OBJECTIVE: The objective of this study was to determine the effects of preventive care at home on child development and primary caregiver mental health at 2 years of age.
METHODS: A total of 120 very preterm infants (<30 weeks) were assigned randomly to intervention (n = 61) or control (n = 59) groups. The intervention group received the preventive care program (9 home visits over the first year from a physiotherapist and a psychologist, focusing on the parent-infant relationship, the parents' mental health, and the infant's development); and the control group received standard care. At corrected age of 2 years, developmental outcomes were assessed, and primary caregivers completed the Infant-Toddler Social and Emotional Assessment. The mental health of the primary caregivers was assessed with the Hospital Anxiety and Depression Scale.
RESULTS: At 2 years of age, 115 children (96%) were assessed with the Bayley Scales of Infant and Toddler Development III and 100 children (83%) with the Infant-Toddler Social and Emotional Assessment; and 91 (86%) of 106 caregivers completed the Hospital Anxiety and Depression Scale. There were no statistically significant differences in cognitive, language, or motor composite scores between the treatment groups. However, children in the intervention group were reported by their primary caregivers to exhibit less externalizing and dysregulation behaviors and increased competence, compared with control subjects. Primary caregivers in the intervention group reported less anxiety and depression.
CONCLUSION: A preventive care program for very preterm infants and their families improved behavioral outcomes for infants and reduced anxiety and depression for primary caregivers.
WHAT'S KNOWN ON THIS SUBJECT:
There have been very few randomized, controlled trials of early developmental interventions for preterm infants that have had long-term effects on developmental outcomes for children born preterm. Furthermore, it is important to target caregiver well-being.
WHAT THIS STUDY ADDS:
This study demonstrated that an early preventive care program for very preterm infants (<30 weeks of gestation) and their families improved infants' behavioral outcomes and reduced primary caregivers' anxiety and depression at corrected age of 2 years.
Because survival rates for very preterm infants (<32 weeks of gestation) have stabilized, more research should be directed toward strategies to reduce the high rates of neurobehavioral disabilities among survivors.1 Recently, several treatments before hospital discharge, including magnesium sulfate given to women at risk of very preterm birth2,3 and caffeine4 and docosahexanoic acid5 administered after birth, have been reported to improve outcomes for surviving, very preterm infants. However, there are many other potential factors influencing development, such as parenting6 and early environmental experiences.7
Very preterm infants at high risk of abnormal development should commence intervention programs as soon as possible, to take advantage of the increased plasticity of the developing brain.1 However, the evidence of effectiveness of intervention for all very preterm infants is inconclusive. Early developmental interventions after hospital discharge for preterm infants have been shown to improve cognitive development in infancy and at preschool age but not later in childhood,8 with the exception of one study that demonstrated a benefit at 18 years of age for infants born weighing >1500 g.9 There is little evidence of effects on motor development. It is important that intervention programs focus not only on the children but also on the parents, because caregivers of preterm infants are at increased risk for emotional distress, which is associated with short- and long-term consequences for their children.10,11 In summary, the benefits of early intervention for very preterm infants are not fully established, and additional, high-quality, randomized, controlled trials focusing on both parents and children are needed.
The aim of this randomized, controlled trial was to determine whether a specifically designed, early intervention program after hospital discharge for families of very preterm infants enhanced child development and caregiver mental health over the first 2 years of life. It was hypothesized that the program would improve both developmental outcomes (cognitive, language, motor, and social-emotional behavior) and caregivers' mental health.
A comprehensive outline of the research protocol was published previously.12 Infants born at gestational ages (GA) of <30 weeks, with no major congenital brain anomalies associated with poor neurodevelopmental outcomes, were eligible for inclusion in this study. Infants were excluded if the family did not live within a 100-km radius of the hospital or spoke no English. Most infants were recruited from the Royal Women's Hospital (Melbourne, Australia), with a small number of infants (n = 6) cared for elsewhere.
Infants were recruited at term-equivalent age, with consent to participate.12 The study was approved by the Royal Women's Hospital and Royal Children's Hospital research ethics committees.
