Early Childhood Development Caregiver Training and Neurocognition of HIV-Exposed Ugandan Siblings

Abstract

Objective:

Early childhood development (ECD) programs can enhance neurocognitive development outcomes through caregiver training. This study explores whether school-age siblings benefited from a program provided to HIV-infected caregivers and their preschool-aged target children.

Methods:

Siblings of target 2- to 3-year-old children in ECD intervention households were evaluated at school age (5–12 years) on neurocognitive outcomes with the Kaufman Assessment Battery for Children (KABC), computerized Test of Variables of Attention, Behavior Rating Inventory for Executive Function (BRIEF; parent), and attention-deficit/hyperactivity disorder rating inventory (ADHD-R)-IV (parent). Households from 18 geographic clusters in eastern Uganda were randomized to individualized biweekly sessions of either (1) Mediational Intervention for Sensitizing Caregivers (MISC) training emphasizing cognitive stimulation/enrichment or (2) health/nutrition/development [Uganda Community-Based Association For Women & Children Welfare (UCOBAC)] program. Siblings with baseline and at least 1 follow-up assessment (n = 216) were included in the analysis. Three repeated postbaseline measures of sibling neurocognitive outcomes were analyzed using the linear mixed-effects model while adjusting for socioeconomic status and behavioral outcome at baseline.

Results:

Siblings in the MISC arm had better performance on KABC sequential processing at 6 months (p = 0.02) and simultaneous processing at 12 months (p = 0.03). MISC mothers rated their children as having significantly more problems on the BRIEF and ADHD-RS-IV (p < 0.01) than UCOBAC mothers across all time points.

Conclusion:

Mediational Intervention for Sensitizing Caregivers training resulted in some short-term neurocognitive benefits for school-aged siblings, but these differences were not sustained at 1-year follow-up. Exploring potential impacts of parenting programs on other children in the home is an important development for the field.

Home-based parenting programs aimed at improving children’s health and well-being have demonstrated success across a range of outcomes in rigorous controlled trials. Many such early childhood development (ECD) programs focus on giving parents and caregivers skills to intervene early in a child’s life to reduce the harmful impact of negative prenatal parental health-related behaviors, dysfunctional infant caregiving, and stressful environmental conditions. Given that 250 million children (43%) younger than 5 years in low-income and middle-income countries are at risk of not reaching their developmental potential, the focus of these programs on their health, nutrition, security and safety, responsive caregiving, and early learning is warranted.¹

However, what has not been explored in these studies is the degree to which other children in the home (i.e., the “nontarget” children) also benefit from these programs. There are 2 potential pathways for other children to benefit from these programs as follows: (1) through the caregiver using their new learned skills with other children and (2) through improvements in the family context and family functioning. Although few researchers have collected data from other siblings living in the homes of children participating in ECD programs, there is some evidence that the benefits of these programs can disperse out to nontarget children. For example, a program aimed at preventing conduct problems in preschool-age siblings of adjudicated youths resulted in reduced antisocial behaviors in nontargeted adolescent siblings.² An early study of a parent-focused program that focused on positive parental receptivity to children and reducing stressful interactions found that the intervention effect extended to both current children and those born after the intervention.^3,4 For physical development outcomes, Ver Ploeg⁵ reported that older siblings of target children who were eligible for food benefits through the Women, Infants, and Children program had healthier eating habits than children in nonparticipating families.

In Uganda, research has documented that children living in HIV-affected households are at a pronounced disadvantage neurodevelopmentally, socially, and psychologically at least in part because of instability in caregiving.⁶ In our preliminary research, caregivers and young children (aged 2–5 years) who received the “parent-child interaction intensive” ECD [Mediational Intervention for Sensitizing Caregivers (MISC)] displayed enhanced child learning and attention, improved parenting, and reduced caregiver depression symptoms compared with other types of enhanced caregiving.^7,8 When MISC was compared with another ECD treatment as usual (TAU) program, the “value-added” neurodevelopmental benefits of MISC for Ugandan preschool children were limited and seemed to depend in part on the health and functionality of the caregiver (mothers with HIV) who participated in the program.^6,9

The aim of the present study is to explore whether the school-age children living in the homes where MISC was provided received any benefit from the caregiver learning the MISC skills as compared with school-age children living with caregivers receiving a comparison ECD program (TAU).

