Abstract

Objective. The study formulates context-sensitive, evidenced-based policy recommendations that guide targeted interventions for reducing child stunting in Suriname. It identifies key variables associated with stunting, including child dietary adequacy, socioeconomic factors, ethnicity, and WASH conditions. Design. Explanatory sequential mixed-methods study in which quantitative analysis of nationally representative survey data preceded and informed semi-structured key-informant interviews. Quantitative and qualitative findings were integrated during interpretation to identify convergence, contextual explanations, and implications for policy. Setting. Suriname, using data from the 2018 UNICEF Multiple Indicator Cluster Survey (MICS) and key-informant interviews conducted in Paramaribo. Participants. Quantitative analyses included 1,212 children aged 0-23 months with complete anthropometric and feeding data from the MICS 2018 Suriname dataset. Qualitative data were obtained from 13 purposively selected stakeholders, including policymakers, researchers, health workers, and community leaders involved in food systems, nutrition, and child health. Results. Multivariate logistic regression analysis shows that male children and children living in households with unimproved sanitation had significantly higher odds of stunting. Children from Maroon backgrounds had significantly lower odds of stunting than Indigenous children, while children from Creole and Hindustani backgrounds showed a tendency toward lower odds of stunting compared with Indigenous children, although the associations did not reach conventional statistical significance. Treated drinking water also showed a suggestive inverse association with stunting. Qualitative findings highlighted economic constraints, geographic isolation, environmental threats, cultural feeding practices, and seasonal food shortages as important contributors to limited dietary diversity and growth faltering. Conclusions. Findings support an integrated approach to reducing child stunting in Suriname, combining improvements in WASH, child dietary adequacy, and access to nutritious foods. Multisectoral coordination and targeted support for geographically and socially vulnerable communities may help address the environmental, dietary, and structural factors identified in this study.

1. Introduction

Food insecurity continues to cast a long shadow over the public health sector in Suriname, with a staggering 35.8% of the population lacking reliable access to safe and nutritious food in 2023 ¹. Food insecurity is driven by a multifaceted set of determinants that compromise one or more of the four food security pillars: availability, access, utility and sustainability ². These broader food-system conditions can influence child nutrition through the foods available to families, caregiving and feeding practices, illness exposure, and access to health and sanitation services. Stunting (low height-for-age) reflects chronic or recurrent constraints on growth and is associated with adverse consequences for health, cognitive development, school performance, and later economic productivity ^{2, 3, 4, 5}.

The prevalence of stunting among children under five in Suriname has declined only slightly, from 9.7% in 2008 to 9.1% in 2024, while undernourishment among the overall population increased from 7.9% to 10.1% over the same period ⁶. Suriname also shares vulnerabilities common to Small Island Developing States (SIDS), including high food costs, import dependence, exposure to climate-related shocks, and limited domestic production capacity ^{7, 8, 9, 10, 11}. Within Suriname, these constraints are particularly relevant for rural and Indigenous communities, where distance, infrastructure, environmental pressures, and access to services can further shape nutrition-related conditions ¹².

Given the limited progress in reducing stunting, rising undernourishment, and the fragility of Suriname’s food system, there is a clear need for context-specific nutrition policies grounded in robust evidence. Understanding the factors associated with child stunting is therefore important for identifying appropriate areas for intervention. However, evidence on how child dietary adequacy, socioeconomic inequalities, ethnicity, and water, sanitation, and hygiene (WASH) conditions relate to stunting in Suriname remains limited. Moreover, relatively little is known about the broader food-system, environmental, and service-related conditions that may contribute to these patterns or how key stakeholders perceive and respond to them. This study addresses these gaps using an explanatory sequential mixed-methods design, combining quantitative analysis of the nationally representative 2018 Multiple Indicator Cluster Survey (MICS) with semi-structured key-informant interviews. It is guided by the central research question: “How are child dietary adequacy and socioeconomic and environmental conditions associated with stunting among young children in Suriname, and what contextual factors may help explain these patterns?” By integrating quantitative and qualitative evidence, the study aims to generate evidence-informed, context-sensitive recommendations for nutrition policy and practice in Suriname.

2. Methods

This study used an explanatory sequential mixed-methods design, in which a quantitative phase was followed by a qualitative phase. In the first phase, secondary data from the 2018 Suriname Multiple Indicator Cluster Survey (MICS) ¹³ were analysed to identify patterns in child stunting and its associations with child dietary adequacy, socioeconomic characteristics, ethnicity, geography, and WASH conditions.