Study Time Line and Protocol
Baseline Data Collection
Extensive baseline perinatal data were collected from medical records and from parent questionnaires. Before random assignment, the primary caregiver completed the Edinburgh Postnatal Depression Scale13 and questionnaires on sociodemographic information, from which families were categorized as being at lower or higher social risk.14
MRI and Qualitative Scoring
At postmenstrual ages of 38 to 44 weeks (before random assignment), infants underwent MRI to assess the presence or absence of moderate/severe white matter abnormalities, with the use of previously published protocols.15,–,17
After they consented to participate, participants were assigned randomly to either the intervention group or the control group, which received standard care targeted for families of preterm children. Random assignment was stratified according to the severity of white matter injury on MRI scans (moderate/severe injury present or absent or no MRI data) and singleton versus multiple births, to ensure equal distribution of important brain injuries and multiple births between treatment groups. Children from multiple births were assigned randomly to the same group because the intervention was family-based. The allocation sequence for each stratum was computer-generated independently, and assignments were concealed by using opaque envelopes.
The intervention designed by the Victorian Infant Brain Studies (VIBeS) team, called VIBeS Plus, consisted of 9 visits over the first year of life, conducted by 2 teams comprising a psychologist and a physiotherapist who were specially trained to deliver the intervention program. The intervention aimed to educate the primary caregivers about infant self-regulation and techniques for improving postural stability, coordination, and strength and to support the parents' mental health and parent-infant relationship throughout the first year. Each session lasted ∼1.5 to 2.0 hours and was conducted in the family home, with a few exceptions in which the infants were seen in the hospital. The program was designed on the basis of the literature on infant development, parent mental health, and the parent-infant relationship and incorporated components of previously successful interventions.18,–,20 Additional details of the intervention are available in the protocol publication.12
Both the intervention and control groups received standard follow-up care, which was not systematic; however, each family had access to a maternal and child health nurse in the community, who assessed the developmental progress of the child among other well-child health tasks, such as weighing and immunizations. Although the majority of children were recruited from the Royal Women's Hospital, they might have been transferred to a local hospital for ongoing care. Standard follow-up care, including surveillance and early intervention, varies greatly between hospitals in Victoria. Referral to early intervention services might have been made by the infant's health care team at any time during the study. To account for this variability in follow-up care, a record sheet for visits to doctors, nurses, and other health care professionals was given to families, so that we could document access to other interventions.
The primary outcomes were cognitive, language, and motor development at corrected age of 2 years, as assessed with the Bayley Scales of Infant and Toddler Development III,21 which were administered by examiners who were masked to group allocation. The Bayley Scales of Infant and Toddler Development III are normative value-referenced assessments, with means of 100 and SDs of 15 points. Children whose scores were 2 SDs below the normative value in a domain were classified as having a significant delay in that domain.
Secondary outcomes were child behavior and emotional regulation at corrected age of 2 years, which were assessed by using the Infant-Toddler Social and Emotional Assessment, a parent-report questionnaire suitable for children 12 to 36 months of age.22 The primary caregiver completed this questionnaire, which assesses 4 broad domains of behavior, namely, behavioral dysregulation, externalizing behavior problems, internalizing behavior problems, and competencies. Higher scores are worse for dysregulation, externalizing behavior problems, and internalizing behavior problems, and lower scores are worse for competencies. Mean scores of ≤10th percentile for competencies and ≥90th percentile for externalizing behavior problems, internalizing behavior problems, and dysregulation indicate at-risk findings.
At child corrected age of 2 years, the mental health of primary caregivers was assessed by using the Hospital Anxiety and Depression Scale,23 which includes 14 questions assessing the mental health of the respondent. Higher scores indicate more-severe symptoms of anxiety or depression. For both anxiety and depression scales, scores of >7 indicate at-risk symptoms.23
Sample Size Estimation
The sample size calculation was based on the primary child outcome of cognitive outcomes at corrected age of 2 years. The research team considered an important clinical difference to detect would be an improvement of ≥0.4 SD (6 points) on the cognitive scale of the Bayley Scales of Infant and Toddler Development III.21 With a type I (α) level of .05 and 80% power, we required 100 subjects in each group (ie, a total sample size of 200). Because of funding restrictions, however, the sample size was limited to 120, which, with allowance for strong correlation between the 14 twin pairs in this data set (intraclass correlation coefficient of 0.7 for Bayley Scales of Infant and Toddler Development III cognitive scores), provided 53% power to detect a difference of 0.4 SD. Although this was not ideal, 120 subjects provided 70% power to detect a difference of 0.48 SD and 80% power to detect a difference of 0.54 SD, which are reasonable differences that may arise.
