|Home | About | Journals | Submit | Contact Us | Français|
To determine if early developmental intervention (EDI) improves developmental abilities in resuscitated children.
This was a parallel group, randomized controlled trial of infants unresponsive to stimulation who received bag and mask ventilation as part of their resuscitation at birth and infants who did not require any resuscitation born in rural communities in India, Pakistan, and Zambia. Intervention infants received a parent-implemented EDI delivered with home visits by parent trainers every other week for 3 years started the first month after birth. Parents in both intervention and control groups received health and safety counseling during home visits on the same schedule. The main outcome measure was the Mental Development Index (MDI) of the Bayley Scales of Infant Development, 2nd edition, assessed at 36 months by evaluators unaware of treatment group and resuscitation history.
MDI was higher in the EDI (102.6±9.8) compared with the control resuscitated children (98.0±14.6, one-sided p=0.0202) but there was no difference between groups in the non-resuscitated children (100.1±10.7 vs. 97.7±10.4, p=0.1392). The Psychomotor Development Index (PDI) was higher in the EDI group for both the resuscitated (p=0.0430) and non-resuscitated children (p=0.0164).
This trial of home-based, parent provided EDI in children resuscitated at birth provides evidence of treatment benefits on cognitive and psychomotor outcomes. MDI and PDI scores of both non-resuscitated and resuscitated infants were within normal range, independent of early intervention.
Failure to initiate or sustain spontaneous breathing at birth, also called birth asphyxia,1 is a leading cause of perinatal mortality, neonatal encephalopathy, intellectual disability, cerebral palsy, and other childhood neurodevelopmental disorders,2,3 particularly in low- and middle-income countries (LMIC).4 Birth asphyxia accounts for about 23% of the 3.5 million neonatal deaths that occur each year worldwide, 98% of which occur in LMIC.5 About 30% of infants who survived following birth asphyxia in a LMIC had abnormal neurological exams at 2 months, accounting for 50% of the infants referred for an abnormal neurological examination at this age.4 An estimated one million children who survive birth asphyxia each year develop problems such as learning difficulties, cerebral palsy, and other disabilities.6 Birth asphyxia is estimated to result in a loss of over 41 million disability adjusted life years, one of the leading causes for all ages worldwide.7 Mortality and morbidity from birth asphyxia disproportionately affect more infants in LMIC.
About 6 to 10% of all infants fail to initiate or sustain spontaneous breathing at birth and need some assistance to establish normal breathing.8,9 Resuscitation at birth decreases fresh stillbirths10 and early neonatal mortality.9,11,12 Although neonatal resuscitation training programs for perinatal health care providers are standard in high- income countries, these programs have had limited penetration in many LMIC, in part because of the concern of saving infants at risk for neurodevelopmental disorders.
It has been estimated that more than 200 of the approximately 700 million children under five years of age fail to reach their potential for cognitive development.13 Programs of early developmental intervention (EDI) use structured experiences in an attempt to prevent or limit impaired cognitive function early in life. Positive effects of EDI have been demonstrated in numerous controlled trials,14,15,16 but few studies have been conducted in infants with birth asphyxia and in LMIC. Although modifiable biological and psychosocial risk factors and possible interventions have been determined,17 a panel of international experts recommended trials to determine effective and scalable parental EDI strategies with larger and more diverse patient populations using strategies of outreach to disadvantaged children.16 Furthermore, the most effective EDI programs appear to be those that provide direct learning experiences to children and families, are targeted toward younger and disadvantaged children, are of longer duration, high quality, and high intensity, and are integrated with family support and health and nutrition counseling,16 but there is no consensus on whether or not EDI programs should be recommended for these patients in LMIC.
The current trial, the Brain Research to Ameliorate Impaired Neurodevelopment: Home-based Intervention Trial (BRAIN-HIT; registered at ClinicalTrials.gov: NCT00639184), was designed to test the primary hypothesis that a well designed16 home-based, parent-implemented early intervention program improves cognitive abilities as indicated by a higher Mental Developmental Index (MDI) at 36 months on the Bayley Scales of Infant Development 2nd edition (BSID-II) among a group of infants who received bag and mask ventilation as part of their resuscitation but did not have severe encephalopathy during the neonatal period compared with a control group. To enable comparison, a group of infants from the same communities who did not require resuscitation at birth were also randomized concurrently to the same intervention or control home visits.
