Scalable AI Platform for Underserved Communities

It is increasingly recognised that health data has a significant role to play in improving global health outcomes by identifying health risks, improving decision-making, and evaluating interventions. 

Over the last decade, an abundance of data has been created from varied sources including electronic health records, and disease surveillance systems and this is increasing due to the proliferation of mobile technology. However this data is fragmented, creating difficulties for accurate decision making and meaningful insights and an increased security risk. 

Current health data is often stored in different formats, scattered across various databases, or even in different geographic regions. This makes it challenging to consolidate and obtain a comprehensive view of the information. Organizations have to spend a considerable amount of time and effort integrating the data, cleaning it, and ensuring data quality before they can analyze it. This is if they can even support a data science team or single data science professional.

Disconnected health-related data can result in fragmented care, as healthcare providers may not have access to a patient's complete medical history or current health status. This can lead to misdiagnosis, inappropriate treatment, and poor health outcomes. Additionally, disconnected data can create administrative burdens, as healthcare providers may spend significant time manually searching for and consolidating patient information.

Compounding the problem is the shortage of healthcare professionals globally, especially in LMIC and fragile environments. According to the Global Health Observatory (WHO) it is estimated that 18 million additional health workers are needed in order to achieve Universal Healthcare by 2030. The current shortage is driven by underinvestment in the education and training of health workers and is complicated by the difficulties in health workers accessing rural areas. 

The shortage of health professionals makes it difficult for patients to access the care they need and leads to long wait times. Health workers will have difficulty prioritizing patients, leading to unnecessary illness and death. 

We need to do more with fewer resources. We need to leverage the tools that we have and the data that they are collecting in smarter ways.

We believe that the next innovation in healthcare will not come from a new vaccine or medication but from behavioral change resulting from increased engagement in healthcare for both patients and providers. 

We research and deploy algorithms to improve personalization for health digital services. Our ML product supports the work of medical care teams, helps patients self-manage, and become more engaged. Our technology, when integrated into a digital app used by healthcare providers (e.g., a job aid for CHWs), patients (e.g., HIV patients), and pharmacists, enables use cases such as the following:

• Example 1: HIV patients receive personalized nudges to help them adhere to treatment and follow-up visits. Our technology uses the profile and app behavior data to optimize the timing and content of the nudges. 

• Example 2: Healthcare workers likely to decrease their productivity receive personalized motivational nudges. It predicts which workers are most likely to decrease their performance and optimizes the motivational messages to increase their performance.

• Example 3: Healthcare workers receive a recommendation to call/visit a list of patients who might miss an appointment. Our technology optimizes the allocation of a limited resource (healthcare workers’ time to call/visit patients) by identifying and recommending the patients with the highest risk of missing an appointment who would benefit most from the call/visit.

All these features are delivered through our AI Platform in the modules described below

• Analytics Module. To visualize the data, KPIs, etc., from the digital app events captured through the SDK integration (e.g., daily/weekly/monthly number of consultations). 

• Segmentation Module. To easily create cohorts of app users, facilities, etc., based on static traits (e.g., age), dynamic traits (e.g., # of patients seen during the last week), or model outputs (e.g., providers with an 80% probability of not using the app). These cohorts, once created, can be used across the other platform modules (e.g., to target an intervention to a subset of users).

• Model Module. An easy-to-use tool with machine learning algorithms to predict outcomes across different areas (e.g., patients’ risk of not adhering to treatment). Includes a tool for visualizing the algorithm's performance.

• Intervention Module. provides a tool for designing, implementing, and tracking digital interventions for app users (e.g., nudges, content, recommendations). Using reinforcement learning algorithms that leverage the individuals’ contextual data (e.g., app usage patterns) to optimize an output of interest (e.g., total number of weekly consultations by the CHW). 

• Assistant Module. An LLM-driven assistant that allows the user to query the platform's data (including the modeling results).

 By using a domain specific SDK (a software code compatible with the technologies used by most apps), our partners easily integrate their digital apps into the AI platform, convert the logs and clinical information into AI ready data, and bring machine learning capabilities that are otherwise not accessible to them. Our all-in-one platform tracks and organizes provider and patient data to use real-time and predicted behavior to deliver adaptive interventions. It allows massive iterative experimentation with rapid cycles of intervention deployment and optimization through Reinforcement Learning.

We focus our work in regions with over-stretched healthcare services, where healthcare workers struggle to identify those patients most in need of urgent care and where patients must travel long distances to visit pharmacies repeatedly waiting for drugs that are out of stock, causing treatment delays and possible deterioration. 

We partner with organizations working on the front lines of global healthcare, through whom we have interviewed people to understand the pain points and identify how we can bridge the resource shortage. 

Our solution fosters patient-centric solutions by supporting healthcare providers and funders in global health to generate more flexible, adaptable, and personalized patient journeys. 