The quantitative findings directly informed the semi-structured interview guide. Findings related to dietary indicators prompted questions about dietary diversity, complementary feeding, food affordability, and cultural feeding practices. Regional and ethnic differences informed questions about geographic isolation, market access, and conditions in interior communities. Associations with water treatment and sanitation informed questions about drinking-water quality, sanitation infrastructure, river use, environmental contamination, and access to WASH services. Socioeconomic variables informed questions about household costs, employment, caregiver time, and the affordability of nutritious foods. Open-ended questions were retained so that participants could identify additional contextual factors not captured by the survey.

The qualitative phase consisted of semi-structured key-informant interviews with stakeholders involved in food security, nutrition, child health, agriculture, environmental health, research, policymaking, and community programmes in Suriname. This phase was designed to contextualise, explain, or qualify patterns observed in the quantitative analysis and to explore social, environmental, geographical, and institutional pathways that may shape child dietary adequacy and growth.

Integration occurred during interpretation. Quantitative associations were compared with qualitative themes to identify convergence, complementarity, and areas where stakeholder accounts suggested additional explanations or cautions. The quantitative component therefore provided population-level evidence on measured associations, while the qualitative component provided contextual evidence on possible mechanisms and lived or programme-level conditions. Neither component was used to claim causal pathways that were not directly tested.

Population and sampling. MICS 2018 provides a nationally representative sample of households in Suriname, covering all 10 districts ¹³. A two-stage stratified design was used: urban and rural “ressorten” within each district served as strata, and 470 enumeration areas were systematically selected with probability proportional to size. Within each enumeration area, 20 households were sampled, resulting in a final sample of 9,508 households. Sample weights provided in the dataset were applied to support population-representative estimates ¹³. For the present study, the analytical sample was restricted to children aged 0–23 months with complete anthropometric and feeding data, resulting in a sample of 1,212 children.

Data collection. Data were collected as part of the 2018 Suriname MICS using standardised questionnaires adapted to the Surinamese context and administered in Dutch. Children aged 0–23 months were included in the overall analysis of stunting because growth faltering can already be present during early infancy and may reflect prenatal, demographic, socioeconomic, and environmental influences that operate from birth onwards. However, dietary adequacy was assessed only among children aged 6–23 months, as the Minimum Acceptable Diet (MAD) indicator is specifically defined for this age group ¹³. Infants aged 0–5 months were therefore retained in analyses of stunting and its non-dietary correlates but were not classified according to MAD.

Child dietary adequacy was assessed using the Minimum Acceptable Diet (MAD), an individual-level infant and young child feeding indicator for children aged 6–23 months ¹³. In the 2018 MICS, MAD combines age- and breastfeeding-specific criteria for Minimum Dietary Diversity (MDD) and Minimum Meal Frequency (MMF). MDD was achieved when a child consumed foods from at least five of eight defined food groups on the previous day: breastmilk; grains, roots and tubers; legumes and nuts; dairy; flesh foods; eggs; vitamin A-rich fruits and vegetables; and other fruits and vegetables ¹³. MMF determined whether children received the required number of feedings per day, distinguishing between breastfed and non-breastfed children. The thresholds were ≥2 feeds for breastfed 6-8 months, ≥3 feeds for breastfed 9-23 months, and ≥4 for non-breastfed 6-23 months including milk feeds (13). A child met the MAD if they satisfied both the MDD and MMF ¹³.

Stunting was based on height-for-age z-scores calculated using the WHO 2006 Child Growth Standards and coded as “stunted” (HAZ < -2.0 SD) and “not stunted” (HAZ ≥ -2.0 SD) ⁵ Covariates included the child’s age and sex, characteristics of the household-head, maternal education, household size, area and region of residence, ethnicity, household wealth, electricity and internet access, land for agriculture, animal ownership, drinking-water source and treatment, and sanitation conditions ^{13, 14}.

