Development, validation and application of a machine learning model to estimate salt consumption in 54 countries

  1. Wilmer Cristobal Guzman-Vilca
  2. Manuel Castillo-Cara
  3. Rodrigo M Carrillo-Larco  Is a corresponding author
  1. Universidad Peruana Cayetano Heredia, Peru
  2. Universidad de Lima, Peru
  3. 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")

The following previously published data sets were used

Article and author information

Author details

  1. Wilmer Cristobal Guzman-Vilca

    Universidad Peruana Cayetano Heredia, Lima, Peru
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2194-8496
  2. Manuel Castillo-Cara

    Universidad de Lima, Lima, Peru
    Competing interests
    The authors declare that no competing interests exist.
  3. Rodrigo M Carrillo-Larco

    Imperial College London, London, United Kingdom
    For correspondence
    rcarrill@ic.ac.uk
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2090-1856

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.

Reviewing Editor

  1. Gian Paolo Rossi, University of Padua, Italy

Version history

  1. Received: August 9, 2021
  2. Preprint posted: September 2, 2021 (view preprint)
  3. Accepted: December 15, 2021
  4. Accepted Manuscript published: January 5, 2022 (version 1)
  5. Version of Record published: January 25, 2022 (version 2)

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

  • 763
    Page views
  • 100
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

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)

  1. Wilmer Cristobal Guzman-Vilca
  2. Manuel Castillo-Cara
  3. Rodrigo M Carrillo-Larco
(2022)
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

    1. Epidemiology and Global Health
    Victoria P Mak, Kami White ... Loic Le Marchand
    Research Article Updated

    Background:

    The Coronavirus Disease of 2019 (COVID-19) has impacted the health and day-to-day life of individuals, especially the elderly and people with certain pre-existing medical conditions, including cancer. The purpose of this study was to investigate how COVID-19 impacted access to cancer screenings and treatment, by studying the participants in the Multiethnic Cohort (MEC) study.

    Methods:

    The MEC has been following over 215,000 residents of Hawai‘i and Los Angeles for the development of cancer and other chronic diseases since 1993–1996. It includes men and women of five racial and ethnic groups: African American, Japanese American, Latino, Native Hawaiian, and White. In 2020, surviving participants were sent an invitation to complete an online survey on the impact of COVID-19 on their daily life activities, including adherence to cancer screening and treatment. Approximately 7,000 MEC participants responded. A cross-sectional analysis was performed to investigate the relationships between the postponement of regular health care visits and cancer screening procedures or treatment with race and ethnicity, age, education, and comorbidity.

    Results:

    Women with more education, women with lung disease, COPD, or asthma, and women and men diagnosed with cancer in the past 5 years were more likely to postpone any cancer screening test/procedure due to the COVID-19 pandemic. Groups less likely to postpone cancer screening included older women compared to younger women and Japanese American men and women compared to White men and women.

    Conclusions:

    This study revealed specific associations of race/ethnicity, age, education level, and comorbidities with the cancer-related screening and healthcare of MEC participants during the COVID-19 pandemic. Increased monitoring of patients in high-risk groups for cancer and other diseases is of the utmost importance as the chance of undiagnosed cases or poor prognosis is increased as a result of delayed screening and treatment.

    Funding:

    This research was partially supported by the Omidyar 'Ohana Foundation and grant U01 CA164973 from the National Cancer Institute.

    1. Epidemiology and Global Health
    Gayathri Nagaraj, Shaveta Vinayak ... Dimpy P Shah
    Research Article Updated

    Background:

    Limited information is available for patients with breast cancer (BC) and coronavirus disease 2019 (COVID-19), especially among underrepresented racial/ethnic populations.

    Methods:

    This is a COVID-19 and Cancer Consortium (CCC19) registry-based retrospective cohort study of females with active or history of BC and laboratory-confirmed severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection diagnosed between March 2020 and June 2021 in the US. Primary outcome was COVID-19 severity measured on a five-level ordinal scale, including none of the following complications, hospitalization, intensive care unit admission, mechanical ventilation, and all-cause mortality. Multivariable ordinal logistic regression model identified characteristics associated with COVID-19 severity.

    Results:

    1383 female patient records with BC and COVID-19 were included in the analysis, the median age was 61 years, and median follow-up was 90 days. Multivariable analysis revealed higher odds of COVID-19 severity for older age (aOR per decade, 1.48 [95% CI, 1.32–1.67]); Black patients (aOR 1.74; 95 CI 1.24–2.45), Asian Americans and Pacific Islander patients (aOR 3.40; 95 CI 1.70–6.79) and Other (aOR 2.97; 95 CI 1.71–5.17) racial/ethnic groups; worse ECOG performance status (ECOG PS ≥2: aOR, 7.78 [95% CI, 4.83–12.5]); pre-existing cardiovascular (aOR, 2.26 [95% CI, 1.63–3.15])/pulmonary comorbidities (aOR, 1.65 [95% CI, 1.20–2.29]); diabetes mellitus (aOR, 2.25 [95% CI, 1.66–3.04]); and active and progressing cancer (aOR, 12.5 [95% CI, 6.89–22.6]). Hispanic ethnicity, timing, and type of anti-cancer therapy modalities were not significantly associated with worse COVID-19 outcomes. The total all-cause mortality and hospitalization rate for the entire cohort was 9% and 37%, respectively however, it varied according to the BC disease status.

    Conclusions:

    Using one of the largest registries on cancer and COVID-19, we identified patient and BC-related factors associated with worse COVID-19 outcomes. After adjusting for baseline characteristics, underrepresented racial/ethnic patients experienced worse outcomes compared to non-Hispanic White patients.

    Funding:

    This study was partly supported by National Cancer Institute grant number P30 CA068485 to Tianyi Sun, Sanjay Mishra, Benjamin French, Jeremy L Warner; P30-CA046592 to Christopher R Friese; P30 CA023100 for Rana R McKay; P30-CA054174 for Pankil K Shah and Dimpy P Shah; KL2 TR002646 for Pankil Shah and the American Cancer Society and Hope Foundation for Cancer Research (MRSG-16-152-01-CCE) and P30-CA054174 for Dimpy P Shah. REDCap is developed and supported by Vanderbilt Institute for Clinical and Translational Research grant support (UL1 TR000445 from NCATS/NIH). The funding sources had no role in the writing of the manuscript or the decision to submit it for publication.

    Clinical trial number:

    CCC19 registry is registered on ClinicalTrials.gov, NCT04354701.