Article Text
PDF
Abstract
Introduction Neonatal mortality is a global public health challenge. Guatemala has the fifth highest neonatal mortality rate in Latin America, and Indigenous communities are particularly impacted. This study aims to understand factors driving neonatal mortality rates among Maya Kaqchikel communities.
Methods We used sequential explanatory mixed methods. The quantitative phase was a secondary analysis of 2014–2016 data from the Global Maternal and Newborn Health Registry from Chimaltenango, Guatemala. Multivariate logistic regression models identified factors associated with perinatal and late neonatal mortality. A number of 33 in-depth interviews were conducted with mothers, traditional Maya midwives and local healthcare professionals to explain quantitative findings.
Results Of 33 759 observations, 351 were lost to follow-up. There were 32 559 live births, 670 stillbirths (20/1000 births), 1265 (38/1000 births) perinatal deaths and 409 (12/1000 live births) late neonatal deaths. Factors identified to have statistically significant associations with a higher risk of perinatal or late neonatal mortality include lack of maternal education, maternal height <140 cm, maternal age under 20 or above 35, attending less than four antenatal visits, delivering without a skilled attendant, delivering at a health facility, preterm birth, congenital anomalies and presence of other obstetrical complications. Qualitative participants linked severe mental and emotional distress and inadequate maternal nutrition to heightened neonatal vulnerability. They also highlighted that mistrust in the healthcare system—fueled by language barriers and healthcare workers’ use of coercive authority—delayed hospital presentations. They provided examples of cooperative relationships between traditional midwives and healthcare staff that resulted in positive outcomes.
Conclusion Structural social forces influence neonatal vulnerability in rural Guatemala. When coupled with healthcare system shortcomings, these forces increase mistrust and mortality. Collaborative relationships among healthcare staff, traditional midwives and families may disrupt this cycle.
- Paediatrics
- Obstetrics
- Other study design
- Child health
Data availability statement
Data may be obtained from a third party and are not publicly available. Quantitative data may be obtained from a third party and are not publicly available. Qualitative data are not available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Neonatal mortality rates disproportionately affect Indigenous communities in Latin America.
WHAT THIS STUDY ADDS
A deeper understanding of the interplay between sociopolitical, cultural and biological factors associated with fetal/neonatal demise. The compounded effect of neonatal vulnerability due to structural violence and health system shortcomings creates a cycle of mistrust, blame-shifting and fatal outcomes.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study highlights the relevance of addressing the intersectional structural forces contributing to neonatal vulnerability and the need to foster collaborative and trusting relationships between the Ministry of Health and traditional Maya midwives.
Introduction
Neonatal deaths and stillbirths remain a critical public health challenge globally, as they account for almost half of all under-5 deaths worldwide.1 Close to 2 million stillbirths and 2.4 million neonatal deaths occur annually.2 3 Despite efforts to increase skilled attendance at delivery, more than half of births worldwide occur outside healthcare facilities.4 In Latin America, economic growth and health system improvements have decreased neonatal mortality. However, these improvements have not been equally distributed.5
Indigenous people in LMICs (low- and middle-income countries) and HICs (high-income countries) alike face significant health disparities, most notably higher rates of infant and maternal mortality than their non-Indigenous peers.6–9 Despite diverse cultural and linguistic backgrounds, they share a common experience of material poverty.10 In Guatemala, where approximately half of the population is considered Indigenous Maya, Xinca or Garífuna, neonatal mortality rates are up to 20% higher among rural Indigenous communities.11 The sociopolitical landscape is characterised by centuries of exclusion from civic and political discourse, resulting in subpar opportunities for economic development and access to healthcare.12 Qualitative research conducted in rural Guatemala revealed logistical challenges that discourage mothers from delivering at hospitals, along with disempowerment and concerns about low-quality and discriminatory care.13–15
Biosocial, mixed-methods research on neonatal mortality among Indigenous populations in Latin America is limited. Previous studies conducted in Asia and Africa using mixed methods have shown that poverty, under-resourced healthcare facilities, and certain cultural beliefs and practices increase neonatal vulnerability.16–19 Most published research about neonatal mortality in Latin America focuses on biological determinants and quantitative analyses.20–22 Studies in Brazil and Perú have highlighted the relationship between racial, socioeconomic, and structural factors and neonatal mortality.23–25 None of these studies have focused on Indigenous populations.
Previous qualitative studies on infant mortality in Mexican Indigenous populations highlighted insufficient human and material resources and mistrust in non-Indigenous healthcare providers as significant barriers to facility-level care.7 26 A comprehensive systematic review of integrated health service delivery networks for maternal and infant care in Latin America revealed similar structural barriers.27 In-depth research exploring how neonatal vulnerability is influenced by sociopolitical forces among Indigenous populations in Latin America is limited. To our knowledge, mixed-methods studies have not been done in Guatemala to explore the causes of neonatal mortality through a biosocial approach. We conducted this study to understand better the interplay of biological, sociocultural and economic factors contributing to inequitable neonatal vulnerability among Indigenous Maya Kaqchikel communities in Guatemala.