METHODS

Randomization and Masking

The sampling for this study was based on the caregiver-child dyads included in the main trial with target children aged 2 to 5 years; all target children were perinatally HIV exposed but uninfected (HEU).¹⁰ For the current study, we invited caregivers to enroll up to 2 additional children aged 0 to 12 years currently living in the same household as the target child. If more than 2 siblings were living in the household at the time of recruitment, the 2 siblings closest in age to the target child were included. Caregivers enrolled 294 siblings from 221 study households, of whom those age 5 to 12 were included in the present analysis. Children with a documented medical history of traumatic brain injury, birth complications, severe malnutrition, central nervous system infection requiring hospitalization (e.g., meningitis, encephalitis, and severe malaria), or recurring seizures were excluded. In addition, children who regularly (50% of the time or more) lived in another home with another caregiver or family were excluded. Siblings who were age 5 years or older were evaluated with our school-age assessment protocol, described later. The recruitment and follow-up of the siblings took place between March 2012 and April 2014, and caregivers provided written informed consent for themselves and their children. During this study period, the caregivers and the preschool-age target child participated in 1 of the 2 caregiver training programs described below. Method and findings for the early childhood development intervention trial with the target child have been previously described.¹⁰

Caregiver Training Interventions

Mediational Intervention for Sensitizing Caregivers

Mediational Intervention for Sensitizing Caregivers (MISC) is a model for training caregivers to enhance their children’s cognitive development and is based on Feuerstein’s theory of cognitive modifiability.^11–13 MISC was previously adapted to the Ugandan context by the authors, and details of the intervention have been described in detail elsewhere. Briefly, MISC learning is accomplished by training caregivers through regular weekly sessions with the caregiver and their young child in mediational processes as focusing (gaining child’s attention), exciting (communicating affection), expanding (transcending learning experiences), encouraging (providing emotional support), and regulation (direct and shape child’s behaviors).¹⁴ For a more detailed description of the theory and content of the MISC curriculum, see Klein et al.¹⁵ They also personally trained the field team of MISC trainers consisting of 4 Makerere University psychology or social work bachelor’s degree graduates, and these 4 trainers consistently worked with the same caregiving clients throughout the 1-year training period. MISC training of the Ugandan field trainers was done by Bar-Ilan University faculty members Professor Pnina Klein and then by Dr. Cilly Shohet and Deborah Givon in Kampala and at the Tororo study site for several 1-week periods during the initial phase of the study preparations.

As detailed previously,¹⁰ each caregiver was expected to receive 24 trainings at intervals of 2 trainings each month, alternating between office and home every 2 weeks. Each training lasted about 45 minutes. At study enrollment and about every 3 months thereafter, an appointment was given for home video collection and administration of the Home Observation for the Measurement Environment (HOME) Inventory and caregiver assessment. The video was recorded from the home of the caregiver. The recording captured about 5 to 8 minutes of the following 3 activities between the mother and target child: bathing, feeding, and work/play. Before each subsequent training session, a video clip from the most recent home-based video recording was selected and reviewed, and strengths were identified from which MISC lessons were drawn up by the field trainer.

During the training session, the selected clip was presented to the caregiver on a laptop, and 1 or 2 MISC components were taught. The trainer avoided giving more than 3 concepts at a single training. Ideally, only the caregiver (usually the mother) attended the office training, not the participating child. For the home-based training sessions, the trainer witnessed at least 1 practical interaction between the target child and caregiver, choosing 1 or 2 MISC components to share depending on the caregiving interaction quality. The strengths and weaknesses identified during this interaction were noted in the logbook for that child/caregiver dyad. Before starting a new topic, the weaknesses observed in the interaction and how to improve them were discussed.

The trainer used the logbook to plan lessons in advance and to record a summary of the training session including their observations. The questions on the log form were used to guide what was entered in the logbook. The logbooks were used to document fidelity of MISC intervention training, and only older siblings in MISC households in which the caregiver had completed all 24 training sessions were included in the present analysis (over 90% of participating HEU households).¹⁰

Uganda Community–Based Association for Child Welfare Program

The comparison treatment as usual caregiver training arm (henceforth referred to as Uganda Community–Based Association for Women & Child Welfare [UCOBAC]) was a manualized nutrition and hygiene information program designed and originally implemented by the UCOBAC http://ucobac.org/) that met the minimum standard of care for families affected by HIV in Uganda.^7,8 UCOBAC caregiver training was structured in a similar manner to MISC in terms of frequency and duration, with biweekly sessions for 1 year alternating between the dyad’s home and the study office.^7–10 As with MISC, UCOBAC training was supervised by Makerere University graduates. Video segments were not used as part of the UCOBAC caregiver training, but logbook documentation of training sessions was used to gauge fidelity of UCOBAC training, and again, only siblings in caregiver households in which all 24 training sessions had been completed were included in the present analyses.

