Artificial Intelligence Is the Modern Antibiotic
The future of healthcare, like the future of most industries, is inextricably tied to the tech industry. Tech is important to not just the level of patient care, but also the operation of systems delivering this care, arguably most importantly, patient access to health services.
But timing plays a key part in success. The right solution offered at the right time can generate a sea change.
Healthcare: Opportunity, Discovery & Progress
Today, healthcare systems in low- and middle-income countries are in a nascent, promising stage. Unlike the U.S., a nation in which healthcare is treated more like a privilege for the elite than a critically important human right, these nations, which have much less established and weaker healthcare systems, understand the critical importance of healthcare to their citizens’ well-being.
These countries, in which government-provided universal healthcare is becoming the norm, have an ideal opportunity to create a medical system that works better and more efficiently than in countries with older healthcare systems bogged down by incompatible processes and red tape. Such progress not just allows dwellers of these countries to enjoy their lives more; but also to be more productive for longer, improving society as a whole. Additionally, such systems provide a working model for other countries, contributing to a global body of knowledge on how to run effective public health systems with modern technology.
Thus, opportunities for making radical innovations are even greater in low- and middle-income nations because of the ability to leapfrog ingrained habits based on now-obsolete technology. An example is the early rise of mobile payments in Kenya (through mPesa), which bypassed traditional forms of cashless payment, such as credit cards.
The unique incentive, and often necessity, to solve problems with fewer resources and fewer constraints from entrenched interests, provides an invaluable opportunity for technology to make a deep and lasting impact on healthcare.
The Ona team has been working on technology for frontline health workers for over a decade. We are building OpenSRP to support these technologies. The advances we have made — and continue to make — are calculable in numbers of lives saved.
With technology-enabled remote health services, healthcare workers are able to capture data, run decision-making algorithms, visualize real-time indicators and plan service schedules armed only with a smartphone. Together with sensors, health records and other data sources, relatively unskilled workers in inaccessible areas are able to perform the low-risk tasks traditionally done by medical professionals.
Possibilities & Potential, Explained
Whereas countries such as the U.S. are struggling with legacy systems and electronic medical records, low- and middle-income nations with forward-thinking Ministries of Health are able to implement technology solutions from the ground up. Having access to sophisticated technology during the developmental stage of a healthcare system helps ensure the system’s strength and longevity — critical factors in areas with few health professionals per capita.
Some examples we’ve deployed with our partners, or are in the process of deploying include:
Smartphones with Sensors
Physical healthcare devices, sphygmomanometers, microscopes, etc., are being replaced or augmented with software that runs on smartphones using their built-in components, improving accuracy and fidelity.
The sensors and computer chips inside of cell phones are often comparable to the components inside of more expensive and less portable medical testing equipment. Costs of delivering care can be greatly reduced by foregoing the purchase of this equipment in favor of using smartphone technology. In instances in which smartphone technology is inadequate for our purposes, we may be able to combine the smartphone with external components for a fraction of the cost of purchasing medical equipment with this capability.
Such technology can even be used by patients themselves. While this type of self-testing is not yet widespread, patients have long been able to test their own blood sugar and correct imbalances, take their own blood pressure, run tests for ovulation, pregnancy and more.
The possibilities in this area are dramatic and promising.
Data Interoperability Standards
The more data that can be aggregated and the better the standards that can be adopted in the early stages of creating healthcare systems, the easier it will be to extrapolate for addition, expansion, and automation down the road. Adopted and implemented early, these standards facilitate the wider and more efficient use of artificial intelligence.
Our task now is to get program managers to understand that using standards-based healthcare, while requiring an investment upfront, will result in a significant payback that compounds over time. The time and effort it takes to establish new methods and standards for each project is wasteful in itself, and further puts up roadblocks for comparing results down the road. Once everyone is on board with the standards and they are adopted and adhered to, we will be able to build more and better tools to interpret the data, saving thousands of person hours of time and significantly reducing errors in interpretation.
This advanced step feels almost miraculous, yet it is within our grasp. Standardization can lead to the type of automation that allows machines to use learning to facilitate foresight and perform tasks without specifically being instructed to do so. For example, this type of technology was used to learn to predict when water wells were malfunctioning, thereby greatly reducing wait times for repair.
We see four immediate ways machine learning can have a significant impact:
Predictive analytics — Machines can step in and do much of the work that used to require the services of a doctor or nurse. This technology allows patients to get care that otherwise would be inaccessible to them by making it possible for trained workers to run tests and make diagnoses that were previously performed by doctors. Better care is delivered faster.
Image processing — Advances in this area allow for camera-based blood pressure measurement, pneumonia detection and other critical tasks.
Cluster analysis — Health managers can track critically important patient data and worker actions to improve processes and outcomes. For example, a healthcare manager might reduce the number of days after a missed prenatal visit until workers visit a client’s home, because the data shows that health workers with fewer days to visit have reduced maternal mortality in the clients they serve.
Pattern detection — The use of data collection and aggregation allows for early pattern detection by AI that saves lives. For instance, an AI algorithm that learns of a recent nearby outbreak of malaria can help prevent it from becoming an epidemic by doing early tests on likely infection carriers. Although it may suffer from other biases, AI systems can be tuned so they are not hampered by the type of bias shown by humans who may choose to forego a test in the absence of some symptoms they may have expected to see but didn’t.
These and other technological advances will help achieve the goals of health systems as described by the WHO:
- Improve populations’ health.
- Improve the health system’s responsiveness to the population it serves.
The Future Depends on Us
Technology allows for infinite innovation — so much so that it is impossible to fully imagine the potential benefits it can bring to the healthcare field. For this reason, hope and expectations flourish, almost unchecked. The danger comes in the rush to the finish line and the expectations that may be left unfulfilled. These can lead to opportunism, pilotitis, fragmentation — and all are symptoms of a different type of disease that prevents progress in global health. To cause sustainable impact and change, these innovations must be applied under controlled and carefully monitored conditions.
While the standards of living in low- and middle-income countries are improving, healthcare systems are struggling to keep pace. Traditional and mobile healthcare has helped, but only incrementally. The way to break away and make real progress in developing the most impactful healthcare systems is by combining:
- Expertise based on years of study
- Mobile health tools
- Transformational standards-based AI tech
Through our work building these software applications, we are revolutionizing the way healthcare is managed and delivered in low- and middle-income nations.