The COVID-19 pandemic has exposed pregnant women to an unprecedented level of stress, resulting in higher levels of prenatal anxiety, depression, and stress. The impact of COVID-19 on maternal mortality and maternal and neonatal morbidity is well-documented (16,66–70). The impact is not exclusively related to the biological effects of SARS-CoV-2 infection but is also associated with accumulating stressors that negatively affect the health of mothers and neonates (16,25,33). The complexity of the pandemic period for pregnant women invites us to build a comprehensive conceptual framework that integrates knowledge of stress physiology, neurobiology, and pregnancy physiology (71). Including the concept of AOL as an exposure variable enables us to take a step towards constructing this comprehensive model. It allows us to consider factors beyond the current gaps in knowledge of maternal health and merge both the adaptive physiological processes of pregnancy, labour, and birth, as well as the increased susceptibility to long-term pathologies that require diagnosis, prevention, and treatment. The expected results of this study will be useful in determining if the AOL evaluation during pregnancy could be a good tool for understanding the physical and mental health consequences of dealing with prolonged stress (3,72). Including AOL in clinical records may serve as an essential health indicator, especially for pregnant women living in more vulnerable conditions, who are at high risk of suffering stress and AOL.

AOL is a holistic concept that incorporates a variety of harmful elements that were exposed to pregnant mothers during the pandemic, affecting their physiological adaptations during pregnancy. As this study protocol proposes to determine, the complexity of AOL is also related to the health outcomes affected, including the mother’s cardiovascular health, the neurodevelopment of children, and the maternal-child relationship. This study will use quantitative and qualitative methods to determine the exposure, combined with sophisticated statistical analysis, including machine learning tools, to comprehensively understand the phenomenon experienced by pregnant women during the COVID-19 pandemic and its consequences.

In this protocol study, the AOL will be determined by examining the clinical records that were previously collected during pregnancy. These records include AOL markers like mean arterial pressure, gestational weight gain, glycemia, mental health and psychosocial risk, reduction or absence of maternity leave, gender violence, economic difficulties, isolation in prenatal controls or birth, lack of maternal-newborn bonding, prenatal anxiety and depression, among others, all variables that could have been significantly altered by the COVID-19 pandemic conditions. Mothers may have different capacities for coping with these difficulties, and this protocol was designed to classify and evaluate the consequences of inadequate response to the challenges for the mother-child dyads.

There is evidence that AOL increases cardiovascular risk (5), and perinatal AOL is associated with preterm birth (73), while prenatal depression increases the prevalence of chronic hypertension later in life in women (8). Due to the evidence of increased perinatal mental health disorders during the COVID-19 pandemic in Latin America and Chile (26), we will consider the perinatal anxiety, depression and stress, and factors that could be cause or mediate mental health disorders, like economic difficulties, lack of social support, and gender violence, as relevant factors to determine the presence of AOL during the pregnancy.

Regarding mental health in prenatal controls, the Edinburgh Postnatal Depression Scale (EPDS) is a 10-item scale for the identification of maternal depression, applied in the Chilean public system in the prenatal and postnatal periods. Each item contains short descriptive statements, scoring the most negative description with the highest value of 3 and the most positive with 0. Scores are tallied so that a score <8 indicates depression is not likely, 9–11 indicates depression is possible, 12–13 indicates a reasonably high possibility of depression, and 14 and higher is considered a positive screen for depression and recommends a diagnostic assessment. Additionally, a positive score is identified for question 10, which identifies participants at risk of harm or suicide (74,75). The Abbreviated Psychosocial Assessment (EPsA) is an instrument for early detection and intervention of some risk factors that introduce inequity in people’s development. This evaluation is considered a priority strategy for monitoring and supporting pregnancy in the Chilean pregnancy and early childhood protection system, "Chile Crece Contigo (ChCC)" (76). These data, more other information will be collected in the clinimetric survey as is shown in table 1. Following the original protocol from Fava (3) and modifications proposed by Peng (77), it is possible to determine a score of AL for each participant. The possibility of determining a score of AL and evaluating the life experience associated with AOL to determine exposure is a strategy to overcome the weaknesses of quantitative or qualitative methods themselves (36). By applying data triangulation, we can gain diverse perspectives on the COVID-19 pandemic and achieve a deeper understanding of this phenomenon. This approach allows us to effectively distinguish between women who experienced allostatic overload and those who managed to cope adequately with the challenges presented during the pandemic.

One advantage of this study design is the ability to assess the long-term effects of COVID-19 during pregnancy. Long COVID, also known as post-COVID-19 syndrome, is estimated to affect tens of millions of people. However, there is still a lack of information regarding the mechanisms or biomarkers that could aid in the classification and treatment of this condition (78). Among the potential consequences of long COVID, cardiovascular diseases have surfaced as a significant concern for the healthcare system, particularly for patients who already have pre-existing conditions such as obesity, diabetes, or other risk factors for cardiovascular disease (79). Given the characteristics of Chilean pregnant women—approximately 30% are obese (80), and around 10% experience gestational diabetes in the Biobío region (81)—there is a possibility that we will observe a percentage of participants that experienced symptoms of long COVID, especially those associated with cardiovascular disease. This will be an additional finding of our study, which will contribute to the understanding of cardiovascular biomarkers related to long COVID (82), particularly in a unique population like women who had a pregnancy with SARS-CoV-2 infection.

Finally, we will use machine learning (ML) to analyse the data besides classic statistical tools. ML corresponds to various methods that use mathematics, statistics, and computational science to learn from multivariate data. Employing pattern recognition performed on various measured variables, different algorithms can find correlations and accurately predict different conditions, such as an individual belonging to a particular group or class or the concentration of a specific biomarker in a sample of interest (65,83). ML approaches are available for managing large and heterogeneous data sources, identifying intricate and occult patterns, and predicting complex outcomes (64). The pregnancy window of time represents an excellent opportunity to exploit all the advantages of using ML, especially for accurately predicting short- and long-term adverse outcomes.