Data processing. Statistical analyses were conducted using IBM SPSS Statistics version 29. Descriptive statistics summarised demographic, socioeconomic, dietary, and environmental characteristics. Associations with stunting were examined in two stages. First, separate bivariate logistic regression models were fitted for each candidate predictor. Second, candidate variables were entered into a multivariable logistic regression model and a backward stepwise procedure was used, with variables with p > 0.10 sequentially removed according to the prespecified selection rule ¹⁵. For interpretation, p < 0.05 was treated as statistically significant; p-values from 0.05 to <0.10 were described as suggestive or marginal rather than statistically significant. A formal explanation of the data logit estimation technique is as follows. Response variables take on two possible outcomes ‘stunted’ of ‘not stunted’.

(1)

With a logit model we aim to model the probability that y=1 as a function of one or more independent variables: x₁, x₂,…,x_n

Problems of Linear Probability Models (e.g. outcomes <0 and >1, heteroskedasticity) are avoided by: 1) transforming probability into odds, 2) take the log of odds and 3) transform log odds into a probability model.

Stepwise this results in:

(2)

(3)

next, accommodate log odds in a logistic function to obtain:

(4)

and, finally, solving for P(y=1) we get a valid probability model

(5)

which always outputs values between 0 and 1. Parameters are log odds, hence, a unit increase in increases the log-odds of the outcome by holding other variables constant.

Ten-fold cross validation. We used a 10-fold cross-validation ¹⁶ to test the model’s (Figure A1) stability of parameter estimates and sensitivity for inclusion or exclusion of observations. Following protocol, data were randomly partitioned into 10 approximately equal-sized subsets, and the model was estimated 10 times with 9 subsets of the data. Stability was assessed by comparing the resulting parameter estimates across iterations.

Population and sampling. The qualitative component consisted of 13 semi-structured key-informant interviews with stakeholders who had professional, policy-related, research, clinical, programme, or community-based expertise relevant to food security, nutrition, child health, agriculture, environmental health, or related fields in Suriname.

Participants were recruited using purposive and snowball sampling. Relevant governmental ministries and organisations were contacted by email and informed about the study. These organisations were asked to identify individuals within their ministry or organisation who had relevant expertise and could provide informed perspectives on the study topics. Individuals identified through this process were subsequently invited to participate in an interview.

Snowball sampling was used to broaden the range of perspectives included in the study. Participants who agreed to be interviewed were asked whether they knew other stakeholders with relevant expertise who might also be willing to participate. Potential participants identified through these referrals were contacted and invited to participate.

Recruitment aimed to include stakeholders from different sectors and professional roles rather than to obtain a statistically representative sample. Individuals were eligible to participate if they were aged 18 years or older, had relevant professional or community-based knowledge of one or more of the study topics, and were able and willing to participate in an interview in Dutch or English. In total, 13 stakeholders participated, including policymakers, researchers, health professionals, programme staff, and community representatives.

Data collection. Semi-structured interviews were conducted in Dutch or English in Paramaribo between March and May 2025. Interviews lasted approximately 30–90 minutes and were conducted in person or online. The interview guides combined questions derived from the study objectives with prompts linked to the preceding quantitative findings. Participants were also encouraged to raise additional issues relevant to child diets, food access, WASH, environmental conditions, and child nutrition in Suriname.

Interview topics included food availability and affordability, access to nutritious foods, dietary diversity, infant and young child feeding practices, household socioeconomic constraints, geographical inequalities, water and sanitation conditions, environmental threats, access to health and nutrition services, and the organisation of food-security and nutrition policies and programmes.

The semi-structured format allowed the interviewer to ask follow-up questions and explore examples provided by participants. All interviews were conducted by the same researcher, audio-recorded with participants’ permission, and subsequently transcribed for analysis. Interviews conducted in Dutch were transcribed and analysed in Dutch to preserve the meaning and context of participants’ accounts. Quotations selected for inclusion in the manuscript were subsequently translated into English by the first author. Interviews conducted in English were transcribed and analysed in English, and quotations from these interviews therefore did not require translation.

Data analysis. Interview data were analysed in ATLAS.ti using a combined deductive and inductive thematic approach. Deductive coding was informed by the study objectives and the IPC Integrated Food Security and Nutrition Framework ¹⁷, including concepts related to food availability, access, utilisation and stability, nutrition and care practices, acute shocks, underlying vulnerability, and access to resources.

Inductive coding was used alongside this framework to identify issues emerging from participants’ accounts that were not adequately captured by the initial codes. Transcripts were first read in full for familiarisation, after which relevant sections were coded, and additional codes were created as new concepts emerged.

