As people all over the world are experiencing a sustained long period of adverse experiences which keeps their level of stress high, researchers are overwhelmingly focusing on mental health consequences of the pandemic [1–3]. A particularly vulnerable group to various form of adversities are children due to their limited understanding of an adverse event, abrupt withdrawal from school, social life and outdoor activities, social isolation and separation from caregivers due to quarantine policies, increased domestic violence, changes in diets and dietary patterns, among other reasons [4–7]. While the short-term psychological and psychiatric repercussions of these adverse conditions on children have been widely discussed, the long-term consequences are to be known only after several years. Early-life experiences, both positive and adverse, lay foundations for health and wellbeing in later life, making early-life experiences an important public health issue [8–11]. Children exposed to various forms of adversities early in life are at increased risk for a broad range of developmental disorders including delayed brain (cognitive) development [12] and poor mental health [13]. They also face an increased risk of morbidity as adults and early mortality [14]. Children who experience adversities before the age of 3 are more vulnerable to the effects of adversities compared to children who experience the adversities at older ages [13].

Actions to prevent and reduce the negative consequences of early-life adversities must be the obligation of every society and is in accordance with the Sustainable Development Goals of the United Nations [15, 16]. And it is important that these actions are based on scientific evidence. Most past research focused on investigating the effect of early-life adversities on adult outcomes, and relatively fewer studies examined the link between early-life adversities and childhood outcomes. However, for policy purposes, studying the effects of early-life adversities on childhood outcomes is preferred over studying their effects on adult outcomes because childhood outcomes manifest early in life and, therefore, are more amenable to policy interventions [17]. Moreover, much of the evidence on the impact of early adversities on outcomes later in life is based on data from the developed countries, and there is a dearth of scientific evidence in the context of low and middle income countries (LMICs) [14]. The current context sheds light on why the findings based on developed counties could not be generalized to LMICs. While most types of adversities faced by a child living in a developed country are also shared by children in LMICs, their burden of economic shocks such as the loss of family income and food insecurity arising from crop damage and loss of livestock are not common in the developed world. In addition, a development disorder of a child living in a LMIC is less likely to be diagnosed and treated than a disorder of a child living in a developed county and therefore the negative consequences of an adverse event during the early childhood on cognitive development are more likely to long prevail.

This study, using multi-country longitudinal data from four LMICs (Ethiopia, India, Peru, and Vietnam), examined whether exposure to adverse experiences during the first year of life (age 1) is associated with early-childhood (age 5) and mid-childhood (age 8) cognition, measured by the Peabody Picture Vocabulary Test (PPVT). In addition, the study examined whether the change in cognition between age 5 and age 8, measured using the unpredicted change in PPVT scores between those ages, is associated with adversities in infancy. Understanding these associations is essential because childhood cognitive development is a strong predictor of adulthood cognition and human capital [8]. Our measure of adversity is broad-based and covers a variety of adverse events relevant in a LMIC context including economic, environmental, emotional and health shocks. Our findings shed light on an important international public health issue; how adverse effects of early childhood experiences propagate over time retarding the cognitive development of children in developing country settings. These research findings can be helpful in formulating public health policies which would mitigate long-term risks of adverse conditions many children in LMICs are currently facing through cost-effective policy interventions.

The pooled data from all countries resulted in a sample of 8,062 children, approximately 2,000 children from each country. After dropping children with missing observations in any of the explanatory variables used in our main analysis, a total sample of 7,327 children was included in the pooled model. Over one third (34.5%) of the children had an AE-1 score of zero, with the remaining 65.0% having AE-1 scores of 1 or more. About half of the children (48.0%) are female and the remaining 52.0% are male. Over one-third of the children (40.6%) had a mother with no education, and nearly two-third (62. 6%) lived in rural areas.