Statistical analyses were performed by using Stata 10.0 (Stata, College Station, TX). Univariable linear regression models were used to examine the differences in child outcomes between the 2 treatment groups, with the use of robust (sandwich) estimators of SEs to allow for clustering of twins. Results are presented as means and 95% confidence intervals (CIs) for the intervention effect for each outcome. The possibility of differing treatment effects according to gender and social risk was explored with the inclusion of main effects and interaction terms for each outcome. Odds ratios (ORs) and 95% CIs are presented from logistic regression models for the binary outcomes of abnormal cognitive, motor, and language scores (cutoff values of <2 SDs) and abnormal scores in the externalizing, internalizing, and dysregulation (cutoff values of ≥90th percentile) and competence (cutoff value of ≤10th percentile) domains, again using sandwich estimates of SEs to allow for clustering of twins. In sensitivity analyses, all analyses were repeated with adjustment for white matter injury and social risk. Fifteen infants did not undergo MRI to classify white matter injury; however, all underwent cranial ultrasonography before term age, which showed no abnormalities, and therefore these infants were grouped with the children with no moderate/severe white matter injury for these analyses.
Caregiver outcomes were compared between groups by using simple linear regression analyses, with parents of multiple births being included in the analyses only once. In sensitivity analyses, analyses were repeated with adjustment for baseline depression (Edinburgh Postnatal Depression Scale scores of >12) and social risk. ORs and 95% CIs are presented from logistic regression models for the binary outcomes of caregiver anxiety and depression (cutoff values of >7).
Participants and Demographic Characteristics
Between January 2005 and January 2007, 343 infants born at GAs of <30 weeks were assessed for eligibility, and 152 were ineligible (Fig 1). Of the 191 eligible infants, 12 infants were not recruited because of workload issues and 59 families declined to participate, which resulted in 120 infants being recruited, including 14 sets of twins, and 106 families. There were no significant differences in GAs or birth weights between the 120 infants who were recruited and the 71 nonparticipants.
Of the 120 recruited infants, 105 (88%) underwent MRI (Fig 1). Infants did not undergo MRI because of lack of parental consent (n = 13) or because they were too ill (n = 2). In the twins 12 sets had the same MRI classifications and 2 sets had discordant classifications. The baseline characteristics were similar in the 2 groups, with a slightly larger proportion of primary caregivers with elevated depression scores in the control group, compared with the intervention group (Table 1).
Fifty-seven infants (95%) in the treatment group received all 9 intervention sessions. Two families completed 6 or 7 sessions (because of illness or family vacations), 1 child died at corrected age of 2 months (as a result of a congenital cardiac defect), and 1 infant withdraw from the study (primary caregiver returned to full-time employment after the first visit). Another infant, who completed the intervention program, died prior to the 24 month assessment. There were no withdrawals or deaths in the control group during the first year of life (period of the intervention program). At the end of the intervention, the families of 92 (77%) of the 119 surviving infants returned the questionnaire that monitored visits to other health care professionals and other early interventions. There was no evidence of a difference between the control and intervention groups with respect to access to other early interventions, such as physiotherapy (intervention, 32%; control, 42%; χ2 = 0.96; P = .328).
At corrected age of 2 years, 115 (97%) of 118 survivors returned for follow-up evaluations (Fig 1). Of those 115 infants, all completed the cognitive and motor composites of the Bayley Scales of Infant and Toddler Development III, and 111 completed the language composite (which was not administered if English was not the primary language spoken at home). There was little evidence of differences in cognitive, language, or motor composite scores between the intervention and control groups (Table 2); the mean difference in cognitive scores was an improvement of 0.26 SD with the intervention. There also was no evidence of differences between groups in the proportions of children classified as having delays in motor, cognitive, or language domains. Adjustment for white matter injury and social risk altered no conclusions.
