Development, validation and application of a machine learning model to estimate salt consumption in 54 countries
- Universidad Peruana Cayetano Heredia, Peru
- Universidad de Lima, Peru
- Imperial College London, United Kingdom
Abstract
Global targets to reduce salt intake have been proposed but their monitoring is challenged by the lack of population-based data on salt consumption. We developed a machine learning (ML) model to predict salt consumption at the population level based on simple predictors and applied this model to national surveys in 54 countries. We used 21 surveys with spot urine samples for the ML model derivation and validation; we developed a supervised ML regression model based on: sex, age, weight, height, systolic and diastolic blood pressure. We applied the ML model to 54 new surveys to quantify the mean salt consumption in the population. The pooled dataset in which we developed the ML model included 49,776 people. Overall, there were no substantial differences between the observed and ML-predicted mean salt intake (p<0.001). The pooled dataset where we applied the ML model included 166,677 people; the predicted mean salt consumption ranged from 6.8 g/day (95% CI: 6.8-6.8 g/day) in Eritrea to 10.0 g/day (95% CI: 9.9-10.0 g/day) in American Samoa. The countries with the highest predicted mean salt intake were in Western Pacific. The lowest predicted intake was found in Africa. The country-specific predicted mean salt intake was within reasonable difference from the best available evidence. A ML model based on readily available predictors estimated daily salt consumption with good accuracy. This model could be used to predict mean salt consumption in the general population where urine samples are not available.
Data availability
This study used nationally-representative survey data that are in the public domain, which was requested through the online repository (https://extranet.who.int/ncdsmicrodata/index.php/home). We provide the analysis code of data preparation and data analysis as supplementary materials to this paper (Source Code File - "Analysis Code | Python and R")
Article and author information
Author details
Wilmer Cristobal Guzman-Vilca
Rodrigo M Carrillo-Larco
Funding
Wellcome Trust (214185/Z/18/Z)
- Rodrigo M Carrillo-Larco
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Copyright
© 2022, Guzman-Vilca et al.
This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 841
- views
-
- 114
- downloads
-
- 0
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Development, validation and application of a machine learning model to estimate salt consumption in 54 countries
eLife 11:e72930.
https://doi.org/10.7554/eLife.72930
Further reading
-
- Epidemiology and Global Health
Background: The role of circulating metabolites on child development is understudied. We investigated associations between children's serum metabolome and early childhood development (ECD).
Methods: Untargeted metabolomics was performed on serum samples of 5,004 children aged 6-59 months, a subset of participants from the Brazilian National Survey on Child Nutrition (ENANI-2019). ECD was assessed using the Survey of Well-being of Young Children's milestones questionnaire. The graded response model was used to estimate developmental age. Developmental quotient (DQ) was calculated as the developmental age divided by chronological age. Partial least square regression selected metabolites with a variable importance projection ≥ 1. The interaction between significant metabolites and the child's age was tested.
Results: Twenty-eight top-ranked metabolites were included in linear regression models adjusted for the child's nutritional status, diet quality, and infant age. Cresol sulfate (β = -0.07; adjusted-p < 0.001), hippuric acid (β = -0.06; adjusted-p < 0.001), phenylacetylglutamine (β = -0.06; adjusted-p < 0.001), and trimethylamine-N-oxide (β = -0.05; adjusted-p = 0.002) showed inverse associations with DQ. We observed opposite directions in the association of DQ for creatinine (for children aged -1 SD: β = -0.05; p =0.01; +1 SD: β = 0.05; p =0.02) and methylhistidine (-1 SD: β = - 0.04; p =0.04; +1 SD: β = 0.04; p =0.03).
Conclusion: Serum biomarkers, including dietary and microbial-derived metabolites involved in the gut-brain axis, may potentially be used to track children at risk for developmental delays.
Funding: Supported by the Brazilian Ministry of Health and the Brazilian National Research Council.
-
- Epidemiology and Global Health
Given the rapid cross-country spread of SARS-CoV-2 and the resulting difficulty in tracking lineage spread, we investigated the potential of combining mobile service data and fine-granular metadata (such as postal codes and genomic data) to advance integrated genomic surveillance of the pandemic in the federal state of Thuringia, Germany. We sequenced over 6500 SARS-CoV-2 Alpha genomes (B.1.1.7) across 7 months within Thuringia while collecting patients’ isolation dates and postal codes. Our dataset is complemented by over 66,000 publicly available German Alpha genomes and mobile service data for Thuringia. We identified the existence and spread of nine persistent mutation variants within the Alpha lineage, seven of which formed separate phylogenetic clusters with different spreading patterns in Thuringia. The remaining two are subclusters. Mobile service data can indicate these clusters’ spread and highlight a potential sampling bias, especially of low-prevalence variants. Thereby, mobile service data can be used either retrospectively to assess surveillance coverage and efficiency from already collected data or to actively guide part of a surveillance sampling process to districts where these variants are expected to emerge. The latter concept was successfully implemented as a proof-of-concept for a mobility-guided sampling strategy in response to the surveillance of Omicron sublineage BQ.1.1. The combination of mobile service data and SARS-CoV-2 surveillance by genome sequencing is a valuable tool for more targeted and responsive surveillance.