This parallel design randomized controlled trial was implemented in two populations: (1) infants with birth asphyxia unresponsive to stimulation and the initial steps of resuscitation who received bag and mask ventilation; and (2) infants who did not require any resuscitation. Infants in each cohort were randomized individually to one of two trial conditions (EDI plus health and safety counseling or control which included health and safety counseling only) using 1:1 concealed parallel allocation, matched for country and chronological time using variable block sizes to assure allocation concealment (Figure; available at www.jpeds.com). There were no changes in trial design or outcome measures following commencement of enrollment other than home visits were continued every other week between 12 and 36 months of age instead of the planned every fourth week visits to increase the intensity of the intervention. The trial was approved by the institutional review boards at the University of Alabama at Birmingham, Research Triangle Institute (RTI) International, and each participating clinical site. Details on the trial design have been published.18
Infants who received bag and mask ventilation for resuscitation at birth, in rural, poor communities in three sites in India, Pakistan, and Zambia during and immediately following the First Breath Trial10 were screened for enrollment into this trial. Birth asphyxia was defined as the inability to initiate or sustain spontaneous breathing at birth using the WHO definition.1 Infants were ineligible if they met any of the following exclusion criteria: 1) the birth weight was less than 1500 grams at birth, 2) their neurological examination at seven days was severely abnormal (grade III by Ellis classification),19 3) the mother was < 15 years of age or unable/unwilling to participate, or 4) the mother was not planning to stay in the study communities for the following three years. Infants with birth asphyxia (resuscitated) and infants without birth asphyxia or other perinatal complications (non-resuscitated) matched for country and chronological time born from January 2007 through June 2008 were randomly selected during the 7-day follow-up visit after birth from infants enrolled in the First Breath Trial.10 A list of potential enrollees was distributed to the investigators in each country to obtain consent for the trial. Written informed consent was obtained during the second week after birth following the seven day neurological assessment and before randomization.
A home-based, parent-implemented EDI model was selected to strengthen parent-child interaction. The Partners for Learning 20 curriculum and supplemental materials were used by parent trainers to introduce playful interactive learning activities depicted on cards given and modeled to the parents during home visits using the Portage Model.21 Partners for Learning covers a full spectrum of competences, organized into the four areas: 1) cognitive and fine motor, 2) social and self-help, 3) gross motor, and 4) language skills. Investigators at each research site selected EDI parent trainers (high school graduates in two sites and four year graduates after high school in one site) who were trained in an initial five-day workshop at each research site. A second workshop was conducted before participating children began to reach 18 months of age to adapt the approach to children up to 36 months. Each parent-child pair was assigned to the same trainer throughout the trial whenever possible. Home visits were conducted every other week from enrollment (at 1–2 weeks) thru 36 months. During each visit, the trainer presented one or two playful interactive learning activities using cards. Each activity targeted a developmentally appropriate competence. The parent practiced the activity in the presence of the trainer who provided feedback to the parents. The activity cards were left with the parents who were encouraged to apply the targeted activities by integrating them into daily life with the child until the next home visit. Cycles of use allowed the parents to implement several activities and move on to new activities as the child mastered each competence. The trainer introduced new activities to match and enhance the child’s developmental competences. The trainers used a log book to record their impressions about each visit. To review progress with the activities, the trainers were supervised with observations during actual home visits and constructive feedback was provided. Adherence to the intervention was measured by home visits completed on schedule within its assigned two week window following the preceding visit and parental report of frequency of implementation of the activities between home visits
Parents in both the control and intervention groups received health and safety counseling during every home visit. Health and safety counseling content was based on the WHO health education curriculum,22 which included breast feeding, nutrition, hygiene, safety in the home and community, awareness of danger signs, management of diarrhea, and well-child checkups and vaccinations. The health and safety counseling was provided by both the intervention and control trainers who had participated in a 5-day training conducted separately from the EDI training at each research site.