Measures and Outcomes

Study measures were administered at home (all caregiver and family outcomes) and at the study office (all child outcomes) in 1 of 3 local languages (Japadola, Ateso, or Luganda). Data were collected at baseline, 6 months (midway through caregiver training), 1 year (posttraining), and 2 years (12 months after the intervention). Tests used to assess the neurocognitive and behavioral performance of children have been previously adapted to the present study context in which they demonstrated stable construct validity and were sensitive to education exposure and health indicators.^16–18

Demographics

Child demographics were recorded at baseline and included age, sex, and physical growth (weight, height, and upper-arm circumference). The World Health Organization norms were applied to derive age- and sex-standardized scores for height [height-for-age z-score (HAZ)] and weight [weight-for-age z-score (WAZ)] (https://www.who.int/childgrowth/standards/en/). Caregivers reported on their age, marital status (married/unmarried), education (any/none), and relationship to the study child (mother/other). A “wealth index” of socioeconomic status (SES) was also used as a covariate in the analyses and is described elsewhere.^16,19 We also included scores on the HOME²⁰ as a covariate in the analyses because scores differed by the treatment arm. The HOME is a composite measure that assessed the quality and quantity of stimulation that the target study child was exposed to in their home environment.

Kaufman Assessment Battery for Children, Second Edition

The Kaufman Assessment Battery for Children, Second Edition (KABC-II) is the principal test for cognitive ability outcomes from a neuropsychological perspective.²¹ It is validated in sub-Saharan African HIV-affected children aged 5 years and older.^18,22 Using the Luria model for neuropsychological assessment within the KABC-II, the primary outcome variables were the global scores of Sequential Processing (memory), Simultaneous Processing (visual-spatial processing and problem solving), Learning (immediate and delayed memory), Planning (executive reasoning), Delayed Recall, Nonverbal Index (subtests not dependent on the understanding of instructions in English), and Mental Processing Index (MPI) (a composite of the principal cognitive performance domains).

Test of Variables of Attention

The Test of Variables of Attention (TOVA) is a computerized visual continuous performance test used to screen, diagnose, and monitor children and adults at risk of attention-deficit/hyperactivity disorder (ADHD)²³ and adapted for pediatric HIV research in Uganda.¹⁸ The TOVA consists of the rapid (tachistoscopic) presentation of a large geometric square on the computer screen with a smaller dark box either in the upper position (signal) or lower position (nonsignal). The child is asked to press a switch with the hand as fast as possible in response to the signal (measuring vigilance attention) but to withhold responding to the nonsignal (measuring impulsivity). TOVA’s primary outcome variables are response time variability (a sensitive indication of inattention), response time, percent commission errors (impulsivity), percent omission errors (inattention), an ADHD index score (missed signals in proportion to incorrect responses to nonsignal), and a D-prime signal detection measure of overall test performance (correct signal “hits” in proportion to correct nonresponses to nonsignal).

Behavior Rating Inventory of Executive Function—School-Age Parent Version

The Behavior Rating Inventory of Executive Function—School-Age Parent Version (BRIEF)²⁴ is a questionnaire assessing behavior, attention, and cognitive problems related to disruption of executive functions as reported by the principal caregiver. Indices used in this trial include scales that measure metacognition and behavior regulation indices and are combined into a Global Executive Composite score.

Attention-Deficit/Hyperactivity Disorder-Rating Scale IV

Attention-deficit/hyperactivity disorder symptoms were assessed using parent report on the ADHD-RS-IV, an 18-item scale that assesses the presence and severity of DSM-IV symptoms of ADHD, including a 9-item subscale for inattention and a 9-item subscale for hyperactivity/impulsivity.²⁵