Infant-Toddler Social and Emotional Assessment questionnaires were returned for 100 children (87%); 3 questionnaires had missing data related to the competence domain, which resulted in 97 children (84%) having complete data for all domains. On average, children in the intervention group had lower (better) scores for externalizing and dysregulation behaviors and higher (better) competence scores, compared with those in the control group (Table 3), with children in the 2 groups having similar results for the internalizing domain. Adjustment for social risk and white matter injury altered no conclusions. Twenty-four percent of children in the control group were classified as being at risk with respect to the dysregulation domain, compared with 0% in the intervention group (Fisher's exact test, P < .0001). Although the odds of being classified as being at risk in the internalizing, externalizing, and competence domains for the intervention group were not statistically significant because of the small numbers of children with these problems, children who received the intervention had lower odds of having a social-emotional behavioral problem in each domain. Children in the intervention group had much lower odds of being classified as at risk in ≥1 domain, compared with children in the control group (21% vs 50%; OR: 0.26 [95% CI: 0.10–0.70]; P = .007).
At baseline, the Edinburgh Postnatal Depression Scale was completed by 98 (92%) of 106 primary caregivers, with slightly more parents with elevated scores in the control group (Table 1). Ninety primary caregivers (85%) completed the Hospital Anxiety and Depression Scale at 2 years. On average, the primary caregivers in the intervention group reported lower (better) levels of anxiety and depression on the Hospital Anxiety and Depression Scale, compared with those in the control group (Table 4). This was also reflected by fewer caregivers having scores in the at-risk range for both anxiety and depression. The results showed similar estimates with adjustment for depression scores at baseline and social risk (data not shown).
The major findings from this study are that this program of post–hospital discharge intervention throughout the first year for families of preterm infants resulted in less dysregulation and externalizing behavior problems, as well as improved competence, for the children in the intervention group; however, it did not improve Bayley Scales of Infant and Toddler Development III cognitive, language, or motor composite scores significantly. Although the change in cognitive scores was not statistically significant, the 0.26-SD increase with the intervention was better than that observed in 2 of the 8 randomized, controlled trials contributing to the Cochrane review of early intervention and was consistent with the overall increase of 0.46 SD favoring intervention in that meta-analysis.8 Importantly, primary caregiver anxiety and depression also were reduced in the intervention group, compared with the control group.
The VIBeS Plus program was purposely targeted to less-mature preterm infants (gestational ages of <30 weeks), compared with previous trials.8 This population is at greater risk of adverse neurodevelopmental outcomes than are infants born at ≥30 weeks.24 Studies by Kaaresen et al25 and Koldewijn et al26 also lowered the mean gestational age range of recruitment, by restricting entry to infants with birth weights of <2000 g and <1500 g, respectively. The study by Kaaresen et al25 was based on a modified version of the Mother-Infant Transaction Program27 and focused on the interaction between parents and their preterm infants, with the aim of preventing cognitive and behavior problems for the children and reducing parents' stress. The authors reported that, with the intervention, which consisted of inpatient sessions and 4 visits after discharge, parents were less stressed at corrected age of 2 years; however, there were no differences in child outcomes, including motor, cognitive, and behavioral development.28 Koldewijn et al26 reported promising findings in favor of their intervention, which consisted of intervention in the hospital and 6 to 8 sessions after discharge up to 6 months of age, with improvements in cognitive, motor, and behavioral outcomes for very preterm infants at 6 months. The length of the intervention after discharge may influence outcomes, particularly for the children, because both our study and the study by Koldewijn et al26 demonstrated improvements in outcomes when the intervention was delivered until ≥6 months after term age.
A major strength of our intervention is that it targets both infant development and primary caregiver mental health. Infants and parents share a unique relationship, and it is essential that interventions focus on the infant/parent dyad. Our study is one of the first to report improvements in both child behavior and primary caregiver mental health at corrected age of 2 years. Another strength of our study was the use of MRI to stratify cases according to brain injury severity, which ensured that the 2 groups had equal numbers of infants with potential for adverse development. The inclusion of infants with severe brain injury and those from multiple births improved the generalizability of the results. The VIBeS Plus program had very high follow-up rates (97%), compared with other studies of early developmental interventions after discharge. However, the questionnaire return rate at 2 years was lower.