There were significant variations across nations over most variables. Ethiopia and India have lower proportions (16.1% and 32.0%) of children with AE-1 score of “zero”, compared with 41.6% and 48.0% for Peru and Vietnam, respectively. Conversely, smaller proportions of children in Peru and Vietnam experienced severe adversities, compared to those in Ethiopia and India. The raw PPVT scores also vary, with PPVT-5 scores ranging from 21.4 in Ethiopia to 37.0 in Vietnam, and PPVT-8 scores ranging from 58 in India to 94 in Vietnam. However, as discussed in Variables section , our standardization approach eliminates these differences and makes the PPTV scores comparable across counties. Cross-contextual and socioeconomic differences were also seen in socioeconomic covariates, including maternal education, rural/urban status of residence, and wealth. Minor differences were noted in the average family size and average age of children (Table 2).

The association between adverse experiences in infancy and childhood cognitive development is shown in Table 3. We find that children with adverse experiences in infancy achieved lower test scores in childhood compared with children who had no adverse experiences, the reference group. We also find an inverse relationship between the test scores and the level of adversity. Compared with children with no adversity in infancy, those with mild, moderate, and high adversities as infants had lower PPVT-5 scores, with estimated coefficients of −0.06, −0.09, and −0.09, respectively. Children with “High” adversity level as infants had a lower PPVT-8 and PPVT-Change scores, with estimated coefficients of −0.15 and −0.16, respectively. The alternative measure without potentially endogenous components produces comparable results (Table 3). Results from the measure with only the endogenous variables are given in the Supplementary Material.

Results from the analysis that controls for adversities both at age 1 and age 5 is given in Table 4 Results from this analysis showed similar effects, with higher adversity scores in the first year of life associated with lower scores in cognitive tests during childhood. Compared with children with no adversity in infancy, those with mild, moderate, and high adversities as infants had lower PPVT-5 scores, with estimated coefficients of −0.06, −0.08, and −0.08, respectively. Children with the “High” adversity level as infants had a lower PPVT-8 and PPVT-Change scores, with estimated coefficients of −0.13 and −0.14, respectively. Again, the alternative measure without potentially endogenous components produces comparable results (Table 4). Results from the measure with only the endogenous variables are given in the Supplementary Material.

The significant effects of the early life adversities on childhood cognition were also seen in country-stratified models for India and Vietnam, but not for the other countries. Both PPVT-5 and PPVT-8 had a statistically significant negative associations with adversities in infancy for India. Adversity in infancy was statistically significantly associated with only PPVT-5 for Vietnam. When the alternative adversity measure without the endogenous components was used, results for Ethiopia at age 8 turned significant while the results for India turned insignificant. The associations for Vietnam remained significant (Table 5). Results from the measure with only the endogenous variables are given in the Supplementary Material.

This study examined the association between adversities in early infancy and childhood cognitive development using a harmonized multi-country dataset. We found a statistically significant association between the adverse experiences at age 1 and the level of cognitive development at ages 5 and 8 measured using standardized PPTV scores, as well as the cognitive development between the two ages. Children who faced adversities as infants had lower PPVT scores during childhood, with PPVT scores getting worse as the severity of the adversities increased. Severe adversity in infancy was also associated with change in cognition between early and middle childhoods. The association remains robust after controlling adversities later in childhood. These results suggest that early adversities have long-lasting effects on childhood cognitive development.

The association of high AE-1 scores with the unpredicted change in PPTV scores in between the ages of 5 and 8 is particularly interesting and policy relevant. This model investigates how early adversities affect cognitive development during later years, net of the same effect through contemporaneous adverse experiences. The children who face adversities during infancy are vulnerable to similar events in their later life and the association between PPTV and AE-1 scores may show this effect rather than the effect of AE-1 score itself. The difference between the actual PPTV score at 8 and the predicted scores using all available information at age 5 including the AE-5 score rules out this possibility. Similarly, the results of our robustness checks using on PPTV-5 and PPTV-8 after controlling for adversities faced at age 5, rules out the possibility of this effect. Another potential limitation is the endogeneity bias in parameter estimates due to some of the constituting components of the AE-1 index [27]. To circumvent the issue, we estimated all our models using an alternative measure which excludes those components. The results rule out any significant bias.

