PROJECTS
Case Study: Our Impact
Location: Oyo State, Nigeria
Population: 10m
We have delivered 3 fully linked, diverse, and independent campaigns in Oyo State, Nigeria
ANC WORKFLOW
The ANC workflow is followed at PHCs when a pregnant woman arrives for care throughout her pregnancy with the goal to cover approximately 10% of the state, which is 3-5 Local Government Authorities (LGAs).
This has been deployed in 100 clinics.
MALARIA NET SURVEY
A Malaria Net Survey and distribution with a goal to deploy software across all of Oyo State (33 LGAs and 351 wards).
FAMILY SURVEY
A Family Survey as part of Oyo State’s Tomotiya health campaign where a goal is to use the software at approximately 25,000 - 50,000 randomly-chosen households in approximately 17 LGAs.
Success Built on Metrics
ANC Example
Digital Health Integration identifies key metrics and implements dashboards to monitor impact. Quick access to trends allows us to make adjustments to programs in real-time to maximize impact.
KEY METRICS FOR ANC
ANC4/ANC1 -
Proportion of women who completed 4 ANC Visits.
IPTp2/ANC1 -
Proportion of women who have received 2 doses of IPT and have attended one ANC Visit.
LLIN/ANC1 -
Proportion of women with bed nets who have at least one ANC visit.
Parasitological Tests –
Number of women receiving parasitological tests.
Sample ANC Dashboard
PATIENT ID
The ability to track the diagnosis and treatment of individuals with HIV, TB and other diseases is key to both cost efficiency and populations-health improvements in the developing world.
Cost efficiency is achieved by the avoidance of unnecessary tests or treatment, while population health improvement is achieved by the longitudinal tracking of patients, avoiding poorly tracked TB treatment, allowing inappropriate treatment to cause drug resistance.
Tracking an individual accurately is remarkably difficult however, especially in the developing world where infrastructure and record-keeping is fragmented and where patients routinely visit different clinics.
This project relates to developing and deploying a biometric solution that enables cross-clinic Unique ID across a population, geography and time, without centralized coordination or politically-sensitive centralized warehousing of biometric data.
SCALING MACHINE GUIDED DIAGNOSTICS
Disease prevalence and context of symptoms are often minimized or ignored in machine guided diagnosis systems. The information available at scale today is usually seasonal or geographical on a large scale.
This project relates to developing methods based on machine learning that uses data captured from local micro-regions on an ongoing basis to produce regularly-updated context / priors for diseases such as malaria specifically for each micro-region, to enable machine guided diagnosis systems deployed in those regions to perform with much higher accuracy.
As well as improved diagnosis rates and health outcomes, it is expected that this will greatly improve the efficiency and reduce the cost of testing triage protocols.
DATA VALIDATION
Data validation helps reduce a healthcare system’s reliance on highly-curated, centralized databases where health records can typically only be consumed and managed by clinicians.
Data validation adds provenance to health data, allowing patients to share an accurate, longitudinal history with healthcare providers, critical to managing chronic diseases like HIV and TB.
This patient-centric approach is designed to improve the quality of care as well as avoid expensive and unnecessary re-testing.
AT HOME TESTS + ASSESSMENTS
At-Home medical test kits are increasingly being considered as part of complete healthcare solutions since they can provide basic but important information on a patient’s health without the cost and logistical barriers of visiting a clinic.
There is a potential gap however between the results of the test and the patient connecting with care. The project is targeted at allowing the patient to provide personalized context using their own mobile phone that is used in conjunction with the test result to perform a preliminary automatic assessment with the goal of triaging the significance of the test result for the individual as well as the population around them.
The patient is then either encouraged to seek the appropriate healthcare as needed, or, depending on the severity of the assessment and the patient’s permission, healthcare workers can seek out the patient directly.
What We're Working On
Digital Health Integration is working on several key projects including patient ID for longitudinal tracking of patients across clinics, enabling the scaled use of machine-guided diagnosis systems using:
-
Context and objective measures for diseases like malaria
-
Data validation to ensure the integrity and heritage of previously-acquired data to improve diagnoses and avoid expensive re-testing
-
Subjective measurement and assessment on personal mobile-phones for diseases like HIV to help bridge the gap between performing an at-home test and seeking healthcare
Each project requires the incorporation of many component modules developed by DHI as well as many leading capability providers. A goal is to develop standards and APIs for these modules so that they can be reused across these and many other projects, whether they are developed by DHI or other groups.
