How we are using AI to help healthcare providers make better clinical decisions in Tanzania.
AI in Healthcare (A Brief Overview)
Technologies that utilize Machine Learning and Artificial Intelligence are transforming healthcare. These technologies allow us to generate unprecedented advancements in disease detection, cancer diagnosis and treatment, radiology, epidemic outbreaks, patient triage, and personalized healthcare. Companies like SkinVision, for example, are using image recognition to detect early signs of skin cancer growth and track patient risk over time. IBM, through Watson for Genomics and Watson for Oncology, is integrating machine learning and healthcare in an attempt to develop better diagnostics and treatments for patients with cancer. Algorithms like IDx-DR are now officially approved by the FDA and are in the market for specialist-level diagnosis of diabetic retinopathy.
Although these technologies are not yet a “mic-drop” solution in most cases (in fact, this article discusses some of the challenges faced by IBM Watson’s system), they are demonstrating that, given enough data, they can perform as well as highly-trained clinicians at doing some very specific tasks. And, they are currently being tested and implemented around the world.
Disclaimer: This, of course, does not mean that we will need less care providers or that AI tools will make human medical practitioners obsolete. I figured we would make that clear from the beginning, in case anyone had thoughts otherwise.
At the current rate of technological advancement and global uptake, these technologies are quickly permeating international borders — they are not confined to the US/European/ Chinese markets. Developers, entrepreneurs, researchers, clinicians, and scientists across the African continent in particular are exploring the use of these technologies to support the growing demand for physicians and to achieve overall better health outcomes for their citizens.
Healthcare in Tanzania
Under the leadership of the current Minister of Health, Hon. Ummy Mwalimu, Tanzania is progressing rapidly in all areas of healthcare. The national health objectives are clear and implementation strategies are well defined with metrics being closely reported and monitored.
Despite the recent clinical and infrastructure breakthroughs however, overall health outcomes in Tanzania remain poor and a variety of health challenges inhibit the ability of the country to achieve vital health outcomes. Contributing to this, Tanzania has one of the worst doctor-to-patient ratios in the world (1:20,000) according to the WHO, and over 75% of the country’s doctors live in urban areas. The country lacks specialist-care options, and patients in rural areas often face long wait times, poor emergency health services, and a fragmented care system. As a result, many patients self-diagnose and self-medicate, leading to prolonged sickness and and overuse of antibiotics.
Healthcare workers on the other hand have limited resources and tools to support their patients, especially in rural areas. They do not always have the extensive training required to treat complex conditions, and are often unable to utilize tests as a way to diagnose their patients. To combat the low number of physicians and specialists, Tanzania has, in the past decade, significantly increased the number of Clinical Officers, Assistant Medical Officers, and Community Health Workers in the country who provide services at the village and ward levels.
Workforce challenges significantly impact the nation’s ability to achieve both nationally mandated health outcomes, and the Sustainable Development Goals. According to the SDG Index, Tanzania has a maternal mortality rate of 398 deaths per 100,000 live births, and an under-five mortality rate of 56.7 per 1,000 live births. Although both of these measures are improving year-to-year, they indicate that the country is far from achieving vital health goals.
Building Dr. Elsa
In response to some of the challenges faced in Tanzania, we’ve built Dr. Elsa.
Dr. Elsa is a Health Assistant tool that supports healthcare workers (HCWs) in rural areas through diagnostic decision support, next step recommendations, and predicting disease outbreaks. Our tool is delivered to HCWs through a mobile application, and acts like a general physician. The HCW inputs patient demographics and symptoms, and the tool outputs differential diagnoses and next steps, allowing the HCW to make more informed care decisions for their patients.
The Health Assistant is also able to analyze existing health data to predict infectious disease outbreaks six months months in advance, meaning health providers and governments can better prepare for these outbreaks.
We are currently building Dr. Elsa for use in three areas: infectious disease, non-communicable disease (including cancer), and pediatric care. If successful, healthcare providers will be able to make more informed care decisions about their patients, therefore decreasing the number of misdiagnoses, decreasing the use of unnecessary antibiotics, decreasing the amount of time and money a patient has to spend in the care system, and increasing the healthcare workers level of support and respect in the community.
With our tools, we hope to see better diagnostics at the first point of care, better linkage to care, and more appropriate allocation of resources as a result of accurate infectious disease outbreak predictions. All of this is in support of SDG 3 (Good Health and Well-Being) and other national priorities.
The Future of AI Health in Tanzania
We are in the beginning stages of a large efficacy pilot in Tanzania to determine the accuracy of our models/ algorithms against the Gold Standard of Care at identifying diseases based on patient history, patient demographics, and current symptoms.
We envision building a complete end-to-end intelligent healthcare system; we hope to put digital tools in the hands of community healthcare workers, nurses, and doctors all over the African continent to connect care providers, improve health outcomes, and support decision making.
Understanding and leveraging Artificial Intelligence and Machine Learning can catalyze immense progress in the future of healthcare — including, but not limited to, healthcare delivery, monitoring, policies, forecasting and better understandings of illnesses and how people are impacted by them. These technologies have the possibility for massive impact, and we have some incredible ideas for expanding from here.
If you would like to learn more about our work with Dr. Elsa, or talk to us about how you think we can better support healthcare workers in East Africa, we would love to chat! You can email us at email@example.com or follow Dr. Elsa on Facebook, Twitter, and Instagram.