Methods
Study design and setting
We conducted an explanatory sequential mixed-methods study with the support of Maya Health Alliance (MHA), a healthcare organisation based in Tecpán, Chimaltenango, Guatemala. The quantitative phase was a secondary analysis of data from the Global Maternal and Newborn Health Registry (MNHR) collected in Chimaltenango between 2014 and 2016. The second qualitative phase consisted of semistructured interviews with caregivers and healthcare workers within the municipality of Tecpán conducted between August and November 2022. Chimaltenango is a department (administrative unit) in rural Western Guatemala (pop. 615,776).11 Tecpán is 1 of 15 municipalities in Chimaltenango (pop. 91,927).11 78% of the Chimaltenango population identifies as Indigenous Maya and 96% of those belong to the Maya Kaqchikel linguistic community.11 In this setting, obstetrical care has traditionally been the responsibility of traditional Maya midwives (TMM) (This term is used throughout the article instead of the more commonly used ‘Traditional Birth Assistant’, which does not accurately represent the societal role, cultural practices, and cosmovision of Maya midwives in Guatemala). Since 2015, MHA has collaborated with local TMMs to promptly identify high-risk obstetrical cases and refer them to advanced care.28
Quantitative methods and data analysis
Quantitative population
We performed a secondary analysis of data from the MNHR, a prospective, observational, population-based study that evaluates obstetrical and newborn health outcomes from eight different countries, including Guatemala.29 These data were collected between 2014 and 2016. We chose this approach because of the geographical overlap between the quantitative and qualitative populations and used the most recent MNHR data publicly available at study start. All pregnant women identified in the recruitment area were enrolled at any moment during prenatal care or delivery. After providing consent, they were followed through 42 days postdelivery.29 Participants derived from 17 geographically defined clusters within Chimaltenango, with at least 300 deliveries per year. Women who did not reside or deliver within the cluster area, unless transferred to a facility outside the area during delivery, were excluded. More detailed procedures have been described elsewhere. Women who were lost to follow-up had a miscarriage or had a medical termination of pregnancy were also excluded from our analysis.
Outcome measures
We investigated perinatal mortality and late neonatal mortality. Perinatal deaths encompassed all fetal deaths occurring at 20 weeks or more of gestation and deaths within days 0–6. We estimate the perinatal mortality rate by dividing the total perinatal deaths by the total births reported (live and stillbirths). Late neonatal deaths were those occurring between days 7 and 28 of life. We estimated the late neonatal mortality rate by dividing late neonatal deaths by live births.
Quantitative data analysis
We derived descriptive statistics of participants in the dataset (table 1). We estimated the bivariate associations between the predictors of interest and perinatal or late neonatal mortality, following a conceptual framework (online supplemental figure 2). We excluded covariates that had 20% or more missing data. We then constructed a final hierarchical mixed effects logistic regression model. We included variables with a statistically significant association in the bivariate analysis (p≤0.05), along with variables of known importance based on literature review, regardless of statistical significance, as fixed effects.30 Cluster ID was included as a random effect. Given that this was a secondary analysis of already-collected data, we did not conduct a power calculation but analysed all the available data. We performed all statistical analyses using STATA/BE V.17.0.
Supplemental material
Characteristics of mothers and newborns in the MNHR dataset
First integration
Regression output was leveraged for designing qualitative interview guides. We crafted qualitative questions to expand our understanding of the relationships between covariates with a statistically significant association and our outcomes.
Qualitative methods
Qualitative data collection and sampling
We conducted 33 semistructured in-depth interviews with TMMs, mothers, nurses and physicians. All interviews were conducted by the first author with the help of a research assistant (MS) who acted as interpreter. The interpreter is from a nearby community, identifies as Maya Kaqchikel, and has previous experience in qualitative research. 30% of interviews were conducted in Spanish, and the remaining in Maya Kaqchikel, with simultaneous Spanish translation. The interviewers did not have a prior relationship with the participants. Interviews took place in participants’ homes or places of work, lasted 35–120 min, and were audiorecorded with the participant consent. We recruited TMMs from the MHA collaborator network. We purposefully sampled TMMs from various Tecpán villages and a wide range of experiences. To engage mothers, we reviewed MHA’s patient records, specifically targeting those who had given birth in the past 12 months and inviting them to participate. We purposefully sampled for age, parity and location diversity among mothers. To enrich for material related to the focus area, we prioritised women with at least one neonatal death. We purposefully sampled nurses and paediatricians at the local hospital within the paediatric and labour-and-delivery areas. Additionally, we interviewed a local physician with extensive administrative experience in Tecpán.
