The use of sustainable and scalable health care technologies in developing countries

Figures & data

Figure 1 Schematic diagram of a crowd-sourcing framework for medical data.

Notes: A consult can happen either in person with health care worker X, or remotely via a webserver with a health care worker or clinician. The mobile application uploads the data to a centralized server which notifies multiple individuals (both other health care workers with expertise and semitrained data reviewers) that there is medical data to label. As each individual submits their label (via an application or web browser) the webserver aggregates the labels using a Bayesian framework, and estimates the error in the aggregate label or diagnosis. When the error is low enough, a diagnosis is sent to the health care worker or patient with recommendations for treatment or further tests. Each data labeler is paid according to their accuracy.
Abbreviations: SMS, short message service; GPRS, general packet radio service.