While the main findings of this study are applicable to all the countries, some important differences across countries should be highlighted. In Ethiopia, the association between AE-1 and the PPVT scores was negative but not statistically significant. The association was negative and statistically significant for India. In Peru, the association was statistically significant for PPVT-8 and PPVT-Change, while it was statistically significant only for PPVT-5 in Vietnam. The results show that the negative effects of the adversities in the first year of life on childhood cognitive development were weaker in Ethiopia, stronger in India, strengthened over time in Peru, and weakened over time in Vietnam. This could be due to contextual or policy related differences across the countries towards addressing various adversities children experience in their early lives. While the overall findings provide evidence for the effects of early-life adversities, countries would need to assess their unique cultural or social conditions in interpreting the findings and implementing effective policy interventions [28]. One previous study finds an association between environmental shocks and higher PPVT scores of children in Ethiopia and Peru. The results for India and Vietnam, however, are either insignificant or in the opposite direction [29].

Although the primary objective of the study was to examine associations between the AE-1 and the PPVT scores, it is also noteworthy to reflect on the socioeconomic and demographic factors impacting childhood cognitive development. Female children scored less in the PPVT test at both age 5 and age 8. Residing in rural areas was also negatively associated with the PPVT scores. Children who live in households with large family sizes scored less in the cognition tests. Similarly, children of less-educated mothers scored less in the PPVT tests. These findings are largely consistent with what has been reported in the literature, suggesting that these factors should be seriously considered in efforts to reduce the negative impacts of early-life adversity experiences on childhood cognitive development [30, 31].

This study is not without limitations. First, in the study, we constructed measures for AE-1 scores based on items available from the survey. Even though our AE-1 score was based on an extensive set of variables, these factors were by no means exhaustive. However, we believe that they provide considerable information for understanding the early-life experiences of children given wide dimensions of questions assessed, such as health, poverty, neglect, environments, and items relevant to measure various forms of adversities a child could experience at age 1. Not every aspect of childhood cognition is captured by the PPVT test as it mainly measures verbal cognition. However, measures of cognitive ability such as PPVT are preferred measures of educational attainment to measures such as years of schooling because cognitive skills are a stronger determinant of adulthood outcomes than school attainment [32]. Even though we controlled for a range of confounding factors, some unobservable factors could still confound the relationship between AE-1 and cognitive development. For example, the unobservable factors, such as household characteristics, institutional help during crisis, urban/rural residence, and caregiver’s education, could explain some of the differences in the PPVT test scores [33]. Finally, cohorts in the YL survey may not be nationally representative as the survey participants were recruited from certain communities in the countries. The sample from India, for example, is unlikely to capture the heterogeneity across states, because the YL children from India came from 20 communities in two of the Indian states (states of Andhra Pradesh and Telangana) [22]. Further, although YL study has taken measures to address limitations in the study design, still there may concerns around the adaptation of measures in cross-cultural settings, translation, and cross-cultural equivalence [20].

Despite the limitations, the data from these four countries provide critical findings for policy implications, especially in developing countries, where many infants are exposed to various forms of adversities. Given that adverse experiences in early life are disproportionately high in developing countries, estimating its negative effects on childhood cognitive development will inform policymakers to design interventions that can reverse the negative effects of early adversities. Since the topic has been understudied in developing country settings, findings in this study may be novel in those contexts. Our findings suggest that early interventions to improve neonatal conditions through policy interventions could result in better childhood outcomes, which will have practical implications for improved adulthood outcomes, too. Additionally, policymakers are advised to seriously consider parental education and remedial education policies that could reverse the adverse effects of poor neonatal conditions.

Early life cognition is an essential component of health and wellbeing across a person’s life course. Our findings suggest that policies that aim to successfully manage a society’s future could do so by effectively addressing early life adversities.