Telehealth Home Monitoring and Postcardiac Surgery for Congenital Heart Disease

This randomized clinical trial was conducted between 2012 and 2016 at 3 university-affiliated pediatric cardiac centers. Parent-infant dyads were randomly assigned to receive the REACH intervention or usual care. Randomization was performed by using a permuted-block design stratified by SV and biventricular physiology. The study was approved by each center’s institutional review board and written informed consent was obtained from each participating parent.

Infants with complex CHD (Risk Adjustment for Congenital Heart Surgery [RACHS-1] category¹¹  ≥2) were included if they underwent cardiac surgery during the first 21 days of life and were expected to be discharged from the hospital within the first 2.5 months of life. We excluded infants with genetic disorders and other syndromes (except DiGeorge syndrome) and/or cardiomyopathy and/or those awaiting heart transplant. Primary caregivers were included if the family had a cellphone or home Internet access, were English speaking and reading, and were ≥18 years of age. Parents were approached for consent after their infant had undergone surgery and was extubated. Mothers were enrolled primarily, unless the father was the primary care provider.

All families in both treatment groups were discharged from the hospital with a digital scale, pulse oximeter, and access to an online educational Web site. All families were instructed to record their infant’s weight, caloric intake, oxygen saturation, and heart rate daily. All families received usual postdischarge care; specifically, infants with biventricular physiology were scheduled for a follow-up visit within 2 weeks after hospital discharge, and infants with SV physiology were managed by the SV home monitoring program at each center. Included in the SV home monitoring program were a weekly phone review by an advanced practice nurse (APN) of daily weight, caloric intake, oxygen saturation, and heart rate, access to the APN as needed, and once or twice monthly cardiology visits alternating with visits with the pediatrician. “Red flags,” which indicated that there was a significant change in the infant’s status and should be addressed, were predefined by the SV home monitoring program. The SV home monitoring programs at all 3 recruiting sites had at minimum the inclusions noted.

In addition to usual care, REACH intervention primary caregivers received a daily text message, e-mail, or phone call with twice weekly videoconferencing with an APN for 4 months after hospital discharge or until readmission for second stage surgery for infants with SV. Each morning for the first month and weekly thereafter, the parents received the automated communication through the Buddy Check Network (Caryl Technologies, Bryn Mawr, PA) asking about the infant’s status, including total feeding intake, vomiting during feedings, morning weight, heart rate, oxygen saturation, and activity. The automated communication also inquired whether the parent desired a phone call from the study APN. Red flags, which indicated that there was a significant change in the infant’s status and should be addressed were predefined for the REACH intervention protocol. The APN at each site was notified automatically by text when infant data were sent. Parents were contacted by the APN if data did not meet preset criteria for infant stability. The APN triaged any problems and consulted as needed with an on-call study pediatric cardiologist.

The REACH virtual home visits were conducted twice weekly by study APNs using either FaceTime or Skype. Study protocols were used to detail the content of virtual home visits, which included a review of daily and weekly infant data, answers for parent questions about cardiac and newborn care, reconfirmation of the parents’ knowledge of care issues, and decision trees for unscheduled parent calls. To maintain treatment fidelity, the study APNs were trained to conduct the REACH virtual home visits using case-based scenarios, and 20% of all virtual home visits were audited during the weekly team meetings for adherence to study protocols. The full REACH protocol has been published previously.¹² 

Family and infant demographics, medical history, and hospital course variables were extracted from the infant’s medical record. Families were given the Posttraumatic Diagnostic Scale (PTDS) and Ulm Quality of Life Inventory for Parents (ULQIE) questionnaires before discharge to be completed during the first 2 weeks at home. End of the study questionnaires were mailed to the family 4 months after hospital discharge. Physiologic data were uploaded to the Buddy Check Network. Biweekly virtual home-visit interviews collected qualitative parental stress and infant status data.