Analysis

Linear mixed-effect (LME) models were used to analyze outcomes at 6, 12, and 24 months while adjusting for SES, and the z-score computed for behavioral outcome at baseline (BRIEF). The LME models used in the analysis allow for a missing at random (MAR) mechanism, so all siblings (aged 5 years or older) with at least 1 postbaseline assessment (N = 216) were included under the MAR assumption. Correlations arising from repeated measures for each child nested within the household were accounted for by specifying the corresponding nested random effect and the autoregressive covariance structure. Inclusion of a random effect for subcounty (unit of randomization) was explored, but the resulting intraclass correlation coefficients were virtually zero across outcomes, indicating that the cluster randomization did not result in an appreciable dependence of outcomes within clusters. This was also the case in the trial with target children as reported earlier.¹⁰

To model potentially nonlinear longitudinal patterns, follow-up time point was specified as a categorical variable. Time-by-intervention arm interaction was included to capture potential changes in differences by the intervention arm over time. The least squares (adjusted) means for each time point and trial arm were output from the LME models, and differences by the trial arm were tested. SAS 9.4 was used for all analyses. Because the UCOBAC intervention focused on nutrition, we also explored the role of anthropometric scores in relation to the KABC MPI by adding HAZ and WAZ (1 at a time) to longitudinal models as time-varying covariates.

RESULTS

Of the 304 siblings in the Mediational Intervention for Sensitizing Caregivers (MISC)/Uganda Community-Based Association For Women & Children Welfare (UCOBAC) HIV-exposed but noninfected cohort households¹⁰ that underwent eligibility screening, 10 were excluded because of a history of hospitalization from infectious disease (e.g., severe malaria and meningitis) or severe malnutrition (Fig. 1). Of the 294 eligible siblings assessed at baseline, 204 were 5 years of age or older and assessed with the school-age battery [Kaufman Assessment Battery for Children, Second Edition (KABC-II), Test of Variables of Attention (TOVA), and Behavior Rating Inventory for Executive Function (BRIEF)]. In addition, 25 siblings who were preschool aged at baseline completed at least 1 school-aged assessment during follow-up and are included in this study. Included in this analysis are siblings with baseline and at least 1 postbaseline assessment (n = 216) (Table 1). Within this sample, there were 181 children with all 3 repeated measures (93 in the UCOBAC arm and 88 in the MISC arm), so that differences between arms of 0.42 of the SD or higher would be detectable at each time point with a power of 0.80 in 2-tailed tests at the 0.05 level of significance. The linear mixed-effects models included all children with one or more nonmissing outcome measure, and power to detect the effect size of 0.42 exceeded 0.80 among N = 216 children analyzed. Study caregivers of the present study siblings were on average 34 years of age, predominantly married (72%), the study child’s biological mother (>90%), and had at least primary education (82%).

Figure 1.

This is the CONSORT diagram for the cluster randomization of dyads to either the MISC training intervention for child cognitive stimulation or the UCOBAC-manualized healthy nutrition caregiver training curriculum. CONSORT, Consolidated Standards of Reporting Trials; MISC, Mediational Intervention for Sensitizing Caregivers; TAU, treatment as usual; UCOBAC, Uganda Community-Based Association For Women & Children Welfare.

Table 1.

Child and Caregiver Characteristics and Outcomes at Baseline by the Trial Arm

Characteristic	Health and Nutrition Curriculum (UCOBAC), N = 109		MISC, N = 107
	N (%)		N (%)
Child
Sex
Male	43 (39)		53 (50)
Female	66 (61)		54 (50)
Wealth group
1	32 (29)		26 (24)
2	52 (48)		71 (66)
3	25 (23)		10 (9)
Characteristic	Health and Nutrition Curriculum (UCOBAC), N = 109		MISC, N = 107
	Mean	SD	Mean	SD
Age, yr	7.84	2.12	7.68	2.07
Height-/length-for-age z-score	−0.71	1.24	−0.52	1.29
Weight-for-age z-score	−1.25	1.00	−0.96	1.03
Score	20.00	2.15	20.67	3.22
KABC Mental Processing Index	60.64	7.95	61.28	6.86
KABC Sequential Processing	71.95	10.04	74.91	10.48
KABC Simultaneous Processing	64.48	9.70	64.42	9.31
KABC Nonverbal Index	59.20	8.89	58.54	8.93
KABC Learning	67.93	10.71	68.15	9.94
KABC Planning	60.37	7.14	59.67	5.40
KABC Delayed Recall	69.01	10.61	69.90	9.41
TOVA response time	690.7	168.6	699.3	181.4
TOVA response time variability	264.2	96.82	259.2	98.72
TOVA D-prime	2.24	1.37	2.33	1.34
TOVA commission errors	12.27	11.78	13.48	14.22
TOVA omission errors	24.82	26.46	22.76	21.61
TOVA ADHD	−3.01	4.12	−2.82	3.21
BRIEF global executive composite school age	49.73	10.29	55.55	12.15
BRIEF global executive composite z-score	−0.25	0.87	0.19	1.03
ADHD symptoms	0.60	0.45	0.72	0.49
Caregiver	Health and Nutrition Curriculum (UCOBAC), N = 109		MISC, N = 107
	N (%)		N (%)
Relationship to child
Mother	98 (90)		98 (92)
Other	11 (10)		9 (8)
Marital status
Married	79 (72)		75 (70)
Other	30 (28)		32 (30)
Education level
No education	21 (19)		20 (19)
Some education	88 (81)		87 (81)
Caregiver age, yr	34.10 (7.64)		34.47 (8.62)

