Inspired Ideas: AI powered decision support platform for healthcare providers
The UNICEF Innovation Fund is proud to see a portfolio member Inspired Ideas graduate. They’ve come a long way – from numerous product iterations to deep diving into understanding their ecosystem better and strengthening their business model. They’re now ready to collaborate at a larger scale – as they find new pathways to work with partners, investors, and the open source community.
The past 12 months have been filled with excitement, adventure, and doing what we love most at Inspired Ideas: impacting the lives of children by optimizing health decisions and improving access to high-quality healthcare. An investment from the UNICEF Innovation Fund allowed our team to build, test, and deploy an AI-powered tool for lower-level healthcare providers in Tanzania so that they can better address pediatric patients under the age of 14 at primary points of care.
In many parts of Tanzania, there are no medical specialists or pediatricians; there is limited diagnostic equipment and resources to support the dedicated front-line healthcare providers who serve these communities; and, there are high mortality rates among children, especially under the age of five, caused by preventable diseases.
Our solution, the Elsa Health Assistant, is a clinical decision support platform that puts specialist-level decision making into the hands of healthcare providers across emerging markets. Powered by artificial intelligence, data, and expert knowledge, Elsa supports providers in identifying the cause of a patient's condition, predicting medication adherence, and providing actionable next steps based on international guidelines. Our AI models - “Elsa’s Intelligence” - are built alongside medical experts, who help us map causality between symptoms, risk factors, geographic location, and temporal information with likely conditions. The Elsa Health Assistant shows healthcare providers evidence-based insights: the likelihood of their patient having malaria, for example, and the safest course of action to ensure a positive outcome for the patient.
We’re the only AI-powered clinical decision support tool deployed in primary health facilities in Tanzania - and we don’t take that lightly.
We know that the introduction of innovative technologies in healthcare has the potential to significantly impact lives across Sub-Saharan Africa, and we are excited to continue developing our tool, expanding the number of diseases we cover, and ensuring healthcare providers across various levels of care have access to Elsa’s Intelligence.
PROTOTYPING THE ELSA HEALTH ASSISTANT TOOL
Our team employs a wide range of human-centered design activities, working closely with end users, medical specialists, and health stakeholders to co-create a solution that is both scientifically rigorous and feasible for integration into the primary care system. Over the course of the investment, we completely re-thought and re-designed our user interface as we learned more about the clinical decision making process and how we can best offer support without increasing cognitive load on the users. During the process of prototyping, we try to go from brainstorms and thoughts to paper prototypes to rough, minimum viable products within hours so that our users can have a hands-on experience trying the different suggestions and providing feedback.
One particularly interesting set of iterations this year were around the explainability of our models: how do we best show healthcare providers with varying levels of clinical and technical ability which conditions are most likely, why we think that is true, and how certain we are about that prediction? We rely heavily on visualizations and graphs, coupled with verbal explanations in multiple languages, to accomplish this.
ELSA IN THE FIELD
Over the last year, we have piloted our pediatric-specific solution with 10 drug dispensers, 20 clinical officers, and 3 pediatricians across Tanzania - all of whom are still using the tool in their facility. Elsa supports clinical decisions for roughly 5,000 patients per month. Our users report that they like the Elsa Health Assistant because it helps them better identify a patient’s condition, better assess the symptoms of a patient, and feel more confident in their decisions.
We are also focused on assessing the performance of the tool and its ability to support clinical decision making through ongoing national studies. Preliminary results show that Elsa’s Intelligence is able to accurately predict the true cause of a patient’s illness 85% of the time across a set of 50 conditions - an exciting accomplishment given the high rates of misdiagnoses made by primary healthcare providers in Tanzania. We’ve also measured that Elsa helps providers more effectively dispense medication to their pediatric patients: providers using Elsa are 3x more likely than those not using Elsa to appropriately give ORS+Zinc to children under five with diarrhoeal disease.
Piloting Elsa has also brought to light a number of challenges. Key among them are regulatory uncertainties and constantly changing policies in markets where the introduction of AI and other emerging technologies is still new. We’re honored to work alongside regulators and stakeholders to share our insights and create sandbox environments that allow for safe testing and deployment. In addition, introducing novel technologies in a sector that is dominated by paper-based reporting sometimes feels like an uphill battle. In an environment where there is a general lack of technical infrastructure and low technical literacy, we have made it part of our mission to upskill healthcare providers at all levels so that they are able to benefit from the future of technological advancement. We’re excited to take on these challenges in the coming year.
ON BEING OPEN SOURCE
Our open source journey, supported by the UNICEF Innovation Fund team, has changed the way we look at our solution. We’ve developed better coding and documentation practices, adjusted our business model, and developed a deeper appreciation for others who manage open source projects.
Open source is critical for our growth
Many solutions in healthcare that are taken up by governments and impact organizations are built in the open for greater accessibility and transparency.Much of our work over the last year has been focused on research and development, and we are now reaching a pivotal moment where we can build a community of like-minded technical and non-technical contributors to strengthen, grow, and maintain our solution. In addition to building a community of developers, we’ve also recently started developing tools that health experts and researchers without programming experience can use to improve our disease models. As we expand to new geographies, with different disease burdens, this is important.
We recognize that building this community is a challenging labor of love, and we’re excited about engaging more people to take our work to the next level.
Our business model has been constantly evolving, and the UNICEF Innovation Fund team has helped us reach greater clarity on our path towards sustainability and growth. We’ve been able to identify ways that companies or organizations looking to support health decisions in their own work can leverage Elsa’s Intelligence through a subscription service, and we were able to secure our first telemedicine customer, who is using Elsa-powered insights to provide better care for patients. We’re also looking forward to working with large-scale impact organizations and governments to integrate our solution into national health information systems for broader health system improvements.
We’re humbled by the individuals and organizations who have passionately collaborated with us as we develop and pilot the Elsa Health Assistant, including health research institutes, international and local non-profit organizations, healthcare providers, private-sector technology companies, and district governments. We will continue working with these partners to provide evidence of the impact of clinical decision support tools across the continent. We also hope to further engage UNICEF Tanzania as a strategic partner given their influence and work to improve health outcomes for children across the country.
NEXT SET OF GOALS
We’re committed to improving health outcomes for children globally, and we’re excited about our next steps towards scalability and sustainability. Over the next year, we will continue to strengthen our decision support algorithms and human-technology-interaction experience, expand our disease coverage to 500 diseases, and conduct additional research studies to prove Elsa’s efficacy. We also look forward to collaborating more closely with the Tanzanian government and Ministry of Health in order to scale Elsa across the country and acquire the needed regulatory approvals to make this possible. As an open source project, we’re doubling down on our efforts to grow a community of passionate contributors. And, of course, we are always pursuing follow-on funding and additional financial support that will allow us to put Elsa in the hands of healthcare providers everywhere.
UNICEF INNOVATION FUND
We’re forever indebted to the UNICEF Innovation Fund team for their genuine support, boundless insight, and constant guidance. From monthly check in calls that sometimes doubled as therapy sessions, to connecting us with strategic partners and potential investors, to pushing us to explore new opportunities with an open mind, being an Innovation Fund portfolio company has opened doors we didn’t even know existed. The team helped us focus on scalability and commercialization of our solution, as well as helped us navigate changing policy landscapes, evaluations of our technology, and a health environment that was disrupted by a global pandemic. The entire Innovation Fund team will always be a part of the Inspired Ideas family.