Qualitative data analysis
Audiorecordings of interviews were transcribed and translated into Spanish by AVS and MS. We used an inductive content analysis approach.31 The lead author open-coded four interviews to identify common phrases and drafted a codebook. The first and senior authors trialled the codebook on five more interviews before finalising it. The lead author coded all interviews using Dedoose V.9.0.86 software with spot-checking by the senior author. We identified relevant themes and subthemes by iteratively reviewing relevant excerpts. Finally, illustrative quotes were translated into English.
Final integration and member-checking
After separate quantitative and qualitative results were obtained, we identified areas of divergence and convergence and visualised overall results with a joint display (see figure 1). After completing data analysis, to optimise data validity, the first author conducted member checking with TMMs from the MHA collaborative network to present the study results and elicit their feedback, comments and questions. We presented major quantitative findings and qualitative categories, which resonated with the attendees. The participants concurred and corroborated our findings by providing additional examples from their lived experiences that supported each category.
Cycle of biosocial interaction of factors associated with neonatal mortality.* *Illustrates the integration of our quantitative and qualitative results. Qualitative themes are listed in grey, associated with quantitative findings in white/lighter grey. For more detailed explanations of these relationships, see table 4. Structural forces increase neonatal vulnerability at the individual and community level, which, when paired with health system inadequacies, feed a cycle of fatal outcomes, mistrust in the hospitals, and a culture of blame. ANC, antenatal care; aOR, adjusted OR.
Public and patient involvement
The involvement of our qualitative participants began at the recruitment stage. For over 15 years, the Centre for Indigenous Health Research at MHA has adapted its consent, recruitment and data collection methodologies to align with local customs, emphasising cultural respect and confidentiality. Coauthor MS, a Maya Kaqchikel member, codesigned interview guides to reflect local values.
Results
Quantitative results
Summary of mothers’ and newborns’ characteristics
Table 1 shows the characteristics of mother–newborn dyads in the dataset. 17% of mothers were teenagers, and 12.4% were illiterate. Mean maternal height was 147 cm (±5.4), with 10.6% of participants at 140 cm or less. 55.5% of births occurred at a health facility, and 44.5% at home. The rate of prematurity was 12.2%, and 16.8% of newborns weighed ≤2500 g at birth. The maternal mortality rate was 86 per 100 000 live births.
Key outcomes
There were 1265 (38 per 1000 births) cases of perinatal mortality (including 670 stillbirths and 595 deaths in the first week) and 409 (12.6 per 1000 live births) late neonatal deaths.
Predictors of perinatal mortality
Table 2 shows bivariate and multivariable mixed-effects logistic regression models. Maternal education was associated with a lower risk of perinatal mortality, with university education conferring the largest risk reduction (adjusted OR (aOR): 0.44, 95% CI (0.29 to 0.67)). Caesarean section was also associated with lower odds of perinatal mortality (aOR: 0.50, 95% CI (0.41 to 0.60)). Some of the factors associated with an increased risk of perinatal mortality were primiparity (aOR: 1.31, 95% CI (1.09 to 1.57)), multiparity (aOR: 1.30, 95% CI (1.10 to 1.54)), maternal height under 140 cm (aOR: 1.24, 95% CI (1.03 to 1.50)), low body mass index (BMI) (aOR: 2.35, 95% CI (1.35 to 4.09)), presence of multiple obstetrical complications, prematurity (aOR: 6.03, 95% CI (5.27 to 6.90)) and male sex (aOR: 1.23, 95% CI (1.08 to 1.39)). Notably, a heightened risk was also associated with delivering at a health facility compared with home (aOR: 1.62, 95% CI (1.09 to 2.42)). While the odds of perinatal mortality did not differ significantly between deliveries attended by physicians or nurses and those attended by midwives, birthing without the help of any skilled attendant other than family was associated with a fivefold increased risk (aOR: 5.14, 95% CI (2.84 to 9.29)). The effect of between-cluster differences was 0.7%.
Predictors of perinatal mortality (stillbirths and deaths on days 0–6) in bivariate† and multivariable‡ (N=30 924) logistic regression
Predictors of late neonatal mortality
Table 3 shows bivariate and multivariable mixed-effects logistic regression models. Newborns of mothers who attended only one prenatal care visit, as compared with four, had a higher risk of death (aOR: 1.98, 95% CI (1.31 to 2.98)). Delivery mode or attendant did not have statistically significant associations. Breast feeding within the first hour after delivery (aOR: 0.41, 95% CI (0.31 to 0.54)) was associated with a lower risk. Factors associated with a higher risk were multiple gestation (aOR: 2.16, 95% CI (1.36 to 3.43)), birth weight below 2500 g (aOR: 2.80, 95% CI (2.15 to 3.65)), prematurity (aOR: 2.07, 95% CI (1.58 to 2.72)) and the presence of congenital anomalies (aOR: 37.98, 95% CI (23.09 to 62.50)). The effect of between-cluster differences was 1.8%.