ADHD, attention-deficit/hyperactivity disorder; BRIEF, Behavior Rating Inventory for Executive Function; KABC, Kaufman Assessment Battery for Children; MISC, Mediational Intervention for Sensitizing Caregivers; TOVA, Test of Variables of Attention; UCOBAC, Uganda Community-Based Association For Women & Children Welfare

The results are presented separately for each of the components of the KABC and TOVA measures and for the BRIEF and attention-deficit/hyperactivity disorder (ADHD) outcomes (Table 2). For the KABC outcomes, MISC siblings did significantly better at 6 months on KABC Sequential Processing (working memory) (p = 0.02) and on KABC Simultaneous Processing at 12 months (p = 0.03) compared with UCOBAC siblings. However, these probabilities were no longer statistically significant when a Bonferroni adjust was made for the number of KABC domains compared. By 24 months, the 2 arms were similar, although the 2 arms differed on KABC Learning (p < 0.01 unadjusted; p = 0.07 adjusted). None of the TOVA outcomes significantly differed by the treatment arm.

Table 2.

School-Age Outcomes at 6 Months (Midintervention), 12 Months (Postintervention), and 24 Months (12 Months After the Intervention) by the Study Arm: Unadjusted Means and Adjusted Means

	Unadjusted			Adjusted
Outcome	UCOBAC, Mean (SE)	MISC, Mean (SE)	p-Value (95% CI for the Difference)	UCOBAC LS, Mean (SE)	MISC LS, Mean (SE)	p-Value (95% CI for the Difference)
KABC MPI, mo
6	61.41 (0.78)	62.63 (0.75)	0.26 (−3.36 to 0.93)	61.16 (0.79)	63.16 (0.86)	0.08 (−4.22 to 0.21)
12	62.62 (0.86)	62.94 (0.65)	0.77 (−2.45 to 1.82)	62.62 (0.79)	63.67 (0.85)	0.35 (−3.24 to 1.15)
24	62.73 (0.80)	61.44 (0.70)	0.23 (−0.81 to 3.39)	62.44 (0.79)	62.38 (0.85)	0.96 (−2.14 to 2.25)
KABC Sequential Processing, mo
6	72.33 (1.00)	74.35 (1.00)	0.16 (−4.83 to 0.78)	72.27 (0.99)	75.61 (1.08)	0.02 (−6.13 to −0.54)
12	72.22 (0.88)	73.56 (0.95)	0.30 (−3.89 to 1.20)	72.39 (0.99)	74.99 (1.06)	0.07 (−5.37 to 0.17)
24	71.24 (0.95)	71.56 (1.01)	0.82 (−3.07 to 2.42)	71.21 (1.00)	73.17 (1.05)	0.16 (−4.72 to 0.80)
KABC Simultaneous Processing, mo
6	64.93 (1.06)	66.27 (1.05)	0.37 (−4.28 to 1.61)	63.99 (1.12)	66.56 (1.22)	0.11 (−5.72 to 0.57)