The primary parental outcome was parenting stress as measured by using the Parenting Stress Index (PSI) (long form).¹³  The PSI is used to measure parenting stress and comprises 6 child domain subscales and 7 parent domain subscales as well as a total score and a life score. The PSI has norms based on large and diverse nationally representative samples and has been used extensively in multiple nonclinical and clinical populations. Construct validity, incremental validity, and treatment sensitivity have been rated as excellent.¹⁴ 

Secondary parental outcomes included a PTSD screen as measured by using the PTDS,¹⁵  which is a self-report measure based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria. QoL was measured by using the ULQIE, which is used to assesses QoL in parents of chronically ill children.⁴  The primary infant outcomes were World Health Organization weight-for-age z scores at study completion and number of unplanned hospital readmissions during the study period. Growth failure was defined as weight-for-age z score ≤−2.0.

Descriptive statistics were used to characterize all parent and infant measures, including demographic, medical history, and hospital course variables. Means and SDs were used to describe continuous variables. Frequencies and percentages were used to describe dichotomous and categorical variables. Differences in demographic and outcome variables were examined between the REACH intervention and control groups by using Fisher’s exact tests for categorical variables and 2-sample t tests for continuous variables. Secondary analyses were performed within functional SV and biventricular subsets. A Bonferroni adjusted P value <.017 was considered statistically significant. All analyses were performed by using SAS version 9.4 (SAS Institute, Inc, Cary, NC).

Assuming 80% power, a Bonferroni corrected type-I error rate of 0.017 to account for 3 primary outcomes (parental stress [total PSI score], infant weight-for-age z score at 4 months, and number of readmissions in each group) with 2-sided 2-sample t tests, 108 parent-infant dyads were needed to detect clinically meaningful differences equal to one-half of a SD in parental stress and infant weight-for-age z score and a 46% reduction in the number of rehospitalizations in the intervention group.

Two-hundred nineteen parent-infant dyads were enrolled, which represented 63% of those eligible and available to participate; 109 were randomly assigned to the intervention group and 110 to the control group (Fig 1). The primary reason for not enrolling eligible patients was either parental refusal or parents not available to consent. Outcome data were complete on 178 parent-infant dyads (94 [86%] intervention, 84 [76%] control). Attrition was related to patient deaths before hospital discharge (2 in each group), withdrawal of consent (10 intervention and 20 control patients), and hospital discharge >2.5 months postsurgery (2 intervention and 3 control).

Overall, parents assigned to the REACH intervention (n = 94) completed a median of 29 (interquartile range [IQR] 22–34) virtual home visits, 76% of what was expected. Parents of infants with SV CHD (n = 52) completed a median number of 28 (IQR 20.5–33) of expected virtual home visits; 16 of 52 parents (31%) completed all virtual home visits. Parents of infants with biventricular CHD (n = 42) completed a median number of 32 (IQR 26–35) expected virtual home visits; 19 of 42 (45%) completed all expected visits.

Demographic and medical history data were similar between the 2 groups (Table 1). Most infants were term, male (56%), white (79%), and non-Hispanic (90%). Eighty-one percent had a prenatal cardiac diagnosis with an almost equal number of infants with SV and biventricular congenital heart defects. RACHS-1 mortality category scores were high, with 37% assigned a score of 5 or 6, indicating a cohort with significant CHD complexity. Statistically significant differences in ventricular physiology (P < .001) were observed between the treatment groups with a greater proportion of SV infants in the intervention group and a greater proportion of biventricular infants in the control group. No significant differences in the number of unplanned reinterventions, hospital length of stay, or ventilation days were observed between the 2 groups. Parent participants were primarily mothers (97%) in their thirties, were college educated (53%), and had private insurance (71%). Parents had a median of 2 children in their household (Table 1).

There were no significant differences in parenting stress between the 2 groups (see Table 2). At the end of the study, 18% of parents in both groups met criteria for PTSD. The mean total symptom severity score was 10 ± 1. There were no significant differences in parent QoL between groups (P = .88). For all parents, the mean ULQIE score (76 ± 13) was comparable to the scores of parents of children with developmental disabilities.³ 

In total, 65 infants (30%) exhibited growth failure at the end of the study. There were no statistically significant differences in growth failure between treatment groups (25% for the intervention group versus 34% for the control group; P = .26).

