Monitoring Disease Trends using Hospital Traffic Data from High Resolution Satellite Imagery: A Feasibility Study

DSpace/Manakin Repository

Monitoring Disease Trends using Hospital Traffic Data from High Resolution Satellite Imagery: A Feasibility Study

Citable link to this page

 

 
Title: Monitoring Disease Trends using Hospital Traffic Data from High Resolution Satellite Imagery: A Feasibility Study
Author: Nsoesie, Elaine O.; Butler, Patrick; Ramakrishnan, Naren; Mekaru, Sumiko R.; Brownstein, John S.

Note: Order does not necessarily reflect citation order of authors.

Citation: Nsoesie, Elaine O., Patrick Butler, Naren Ramakrishnan, Sumiko R. Mekaru, and John S. Brownstein. 2015. “Monitoring Disease Trends using Hospital Traffic Data from High Resolution Satellite Imagery: A Feasibility Study.” Scientific Reports 5 (1): 9112. doi:10.1038/srep09112. http://dx.doi.org/10.1038/srep09112.
Full Text & Related Files:
Abstract: Challenges with alternative data sources for disease surveillance include differentiating the signal from the noise, and obtaining information from data constrained settings. For the latter, events such as increases in hospital traffic could serve as early indicators of social disruption resulting from disease. In this study, we evaluate the feasibility of using hospital parking lot traffic data extracted from high-resolution satellite imagery to augment public health disease surveillance in Chile, Argentina and Mexico. We used archived satellite imagery collected from January 2010 to May 2013 and data on the incidence of respiratory virus illnesses from the Pan American Health Organization as a reference. We developed dynamical Elastic Net multivariable linear regression models to estimate the incidence of respiratory virus illnesses using hospital traffic and assessed how to minimize the effects of noise on the models. We noted that predictions based on models fitted using a sample of observations were better. The results were consistent across countries with selected models having reasonably low normalized root-mean-squared errors and high correlations for both the fits and predictions. The observations from this study suggest that if properly procured and combined with other information, this data source could be useful for monitoring disease trends.
Published Version: doi:10.1038/srep09112
Other Sources: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4357853/pdf/
Terms of Use: This article is made available under the terms and conditions applicable to Other Posted Material, as set forth at http://nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA
Citable link to this page: http://nrs.harvard.edu/urn-3:HUL.InstRepos:14351334
Downloads of this work:

Show full Dublin Core record

This item appears in the following Collection(s)

 
 

Search DASH


Advanced Search
 
 

Submitters