12	66.75 (1.24)	68.94 (1.16)	0.20 (−5.55 to 1.17)	66.03 (1.12)	69.43 (1.20)	0.03 (−6.51 to −0.29)
24	69.46 (1.16)	68.73 (1.01)	0.63 (−2.30 to 3.76)	68.68 (1.12)	69.66 (1.19)	0.53 (−4.09 to 2.12)
KABC NVI, mo
6	58.29 (0.85)	59.57 (0.94)	0.31 (−3.80 to 1.23)	58.02 (0.91)	60.36 (0.99)	0.07 (−4.91 to 0.22)
12	59.73 (0.91)	57.88 (10.79)	0.19 (−0.94 to 4.64)	59.54 (0.92)	58.89 (0.97)	0.62 (−1.89 to 3.19)
24	60.22 (0.89)	59.11 (0.75)	0.34 (−1.18 to 3.40)	59.63 (0.91)	60.12 (0.96)	0.70 (−3.02 to 2.04)
KABC Learning, mo
6	71.37 (1.08)	72.15 (1.07)	0.61 (−3.78 to 2.22)	71.14 (1.15)	73.15 (1.26)	0.22 (−5.26 to 1.23)
12	74.35 (1.27)	72.57 (1.06)	0.28 (−1.49 to 5.05)	74.40 (1.15)	73.65 (1.25)	0.65 (−2.47 to 3.97)
24	75.72 (1.22)	71.60 (1.14)	0.01 (0.84 to 7.41)	75.76 (1.16)	72.80 (1.24)	0.07 (−0.26 to 6.18)
KABC Planning, mo
6	58.58 (0.66)	60.28 (0.76)	0.09 (−3.69 to 0.27)	59.11 (0.73)	60.43 (0.82)	0.22 (−3.42 to 0.79)
12	60.18 (0.69)	60.04 (0.62)	0.88 (−1.70 to 1.98)	60.76 (0.71)	60.07 (0.77)	0.49 (−1.30 to 2.69)
24	60.61 (0.80)	59.78 (0.55)	0.39 (−1.08 to 2.72)	60.28 (0.67)	59.66 (0.71)	0.51 (−1.25 to 2.49)
KABC Delayed Recall, mo
6	72.32 (1.20)	71.21 (1.13)	0.50 (−2.15 to 4.36)	72.37 (1.16)	72.23 (1.26)	0.94 (−3.13 to 3.40)
12	73.48 (1.15)	71.67 (1.08)	0.25 (−1.30 to 4.92)	73.62 (1.16)	72.82 (1.24)	0.63 (−2.45 to 4.02)
24	73.46 (1.13)	71.59 (10.74)	0.23 (−1.19 to 4.92)	73.19 (1.16)	72.60 (1.23)	0.73 (−2.64 to 3.81)
TOVA response time variability, mo
6	232.0 (8.84)	224.6 (9.35)	0.52 (−17.10 to 33.62)	249.15 (6.70)	227.69 (7.27)	0.20 (−6.44 to 31.36)
12	206.0 (8.45)	203.7 (7.63)	0.84 (−20.17 to 24.80)	212.11 (6.71)	202.31 (7.15)	0.30 (−8.93 to 28.52)
24	198.2 (7.49)	188.7 (7.51)	0.37 (−1 1.45 to 30.39)	198.08 (6.74)	184.97 (7.10)	0.17 (−5.61 to 31.83)
TOVA signal detection D-prime, mo
6	2.77 (0.15)	2.82 (0.15)	0.82 (−0.46 to 0.36)	2.62 (0.12)	2.84 (0.13)	0.22 (−0.56 to 0.13)
12	3.05 (0.16)	3.10 (0.15)	0.81 (−0.48 to 0.38)	2.98 (0.12)	3.18 (0.13)	0.25 (−0.55 to 0.14)
24	3.26 (0.12)	3.34 (0.15)	0.68 (−0.47 to 0.31)	3.23 (0.12)	3.44 (0.13)	0.21 (−0.56 to 0.12)
TOVA percent commission errors, mo
