Guyana Outreach Project
Project DataREACH began in the summer of 2015, with a trip to Guyana,
presenting at hospitals, universities, medical schools, and government
entities (Ministry of Health and National Cancer Institute) about the
potential for the application of data science in medicine, including
machine learning/predictive analytics for risk assessment, as well as
exploring techniques to utilize surveillance data. Project DataREACH
was commended by Guyana's Ministry of Health and National Cancer
Institute for this outreach effort, and fostered numerous connections
from which to explore future projects.
Cameroon World Health Organization (WHO) Smart Surveillance Platform
In December 2016, Project DataREACH partnered with the World Health
Organization (WHO) country office in Cameroon, to develop a data
science platform for effectively analyzing surveillance data for
infectious disease data in Cameroon. A two step smart surveillance
platform was developed, the first step being the implementation of
outbreak detection algorithms in health districts in order to enable
alerts/flags for outbreaks in a timely manner. The second step of the
smart surveillance platform involves a novel application of
multivariate time series analysis to analyze the correlations between
regions by examining cross correlations between their modeled time
series, and using these correlations to extrapolate impact response
conclusions, and understand disease spread and result in efficient
resource allocation to alleviate the casualties resulting from these
infectious diseases.
DataREACH EMR Software
In the spring of 2017, Project DataREACH developed custom software for
data collection, to be used in health systems without electronic data
collection capabilities. Data stored into the server connected to this
platform will be designed before to feed into machine learning
pipeline, developed in partnerships with doctors and other medical
professionals. Advancements in medical data collection can also lead
to cardiovascular risk assessment, through ASCVD algorithm. A demo of
an ML model deployed from this platform to predict artery narrowing
from an assortment of risk factors can be used at db.datareach.org.
Through this software, Project DataREACH hopes to build the first from
scratch machine learning pipelines for medicine in Cameroon.