Mapping the prevalence of severe acute malnutrition in Papua, Indonesia using geostatistical models | BMC Nutrition

Our study is the first to use geospatial modeling to predict SAM prevalence at 1 km spatial resolution from sparse survey data, and our work presents some critical findings. We estimate that approximately 6.3% (95% CI 4.2–10.9%) of all Papuan children under the age of 2 had SAM at the end of 2018. Based on geostatistical analysis, there are areas in Papua that most likely experienced even higher levels of SAM. SAT. Importantly, we used a Bayesian framework to estimate our models, which allowed us to quantify the uncertainty in the predictions. Producing estimates on a gridded surface also allowed for easy visualization of results, providing the flexibility to aggregate gridded estimates into any geographically defined unit, which could be useful from a policy or programmatic perspective. . We have demonstrated this step by aggregating gridded predictions at the Papua district level. The results of the geostatistical model predicted the proportion of Papuan children under 2 with SAM. By combining these estimates with gridded estimates of the population at risk, we were also able to predict the total number of children who experienced SAM. The use of proportions should be weighed against expected estimates (numbers) as they may have different implications for policy responses. For example, although the prevalence of SAM in Asmat district (14%) is higher than in Mimika district (8%), the total number of children affected by SAM in Mimika (2007) was significantly higher. higher than that of Asmat (568 children).

The use of exceedance probabilities to express uncertainty in predictions exceeding SAM thresholds can be particularly useful from a policy-making perspective. For example, parts of Papua Province are likely to be in critical condition, with well over 15% of Papuan children under two suffering from severe acute malnutrition. This has important implications for malnutrition programming in Papua to target those most in need. Our analyzes highlight that significant progress in addressing malnutrition is needed in the province if it is to achieve the WHO Global Nutrition Goal (GNT) to reduce the prevalence of wasting to less than 5% or the United Nations SDGs to end all forms of hunger and malnutrition by 2030. [22].


With respect to this particular study, the analysis has some limitations related to its source datasets. First, the distribution of sample locations is not ideal for geostatistical modeling methods. Geostatistical models derive their strength from the spatial distribution of sample sites and the assumption that areas close to observed samples are more similar. However, in this case study, the primary sampling units are located in a small number of districts, resulting in a low dispersion of observations in the study area. This leaves large portions of the study area to be predicted from distant data points, which can lead to greater uncertainty in predictions and limit our ability to validate results in these areas. Additionally, sites that are very close (

Also, the source data is not representative in the way that a national survey, such as the DHS or national nutrition surveys, would be. The baseline survey data was sampled from households in Papua where the caregiver identified as being of indigenous Papuan origin. [8]. We used this sample to examine geographic variation in SAM, therefore our predicted risk of SAM is most representative of this population of children in Papua. In the absence of native Papuan population estimates, we used total population estimates to approximate the population distribution. If children from different ethnic groups in Papua experience higher (or lower) rates of SAM, then our estimates of the absolute number of children who were SAM – which are based on an estimate of the total population – could be understated ( or over-) estimated. Future studies are needed to understand the distribution of different population groups in Papua and their risk of malnutrition.

In addition, treatment districts for the Child Grant (BANGGA Papua) have been specifically targeted and selected from among the poorest districts in the province. We did not explicitly model this characteristic of the sample, but these factors were taken into account by controlling for accessibility and local context so that predictions in unsampled areas are as accurate as possible. However, with a single source dataset for the Papua analysis, validation options for modeling were also limited. Cross-validation was used to assess out-of-sample precision.

It should also be noted that SAM is a relatively quick indicator of malnutrition. In this regard, SAM or wasting reflects acute or short-term malnutrition, while stunting reflects chronic or long-term malnutrition. [23] and although we can predict SAM at any given time, the prevalence of this indicator may have changed shortly after measurement. Cross-sectional surveys, as used in this study, may not fully capture the rapidly changing risk of SAM. More waves of data at shorter time intervals (e.g. multiple times per season) could help identify “hotspots” with consistently higher SAM risk.

Political Relevance and Benefits

Using modeling methods to combine geospatial data with sparse geotagged survey data to predict health outcomes at high resolution or in unsampled areas offers many potential benefits in program planning and monitoring progress. towards government goals and the SDGs.

As noted earlier, many parts of Papua are very remote and not safely accessible due to outbreaks of violent conflict, making ground-level data collection costly or impossible. [8]. Our findings suggest that this approach could offer real benefits in similar contexts where data collection is not possible or where traditional surveys may experience gaps in coverage, such as remote areas, conflict-affected states or areas with security issues.

While the baseline data for Papua only covered six districts, these modeling techniques allowed us to predict SAM prevalence for the entire province, including districts that were not initially included in the baseline. baseline survey. Some of these districts also had high levels of SAM predicted, illustrating how this approach allowed us to identify SAM hotspots that could be targeted with interventions known to be effective in addressing child undernutrition, such as by example the WHO-recommended approach of Community Management of Acute Malnutrition (CMAM) and Ready-to-Use Therapeutic Foods (RUTF) in community settings [24, 25].

Comments are closed.