Predictors of late neonatal mortality (deaths on days 7–42) in bivariate† and multivariable‡ (N=30 039) logistic regression
Sensitivity analysis
Online supplemental table 5 provides information about missing data and a comparison of basic demographics in included versus excluded data. Congenital anomaly had the most missing observations (1589), followed by BMI (1557) and early breast feeding (1180). For the BMI variable, the missing data appeared to be at random. However, most missing observations from the breastfeeding-in-the-first-hour and congenital-anomaly variables corresponded to stillbirths and perinatal deaths, which led to their definitive exclusion from the perinatal mortality model (table 2). While we did not consider the early breastfeeding data as truly missing data on stillbirths, we hypothesised that the congenital anomalies variable would absorb much of the variance of the model if present. We tested this by running a hypothetical scenario where all the missing data points were assumed to have congenital anomalies present. The association with some variables such as parity, antenatal care visits, maternal age, height and delivery mode became insignificant, while the association between facility delivery and risk of perinatal mortality became stronger. This hypothetical model is compared with our real model in online supplemental table 6.
Additionally, we ran the late neonatal death model without the two variables in question and compared the results. The outcomes, present in online supplemental table 7, remained largely unchanged. However, the relationship between c-section deliveries and late neonatal mortality changed direction and became statistically significant. This alteration likely reflects the influence of congenital anomalies on late neonatal mortality. Without that variable, the estimated effect of caesarean section was likely compounded. Lastly, we tested our hypothesis that the association between facility deliveries and a higher risk of perinatal mortality was due to facility deliveries carrying a higher risk profile. Online supplemental table 8 shows the comparison between perinatal deaths that occurred in health facilities versus at home. The prevalence of obstetrical and neonatal risk factors is substantially different between the two groups.
Qualitative results
Study population
We performed 33 interviews with mothers (n=16), TMMs (n=10), paediatricians (n=3), nurses from the local public hospital (n=3) and one physician with experience in leadership at the Ministry of Health and the nonprofit sector. Online supplemental table 9 shows the detailed sociodemographic characteristics of participants. All interviewed TMMs are licensed by the Ministry of Health and participate in the MHA collaborator network referenced above. Five major themes and 12 subthemes, presented below with selected quotes, describe the social forces and health system-level factors that shape neonatal vulnerability.
Sources of newborn vulnerability
Inadequate maternal nutrition
TMMs emphasised that maternal nutrition is critical for fetal growth and development. Poverty hinders access to nutritious foods, leading to low birth weight and other complications.
It’s due to poor nutrition. That’s also why babies don't develop in the womb… since there isn't enough food, the baby dies and many miscarriages happen.
TMM, for over 15 years, #5.
Lack of maternal education
All participant groups described the positive impact of women’s education on children’s health. Access to education is hindered by cultural beliefs such as machismo (patriarchal culture) and financial constraints.
If a woman has a higher level of education, her family planning is likely to be better. The interval between children is also likely to be longer.
Healthcare provider, #28
Those who attended school receive advice. With those who didn’t, it’s difficult, especially the older women. They prefer to stick to the customs of the past. They have children who don't grow properly, they are born very small, and sometimes they are born with an illness. In my work, it’s easier with those who attended school; they understand the importance of compliance.
TMM for 30-40 years, #18
Psychoemotional distress during pregnancy
Mothers and TMMs cited experiences of psychoemotional distress during pregnancy related to domestic violence, financial difficulties and marital problems as risk factors for prematurity and neonatal deaths.
Doña X [traditional midwife]…said: ‘You've been through so many struggles, so many tests… so much sadness. And all that has affected the baby. That’s why he was born before his time … he was born with a purple color… and he weighed only 4 pounds.
Mother, 40-50 years old, #11
Factors related to healthcare-seeking behaviours
Family dynamics of decision-making
Key family members are involved in deciding to seek hospital-level care for a birthing mother or newborn. Some male partners deter women from seeking care. As one TMM noted:
There are men that say ‘don't go…why are you going to show yourself to a doctor who will look at you there and examine you?’…They are jealous. That’s why sometimes women don't go to the health center.
TMM for over 15 years, #25
Most mothers do not make decisions independently but look for consent and support from other key family members including grandmothers, whose guidance is valued due to their experience caring for children.
First, I have to ask my baby’s father and then my mother-in-law. I cannot go alone with the sick baby without letting them know… so my baby’s father tells me where to take the baby.
Mother, 30-40 years old, #12.
Traditional midwives between a rock and a hard place
TMMs reported finding themselves conflicted about attending high-risk deliveries when the patient or family refuse to go to the hospital. Although TMMs are trusted, they have limited authority over their patients' healthcare-seeking behaviours. This situation places TMMs in a conflict between helping sick patients and avoiding being blamed by health authorities for adverse outcomes.