We work in partnership with governments, not-for-profit organizations, and healthcare organizations. The Causal Foundry platform integrates with existing digital tools and mobile applications or builds end-to-end, natively integrated digital solutions serving patients directly and through pharmacists, community health workers, and health care providers at all levels of the health systems.  

Causal Foundry supports the daily work of healthcare providers by offering, for example, recommendations for drug and treatment adherence, diagnostic tests, patient referrals, and the management of medical supplies based on historical behavior and information gathered in real-time. It can also provide a triage service to identify those most at risk and escalate their cases to the secondary care providers to prioritize those who need special attention. 

In addition, the platform supports an iterative cycle of improvements in digital tools for health workers. Program managers can monitor how apps and tablets are being used by the healthcare providers in the field, assessing the most valuable parts of the platforms and those that are rarely used or used incorrectly. This creates a feedback loop to constantly improve the relevance and usefulness of digital tools for community health workers. 

The platform can also serve funders, both philanthropic and governmental, to transparently monitor the impact and success of health programs and ensure that they are reaching objectives without overburdening program management teams with additional reporting requirements. We aim to develop the capacity to measure the number of lives that are saved or improved by an intervention rather than the number of people involved in a program.

CAUSAL FOUNDRY (CF)  is a Public Benefit Corporation whose experts in reinforcement learning, ML, and adaptive experimentation build AI products to solve the most important health challenges. 

The diverse team represents 11 nationalities and experiences from different sectors, including healthcare and AI. We have actively recruited people with differing backgrounds, including lived experience of the cultures that we wish to serve, to ensure a well-rounded approach.

Our business model is based on a partnership approach, working with organizations active in the field connected to the populations they serve, who provide us with local insight. With each of our key partnerships we have completed at least one field visit, spending time with the local team and end users and using the opportunity to conduct quantitative and qualitative research. We also conduct ongoing UX research to get feedback on our product through the development lifecycle. 

Examples of partnerships with organizations with close proximity to the communities we wish to serve include Mothers to Mothers (M2M) and Aide Chemists. M2M.org, is an organization based in South Africa that trains HIV positive mothers to become community health care workers, providing healthcare by people in the community for people in the community. We are developing an end to end solution incorporating a mobile application and our machine learning platform to provide adaptive nudges to support these healthcare workers. 

Aide chemists has the leading network of pharmacies in Ghana. We have developed a AI-native e-commerce application, called Momentum, integrated with the machine learning platform, providing patients with access to medication anytime anywhere, including prediction of demand to ensure always on-stock essential medicines in every pharmacy.

Our team is a dynamic and dedicated collective of professionals uniquely composed to tackle the challenges at the intersection of AI, technology, and healthcare in low-resource settings. Our expertise spans deep learning, reinforcement learning, statistical learning, generative AI, distributed systems and software engineering.

In addition to specific profiles with longstanding experience in digital health in low-resource settings, our team members have gained a deep understanding of the needs and challenges faced by communities in low- and middle-income countries through our work with local partners in Africa and Asia over the past couple of years. 

Our chief strategy officer, Enric Jané, MD PhD, has a wealth of experience in international development including several years working with the Bill and Melinda Gates Foundation. Eniola Olaleye, from the Data Science team, is Nigerian and permanently based in Kigali, Rwanda, where much of our work with healthcare innovators has been focused.

We have built an AI and reinforcement learning platform (RL), a Software Development Kit (SDK) to convert raw logs into AI ready data of different specific fields, and a frontend to manage adaptive interventions and individual predictions. The RL platform includes a recommendation system based on embeddings and reinforcement learning techniques like contextual multi-armed bandits and restless bandits.

Our implementation across several countries and partners, is currently reaching encompassing 30,000 unique users on monthly basis, has provided us with valuable lessons and insights that have significantly influenced the evolution of our innovation. 

As an organization that designs applied science, we are excited at the opportunity of working closely with MIT, an institution that excels at pushing to the forefront the  business innovations in the area of applied science.

We are applying for the MIT Solve program not only because of the wealth of mentoring and networking opportunities it can offer start-ups but also because it is the right time for Causal Foundry to take full advantage of this opportunity. 

To date, we have been focused on developing the product and integrating the partners. We are funded by the Bill and Melinda Gates Foundation and Google.org. As we move into the next phase in our growth, new challenges are arising, specifically related to making our operation sustainable and developing a marketing and sales strategy. We recognise that being part of a cohort of impact-based peers, supported by expert mentors, will enable us to learn, network, and grow to be more effective. 

Our success is based on partnering with the right organizations. Thus, the exposure possible through MIT solve will be invaluable for reaching the organizations that are best fit for this critical juncture in our development. 

Our innovation is a sophisticated AI Software Platform designed to revolutionize digital applications with state-of-the-art personalization technologies. By integrating data from various digital tools using a domain-specific Software Development Kit (SDK), our platform employs advanced algorithms for predictive analytics, resource allocation and adaptive interventions (reinforcement learning based like linear and contextual bandits, restless bandits for resource allocation, and networked restless bandits for budgeted interventions in communities) to support healthcare workers and managers. We provide just-in-time recommendations through digital devices (already deployed or that we build) to ensure access to vaccines and information to improve healthcare outcomes.

