Kangaroo mother care for preterm infants and child development – evidence from São Paulo, Brazil

ABSTRACT Globally, an estimated 15 million children are born prematurely each year, resulting in a high burden of under-five mortality and neurodevelopmental disability. Kangaroo Mother Care (KMC) is a key intervention to support the development of preterm infants. However, evidence on the impact of KMC in routine care settings remains limited. This paper examines the associations between maternal KMC efforts and child development among preterm infants using data from a prospective cohort study conducted in São Paulo, Brazil. Study outcomes were height-for-age z-scores (HAZ) and cognitive development at age 3 years. Practicing KMC was positively associated with HAZ (+0.91 SD, 95% CI [0.126, 1.695]), and cognition (+0.37 SD, [0.034, 0.806]) of preterm children. The results suggest that KMC can improve children’s cognitive and physical development. New programmes to increase KMC uptake in the setting studied may be beneficial for both mothers and their preterm children.

A growing literature has highlighted that early intervention programmes for preterm infants can positively influence developmental outcomes with cognitive benefits that persist (Spittle, Orton, Anderson, Boyd, & Doyle, 2015). Kangaroo Mother Care (KMC) is increasingly recognized as one of the most effective interventions in this area (Mekonnen, Yehualashet, & Bayleyegn, 2019). The key element of KMC is extensive skin-to-skin contact between mothers or other caregivers including fathers and their infants during the postnatal period at the hospital. Ideally, these activities are then continued at home until the child gains a normal weight. Skin-to-skin contact has been shown to increase attachment, enhance mothers' ability to breastfeed, and stimulate the child (Johnston et al., 2017;Moore, Bergman, Anderson, & Medley, 2016). Several recent studies have underscored the effectiveness and cost-effectiveness of KMC in particular for reducing infant mortality among preterm and low birth weight children, but also for reducing the likelihood of severe illness and infections, and for reducing the average length of stay at the hospital (Campbell-Yeo, Disher, Benoit, & Johnston, 2015;Conde-Agudelo & Diaz-Rossello, 2016;Jefferies, 2012;Lawn, Mwansa-Kambafwile, Horta, Barros, & Cousens, 2010;Watkins, Morgan, Nambuya, Waiswa, & Lawn, 2018). Furthermore, KMC has been shown to improve infant and child health and growth of children (Evereklian & Posmontier, 2017) and to decrease the risk of postpartum depression and maternal stress (Badr & Zauszniewski, 2017;Stockwell, 2017).
In many settings, KMC has been integrated into routine health care to monitor the growth and development of preterm infants. KMC has been used within primary health care systems to enhance breastfeeding, for the management of suspected infection, and for improving child survival for preterm children (Chan, Valsangkar, Kajeepeta, Boundy, & Wall, 2016Evereklian & Posmontier, 2017). KMC has also been useful in enhancing early discharge from the hospital with adequate follow up to the child (Bisanalli, Nesargi, Govindu, & Rao, 2019;Boo & Jamli, 2007;Cattaneo et al., 1998). Despite the presumably large benefits of KMC, limited evidence exists on the use and impact of KMC in routine care settings. This paper aims to address this gap by using newly collected data on infant health and child development within a cohort of children born at the University Hospital of São Paulo between 2012 and 2014. As part of the cohort study, mothers of preterm infants were asked to report on their KMC activities after birth. Children's cognitive and physical development was then assessed through a home-based assessment by trained staff at age 3. In Brazil, the Kangaroo programme is a public policy that encompasses other care practices and includes other family members. Our study is limited to the practices as reported by the mother, thus Kangaroo Mother Care.

Study procedures
The study was designed as an observational prospective cohort study conducted in São Paulo Municipality, Brazil. The São Paulo Western region cohort includes all children born to local residents at the University Hospital of São Paulo between April 1, 2012, and March 31, 2014 (Brentani et al., 2020). Children were revisited and re-assessed as part of a 3-year follow-up conducted between June 2015 and March 2018. Prior to the visits, qualified field interviewers were trained on ethics, expected behaviour during the household visits, and assessment methods. Caregivers were either contacted by phone or through a personal visit to schedule the actual interview. All interviews were conducted at children's homes. The entire visit including interviews and child assessments lasted for about 90 min. Our study was restricted to children born prematurely during this period (born before 37 weeks of gestation). Data on birth outcomes including gestational length were directly extracted from the University Hospital's electronic records. The gestational age of all children was assessed by the delivering health worker using the Capurro method (Capurro, Konichezky, Fonseca, & Caldeyro-Barcia, 1978). The Capurro allows for an accurate assessment of gestational age when data on the last menstrual period or and early ultrasound results are not available (Pereira, Dias, Bastos, da Gama, & Leal Mdo, 2013).
Data on KMC and development at age 3 were collected through interviews during the home visit by trained data collectors. Mothers of preterm children were asked to report on their KMC activities, including the frequencies and duration of each mother-child interaction in the preterm period. Records with missing information on KMC engagement were excluded from the study.