She didn't want to go to the hospital… and I thought, what am I going to do? They [authorities] are going to scold me… I couldn't leave the baby here either, because it was already outside and was strangled like this, so it’s about helping quickly… I couldn't let the baby or her die.
TMM for over 20 years, #26.
Shortcomings intrinsic to the healthcare system
Newborn health is a ‘no man’s land’
It is unclear who is responsible for neonatal health supervision in the rural villages of Tecpán. Local health centres are ill equipped to appropriately evaluate newborns and recognise danger signs. A healthcare provider shared that:
There is a mandate that the newborn has to be visited… within eight days. But who is going to do it? The nursing assistant [at the health center], if they organize well and juggle their time, may have the possibility of visiting. But what are they going to do with the child [newborn]?… I don't want to belittle their work. They are very hardworking, but they don't have training or experience…
- Healthcare provider, #33.
Mothers hesitate to ask TMMs about newborn concerns because they perceive TMMs’ role to end after childbirth. Within this ‘no man’s land’, caregivers are often left alone to triage newborn danger signs and decide on a course of action.
The TMM’s job is only the childbirth…when you have your baby, if there is a need, each person has to decide whether to take the baby to the [private] doctor or a health center.
- Mother, 20-30 years old, #1.
The role of language
Patients who mainly speak Maya Kaqchikel reported feeling disempowered, disconnected from providers and discriminated against by professionals primarily speaking Spanish.
We from the village don't know how to speak in Spanish. They should speak our language to have a dialogue to ask questions about the pain. We, as patients, try to explain to them…But in Spanish not everyone can understand. We are left in solitude with the pain, even though we want to communicate that there is a lot of pain, we can't.
- Mother, 30-40 years old, #10
One doctor reflected on the history of colonialism and oppression of Indigenous peoples, explaining that professionals of Indigenous descent, fearing stigma, avoid speaking their Indigenous language.
Here. many of these [doctors and nurses] are of Indigenous origin. But if you ask them to speak their language in the hospital, they won't do it. This is still a situation of the oppression imposed on Indigenous people, to be ashamed of their origin and ways. So that persists.
- Healthcare provider, #33.
Coercive authority of the healthcare system
Mothers and TMMs described several punitive measures within the Guatemalan healthcare system designed to enforce compliance. For example, if pregnant women do not attend at least four prenatal visits health authorities often withhold the birth certificate. This use of coercive authority at the institutional and interpersonal levels fosters a culture of fear.
I can't force them to go [to the clinic or hospital] if they don't want to. Now, because they're not given the birth certificate, some of them go… Sometimes, if they don’t complete all of their [prenatal] check-ups, after they give birth, they don’t give them the baby’s birth certificate.
- TMM for over 15 years, #25.
Mistrust and blame-shifting
Cycle of mistrust in hospitals
Hospital-level care is the last choice for many families and patients due to mistrust and fear of negative experiences. Delay in care is particularly dangerous for newborns, whose condition can worsen quickly and require transfer to distant tertiary-level hospitals or result in fatalities that further erode trust. As a hospital worker noted:
Our own medicine, they see it as the last option, and since it [home remedies] has worked for them in other times…they first try that, and then if the TMM or the healer can't handle it, then they bring them here…The problem is that when some babies are brought, they are already in bad condition; they die or have to be transferred, so they make that association.: ‘If I go to the hospital, they will send me to Guatemala [capital city] or my baby will die.’
- Healthcare provider, #28.
Culture of blame-shifting
Patients, comadronas and healthcare providers tend to blame one another for fatal outcomes. Hospital staff categorise TMMs as ‘good/responsible’ or ‘bad/irresponsible’ based on the patient’s condition on arrival.
There are responsible TMMs. They send them [patients] to the hospital immediately if they see any signs of danger. But there are TMMs who… only when they see that the woman is dying, and the baby is not moving, they send them to the hospital. Unfortunately, sometimes it’s too late.
- Healthcare provider, #32.
One TMM attributed the accusations to discrimination due to lack of formal education:
They [doctor/nurses] even discriminate against us there, asking them [patients] why they go with a TMM. ‘The TMM doesn't know anything, she is illiterate.’… A doctor or nurse has studied, so they discriminate against someone who doesn't have an education. For them [doctors/nurses], we are nothing.
- TMM for over 15 years, #25.
Patients and families also hold TMMs responsible, or blame healthcare providers at the hospital for poor quality of care.
They [families] mention that the baby died because of the TMM. They don't say it was due to an illness; one hears that it’s because of the TMM.
- TMM for over 20 years, #24.
People say that if he [the newborn] goes to the hospital, he will get worse, they will give him medicine, and he will die right away. Sometimes they treat patients well, but sometimes they don't pay attention and they end up making things worse.
- Mother, 40-50 years old, #8.