6	10.00 (1.20)	9.79 (1.08)	0.90 (−2.99 to 3.41)	10.70 (1.00)	9.69 (1.08)	0.48 (−1.82 to 3.86)
12	8.02 (1.17)	8.15 (1.08)	0.93 (−3.28 to 3.02)	8.40 (1.01)	8.12 (1.06)	0.84 (−2.52 to 3.10)
24	6.06 (0.69)	7.34 (1.14)	0.34 (−3.92 to 1.35)	5.89 (1.01)	7.03 (1.04)	0.42 (−3.94 to 1.65)
TOVA percent omission errors, mo
6	18.61 (2.14)	17.83 (19.92)	0.79 (−5.09 to 6.64)	20.66 (1.84)	18.34 (1.99)	0.38 (−2.86 to 7.51)
12	16.69 (2.17)	15.06 (1.89)	0.57 (−4.05 to 7.32)	18.06 (1.84)	13.70 (1.95)	0.10 (−0.77 to 9.50)
24	12.41 (1.62)	12.06 (1.74)	0.88 (−4.34 to 5.05)	13.24 (1.85)	10.72 (1.93)	0.33 (−2.60 to 7.64)
TOVA ADHD index, mo
6	−2.01 (0.38)	−1.34 (2.89)	0.18 (−1.64 to 0.31)	−2.03 (0.34)	−1.44 (0.37)	0.23 (−1.54 to 0.38)
12	−1.00 (0.36)	−0.91 (0.29)	0.85 (−1.00 to 0.82)	−1.19 (0.34)	−0.89 (0.37)	0.53 (−1.25 to 0.65)
24	−1.28 (0.32)	−0.69 (0.29)	0.17 (−1.43 to 0.26)	−1.41 (0.34)	−0.59 (0.36)	0.09 (−1.76 to 0.13)
BRIEF Global Executive Composite, mo
6	45.87 (1.00)	52.91 (1.16)	<0.01 (−10.05 to −4.02)	47.17 (0.93)	52.84 (1.00)	< 0.01 (−8.28 to −3.06)
12	46.65 (0.97)	51.42 (1.17)	<0.01 (−7.76 to −1.76)	48.41 (0.93)	51.92 (0.99)	0.01 (−6.10 to −0.93)
24	43.62 (0.88)	49.66 (1.03)	<0.01 (−8.71 to −3.36)	45.24 (0.93)	49.92 (0.98)	<0.01 (−7.25 to −2.10)
ADHD-R, mo
6	0.36 (0.03)	0.61 (0.05)	<0.01 (−0.37 to −0.13)	0.41 (0.04)	0.61 (0.04)	<0.01 (−0.32 to −0.10)
12	0.35 (0.03)	0.57 (0.05)	<0.01 (−0.33 to −0.09)	0.41 (004)	0.59 (0.04)	< 0.01 (−0.29 to −0.07)
24	0.30 (0.03)	0.47 (0.04)	< 0.01 (−0.27 to −0.07)	0.32 (0.04)	0.48 (0.04)	< 0.01 (−0.27 to −0.05)
Weight-for-age z-score (WHO norms 2014), mo
6	−0.91 (0.10)	−0.81 (0.10)	0.49 (−0.38 to 0.18)	−0.92 (0.08)	−0.84 (0.08)	0.47 (−0.31 to 0.14)
12	−0.90 (0.11)	−0.79 (0.10)	0.48 (−0.40 to 0.19)	−0.92 (0.08)	−0.79 (0.09)	0.26 (−0.37 to 0.10)
24	−0.95 (0.11)	−1.02 (0.10)	0.65 (−0.23 to 0.36)	−0.99 (0.09)	−1.08 (0.09)	0.48 (−0.16 to 0.34)
Height-for-age z-score (WHO norms 2014), mo
6	−0.84 (0.12)	−0.52 (0.11)	0.05 (−0.64 to 0.00)	−0.76 (0.08)	−0.57 (0.08)	0.10 (−0.40 to 0.04)
12	−0.96 (0.11)	−0.75 (0.12)	0.20 (−0.53 to 0.11)	−0.84 (0.08)	−0.77 (0.08)	0.52 (−0.29 to 0.15)
24	−1.12 (0.11)	−0.90 (0.10)	0.13 (−0.53 to 0.07)	−1.05 (0.08)	−0.90 (0.08)	0.17 (−0.37 to 0.07)

