Dataset Information

Use of earth observation-derived hydrometeorological variables to model and predict rotavirus infection (MAL-ED): a multisite cohort study.

ABSTRACT:

Background

Climate change threatens to undermine recent progress in reducing global deaths from diarrhoeal disease in children. However, the scarcity of evidence about how individual environmental factors affect transmission of specific pathogens makes prediction of trends under different climate scenarios challenging. We aimed to model associations between daily estimates of a suite of hydrometeorological variables and rotavirus infection status ascertained through community-based surveillance.

Methods

For this analysis of multisite cohort data, rotavirus infection status was ascertained through community-based surveillance of infants in the eight-site MAL-ED cohort study, and matched by date with earth observation estimates of nine hydrometeorological variables from the Global Land Data Assimilation System: daily total precipitation volume (mm), daily total surface runoff (mm), surface pressure (mbar), wind speed (m/s), relative humidity (%), soil moisture (%), solar radiation (W/m²), specific humidity (kg/kg), and average daily temperatures (°C). Lag relationships, independent effects, and interactions were characterised by use of modified Poisson models and compared with and without adjustment for seasonality and between-site variation. Final models were created with stepwise selection of main effects and interactions and their validity assessed by excluding each site in turn and calculating Tjur's Coefficients of Determination.

Findings

All nine hydrometeorological variables were significantly associated with rotavirus infection after adjusting for seasonality and between-site variation over multiple consecutive or non-consecutive lags, showing complex, often non-linear associations that differed by symptom status and showed considerable mutual interaction. The final models explained 5·9% to 6·2% of the variability in rotavirus infection in the pooled data and their predictions explained between 0·0% and 14·1% of the variability at individual study sites.

Interpretation

These results suggest that the effect of climate on rotavirus transmission was mediated by four independent mechanisms: waterborne dispersal, airborne dispersal, virus survival on soil and surfaces, and host factors. Earth observation data products available at a global scale and at subdaily resolution can be combined with longitudinal surveillance data to test hypotheses about routes and drivers of transmission but showed little potential for making predictions in this setting.

Funding

Bill & Melinda Gates Foundation; Foundation for the National Institutes of Health, National Institutes of Health, Fogarty International Center; Sherrilyn and Ken Fisher Center for Environmental Infectious Diseases, Johns Hopkins School of Medicine; and NASA's Group on Earth Observations Work Programme.

SUBMITTER: Colston JM

PROVIDER: S-EPMC6650544 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundMost studies of the causes of diarrhoea in low-income and middle-income countries have looked at severe disease in people presenting for care, and there are few estimates of pathogen-specific diarrhoea burdens in the community.MethodsWe undertook a birth cohort study with not only intensive community surveillance for diarrhoea but also routine collection of non-diarrhoeal stools from eight sites in South America, Africa, and Asia. We enrolled children within 17 days of birth, and diarrhoeal episodes (defined as maternal report of three or more loose stools in 24 h, or one loose stool with visible blood) were identified through twice-weekly home visits by fieldworkers over a follow-up period of 24 months. Non-diarrhoeal stool specimens were also collected for surveillance for months 1-12, 15, 18, 21, and 24. Stools were analysed for a broad range of enteropathogens using culture, enzyme immunoassay, and PCR. We used the adjusted attributable fraction (AF) to estimate pathogen-specific burdens of diarrhoea.FindingsBetween November 26, 2009, and February 25, 2014, we tested 7318 diarrhoeal and 24 310 non-diarrhoeal stools collected from 2145 children aged 0-24 months. Pathogen detection was common in non-diarrhoeal stools but was higher with diarrhoea. Norovirus GII (AF 5·2%, 95% CI 3·0-7·1), rotavirus (4·8%, 4·5-5·0), Campylobacter spp (3·5%, 0·4-6·3), astrovirus (2·7%, 2·2-3·1), and Cryptosporidium spp (2·0%, 1·3-2·6) exhibited the highest attributable burdens of diarrhoea in the first year of life. The major pathogens associated with diarrhoea in the second year of life were Campylobacter spp (7·9%, 3·1-12·1), norovirus GII (5·4%, 2·1-7·8), rotavirus (4·9%, 4·4-5·2), astrovirus (4·2%, 3·5-4·7), and Shigella spp (4·0%, 3·6-4·3). Rotavirus had the highest AF for sites without rotavirus vaccination and the fifth highest AF for sites with the vaccination. There was substantial variation in pathogens according to geography, diarrhoea severity, and season. Bloody diarrhoea was primarily associated with Campylobacter spp and Shigella spp, fever and vomiting with rotavirus, and vomiting with norovirus GII.InterpretationThere was substantial heterogeneity in pathogen-specific burdens of diarrhoea, with important determinants including age, geography, season, rotavirus vaccine usage, and symptoms. These findings suggest that although single-pathogen strategies have an important role in the reduction of the burden of severe diarrhoeal disease, the effect of such interventions on total diarrhoeal incidence at the community level might be limited.

Project description:BackgroundThere is no standardized approach to comparing socioeconomic status (SES) across multiple sites in epidemiological studies. This is particularly problematic when cross-country comparisons are of interest. We sought to develop a simple measure of SES that would perform well across diverse, resource-limited settings.MethodsA cross-sectional study was conducted with 800 children aged 24 to 60 months across eight resource-limited settings. Parents were asked to respond to a household SES questionnaire, and the height of each child was measured. A statistical analysis was done in two phases. First, the best approach for selecting and weighting household assets as a proxy for wealth was identified. We compared four approaches to measuring wealth: maternal education, principal components analysis, Multidimensional Poverty Index, and a novel variable selection approach based on the use of random forests. Second, the selected wealth measure was combined with other relevant variables to form a more complete measure of household SES. We used child height-for-age Z-score (HAZ) as the outcome of interest.ResultsMean age of study children was 41 months, 52% were boys, and 42% were stunted. Using cross-validation, we found that random forests yielded the lowest prediction error when selecting assets as a measure of household wealth. The final SES index included access to improved water and sanitation, eight selected assets, maternal education, and household income (the WAMI index). A 25% difference in the WAMI index was positively associated with a difference of 0.38 standard deviations in HAZ (95% CI 0.22 to 0.55).ConclusionsStatistical learning methods such as random forests provide an alternative to principal components analysis in the development of SES scores. Results from this multicountry study demonstrate the validity of a simplified SES index. With further validation, this simplified index may provide a standard approach for SES adjustment across resource-limited settings.