The positive impact of perseverance and collaboration
Hospital worker’s perseverance despite resource limitations
Despite material limitations and social tensions, mothers highlighted the continued willingness of healthcare workers and TMMs to help their patients and families. Paediatric providers leveraged the limited resources available to optimise the quality of care and highlighted instances of newborns surviving despite unfavourable circumstances.
We help them even though we don't have an intensive care unit as such, but we have two mechanical ventilators, which we have used to save several newborns who eventually go home. And that’s what motivates us to be here.
- Healthcare provider, #27.
One mother expressed gratitude towards the hospital providers for their exceptional efforts in assisting her and her unborn child:
In the hospital? Thank God they did attend to me…they brought about four machines and everything…checked everything to see if they could save her, but she was already dead when I arrived… they did try to find a way to do everything.
- Mother, 20-30 years old, #6.
Cooperation and proximity as an antidote to mistrust and blame
TMMs described how cooperative relationships with local health authorities benefited them and their patients, especially with families hesitant to seek hospital care:
We called the nurse from the center, we had her number…I asked her to come and see the baby because I could tell that something wasn't right…Immediately they came with everything and took them [mother and child]…they went to the hospital.
- TMM for over 30 years, #23.
Additionally, one TMM recounted that health authorities defended her against blame for fatal outcomes because of their previous relationship.
They [health center staff and local authorities] said…thanks to her [the TMM] the girl’s life was saved…the midwife had nothing to do with this maternal death… Because they [health center staff] have seen my work, and everyone in the center knows me and how I work, and they supported me.
- TMM for over 15 years, #26.
Integration of quantitative and qualitative results
Integrated findings are visually depicted in joint displays (figure 1, table 4). We identified several areas of congruence and one divergent finding. Quantitative analysis underscored the heightened risk of perinatal and late neonatal mortality associated with prematurity, low birth weight, maternal height and BMI, and suboptimal maternal education. Qualitative insights illuminated poverty and patriarchal culture as underlying factors influencing women’s education, nutrition and fetal development.
Associations between qualitative and quantitative findings
The unexpected link between delivering at a health facility and increased perinatal mortality emerged. Qualitative data suggested this might be partly attributed to the high-risk profile of facility-delivered patients due to delayed presentation, elevating the risk of intrapartum hypoxia. Mistrust in healthcare institutions and fear of negative experiences were drivers of these delays. While quantitative results indicated lower neonatal mortality with increased prenatal care visits, qualitative data exposed coercive measures enforced by authorities, potentially leading to short-term antenatal visit escalation but perpetuating mistrust in the long run. This emerged as a non-congruent finding.
Discussion
Our study sheds light on the interplay of structural forces contributing to the vulnerability of excess neonatal mortality (figure 1 and table 4). We leveraged our merged qualitative and quantitative results to build an explanatory framework, which is illustrated in a joint display (figure 1).32 Our results illuminate a complex cycle of neonatal mortality, blaming and mistrust that is enabled on one side by pervasive individual and community-level structurally induced vulnerability and on the other by health system shortcomings. We argue that the high rates of neonatal mortality in Maya Kaqchikel communities are intrinsically related to historical and sociopolitical forces such as colonialism, extractive capitalism and neoliberalism which, in our setting, translate to discrimination of Indigenous people, underfunded health systems and concentrated poverty. In other settings with severe health inequities and high neonatal deaths, it is relevant to take these structural forces into account and analyse how they play out in each setting.
The study highlights the crucial role maternal nutrition plays in neonatal mortality. We found that low maternal height and BMI were associated with perinatal mortality, consistent with evidence from other LMIC countries.33 34 Short maternal stature from early life nutritional insults is a known predictor of poor perinatal outcomes.34–36 Short maternal height is associated with smaller pelvic girdle size and increased risk of cephalon-pelvic disproportion, obstructed labour and intrapartum hypoxia.37 The connection between prepregnancy underweight (BMI≤18.5 kg/m²) and an increased risk of prematurity and inadequate birth weight has been previously documented in existing literature.38 Our analysis, however, did not yield statistically significant associations between neonatal mortality and overweight or obesity, despite their established roles as predictors for obstetric and neonatal complications.38 39 In the context of Guatemala’s alarming stunting rates,40 41 the maternal nutritional status, particularly height, plays a pivotal role in evaluating obstetrical risk. Implementing early referrals for mothers with a height <140 cm and those with underweight BMI to deliver in a hospital could potentially yield advantageous outcomes.
Our qualitative participants highlighted the relationship between maternal nutrition and financial destitution. Although stunting is multifactorial, it is closely tied to the effects of colonialism.41 Across all wealth quintiles, Indigenous Guatemalans have higher rates of stunting than their non-Indigenous fellow citizens.40–42 The concentration of poverty and malnutrition among Indigenous communities in Guatemala stems from a history of systematic dispossession of ancestral territories and the erasure of Indigenous practices.12 43 These results highlight the relevance of addressing the root causes of malnutrition, including land and resource redistribution.