Biweekly virtual home visits and end of study interviews revealed high levels of stress and anxiety in parents. Parents described their worries about their infant’s health, lack of weight gain, and impending second stage surgery for the SV infants as well as daily management tasks.

There were significant differences in the number of hospital readmissions between groups. Fewer infants in the control group were readmitted (66% vs 57%), and more infants in the intervention group had >1 hospital readmission (15% vs 7%; P < .001).

When several post hoc analyses of parent and infant subgroups were performed, the key findings did not change. Subgroup analyses including SV or biventricular cohort (Tables 3 and 4), race, parental education, income, number of children, weight-for-age z score, number of medications, and feeding mode were performed.

In this prospective randomized trial, we confirmed that the level of stress remains moderately high in parents of infants undergoing surgery for complex CHD. Nearly 20% of mothers met the criteria for PTSD, and their QoL is similar to other parents of children with developmental disabilities and pediatric chronic diseases like cancer.³  In addition, the infants with complex CHD continue to struggle with slow weight gain after hospital discharge, and one-third experience growth failure. Additional support with a telehealth intervention at home did not improve parent or infant outcomes.

In previous studies, it has been noted that parent anxiety often peaks in the days and weeks after hospital discharge because families typically no longer have immediate access to the supportive nursing and medical staff.¹⁶  Several previous studies revealed that parents of neonates who undergo cardiac surgery experience high levels of stress.^17,18  In these studies, researchers indicate that the demands and anxiety of caring for an infant postcardiac surgery is often overwhelming. This is critically important because long-term parental stress is thought to negatively affect one’s capacity to optimally parent both the infant with CHD and the other children in the household.⁸ 

This study was based on Brooten’s transition to home model of care conceptual model.¹⁹  She reported that premature infants and their families successfully transitioned to home earlier with fewer infant complications and parental stress with the support of APN home visits. We believed that our adapted transition model with the use of telehealth would also decrease maternal stress as a result of the APN interactions during a biweekly “virtual home visits,” vigilant infant monitoring, and 24/7 availability. It is difficult to explain the difference in the PSI results, mainly moderate to high average subscales, in comparison to the stress described during the biweekly virtual home visits (D.A.F., unpublished observations). In focus groups, mothers describe their daily stress levels as often overwhelming.²⁰  It may be that the PSI was not sensitive enough to capture the stress related to caring for an infant after neonatal heart surgery.

The REACH intervention was used to provide support on a biweekly basis, helping parents cope with an issue or issues that the family was facing at that specific time regarding infant care, particularly feeding and weight gain issues. As reported by families, the use of the REACH intervention decreased stress at a specific time point. However, these intense interactions between parents and the APN over 4 months were not enough to mitigate long-term stress, as measured by using the PSI. Clearly, there is a need for additional alternative interventions, such as counseling or meditation, to further mitigate parental stress and these should be investigated in future studies.

The high rate of PTSD found in our study is consistent with previous investigations.²¹  In general, we are unable to determine which parents were at highest risk for PTSD. With further prospective inquiry, we may be able to identify a PTSD phenotype on the basis of a cluster of socio-behavioral variables unrelated to their infants’ heart defect. Characterizing this phenotype would allow early identification and targeted supportive interventions in the high-risk group. Until we have these data, all parents of infants with complex CHD should be routinely screened for PTSD, and parents who screen positive for PTSD should be referred for counseling.

Using the ULQIE, Goldbeck et al⁴  assessed the QoL of parents of children with CHD ranging from 7 to 20 years old. The parent respondents in our trial in both the intervention and control groups rated their own QoL similarly to the German parents of children with CHD in the original validation study. The parent respondents in our study reported a decrease in QoL, like that of parents of children with cancer and cystic fibrosis. The high levels of stress described by our families likely contributed to the lower QoL scores noted. Given the lack of effectiveness of our intervention on parent stress, it is not surprising that there was no difference in QoL between the intervention and control groups.

