The prevalence of disease in sub-Saharan Africa makes it the best candidate for investing time to save lives. Today, there is too little listening to local expertise to build bottom-up approaches for everything health. As a community, we can be better at clinical trials, targeted interventions, effective implementations, artificial intelligence, data, and applied technology, all while thinking about realistic sustainability and what core assets are needed to create jobs as a side effect. While Africans can act together, Africa is really a huge breadth of people, cultures, languages and individuals. We’ve found often it just takes active listening, bringing the right people together, to make a match: joining motives and resources and people to create transformation. It is possible to make sustainable change here.
AI/ ML/ Big Data
Artificial intelligence (AI), machine learning (ML), data science, big data: it is everywhere in the global North. If applied consistently at global North partners—funders, pharma, biotech, Big Tech, principal investigators’ universities—the ability to ‘activate’ is faster. If applied within low- and middle-income countries, where AI and data science only exists in small blooms today, it can be used to leapfrog missing infrastructure, and show the need for an increase in granular, patient-level digital data beyond images. In the global North, novel data is routinely being plumbed to assess its value as inputs to predictive models and learning to be a ‘data scientist’ and finding a peer community is cheaper and easier each year, and often subsidized by companies, universities or governments. As we solve health problems in LMIC, we desire to increase the ubiquity of AI and ML in LMIC—which may only be possible with increased prevalence of big data in the form of electronic health records.
Global health clinical trials have an amazing opportunity to be better, faster, more generalizable to local populations, and answer more questions. Many clinical trial innovations in the global North are not implemented today in global health trials. Even more trials fail due to a lack of leveraging critically valuable human intelligence from locals on the ground, both health workers and regional principal investigators. At the intersection of clinical trials and a set of modern approaches—AI, data, databases, master protocols, ML, public-private partnerships, to name a few—there is a gap we fill enhancing trials, ensuring better designs and implementations, and building long-term capacity, ‘centers of excellence’, registries and stable assets.
New technology, digital, big data and AI/ ML have opened up new possibilities to identify earlier where and when an epidemic is about to start, or about to spread and move. Working with the World Economic Forum Epidemics Readiness Accelerator, we were able to touch the confluence of advocacy, machine learning data science, public-private partnerships, and open source software, toward using novel data to predict epidemic movement and risk. We first evangelized modern weather forecasting as a predictive exemplar to understand the multi-decade evolution that lies ahead for us: toward precise, current, big data and AI-based models that can be constantly improved.
Open Source Software
This millennium has shown results in the ‘experiment’ that is free and open-source software. In health generally, successful free and open-source projects exist across multiple disciplines. Projects such as deepchem (drug discovery), OpenMRS (electronic health record), and cTAKES (natural language processing for clinical text and doctor’s notes) are a few notable examples. There is a great breadth across projects, with a myriad of use cases. Developers now consider “open-sourcing” as a viable sustainability approach. Most compelling: large masses of global organizations have standardized on single community projects and grown them. The epitome of this is the Observational Health Data Science & Informatics consortium. OHDSI (ohdsi.org ) has hundreds of researchers and collaborators, dozens of competing pharma’s; academic medical centers and CRO’s; and even sovereign nations having standardized on OHDSI software. The creators and sustainers of the software are employees of the organizations using the software!
Precision Public Health
Precision public health gives us a chance to solve more suffering with the same investment. Precision public health uses technology and robotics, big data, data science and personalized medicine. These techniques are used to bring more effective cures to the cohort of those in most need. Examples include use of drones in humanitarian crises, use of geospatial data for improved understanding of what is happening on the ground, hotspot and coldspot predictive analytics mapping, vaccinomics and innovations in ring vaccines, and patient-level risk stratification and intervention.
Real World Evidence
The use of longitudinal, patient level data (anonymized) has grown widely for decades in some sectors of the health economy. While much of this data originally was only digital in the hands of government or insurance company ‘payers’, the advent of electronic health records, patient data protections, and even wearables and connected medical devices have made much more data accessible from syndicated data providers. However, only tiny slivers of digital, purchase-able, anonymized patient-level real world data (RWD) are from patients living in low- and middle-income countries. As such, mining such data for real world evidence (RWE) of outcomes, common treatment pathways, adverse events, and more is not realistic for global health populations of interest.