Our participants linked the exposure to mental, physical or emotional stress during pregnancy to increased risk of prematurity and other adverse outcomes. Sources of stress included domestic violence, financial insecurity and illnesses. Our research findings are consistent with a substantial body of literature documenting the relationship between prenatal stress and prematurity, mediated by inflammatory mechanisms such as the corticotropin hormone cascade and inflammatory cytokines.44–47 Our qualitative data identified patriarchy, ethnic discrimination and poverty as the underlying causes of psychosocial suffering. Schüssler Fiorenza refers to these ‘interlocking systems of oppression’ as ‘kyriarchy’, which can propagate structural violence in an intergenerational pattern affecting mothers and newborns.48 Improving neonatal outcomes requires addressing these intersectional social forces.
In our study, prenatal care was associated with a lower risk of perinatal and late neonatal mortality, consistent with previous research.49 50 However, our participants highlighted how the Guatemalan health system uses coercive tactics to enforce compliance with antenatal care.51 While this measure may improve performance metrics in the short term, our qualitative findings on cycles of mistrust suggest that in the long term, coercion may contribute to the erosion of long-term relationships with the healthcare system.
The surprising finding of the association between facility delivery and increased risk of perinatal mortality is complex. One possible explanation is selection bias, given that mothers who ultimately deliver at the hospital have more complicated pregnancies. Our multivariable models controlled for multiple obstetric complications, including prolonged labour, haemorrhage, malpresentation, hypertension and others; however, residual confounding may persist. Previous studies in Kenya and Bangladesh showed similar associations,52 53 and those authors hypothesised that this may be related to decreased quality of care or delayed presentation.53 We compared perinatal deaths that occurred at health facilities and homes and proved that the risk profiles are substantially different (see online supplemental table 8). Our qualitative data provided valuable insights, revealing that mistrust in the healthcare system and fear of adverse experiences result in significant delays leading to a higher likelihood of fatal outcomes. The lack of clear responsibility for neonatal health supervision and the pre-established mistrust between Indigenous communities and the healthcare system contribute to a blame culture among all involved. These accusations exacerbate the distance between biomedical providers, patients and TMMs. Interventions that bring all stakeholders together are likely to promote lasting change. Some of our qualitative data highlighted positive developments that disrupted the cycle of mistrust and blame-shifting. For example, some TMMs described cooperation with biomedical providers leading to successful referrals to higher levels of care. Another successful example of collaboration is MHA’s maternal health programme, which has resulted in a higher rate of successful specialty referrals by TMMs. Addressing patients’ hesitation to go to the hospital through accompaniment and advocacy has proven beneficial.14 54 Strategies that bridge patients, TMMs and the healthcare system should be prioritised.
Our results are presented in light of a growing recognition of the value of Indigenous languages, cultures and knowledge for the health of societal and planetary health.55–57 Despite this awareness, Indigenous peoples worldwide, including in high-income settings and countries with universal healthcare coverage, experience substandard quality of life and health outcomes.6 9 10 58 59 These disparities highlight the need to delve deeper into structural and historical factors such as colonialism, patriarchy, and racism, and understand their complex interplay.
Future directions and recommendations for change
This research illuminates several areas of opportunity along the maternal–child healthcare continuum spanning different spheres from the community to the societal levels. A very pressing need is to expand neonatal health supervision capabilities at the community level. Potential strategies include incentives for TMMs to perform these routine neonatal evaluations or allocate resources to increase the capabilities of local health post staff to perform neonatal home visits. Home-based newborn care has proven to halter neonatal deaths in multiple settings similar to rural Guatemala.60 MHA plans to strengthen the availability of newborn care resources in the region by implementing neonatal care navigators that perform home visitations and provide accompaniment for high-risk cases. Additionally, recognising the indispensable role of TMMs in maternal–child health in rural Guatemala, we believe that strengthening collaborative ties between TMMs and the public health system will be largely beneficial. Opening channels of communication and streamlining referral processes are good places to begin. At the institutional level, we urge the Ministry of Health to continue its reconciliation efforts to enhance trust from community members. Strategies such as increasing translation capabilities for Maya languages in health facilities and substituting coercive or punitive strategies for compliance for others that offer more value to patients might prove beneficial. Lastly, at the societal level, critical recognition of the legacies of colonialism and their effect on health is urgently needed to guide praxis that leads to equity.