The primary outcome measure was the BSID-II MDI.23 Secondary outcome measures included the psychomotor scale of the BSID-II (PDI), the Ages and Stages Questionnaire, 2nd edition (ASQ),24 the Ages and Stages Questionnaire: Social-Emotional (ASQ:SE), and health data. The BSID-II was selected as the main outcome measure for this trial because it had been used extensively in a number of LMIC. The BSID-II underwent extensive pre-testing at each site to verify validity in the local context, and a few items were slightly modified to make it more culturally appropriate (e.g., image of a sandal instead of a shoe). The BSID-II was administered directly to each child in the appropriate language using standard material. The ASQ, which is a screening tool previously used in LMIC,25 was used to assess parent-reported child development observed in the home environment in the domains of communication, gross motor and fine motor, problem solving, and personal-social development using age-specific forms. To remove issues of literacy, the ASQ was administered in an interview format with the parent. Health data were collected on weight, height, head circumference, breastfeeding, diet, immunization status, hearing and vision impairment, and childhood illnesses. Assessments of outcomes occurred when the children were 12, 24, and 36 months of age and were performed by evaluators trained in 4-day workshops before each yearly evaluation. The evaluators were familiar with local language and culture and unaware of the participants’ resuscitation history and treatment group. Study evaluators made every attempt to successfully evaluate the children on or shortly after the child’s third birthday. If an evaluation could not be completed due to severe behavioral problems uncharacteristic of the child or an acute illness, the evaluation was rescheduled and the child was retested when possible. Outcome evaluation data at 36 months are presented here.
To test the hypothesis that the EDI improves the MDI on the BSID-II, at least 40 children in each of the four groups had to have completed the intervention and three-year assessment (assuming a 5% significance level, a power of 90%, a difference between the groups on the BSID-II of 10 points favoring the early development group, and a standard deviation of 15 using a one-tailed test). Because of high attrition and missing values in our pilot trial of 80 infants (24% lost to follow up, 20% withdrew, and 4% died),4 at least 60 infants (a minimum of 20 per site) were required to be enrolled in each of the four groups.
Maternal and neonatal data were collected by trained birth attendants who were supervised by community coordinators as part of the First Breath Trial.10 In addition, data on family demographics and resources were collected at enrollment in the trial through the administration of a modified World Bank tool in a structured interview and included maternal age, education (none/illiterate, none but literate/primary school, at least some secondary school), family assets, and home living standard. The presence of 11 family assets (e.g., radio, TV, refrigerator, bicycle) were tallied as a Family Resources Index. A Home Living Standard Index was calculated based on status regarding seven indicators (e.g., home building material, number of rooms, water source). Data on each visit were collected by the parent trainers. The data were transmitted electronically to the data coordinating center (RTI). Data edits and inter- and intra-form consistency checks were performed. Adverse events were reported to the Data Monitoring Committee, which monitored the study progress and made reports and recommendations for continuation of the trial.
Descriptive statistics were generated for monitoring enrollment and retention, completion of home visits, and completion of the 36-month evaluation. For the ASQ and ASQ:SE, only subjects who completed the evaluation within the manual-specified windows of 34–39 months for the ASQ and 34–41 months for the ASQ:SE were included in these analyses. Mean differences in 36-month evaluation outcomes were determined between the two treatment groups for each resuscitation cohort using t-tests. All subjects included in the analyses were assigned the treatment to which they were randomized. Additionally, an exploratory analysis combining the resuscitated and non-resuscitated children was performed to assess differences in 36-month evaluation outcomes by treatment group (ie, intervention versus control). This post-hoc analysis was based on general linear models. Although a one-sided t-test (alpha=0.05) as specified by the protocol was employed to test the primary outcome, differences between the means for all other continuous measures were tested using a two-sided t-test (alpha=0.05). Point and interval estimates of relative risk were calculated using standard large-sample formulae for categorical measures. Differences in categorical measures were tested using chi-square and Fisher exact tests.
Cutoff values indicating moderate delay (score <85) and severe delay (score <70) for the BSIDII were set at one and two standard deviations from the instrument standardized mean (100±15 points). Standard cutoffs24 for each ASQ developmental domain were used to distinguish children requiring a referral for further neurodevelopmental evaluation from those operating within normal range.
