Satellite images and machine learning can identify remote communities to facilitate access to health services

Abstract

Objective Community health systems operating in remote areas require accurate information about where people live to efficiently provide services across large regions. We sought to determine whether a machine learning analyses of satellite imagery can be used to map remote communities to facilitate service delivery and planning. Materials and Methods We developed a method for mapping communities using a deep learning approach that excels at detecting objects within images. We trained an algorithm to detect individual buildings, then examined building clusters to identify groupings suggestive of communities. The approach was validated in southeastern Liberia, by comparing algorithmically generated results with community location data collected manually by enumerators and community health workers. Results The deep learning approach achieved 86.47% positive predictive value and 79.49% sensitivity with respect to individual building detection. The approach identified 75.67% (n = 451) of communities registered through the community enumeration process, and identified an additional 167 potential communities not previously registered. Several instances of false positives and false negatives were identified. Discussion Analysis of satellite images is a promising solution for mapping remote communities rapidly, and with relatively low costs. Further research is needed to determine whether the communities identified algorithmically, but not registered in the manual enumeration process, are currently inhabited. Conclusions To our knowledge, this study represents the first effort to apply image recognition algorithms to rural healthcare delivery. Results suggest that these methods have the potential to enhance community health worker scale-up efforts in underserved remote communities.

Public Health and Epidemiology Informatics: Recent Research Trends Moving toward Public Health Data Science

“…The strategy is similar to that of online targeted advertising except that the goal is to deliver dedicated public health, rather than advertising, messages. The second selected best paper applied deep learning on satellite images to identify rural and hard-to-reach remote communities in low-income countries and help community health workers deliver health services [20]. The geographical segmentation of population based on their access to healthcare is needed to organize specific healthcare delivery and to reduce inequalities.…”

Using Satellite Images and Deep Learning to Identify Associations Between County-Level Mortality and Residential Neighborhood Features Proximal to Schools: A Cross-Sectional Study

“…Despite the study limitations, this approach to assess mortality using CNNs has much external applicability and room for improvement. Offshoots of this approach may further explore how the built environment affects mortality in more precise estimates, such as cities, or explore rural areas exclusively [85]. Additionally, our approach demonstrates that built image features of the environment are correlated with demographic characteristics.…”

Les données de santé en France

“…Cela s'explique par le fait qu'il s'agit de données de gestion, qui de plus nécessitent une très bonne connaissance du contexte juridique et technique du remboursement pour être exploitées correctement, car le codage des données est complexe et variable dans le temps. À titre d'exemples, les actes de détartrage dentaire avaient, jusqu'en 2014, le même code que pour une obturation 11 et trois codes différents étaient utilisés pour un frottis du col utérin, selon le lieu où il était réalisé. Ainsi, comment les diagnostics issus du PMSI peuvent-ils être utilisés correctement sans en connaître les règles de codage ?…”

“…Qu'il s'agisse de données issues d'applications de santé mobile (applications destinées aux patients, comme pour le contrôle de la tension artérielle ou de la glycémie), de dispositifs médicaux implantables communicants (défibrillateurs, prothèses connectées) ou d'appareils de bien-être (montres et balances connectées), ces données font l'objet d'investissements très importants par l'industrie de haute technologie et l'industrie pharmaceutique[6,7] et ouvrent des perspectives de recherche en épidémiologie digitale[8]. Des données massives a priori hors champ santé, comme les images satellitaires, trouvent par exemple des applications épidémiologiques sur des questions de santé-environnement[9][10][11]. Des données individuelles de consommation d'énergie recueillies par des comp-6 L'objectif de cette discipline est de mieux caractériser les tumeurs en exploitant de manière approfondie des données directement contenues dans les approches classiques d'imagerie 7.…”