The REACH intervention was novel for infants with CHD.¹²  The telehealth platform performed well, facilitating automated and timely transmission of clinical data between parents and study APNs. This study reveals the feasibility of performing virtual home visits by using widely available telecommunication products in the infant CHD population. Previous home monitoring programs were clinical demonstration programs aimed at improving infant weight gain and reducing infant mortality.^21,22 

Contrary to our hypothesis, the REACH intervention did not reduce parental stress or PTSD, improve the parent’s QoL, or improve infant somatic growth or reduce infant readmission rates. The duration of intervention may have been too short, or the frequency of contacts during the intervention may have been too infrequent. We questioned whether implementing the intervention decreased the burden of home decision-making with the availability of the APNs or imposed more of a burden. We hypothesized that families of infants with SV would benefit the most from participating in REACH, but this was not supported by using our findings. The home monitoring for SV infants in the control group was more intense than the standard of care for infants with biventricular physiology, which may explain similar infant outcomes. The heterogeneity of the biventricular group, with more infants with less complicated postdischarge care demands, may have diluted the benefits of the REACH intervention in this group.

We considered a number of issues that may have contributed to the negative findings of this trial. Given that outcomes were nearly identical between treatment groups, we do not anticipate that these findings would change with a larger sample size. Drop-out rates were low, and compliance with the intervention was good. It is possible that the additional burden of data collection and communication required of parents assigned to the REACH intervention group adversely impacted stress and QoL. Furthermore, although most clinicians are likely reassured by serial data measurement in their patients, it is possible that parents may have greater anxiety and feel a greater burden when obtaining serial clinical information on their own infant. Additionally, although parents assigned to the intervention group were reassured by easy and frequent access to the study APNs, APNs do not have formal education in psychological counseling and stress reduction techniques. Interventions with a strong focus on parental counseling should be included in future studies.

We believe the higher readmission rates in the intervention group were due to the strict guidelines for evaluation (red flags) in the REACH intervention protocol. The advantage of the Buddy Check Network was that the APN was notified immediately when data were received, and the data were, therefore, evaluated daily. Likewise, if the daily data were not received when expected, parents were also contacted to make sure all was well. All of the readmissions were evaluated by the admitting cardiologist to be timely and necessary.

The study was conducted at 3 large pediatric academic cardiovascular centers, and, thus, the findings may not be widely generalizable to all centers. We did not design this study to allow post hoc analysis of who might derive benefit from a telehealth intervention. It may be that parents who are technologically savvy, those with little extended family support, or those living far from treatment centers might benefit. Most parental participants were mothers, and, thus, we cannot comment on whether using the REACH intervention would influence outcomes in fathers.

Using our study results, we confirm that PTSD rates are high in parents of infants who are recovering from cardiac surgery. The intensive telehealth home care REACH intervention was not associated with improvement in parent or infant outcomes when compared with usual care. Further studies are needed to identify impactful interventions.

We acknowledge and thank the members of our data safety monitoring board: Jeffrey Jacobs MD, Caren Goldberg, MD, Patricia O’Brien, RN, MS, and Kathy Mussata, RN, PhD. We also thank members of our research team: Deborah Torowicz; Carrie Alden; Kathleen Van’t Hof; Shobha Natarajan, MD; David Goldberg, MD; and Madelyne Greene. Finally, we thank our families who participated. In addition, we thank our attending physicians who were part of our research team: David Goldberg, MD, and Shoba ha Natarajan, MD.

APN

advanced practice nurse

CHD

congenital heart disease

IQR

interquartile range

PSI

Parenting Stress Index

PTDS

Posttraumatic Diagnostic Scale

PTSD

posttraumatic stress disorder

QoL

quality of life

RACHS-1

Risk Adjustment for Congenital Heart Surgery

SV

single ventricle

ULQIE

Ulm Quality of Life Inventory for Parents