The infants enrolled in this trial were derived from a total of 540 births screened (Figure). Of these, 438 were eligible and 407 (93%) were consented; 36 (9%) infants were not assessed at 36 months because of death (n=16), withdrawal (n=10), loss to follow up (n=8), or other reasons (n=2) with losses comparable across treatment arms within resuscitation cohorts. Six subjects who completed the 36-month evaluation discontinued the home visit treatment prior to the end of the study but their data were included in all analyses. Maternal and infant characteristics were similar in the intervention and control group (Table I; available at www.jpeds.com) except for a higher rate of prenatal care in the control group non-resuscitated infants. Adherence to the home visits through the 36-month evaluation was 96.3% and 97.8% in the early intervention and control groups, respectively. Among the intervention group, parent report of adherence to the implementation of the intervention activities over the three years of the study among those who completed the 36-month evaluation was an average of 62.1% of the days among resuscitated infants and 61.5% of the days among non-resuscitated infants. This equates to practicing the intervention activities an average of 4.3 days per week in both groups.
The BSID-II MDI and PDI scores were successfully measured in 123 resuscitated and 170 non-resuscitated children for a total of 293/371 (79%) of the 36-month evaluations. Two children who scored within normal limits for the 12- and 24-month evaluations did not successfully complete the 36-month BSID-II evaluation. Additionally, evaluations done by one evaluator were excluded from the BSID-II analysis (n=76) because the evaluator consistently scored the children much lower than the other evaluators. Specifically, the evaluator’s mean MDI score was 13 points lower than the next lowest evaluator and 17 points lower than the overall mean MDI. The 76 evaluations were evenly distributed across the treatment groups by resuscitation status. The demographics of the 76 children did not differ from evaluations completed by other evaluators at the affected clinical site. The primary outcome was computed both with and without the evaluator’s records to confirm that removing the data from this evaluator did not change the determination of treatment effect.
The primary outcome measure, MDI at 36 months, among the 123 resuscitated children with a successful BSID-II evaluation was 102.6±9.8 in the early intervention group and 98.0±14.6 in the control group for a mean difference and corresponding 95% confidence interval of 4.63 (0.93, ∞) (one-sided p=0.0202) (Table II). Including the records from the excluded evaluator resulted in a statistically significant treatment effect of slightly lesser magnitude (mean MDI score of 98.3±13.0 in the early intervention group and 94.2±15.6 in the control group; mean difference (95% CI) 4.05 (0.18, ∞), one-sided p=0.0425). Among the 170 non-resuscitated children with a successful BSID-II, MDI was 100.1±10.7 in the early intervention group and 97.7±10.4 in the control group (p=0.1392). A linear model that included effects for treatment, resuscitation cohort, and their interaction showed no evidence of a differential treatment effect across the two cohorts (interaction p=0.4158). A reduced model that included the two main effects showed small, non-significant differences of resuscitation cohort (resuscitated=100.3±1.0 [adjusted mean ± SE], non-resuscitated=98.9±0.86 [p=0.3238]) and a significant treatment effect (intervention arm=101.3±0.94, control arm=97.9±0.93 [p=0.0131]).
The treatment effect remained beneficial after controlling for potential confounders including study site, maternal demographics, socioeconomic status, characteristics of the child’s birth, and the child’s demographics. The results were comparable with the unadjusted results (105.2±4.9 for the intervention arm and 101.4±4.8 for the control arm [p=0.0143]) with both arms well within the normal range. The child’s resuscitation status remained unassociated with MDI score in this model (101.2±5.1 for resuscitated infants and 105.4±5.5 for non-resuscitated infants [p=0.3669]). For the combined group, MDI was higher in the early intervention children (101.1±10.4) compared with the control children (97.8±12.4), p=0.0141.
PDI was over five points higher for the intervention group when compared with the control group among the resuscitated (p=0.0430), non-resuscitated (p=0.0164), and combined (p=0.0022) groups of children (Table II). The ASQ and ASQ-SE scores did not differ between the intervention and control groups for both resuscitated and non-resuscitated children (Table III). At 36 months, severe hearing and vision impairment occurred in one control resuscitated infant and severe vision impairment occurred in one control non-resuscitated infant. The prevalence of infants with one or more anthropometric measure below 5th percentile was comparable between the early intervention and control conditions (39 to 46% for weight, 69 to 74% for height, and 26 to 40% for head circumference). Harm or unintentional effects were not observed.