Codes were compared across interviews and grouped into broader categories and themes. The analysis considered not only which barriers and facilitators were described, but also the relationships participants drew between them. Particular attention was given to pathways linking import dependence, food prices, geographical isolation, infrastructure, environmental contamination, climate-related shocks, feeding practices, WASH conditions, and child nutrition.

The analysis also considered differences between stakeholder perspectives and accounts that contradicted, qualified, or added nuance to dominant findings. Themes were reviewed in relation to the research question, the IPC framework, and the quantitative findings so that the final interpretation reflected both convergence and divergence between methods.

Concerning the Ethical Considerations on the use of the MICS data, these data were collected by UNICEF and made available through the MICS website. For the data collection the Global MICS Programme’s Technical Committee created an Internal Review Board (IRB) whose members received training in ethical guidelines using UNICEF E-modules on ethics and assisted the survey objectives by reviewing, approving, modifying, or disapproving protocols to protect the privacy of research participants ¹³. Furthermore, verbal consent was gained from all respondents and participants were informed about the voluntary nature of participation, confidentiality, and anonymity of the information collected. They also had the right to refuse to answer any questions or stop the interview at any time ¹³.

For the qualitative component, interviews were conducted to gain insights into food insecurity and child stunting. A purposive and snowball sampling approach was used to reach and identify participants. Participants were informed of the purpose of the study and provided informed consent before the interviews. They were assured that their participation was voluntary, that their responses would remain confidential and anonymous, and that they had the right to decline to answer any questions or withdraw at any time. All interview data was securely stored in a protected environment, accessible only to authorized researchers, and was retained for 6 months before being securely deleted.

This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving research study participants were approved by UNICEF’s Internal Review Board (IRB) ¹³. Written and verbal consent was obtained from all subjects/patients. Verbal consent was witnessed and formally recorded.

3. Results

This section reports on the results of the quantitative (3.1) and qualitative (3.2) analyses.

The quantitative analysis presents the descriptive statistics (3.1.1), results of a bivariate analysis (3.1.2) and the findings of the multivariate analysis (3.1.3).

Table A1 (Annex 1) reports the demographic, socioeconomic, dietary, and environmental characteristics of the analytical sample by stunting status.

A full overview of the results of the bivariate analysis is presented in Table A2 (Annex 2). Before multivariable modelling, correlations among candidate predictors were examined using Spearman correlation coefficients; an absolute correlation coefficient above 0.50 was used as a threshold for potential collinearity ¹⁸. The reported Hosmer–Lemeshow statistic was p = 0.961 ¹⁹.

In the bivariate logistic regression, girls had lower odds of stunting than boys (OR 0.555, p = 0.021). Residence in Sipaliwini was associated with higher odds of stunting compared to Paramaribo (OR 2.667, p = 0.029). compared with children from Indigenous households, lower odds were observed for children from Maroon (OR 0.315, p = 0.005), Creole (OR 0.337, p = 0.015), Hindustani (OR 0.364, p = 0.018), and mixed-ethnicity (OR 0.344, p = 0.028). Unimproved sanitation was associated with higher odds of stunting than improved sanitation (OR 4.118, p = 0.032). Not treating household drinking water showed a suggestive association with higher odds of stunting (OR 1.675 p = 0.080). No other bivariate associations met the prespecified p < 0.10 screening threshold.

Table A3 (Annex 3) presents the final multivariate model. Girls had lower adjusted odds of stunting than boys (AOR = 0.271, p < 0.001), and children from Maroon backgrounds had lower adjusted odds than children from Indigenous backgrounds (AOR = 0.202, p = 0.005). Lower adjusted odds were also observed for children from Creole (AOR = 0.362, p = 0.073) and Hindustani backgrounds (AOR = 0.355, p = 0.063), although these associations did not meet the conventional p < 0.05 threshold. Treated drinking water similarly showed a suggestive inverse association with stunting (AOR = 0.530, p = 0.089). Children living in households with unimproved sanitation had substantially higher adjusted odds of stunting compared with those with improved sanitation (AOR = 19.074, 95% CI: 3.724–97.690, p < 0.001). However, this estimate was based on a small subgroup of 15 children with unimproved sanitation, of whom three were stunted (Table A1). The wide confidence interval therefore indicates considerable uncertainty around the magnitude of the association.