Exposure and outcome variables
The main exposure variable of interest in our subsample was the practice of KMC as reported by the caregivers. As part of their interview, caregivers reported whether they engaged in KMC, as well as the number of times per week it was practiced. The primary outcomes for this study were children's cognitive and physical development at 3 years of age. Cognitive development was measured through direct observation using the PRIDI scale. The PRIDI scale has been used widely in Latin America alongside other assessment tools to assess children's developmental outcomes (Altafim, Brentani, Escobar, Grisi, & Sjfe Fink, 2018Thompson, 2016). The PRIDI was developed to generate high quality regionally comparable data on child development. PRIDI measures children's development between 24 -59 months and covers cognition, language and communication, socio-emotional, and motor domains (Verdisco et al., 2015). The PRIDI tool captures domains of child development as well as factors associated with ECD such as contextual variables around the child, the parent, and the community. The PRIDI tool is easy to use and can be administered by trained lay workers. PRIDI scores were corrected for the biological age of children.
Anthropometric information (height and weight) were collected for each of the child by the trained assessors at children's homes. Height-for-age z-scores (HAZ) were computed using the World Health Organization (WHO) Anthro software (WHO, 2010). Stunting was defined as HAZ<−2.

Statistical analysis
Data analysis was completed in three steps. First, we used descriptive statistics to characterize the study population. Socio-demographic and clinical data included a child's age and weight at birth, caregiver's age at delivery, caregiver's education level, caregiver's marital status as well as family receipt of social transfers (bolsa familia). Second, we estimated unconditional associations between KMC practicing and cognitive and physical development. Third, we estimated multivariable linear regression adjusting the estimated associations for caregiver's age (indicators for ages <=20 and ages >=35), childbirth weight (categorized into <1500 g, 1500-1999g, 2000-2499 g and >=2500 g), child's sex, caregiver's schooling, caregiver's higher schooling, receipt of social support, caregiver's marital status, and wealth index. Finally, we stratified the HAZ with the child's weight at birth (<2500 g and > 2500 g) and caregiver's education level (basic education or less and secondary education or higher). Data analysis was performed using STATA 15.0 for Windows (STATA Corporation, College Station, TX) (StataCorp, 2015).

Results
Out of the 3620 children assessed at age three, 254 (7%) were born prematurely. From this subsample of premature children, the KMC questionnaire was administered to 139 mothers. Out of this sample, 26 mothers reported having engaged in KMC, while 113 mothers reported not have engaged in KMC activities. Table 1 summarizes caregivers'-children's socio-demographic characteristics by KMC status. 23% of mothers were under age 20 at the time of birth, and 10% of mothers were 35 or older. 7.9% of children born before week 37 of gestation had a birth-weight of below 1500 grams, while 54.7% of children had a birth-weight in the normal range (>=2500 grams). 51.8% of children were female, and 25.9% of families received social transfers (bolsa familia). On average, mothers practicing KMC were slightly younger and had infants with substantially lower weight at birth. Differences in social transfer receipts were small. Figure 1a and 1.b summarize the frequency of KMC practice. 46% of mothers engaging in KMC reported engaging in this practice every day. 30% of mothers reported engaging in KMC less than 3 days per week. The median duration of each session was two hours. Only 3 mothers reported practicing 8 or more hours of KMC each day. Figure 2 shows estimated kernel densities for height for age z-scores at age 3 by the KMC group. Despite the large negative weight differential at birth, mean HAZ was substantially higher in the KMC group at age 3 (−0.17 vs. −0.96 in the no KMC group), with a pronounced shift of the entire height distribution to the right. Figure 2 shows the estimated kernel densities for PRIDI z-scores. Once again, mean scores were substantially higher for the KMC group (0.50 vs. −0.09 in the no KMC group), with particularly large differences in the left tail of the distribution (z-scores < −2) Figure 3. Table 2 shows crude and adjusted association between HAZ, stunting, PRIDI and KMC exposure. In unadjusted models, practicing KMC was associated with a 0.79 SD increase in HAZ (95% CI [0.22 -    Table 3 shows the results of the stratified analysis. When we stratified the sample by caregiver education, positive associations between KMC and HAZ were primarily found for mothers with limited education. When we stratified the sample by birth weight, protective effects of similar magnitudes were found for children with birth weight above and below 2500 grams, but neither effect was statistically significant due to the relatively large standard errors. To further illustrate the observed positive associations, we show the estimated distributions of HAZ by the KMC group in Figure 4. While only one child (4%) in the KMC group had HAZ < −2, the same was true for 35 children (33%) without KMC.  Notes: All models adjust for the full set of covariates in Table 1. Coefficients displayed based on linear regressions, with 95% confidence intervals in parentheses. ***p<0.01, **p<0.05, *p<0.1.