The results indicate that the children’s cognitive and psychomotor abilities as measured by the BSID-II were significantly higher in the intervention group when compared with the control group. This trial tested an EDI program started soon after birth in children in LMIC at high risk for developmental disorders. The intervention was integrated with family support and nutritional and health education as recommended.16 The effect size (4.6 points in MDI), though smaller than hypothesized, was comparable with that observed in a large trial of children with various perinatal disorders in India (3 points in MDI)26 but substantially smaller than that of a similar trial in children with birth asphyxia in China (14 points in MDI).27 The diversity of participants enrolled in the current multi-country trial and the similar effect size to the Indian trial suggest that EDI may be effective but the benefits are relatively small. It is possible that the control visits every two weeks resulted in some benefits, reducing the effect size.
The study population consisted of survivors of birth asphyxia from a large multi-country trial in which training of birth attendants in neonatal resuscitation and other essential newborn care knowledge and skills reduced stillbirths and perinatal mortality.10 Training of midwives with the same educational program reduced neonatal and perinatal mortality in a large multicenter study in a low-income country.11 It is reassuring that the current trial shows that the overwhelming majority of the resuscitated infants did not have an increased rate of neurodevelopmental impairment at 36 months and that the mean BSID-II scores in survivors were in the normal range, indicating that resuscitation increased unimpaired survivors. Since the time of study inception, the BSID-II has largely been replaced by the BSID-III. The BSID-III covers similar skills and abilities as the BSID-II at 36 months. However, even though the average BSID-II results approximate 100, if we had used the new BSID-III our study may have resulted in higher scores as reported by others.28
A limitation of this study is that the role of the parent trainer in the intervention was not as intense as in other intervention studies conducted in high income countries.14,15,29 In order to integrate the developmental intervention program with other family support and nutrition and health counseling, it was implemented through the parents. It is possible that parental adherence was less intense than desirable but the activities were practiced on average more than 50% of the days throughout the three years of the intervention. We do not have independent observations of the quality of the implementation of the intervention at home but data on frequency of practice of the activities were collected prospectively during each home visit. The intervention lasted only three years, and it is possible that a longer intervention may be necessary.
Another limitation is that infants were evaluated at three years when neurodevelopmental indicators may be less predictive of long term outcomes than later evaluations. Although outcomes may improve from infancy to school age,30,31 studies have found that individual stability of scores is only moderate to high.32 It is possible that larger benefits may be observed at school age or that more survivors will have problems that manifest during childhood and adulthood.33,34 Follow-up of 636 former low birth weight infants who participated in a multicenter early (first three years) intervention trial showed some persistent benefits at 18 years.35 Thus, larger studies and longer follow-up to school age are needed. The control group received health and safety counseling every two weeks and it is possible they received some benefits from the trial which reduced the effect size. However, although this is the largest trial of early intervention in infants with birth asphyxia in LMIC, the small effects of the intervention would require a larger sample size to confirm the effectiveness noted in the combined group in this trial. The high prevalence of infants with anthropometric measures below the 5th percentile may be due in part to the low socioeconomic status of these families in comparison with the reference group of the WHO study, which was purposely designed to produce standards among healthy children living under conditions likely to favor the achievement of their full genetic growth potential.36
Because the definition of birth asphyxia varies markedly between studies, a comparison of results may be difficult. The current study used the inclusive WHO definition of birth asphyxia: failure to establish spontaneous breathing at birth which includes neonatal respiratory depression.1 Thus, the high incidence of survivors with normal outcome is not surprising. With moderate severity of birth asphyxia, there is an increased risk of mortality and the survivors are at increasing risk of neurodevelopmental impairment.33,34 With the most severe birth asphyxia, mortality and morbidity increased substantially.37,38 Thus, although birth asphyxia is a major cause of mortality and disability, the degree of birth asphyxia markedly influences the long-term outcome. The high rate of mild to moderate asphyxia may be due to the effectiveness of the First Breath Trial intervention as well as the high early mortality due to asphyxia in this population.10 This trial documents that the overwhelming majority of infants with mild to moderate degrees of birth asphyxia in LMIC are likely to have intact survival.