A 10-fold cross validation was conducted to test stability of estimated parameters for inclusion and exclusion of observations. The 10 models that were based on a subset of 90% of the observations included all six significant variables. Figure A1 (Annex 4) shows the standardized parameter estimates. Only minor fluctuations are observed except for the ‘Unimproved sanitation’ variable which decreases in the first to the second round from 1,068 to 0,716. Yet, in the following, remaining rounds, we observe also stable almost similar outcomes for the ‘Unimproved sanitation’ variable. These results suggest overall stability of the selected model across the cross-validation rounds, although they do not eliminate uncertainty arising from sparse observations in some categories.

The qualitative analysis, informed by the IPC framework, generated three overarching themes: food security constraints (3.2.1), nutrition and care (3.2.2), and external stressors (3.2.3) ¹⁷.

Food availability. Food availability in Suriname as described as being influenced by limitations in local production, infrastructure, reliance on imports, and high input costs. “We import more than we produce. Inputs are expensive, local produce can’t compete.” (I6). Participants described price dynamics as placing local producers at a disadvantage and making fresh and healthy foods less affordable and accessible for consumers.

In the interior, agriculture mainly produces for personal use, and access to markets is hindered by poor infrastructure and distances. “In the interior, farming is mostly practiced for personal use. They try to sell the extras, but infrastructure poses a major obstacle due to distance.” (I2). Environmental shocks like floods and droughts, along with post-harvest losses, reduce supply even further. “If there is too much rain, crops get destroyed, which drives up the price. Or if there is a prolonged drought, certain vegetables struggle to grow, and prices also rise.” (I6).

Food access. Economic insecurity was identified as a critical barrier to access to nutritious foods. “The barrier to healthy food is money…If you have to feed a family, you tend to choose something cheaper.” (I6). Participants described how limited household budgets encouraged reliance on cheaper, less diverse foods. Financial constraints were compounded by geographical disparities, particularly in interior regions. “Everything we buy here for 50 SRD (Paramaribo), people there (in the interior) buy for 250 SRD.” (I2). Physical access was also described as difficult in remote communities: “...Sometimes, they travel two or three hours just to do their groceries.” (I2). These accounts illustrate how affordability and distance may constrain dietary adequacy during early childhood.

Time poverty among caregivers, working long hours or multiple jobs, limits their ability to prepare meals. “Many parents have two or three jobs and don’t always have time to prepare a nutritious meal for their children.” (I8). Due to insufficient local food sources, people sometimes turn to bordering countries to purchase cheaper food: “In the interior, it is harder to access fruits and vegetables. So, they cross the river to French Guyana and buy cheap food there.” (I2).

Food utilization. Diets were often described as repetitive and cost-driven, with low dietary diversity among young children. “Lots of bread and rice; few vegetables, fruits, nuts and diary, often depending on money or seasonality.” (I13). Lack of nutritional knowledge is another barrier. Families are provided with dietary counselling to improve feeding habits, but sustainable improvements in feeding habits depend on long-term education and support.

Food safety concerns were also raised. Participants described perceived gaps in pesticide regulation and oversight and expressed concern about contamination of imported and locally produced foods. In interior communities, fishing and hunting were described as important sources of food, but participants also raised concerns about mercury contamination of fish and pollution of rivers, potentially affecting both food safety and the reliability of traditional food sources.

Food stability. Environmental variability poses major threats to all three pillars above. “There are periods where there is excessive rainfall, causing all crops to be destroyed.” (I6). The excessive rainfall affects both food availability and affordability for those who rely on seasonal harvests or small-scale farming. Additionally, droughts disrupt ecosystems and reduce access to hunting, farming, and fishing: “The swamps used for fishing dry up. Because the entire system is disrupted, the animals leave, and we must go deeper into the forest to hunt.” (I2). When climate shocks damage crops, drive up food prices, or disrupt storage and preparation, families face not only reduced access to food but also limited ability to use the food they do have effectively.

Food consumption. Interviewees described a double burden: undernutrition in remote villages and overnutrition among older children in urban areas. “The nutritional status in the southern villages is poor. There are no stores and no development here.” (I2). “Studies have shown that among children around twelve years old, nearly thirty percent are already overweight.” (I3). Several interviewees noted an increasing trend toward energy-dense but nutrient-poor diets, driven by cost rather than quality. This illustrates that economic constraints limit access to nutritious food, forcing individuals and families to make decisions where affordability outweighs nutritional value.