Discussion
In this paper, we used a novel data set from São Paulo Brazil to estimate the empirical associations between Kangaroo Mother Care (KMC) and child development among children born prematurely. Our results suggest that practicing KMC may have large protective effects on children born prematurely, both in terms of their cognitive and in terms of their physical development. These findings are consistent with other studies of preterm (Bera et al., 2014Dodd, 2005Evereklian & Posmontier, 2017;Jefferies, 2012;Sharma, Farahbakhsh, Sharma, Sharma, & Sharma, 2019) or low birth weight infants (Samra, Taweel, & Cadwell, 2013). In general, these studies found relatively large benefits of KMC on children's development. However, existing studies associating KMC with cognitive development and have been mostly conducted mainly in high-income countries. Our study shows that positive impacts of KMC on child development, and particularly on cognitive development, can also be expected in low resource settings.
A large literature has documented the high risk faced by preterm infants concerning their cognitive development (Lechner & Vohr, 2017;Lemola et al., 2017;Murner-Lavanchy, Rummel, Steinlin, & Everts, 2018). These risks are well visible in the data presented in this study, with an average HAZ of −0.8 at age 3 among preterm children in the sample. Rather remarkably, these risks seem however almost exclusively restricted to children not benefitting from KMC, with an average HAZ of close to zero among children benefitting from KMC, and an average HAZ of −1.0 among children not benefitting from KMC. These differences observed at age three are definitely not due to the initial advantages of children in the KMC group: the median gestational length in the KMC group was 33 weeks, compared to a gestational length of 36 weeks in the group of children without KMC. Children benefitting from KMC do not only seem to be able to close this initial gap, but also appear to be substantially better developed at age 3. From a behavioural perspective, it is of course, possible that mothers practicing KMC may on average be more engaged with children more generally than mothers opting against it. We controlled for several potential factors in our empirical models, but cannot fully rule out residual confounding in our analysis. Given the rather large associations observed in fully adjusted models, it seems however unlikely that residual confounding would fully explain the observed patterns (Vanderweele & Arah, 2011).
Several limitations of this study should be noted. First, the overall sample size was small, limiting the statistical power of the study. Second, the observational nature of this study does not fully allow us to rule out confounding. While we include several critical confounders in the model, KMC practice may be correlated with other parental factors positively contributing to child development. Third, we only have information about the KMC during the hospital stay and are not able to assess the extent to which KMC was practiced at home. Additionally, the mothers were interviewed three years after birth and thus likely did not remember all of the details of their early child interactions. Notwithstanding these limitations, our results indicate that even relatively limited engagement with KMC during the first days of life may be highly beneficial for preterm infants. Our study does not allow us to directly understand the KMC mechanism, but the main links highlighted in the literature (Campbell-Yeo et al., 2015;Jefferies, 2012;Johnston et al., 2017;Lawn et al., 2010;Moore et al., 2016;Watkins et al., 2018) such as bonding, breastfeeding, stimulation, are likely to apply in this setting as well. Further research will be needed to confirm the results of this study and to identify optimal and minimal dosing of KMC.
Despite the presumably large benefits of KMC and the official hospital commitment to this programme, our findings revealed that only about 20% of the women with preterm infants practiced KMC in the 2012-2014 period. Similarly, low rates of KMC uptake have also been noted in several other countries (Evereklian & Posmontier, 2017). Low uptake of KMC may be related to stress and stigma associated with having a preterm child, lack of standardized measure, and poor understating of KMC benefits (Chan, Labar, Wall, & Atun, 2016Evereklian & Posmontier, 2017), but may also reflect lack of knowledge about KMC by families and healthcare workers and local cultural practice (Chan, Labar, et al., 2016;Charpak & Ruiz-Peláez, 2006). Further research will be needed to better understand the main barriers in this setting as well as to identify the most effective ways to support mothers and their infant children during the post-natal period more generally.

Conclusion
Kangaroo Mother Care is a well-established programme and has been accepted in many countries as an intervention and health policy with potential benefits for preterm children and their caregivers. Our study suggests that the potential to improve children's cognitive and physical development in an urban Brazilian context is large. Despite this, uptake remains low, and major efforts will be needed to further engage mothers with KMC.

Ethics approvals and consent to participate
Ethics approvals were obtained from the University Hospital of São Paulo. Informed consent was obtained from all the participants before the interviews. As part of the consenting process, study purpose and procedures were explained to the caregivers, and families were made aware that they can end the assessment at any point of time or skip any question they were not comfortable with. Caregivers were also provided with the PI's contact information in case they had any questions regarding the study or wanted to obtain further information.

Acknowledgment
We would like to acknowledge and thank the staff at the University Hospital of São Paulo as well as all study subjects for their support of this study.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Availability of data and materials
The datasets used in analysis during the current study are available from the corresponding author on request.

Notes on contributors
Silas Onyango is a doctoral student at the Swiss Tropical and Public Health Institute and the University of Basel. He is also a researcher at the African population and Health Research Center. Alexandra Brentani, is a faculty member in the Department of Pediatrics at the University of São Paulo Medical School.
Günther Fink is Associate Professor at the University of Basel as well as Head of the Household Economics and Health Systems Research Unit at the Swiss Tropical and Public Health Institute.