding difficulty may not be represented. Hospitals assessing future adoption will need to understand precisely which congenital heart disease population generated the evidence.
Why does the adaptive platform offer research efficiency while creating analytical complexity?
An adaptive platform can use shared infrastructure, standardized assessments and participating centers to evaluate several interventions over time. New domains can potentially be introduced as unanswered questions emerge, while ineffective approaches can be revised or discontinued according to prespecified rules.
That flexibility is particularly relevant to family-centered congenital heart disease care, where many possible interventions involve psychology, nutrition, bonding, developmental support and clinical workflow. Studying each component through a separate multicenter trial could take many years and duplicate recruitment and data-collection systems.
Flexibility also requires statistical discipline. Investigators must account for multiple interventions, possible overlap between domains, differences in participant eligibility and changes in usual care over calendar time. Any adaptation must follow rules established before investigators know the comparative results.
The two initial domains also randomize at different levels. HeartGPS assigns individual participants to intervention or usual care, whereas the breastfeeding domain assigns clinical sites. Cluster randomization helps prevent the same hospital from applying a new feeding pathway inconsistently across enrolled families, but site-level differences in staffing, patient mix and existing lactation practices can complicate interpretation.
Recruiting 350 families with infants facing major early cardiac intervention will be another practical test. These families will be approached during periods of substantial medical and emotional pressure. Retention through multiple assessments and the child’s first year could be difficult, particularly when families face repeated admissions, travel or financial strain.
Missing data may not be random. Families experiencing the greatest stress or most complicated clinical courses could be less able to complete counselling sessions and assessments. The eventual analysis will need to show whether attrition differed between groups and whether the conclusions remain credible under appropriate sensitivity analyses.
What could prevent positive BOND results from becoming routine family-centered cardiac care?
If HeartGPS reduces maternal anxiety, implementation would still require medical psychologists trained in perinatal mental health and congenital heart disease. Hospitals would need referral pathways, virtual-care infrastructure, scheduling capacity and plans for families requiring treatment beyond the structured eight-session program.
Financing may be equally important. Research funding can support a controlled trial, but routine delivery would depend on hospital budgets, insurance coverage, telehealth reimbursement and the availability of qualified professionals. Congenital heart centers already managing expensive surgical and intensive care programs may require evidence that psychological support improves outcomes meaningful to families and health systems.
The breastfeeding intervention carries a different resource burden. Consistent delivery may require registered lactation consultants who can work within cardiac intensive care, suitable pumps, staff education and coordinated follow-up after discharge. Smaller centers may struggle to reproduce a model developed across large pediatric heart programs.
A statistically positive result will therefore be only the first step. Hospital leaders will want to know the size and durability of benefit, the number of counselling or lactation contacts required, adherence rates, staffing costs and whether the intervention performs consistently across demographic and clinical subgroups.
The study could also produce mixed findings. HeartGPS may improve maternal anxiety without changing infant developmental measures, while the feeding guideline may increase direct breastfeeding without demonstrating measurable changes in medical outcomes. Such results would still be informative if the endpoints are interpreted according to what each domain was actually designed to test.
Which milestones will determine whether BOND can influence congenital heart disease care?
The first milestone is operational: participating centers must move from registry preparation into active recruitment and demonstrate that families can enrol during the narrow prenatal and neonatal windows. Investigators will then need to show that counselling sessions, assessments and feeding-support procedures can be delivered consistently across geographically dispersed hospitals.
The HeartGPS registry estimates primary completion in April 2029 and full study completion in June 2030. The broader BOND timeline could evolve as the adaptive platform adds, modifies or completes intervention domains, making periodic registry updates important for understanding the program’s actual progress.
The most consequential evidence will be the prespecified comparison between HeartGPS and usual fetal cardiac care on maternal anxiety at four months, followed by transparent reporting of secondary psychological, bonding and developmental outcomes. For the breastfeeding domain, attention will center on direct breastfeeding at four months, implementation fidelity and whether exploratory health measures support further investigation.
BOND is unlikely to resolve every psychological, developmental and feeding challenge associated with congenital heart disease. Its more realistic contribution is to determine whether structured family support can be tested with the same multicenter discipline increasingly expected of surgical and medical interventions.
If the platform produces credible results, it could give pediatric heart centers a stronger basis for deciding when psychological and lactation support should begin, who should deliver it and which outcomes justify continued investment. Until those data arrive, BOND represents an ambitious evidence-building program, not confirmation that either intervention has already changed outcomes.