Strengths and limitations
Our study has several strengths. We use a robust mixed-methods approach to explore neonatal mortality in Indigenous communities in Guatemala. The quantitative sample is large, providing high statistical power. By interviewing mothers, TMMs and hospital staff, we ensured diverse and contrasting qualitative perspectives. Our study also has several limitations. First, our secondary analysis of MNHR data is limited by existing variables in the dataset. Important variables such as self-reported ethnicity and socioeconomic characteristics were not available. Second, many perinatal deaths in rural areas go unreported, which may have led to an underestimation of the perinatal mortality rate among home deliveries. Third, the MNHR data are from one department in rural Guatemala, which limits generalisability to other areas in Guatemala or other LMICs. Furthermore, we acknowledge that our pool of qualitative participants is limited since all TMMs and mothers interviewed were recruited from within the MHA collaborative network and are eligible for substantial support provided by MHA. Therefore, our findings are not representative of the experiences of women in other regions of Guatemala or around the world who do not benefit from this type of service. We anticipate that some of the factors named in our findings might be exacerbated in other settings.
Another limitation is the time discrepancy between the quantitative (2014–2016) and qualitative (2022) data collection. There have been substantial changes to the healthcare and transportation infrastructure within the research context within this period. Improvements in roads and the upgrade of the local centre for medical emergencies to a hospital have facilitated access to emergent obstetric care in the region.61 Contrasting with these improvements, we should also highlight the negative impact of the COVID-19 pandemic on healthcare practices and the availability of services, including those provided by MHA. Despite this discrepancy, we believe that our quantitative data provide a realistic picture of the maternal and newborn landscape in our setting.
Conclusion
The present study highlights important policy and practical implications around stillbirths and neonatal mortality in local and international settings. Some of our quantitative findings can guide the development of an obstetrical and neonatal risk-predicting tool for the researched communities to identify high-risk cases and facilitate early referrals. Our findings demonstrate the critical need to address the historical roots of mistrust between Indigenous populations and governmental institutions. Initiatives that facilitate dialogue between these groups can potentially improve health outcomes. Increasing opportunities for collaboration between biomedical healthcare providers and traditional healers such as TMMs should be prioritised. It is also crucial to recognise that social forces contributing to neonatal vulnerability cannot be addressed in isolation. Instead, their interaction with other systems of oppression should be addressed. Measures that increase the value delivered to patients and foster a trusting relationship with the health system should be prioritised.
Supplemental material
Data availability statement
Data may be obtained from a third party and are not publicly available. Quantitative data may be obtained from a third party and are not publicly available. Qualitative data are not available.
Ethics statements
Patient consent for publication
Ethics approval
We received an exemption from the Harvard Medical School (HMS) and Maya Health Alliance Institutional Review Boards (IRB) to use deidentified data from the MNHR. A data user agreement between HMS and NIH (National Institute of Health)/NICHD (Eunice Kennedy Shriver National Institute of Child Health and Human Development) was also obtained before accessing the MNHR dataset (DAT 22-0274). Ethics approval for the MNHR data in Guatemala was originally obtained from the Instituto de Nutrición de Centro América y Panamá. For the qualitative phase, we obtained approval from the IRB at MHA in Guatemala and HMS in Boston, MA (IRB 22-0105). Before conducting interviews, we obtained informed verbal consent from all qualitative participants.
Acknowledgments
This work was conducted with support from the Master of Medical Sciences in Global Health Delivery programme of Harvard Medical School Department of Global Health and Social Medicine and financial contributions from Harvard University and the Ronda Stryker and William Johnston MMSc Fellowship in Global Health Delivery. The David Rockefeller Center for Latin American Studies of Harvard University also provided funding for this research. The authors thank all the communities, hospitals, health providers, research staff, women and their babies who participated in the Global Network for Women’s and Children’s Health Research Maternal and Neonatal Registry. The previous study was funded by grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development. We acknowledge NICHD DASH for providing the Global Network's Maternal Newborn Health Registry data that was used for this research. Lastly, we extend our gratitude to our qualitative participants for sharing their stories and insights.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Seye Abimbola
X @anahiivenzor
Contributors AVS, HG, ACM, ACCL and PR were closely involved in the study design, methods and data selection. MS helped design the qualitative phase of study including data collection tools, conducted, transcribed and translated most of the interviews, and validated qualitative results. AVS prepared the final manuscript. All contributing authors validated the results and agreed with the manuscript content. AVS is the guarantor and accepts full responsibility for this research.
Funding This research was supported by the Department of Global Health and Social Medicine (N/A) and the David Rockefeller Center for Latin American Studies (N/A) under student research grants.
Disclaimer The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard University and its affiliated academic health care centres. The funding agencies had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.
Competing interests AVS, ACM, HG and MS have no competing interests to report. PR reports research support from: NIH/NICHD, NIH/NIEHS, NIH/NIDDK, the Google Nonprofit Foundation, the World Diabetes Foundation and the Academy of Nutrition and Dietetics. ACCL reports her research work is financially supported by the following institutions: NIH/NICHD, WHO, Bill & Melinda Gates Foundation and Johns Hopkins University. None of these institutions were involved in the design or implementation of this project.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Author note The reflexivity statement for this paper is linked as an online supplemental file 2.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.