There are several limitations to this study. Many of our outcomes, including those that were significant, were obtained through caregiver questionnaires rather than through independent observations. However, the fact that primary caregivers reported improved infant behavior as well as better mental health for themselves is important. Another limitation is that the Bayley Scales of Infant and Toddler Development III may not be sensitive enough to detect subtle but important differences in development. We reported recently that the Bayley Scales of Infant and Toddler Development III underestimated developmental delay in Australian, very preterm and term, 2-year-old children, which may explain the low rates of developmental delay.29 The assessments were conducted before the emergence of many important skills, and reassessment of the cohort at a later age, when higher-order skills can be assessed reliably, is essential to determine whether there are important, long-term, cognitive benefits for the children. An additional limitation was that the recruitment target of 200 subjects was not reached, which reduced the power to find a clinically important improvement of 0.4 SD in the developmental scales and also constrained subgroup analysis because of reduced power.
Although previous literature findings suggested that early developmental interventions improve cognitive outcomes up to preschool age, there is limited evidence for any effect on motor development. This does not mean that preterm infants should not be referred for early developmental interventions to improve motor development but emphasizes the limitations of these programs when applied to all high-risk infants as a prevention program, rather than a specific treatment program to be delivered if a child has a specific developmental problem. Greater selectivity for high-risk populations may identify infants who may benefit most from any specific, targeted form of intervention. Furthermore, although the results favored psychological (behavior and primary caregiver well-being) rather than physiotherapeutic (motor) outcomes, this intervention involved equal input from physiotherapists and psychologists, and we would not recommend excluding physiotherapy from the program. The physiotherapists provided support to families, including information on caregiver-infant interactions, handling, and play, which might influence both general child outcomes and parent confidence.
Our preventive care program of post–hospital discharge intervention for very preterm infants and their families improved parent-reported behavioral outcomes for infants and decreased the rates of primary caregiver-reported anxiety and depression but did not have significant effects on cognitive, language, or motor development of the children at corrected age of 2 years. Reassessment of the children and their families at a later age is vital for determining the longer-term benefits of this program.
We acknowledge support from the National Health and Medical Council (project grant ID 284512 and career development award ID 473840, to Dr Boyd), the Cerebral Palsy Foundation (project grant and postdoctoral fellowship, to Dr Spittle), the Murdoch Childrens Research Institute, the Myer Foundation, Allens Arthur Robinson, and the Thyne Reid Foundation.
- Accepted March 5, 2010.
- Address correspondence to Alicia J. Spittle, PhD, Murdoch Childrens Research Institute, Victoria Infant Brain Studies, 2nd Floor, Royal Children's Hospital, Flemington Road, Parkville, Victoria, Australia, 3052. E-mail:
This trial has been registered with the Australian New Zealand Clinical Trials Registry (identifier ACTRN12605000492651).
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- CI =
- confidence interval •
- GA =
- gestational age •
- OR =
- odds ratio •
- VIBeS =
- Victorian Infant Brain Studies •
- WMI =
- white matter injury
- Wang CJ,
- McGlynn EA,
- Brook RH,
- et al
- Doyle LW,
- Crowther CA,
- Middleton P,
- Marret S,
- Rouse D
- Treyvaud K,
- Anderson VA,
- Howard K,
- et al
- Als H,
- Duffy FH,
- McAnulty GB,
- et al
- Spittle AJ,
- Orton J,
- Doyle LW,
- Boyd R
- McCormick MC,
- Brooks-Gunn J,
- Buka SL,
- et al
- Carter JD,
- Mulder RT,
- Bartram AF,
- Darlow BA
- Boyce P,
- Stubbs J,
- Todd A
- Spittle AJ,
- Brown NC,
- Doyle LW,
- et al
- Inder TE,
- Anderson NJ,
- Spencer C,
- Wells S,
- Volpe JJ
- Inder TE,
- Warfield SK,
- Wang H,
- Huppi PS,
- Volpe JJ
- Dolby R,
- Warren B,
- Heath J,
- Meade V,
- Conney G
- Bayley N
- Carter AS,
- Briggs-Gowan MJ
- Kaaresen PI,
- Ronning JA,
- Ulvund SE,
- Dahl LB
- Copyright © 2010 by the American Academy of Pediatrics