Caring and feeding practices. Exclusive breastfeeding rates were described as very low (~9%), largely due to short maternity leave, mothers juggling multiple jobs, and unsupportive work environments (I1, I13). Interviewees reported suboptimal complementary feeding and strong cultural traditions that sometimes override clinical advice. “Some communities use honey to clean a child’s tongue… my parents always did it this way, so I’ll continue doing it.’ (I1). “Some also give cassava porridge. We explain that it doesn’t contain enough nutrients for such young children. But they say, ‘I gave it to all my five children, and nothing happened to them’.” (I4). In both cases, the families rationalized that it has not caused problems in past generations and would therefore continue these practices.

Health services and environmental health. In interior regions, access to paediatric care is minimal and preventive health services are lacking. “The government has developed a system, ‘Medische Zending’, which is the only form of acute medical care available in villages. (I2). In emergency situations, access remains difficult. “If a child is ill, you bring the child for a check-up. … if necessary, the child is flown to the city.” (I13).

Beyond healthcare, environmental conditions also threaten nutritional health. Many indigenous and tribal populations reside along rivers, which serve as sources of food and water. “The indigenous villages are located along rivers. These rivers are polluted due to gold mining. The fish now contain very high levels of mercury, directly affecting health.” (I2). During dry season, water shortages worsen contamination and spread water-borne diseases. “In 2024, we had dry periods, causing rivers to dry up. The water became greenish and slimy because it was near the bottom and in contact with algae, diseases emerged as a result.” (I2).

Acute events or ongoing conditions. Acute shocks repeatedly disrupt food systems. During the early stages of COVID-19, supply chain disruptions worsened food access. “We were worried more about food security than we were about getting COVID or dying from COVID.” (I5). A brief uptick in home gardening expanded household supply but faded after the lockdown ended.

Flooding and other extreme weather events were also described as damaging crops, while farmers faced difficulties obtaining insurance. “Because nationally, the insurance companies were not willing to insure farms. They say it’s too much of a risk.” (I5). Participants linked these conditions to the possibility of losing harvests without compensation and to wider concerns about preparedness for climate-related shocks. “Preparedness is weak. We are highly vulnerable because there is almost no preparedness. Actions only start when something happens.” (I1).

Vulnerability, resources and control. Interviewees described situations in which single mothers were responsible for large households with limited financial resources and support. They linked the combination of financial constraints and limited caregiving time to repetitive meals with low dietary diversity.

A lack of qualified agriculture personnel, driven by low pay and limited incentives, suppresses local production and sustains reliance on imports. “It is getting the people who are willing, available to do that sort of work… So, either they're not there or they have no interest. The compensation is not good. The salaries are low.” (I5). In health clinics, practitioners report seeing children receiving inadequate or poor-quality food, often due to a lack of knowledge and financial hardships. “Once at the clinic, you see children who really get far too little food. This is partly due to lack of knowledge, but certainly also due to financial problems.” (I13).

4. Discussion and Conclusions

This mixed-methods study examined factors associated with child stunting in Suriname and explored the broader social, environmental, dietary, and structural conditions that may help contextualise these associations. The quantitative findings identified sex, ethnicity, and sanitation as important correlates of stunting, while the qualitative findings highlighted constraints related to food affordability and availability, geographic isolation, child-feeding practices, environmental contamination, access to services, and climate-related disruptions. Together, the findings suggest that child growth in Suriname is shaped by interacting biological, environmental, dietary, and social conditions rather than by a single determinant.

Girls had substantially lower adjusted odds of stunting than boys. This finding is consistent with evidence from low- and middle-income countries showing a higher prevalence of undernutrition and linear growth faltering among boys in early childhood ^{20, 21, 22}. Proposed explanations include sex-related differences in biological vulnerability, susceptibility to infectious disease, and nutritional requirements, although the underlying mechanisms remain incompletely understood ^{21, 22}.

Ethnicity was also associated with stunting. Children from Maroon backgrounds had significantly lower adjusted odds of stunting than Indigenous children. The estimated odds ratios for Creole and Hindustani children were also below one, although these associations did not meet the conventional p < 0.05 threshold and should therefore be interpreted cautiously. These findings are broadly consistent with evidence from Latin America showing persistent inequalities in child stunting between Indigenous and non-Indigenous populations ^{23, 24, 25}. Gatica-Domínguez et al. ²³ found that Indigenous children in several Latin American countries experienced higher levels of stunting even after accounting for socioeconomic characteristics, suggesting that household-level indicators alone may not fully capture the structural conditions underlying these inequalities.