Significant (p < 0.05) group differences are bolded. ADHD-R, attention-deficit/hyperactivity disorder rating inventory; BRIEF, Behavior Rating Inventory of Executive Function; KABC, Kaufman Assessment Battery for Children; LS, least squares; MISC, Mediational Intervention for Sensitizing Caregivers; MPI, Mental Processing Index; NVI, Nonverbal Index; TOVA, Test of Variables of Attention; UCOBAC, Uganda Community-Based Association For Women & Children Welfare.

Analyses of the BRIEF and ADHD outcomes resulted in significant differences across treatment conditions across all time points, with MISC mothers scoring their children with significantly more problems compared with UCOBAC mothers. When added as a time-varying covariate to the longitudinal models, height-for-age z-score and weight-for-age z-score (1 at a time) were significantly positively associated with the KABC Mental Processing Index (p < 0.01).

DISCUSSION

We present data from the school-age siblings of children who participated in a randomized controlled trial (RCT) of a parent-training intervention [Mediational Intervention for Sensitizing Caregivers (MISC)] compared with a nutrition and development program [Uganda Community-Based Association For Women & Children Welfare (UCOBAC)] for preschool-age children for neurocognitive and behavioral outcomes. Compared to children in UCOBAC, MISC school-age siblings performed better on working memory (Sequential Processing) at 6 months and visual-spatial analyses (Simultaneous Processing) during 12-month assessment. However, these benefits were not robust when considering the number of neurocognitive performance outcomes that were evaluated, nor did they persist to 1-year follow-up after training. As was the case in previous reported outcomes for the preschool-age target children,¹⁰ we must be cautious in concluding that MISC training resulted in collateral neurocognitive benefits for the older siblings. Also, the time course of these benefits differed, with working memory (Sequential Processing) differences emerging by 6 months from the start of caregiver training but then dissipating and visual-spatial benefits (Simultaneous Processing) by 1 year and then dissipating at 1-year follow-up.

The MISC intervention has several targets that we proposed could benefit the older siblings living in the same household as the target study children. These include a focus on enhancing the developmental milieu in the home and improving adult-child relationships and attachments.^13,26,27 Although these specific mechanisms were not directly measured in the present study, it is reasonable to suggest that HIV-affected children often have lost or partially lost this “protective shield” by death of a parent(s) or reduced parental capacity due to their HIV illness.²⁸ With the goal of the MISC program to improve this “protective shield” for 1 target child, we hypothesized that such protection could be experienced by other children in the home as well.

Although the “value-added” neurocognitive benefits to the school-age siblings of the MISC study preschool children was limited in our present study findings, this study does address a key gap in the early childhood development (ECD) literature in low-income and middle-income countries. This is because virtually nothing is known about collateral benefits to household children of caregiving interventions in international contexts, in which the need to affect multiple vulnerable children is often most pressing. Such dynamics, however, may depend on the physical and emotional well-being of the caregiver receiving ECD training.^29–31 Caregivers in our previous MISC/UCOBAC clinical trial reported less functional impairment in the MISC arm than those in UCOBAC.¹⁰ Some of the nonsignificant but better developmental outcomes observed among MISC children or the improvements caregivers experienced in the quality of their interactions with their children could have led to less functional impairment.^6,29,30

The exploration of impacts on siblings of the MISC intervention was limited by the lack of strong study findings on the full range of child outcomes from the main MISC study with preschool-age children. In addition, we were limited in the present study by the lack of measures assessing whether the caregiver used her MISC skills with other children in the household. There was also the absence of a true control (no caregiver training) condition. However, given that caregiver training and support should be provided at some level wherever possible as a standard of care,¹ the present study assessed the “value-added” benefits of MISC versus UCOBAC for older siblings. The choice to compare 2 active interventions was made because we did not feel we could ethically withhold both the MISC and UCOBAC interventions to the present study dyads, given the developmental risks experienced by HIV-affected children in previous reported findings of effectiveness of ECD programs for children in such settings. The present supplemental study to a cluster RCT, however, is a rigorous comparison of 2 different caregiver training interventions in which (1) a nutritional supplement was provided to all study children throughout the study and (2) the interventions in both study arms were of comparable structure (in terms of the quality of training sessions and the psychosocial support dimensions provided to the caregivers) and duration. Furthermore, this is the first study to evaluate the comparative neurodevelopmental benefits for the older siblings.

CONCLUSION

Traditionally, research on parenting and early childhood development programs has focused on the participating caregiver-child dyads, ignoring the potential impact parenting programs may have on other children living in the same homes. Although this study did not show substantial benefits to siblings living in the Mediational Intervention for Sensitizing Caregivers intervention homes compared with siblings in the homes receiving the control condition, the framing of the issue of dispersion of program impacts is an important contribution to the literature and deserves further exploration. This topic has far-reaching implications for promoting positive youth development and preventing emotional and behavioral problems, particularly in low-resource settings in which families are large and children often face significant adversity. With this in mind, more studies of this sort are needed within implementation science so that we might better understand the contributions of good nutrition, good parenting, good stimulation, and good caregiver mental health and functionality to overall better child development throughout early and middle childhood in resource-constrained settings.