Even though programs for early intervention are common in high income countries, trials to assess the effect of EDI for infants following birth asphyxia have not been conducted to our knowledge. A relatively small single center randomized controlled trial (n=64) in infants who survived birth asphyxia in China showed an increase (14 points) in MDI (105±15 in the early intervention group versus 91±11 in the control group, p<0.001).27 The effect size was about one standard deviation which is a relatively large effect compared with the effects reported in a meta-analysis. 38 Similar to our study, the intervention was provided by the parents but contacts with trainers occurred less often than in our trial (monthly during the first year and every two months during the second year). A trial of early intervention in infants at risk for impairment in India included 665 infants, 218 of whom had mild or moderate birth asphyxia.26 For the group as a whole, MDI was slightly higher in the EDI group than in the control group (83±14 versus 80±13, p<0.005) but was low in both groups overall. Data were not reported separately for infants with birth asphyxia but the overall effect size was comparable with that observed in the current trial.
In summary, the current trial of an EDI program implemented through parents from soon after birth to three years of age in infants with birth asphyxia provides evidence of a treatment benefit on cognitive and psychomotor outcomes. Both MDI and PDI improved by approximately 5 points in the EDI group compared with the health and safety counseling control group. Survivors of mild or moderate birth asphyxia had neurodevelopmental outcomes in the normal ranges. A review that addresses the role of interventions in children with early developmental difficulties was published by WHO.39 The current trial used methods as proposed by this consensus of experts, including advancing policy, use of common international approaches, increasing local capacity, empowering caregivers, and conducting research in low-resource income settings. Furthermore, the current trial documented benefits in children at low-risk for neurodevelopmental difficulties.
Funded by Eunice Kennedy Shriver National Institute of Child Health and Human Development Global Network for Women’s and Children’s Health Research and the National Institute of Neurological Disorders and Stroke (HD43464, HD42372, HD40607, and HD40636), the Fogarty International Center (TW006703), the Perinatal Health and Human Development Research Program, and the Children’s of Alabama Centennial Scholar Fund of the University of Alabama at Birmingham. W.C. is on the Mednax Board of Directors.
Brain Research to Ameliorate Impaired Neurodevelopment: Home-based Intervention Trial (BRAIN-HIT) investigators include: Jawaharlal Nehru Medical College, Belgaum, India: Roopa M. Bellad, Sangappa M. Dhaded, Niranjana S. Mahantshetti, Bhalachandra S. Kodkany, and Shivaprasad S. Goudar; Aga Khan University, Karachi, Pakistan: Omrana Pasha and Zahid Abbasi; University of Zambia, Lusaka, Zambia: Elwyn Chomba; University of Alabama at Birmingham, Birmingham, Alabama: Waldemar A. Carlo and Fred J. Biasini; Research Triangle Institute, Durham, North Carolina: Hrishikesh Chakraborty, Elizabeth M. McClure, Dennis Wallace, and Vanessa Thorsten; Christiana Care, Wilmington, Delaware: Richard J. Derman; Drexel University, Philadelphia, Pennsylvania: Robert L. Goldenberg; Western Kentucky University, Bowling Green, Kentucky: Darlene Shearer; University of California, Merced, California: Jan Wallander; and National Institutes of Child and Human Development, Rockville, Maryland: Linda L. Wright.
Eunice Kennedy Shriver National Institutes of Child Health & Human Development (NICHD) Global Network for Women’s and Children’s Health Research Investigators include: University of Alabama at Birmingham, Birmingham, Alabama: Waldemar A. Carlo; University Teaching Hospital, Lusaka, Zambia: Elwyn Chomba; Center for Disease Research in Zambia, Lusaka, Zambia: Albert Manasyan; Christiana Care, Wilmington, Delaware: Richard J. Derman; KLE University’s Jawaharlal Nehru Medical College, Belgaum, India: Shivaprasad Goudar and Bhalchandra S. Kodkany; Drexel University, Philadelphia, Pennsylvania: Robert Goldenberg; Aga Khan University Medical College, Karachi, Pakistan: Omrana Pasha; Research Triangle Institute, Durham, North Carolina: Beth McClure, Dennis Wallace, and Vanessa Thorsten.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The other authors declare no conflicts of interest.