The qualitative findings provide further context for the observed ethnic differences. Participants described substantial geographical barriers in interior regions, including long travel distances, limited access to markets and health services, dependence on river transport, and greater vulnerability to environmental disruption. These circumstances may affect access to diverse foods, preventive healthcare, safe water, and sanitation. At the same time, Indigenous communities in Suriname face broader structural challenges related to collective land and resource rights, self-determination, cultural continuity, and dependence on traditional environments for livelihoods and food resources ^{26, 27, 28}. Constraints on access to land, forests, and rivers, or environmental changes affecting these resources, may therefore also influence traditional food systems and livelihood opportunities. However, because land rights, cultural autonomy, and access to traditional food systems were not directly assessed in this study, these factors should be considered contextual explanations rather than demonstrated mechanisms. Future research involving Indigenous caregivers and communities directly would help clarify how these structural conditions, environmental change, and service availability interact to influence child nutrition.

One of the strongest quantitative findings was the association between unimproved sanitation and stunting. Children living in households with unimproved sanitation had substantially higher adjusted odds of stunting than those with improved sanitation. This is consistent with literature linking inadequate WASH conditions with impaired linear growth ^{29, 30, 31}. Poor sanitation may increase exposure to faecal pathogens, repeated enteric infections, and intestinal inflammation, thereby reducing nutrient absorption and affecting growth ^{29, 30}.

The magnitude of this association nevertheless requires careful interpretation. Only 15 children in the sample lived in households classified as having unimproved sanitation, of whom three were stunted. The estimate was therefore accompanied by a very wide confidence interval (AOR = 19.074, 95% CI: 3.724–97.690). Sparse observations within a category can produce unstable or exaggerated odds-ratio estimates in logistic regression ³². The result therefore supports an association between sanitation conditions and stunting, but should not be interpreted as showing that unimproved sanitation increases the odds of stunting by precisely nineteen-fold. Residual confounding also cannot be excluded.

The qualitative findings add context to this association. Interviewees described dependence on rivers for food and water, seasonal water shortages, and contamination associated with gold-mining activities in interior communities. These accounts cannot explain the magnitude of the sanitation odds ratio, but they illustrate the wider environmental conditions in which WASH-related risks occur. Previous research from Suriname has also documented mercury exposure among fish-reliant populations in the interior ³³, reinforcing the need to consider sanitation together with water quality, food safety, and environmental contamination.

At the same time, the strong observational association between sanitation and stunting should not be interpreted as evidence that WASH improvements alone will necessarily reduce stunting. Large cluster-randomised trials in Zimbabwe, Kenya, and Bangladesh have reported limited or inconsistent effects of standalone or basic WASH interventions on linear growth ^{34, 35, 36}. This highlights the multifactorial nature of stunting. WASH improvements should therefore be considered as one component of an integrated nutrition-sensitive strategy that also addresses dietary adequacy, infection prevention, caregiving, and access to healthcare.

Household wealth, maternal education, and urban-rural residence were not statistically significant in the regression analyses. This differs from the broader literature, in which socioeconomic disadvantage, lower maternal education, and rural residence are consistently associated with child stunting ^{20, 23}. However, this should not be interpreted as evidence that socioeconomic circumstances are unimportant in Suriname. The qualitative component repeatedly identified financial constraints as shaping food choices. Participants described families selecting cheaper and less nutritious foods, caregivers working multiple jobs with limited time for meal preparation, and substantially higher food prices and transport barriers in remote areas.

Several factors may help explain this difference. The relatively small number of stunted children may have limited the statistical power to detect more modest associations. In addition, socioeconomic influences may act through more immediate conditions, such as sanitation, geographic access, dietary practices, and environmental circumstances. Household resources, for example, can shape access to safe sanitation, diverse foods, markets, and health services. As these pathways were not formally examined through mediation analysis, they should be regarded as plausible explanations rather than mechanisms established by the present study.

The qualitative findings also highlighted challenges surrounding infant and young child feeding. Participants described limited dietary diversity, reliance on inexpensive staple foods, low exclusive breastfeeding, and complementary-feeding practices influenced by financial constraints, working conditions, knowledge, and cultural traditions. These findings suggest that improving child nutrition requires more than information alone. Caregivers may understand nutritional recommendations but still be unable to follow them when nutritious foods are unaffordable or inaccessible, or when employment conditions limit breastfeeding and meal preparation.

These findings are relevant within the broader nutritional transition occurring in Small Island Developing States and across Latin America. Participants described both undernutrition in remote communities and increasing overweight among older children, alongside greater reliance on energy-dense, nutrient-poor foods. Similar patterns of a double or triple burden of malnutrition have been documented in SIDS, where import dependence, changing food environments, climate vulnerability, and constraints on local production can contribute simultaneously to undernutrition, micronutrient deficiencies, and overweight ^{37, 38, 39}.

A major strength of this study is the explanatory sequential mixed-methods design. The quantitative component provided population-level evidence on factors associated with stunting, while the qualitative component helped explain and contextualise several of these patterns. For example, the sanitation association was complemented by accounts of water insecurity and environmental health challenges, while the absence of independent socioeconomic associations in the regression contrasted with repeated descriptions of affordability, time poverty, and geographic access. This illustrates the value of combining survey data with stakeholder perspectives.

Several limitations should be considered. First, the cross-sectional MICS data prevent causal inference. Second, the relatively low prevalence of stunting resulted in a limited number of outcome events, and some exposure categories were very small. This was particularly relevant for unimproved sanitation and contributed to the wide confidence interval around the odds ratio. Although the ten-fold cross-validation suggested overall model stability, it does not remove uncertainty arising from sparse observations within individual categories.

A further limitation concerns measurement. MAD reflects child dietary adequacy during the preceding day and does not directly assess household food insecurity, longer-term dietary intake, food quantity, or seasonal variation ¹³. Broader conclusions concerning household food-security constraints therefore rely primarily on the qualitative component. In addition, socioeconomic indicators were self-reported and may not fully capture household economic stability. Finally, the qualitative participants were policymakers, researchers, health professionals, programme staff, and community representatives rather than a representative sample of caregivers. Their perspectives are valuable but cannot be assumed to represent all families in Suriname, particularly those living in remote Indigenous communities.

Policy implications and recommendations

• Prioritise WASH improvements as part of an integrated nutrition strategy. The strong association between unimproved sanitation and stunting supports targeted investment in safe sanitation and drinking-water infrastructure, particularly in remote and interior communities. However, because large trials have shown that WASH interventions alone may not improve linear growth, these measures should be combined with support for child dietary adequacy, infection prevention, caregiving, and access to health services rather than implemented as a standalone stunting intervention.

• Strengthen infant and young child feeding support with a focus on dietary adequacy. Community and primary-care programmes should provide practical, culturally appropriate guidance on dietary diversity, complementary feeding, breastfeeding, and hygiene, while recognising barriers such as food costs, caregiver time constraints, and local food availability.

• Improve access to diverse, nutrient-rich foods, particularly in geographically isolated communities. Support for local food production, improved transport and market access, and incentives for nutrient-rich crops could help address the availability and affordability barriers identified in the qualitative findings.

• Strengthen multisectoral coordination and nutrition monitoring. A national nutrition council or task force could coordinate action across health, agriculture, education, environmental management, and social protection, with a focus on child dietary adequacy, WASH, household food-security conditions, and geographic and ethnic inequalities.

Conclusion. This study shows that child stunting in Suriname is associated with a combination of individual, environmental, and social factors. Sex, ethnicity, and sanitation were important correlates, while the qualitative findings highlighted broader challenges related to food affordability, geographic access, child-feeding practices, environmental conditions, and access to services.

The strong association with unimproved sanitation should be interpreted cautiously because of the small number of children in this category, but it reinforces the importance of WASH within a broader nutrition strategy. Overall, reducing child stunting in Suriname will likely require coordinated action that combines improvements in child dietary adequacy, WASH, food access, caregiving support, and targeted attention to vulnerable communities.

ACKNOWLEDGEMENTS

The authors would like to thank Ben Sonneveld for his continuous guidance and support throughout the research process, and Amani for her valuable assistance with editing the manuscript. We are also sincerely grateful to all participants who generously shared their time and perspectives and made this study possible.

Statement of Competing Interests

The authors have no competing interests.

References