These are the capabilities of our innovation:

- Demand Prediction: Our platform can support supply chain systems and healthcare workers at the facility level by optimizing demand forecasts at the delivery point level, ensuring that medication is  available where and when needed.

- AI-based resource allocation system: prioritize and optimize the distribution of medication based on contextual information, demographics,and distributor capabilities. This is to ensure efficient delivery to those most in need. 

- Nudges for Healthcare Providers: Our platform can send healthcare workers personalized nudges that include contextual information, such as local climate data or other relevant information, to guide service delivery planning. For example, these nudges can motivate and guide providers on the timing and execution of necessary immunization and outreach sessions, enhancing the effectiveness of vaccination programs.

- Data Access for Managers: An LLM-driven Assistant simplifies data access for program managers, allowing them to ask specific questions and receive instant insights using a chatbot. This Assistant can access relevant data and provide answers to managers, facilitating quick and informed decision-making. This Assistant complements traditional dashboards that are too often not used by managers.

Through these applications, our AI software platform brings unparalleled personalization and efficiency to immunization programs, directly addressing the challenges of healthcare delivery in dynamic and complex environments. We are happy to provide you with a demonstration to  showcase the capabilities of our platform in detail.

Causal Foundry aims to revolutionize healthcare by improving patient and provider engagement through personalized interventions, leading to significant advancements in healthcare outcomes. 

Assumptions:

• Existing digital health tools have untapped potential to enhance patient and provider engagement.

• Effective engagement between patients and providers leads to better health outcomes.

• Access to understandable information helps patients make better health choices

• Personalized interventions directed to clients, healthcare providers, and pharmacists can effectively nudge behaviors toward healthier and more impactful choices.

Strategies:

Leveraging Existing Digital Health Tools: Causal Foundry will maximize the potential of existing digital health tools to enhance patient and provider engagement. This involves optimizing tool functionality to address gaps in the healthcare ecosystem.

Three Main Use Cases: focus on three primary areas:

• Pharmacy: Improving supply chain and engagement between patients and pharmacists for better medication adherence and management.

• Primary Healthcare: Enhancing communication and collaboration between patients and primary care providers to optimize preventive care and chronic disease management.

• Community Healthcare Worker Support: Empowering community healthcare workers with tools and personalized interventions to improve their performance and support to their communities.

Outputs:

Causal Foundry's machine learning system provides analysis, insights, adaptive interventions, recommendations, and predictions based on an iterative process.

Short-Term Outcomes:

• Pharmacy: Efficiency in supply chain management, prediction of demand, improved traceability of medication. Improved engagement of clients with pharmacies through their apps.

• Primary Healthcare: Monitoring healthcare worker activity, identifying at-risk patients, and triage/prioritization of patients.

• Community Healthcare Worker Support: Connection to senior staff for additional information, educational support through a chatbot, and patient connection for ongoing support.

Medium-Term Outcomes:

• Pharmacy: Elimination of drug stock-outs, increased adherence to treatment by clients.

• Primary Healthcare: Reduction in deaths related to chronic diseases, improved quality of life for patients.

• Community Healthcare Workers: Reduction in maternal and infant mortality through improved access to healthcare and health advice.

Long-Term Outcome:

• Improved health outcomes and reduction in healthcare inequalities globally.

Impact:

Causal Foundry envisions a healthcare ecosystem where personalized interventions and improved engagement drive better health outcomes for all, contributing to a reduction in healthcare disparities and improved global health.

Our impact goals over the next five years are based on an iterative cycle of improvement. With each intervention we get access to better data, creating more effective interventions, creating better data. 

Overall, Causal Foundry aims to impact 500k patients with our solution within the next 5 years, improving the health outcomes for these individuals.  

The impact goals can be divided into three main use cases, pharmacy, primary healthcare, and community healthcare support:

• Pharmacy:

  • Next year: 

    • Improved supply chain management, reducing drug stockouts by 10% for one B2B pharmacy network

  • Five years: 

    • Elimination of drug stockouts for one B2B pharmacy network

    • Working with the public sector health services to predict and prevent disease outbreak 

• Primary health care:

  • Next year:

    • Monitoring and insight into workflows for 2 clinic networks

    • For 2 clinic networks, improved workflows and engagement in digital tools, increased usage of the app by 10% 

  • Five years:

    • Improved patient journey

    • Improved management of non-communicable diseases

    • Reduced burden on primary healthcare

• Community healthcare worker support

  • Next year: 

    • Development Increased engagement in digital tools by 10>#/p###

  • Five years

    • Increased diagnosis of disease and non-communicable disease

    • Reduction in maternal, neonatal, and infant mortality

Each use case is initially developed with strategic partners to solve specific problems within the global health sector. The AI products subsequently generated can then be fine-tuned to get the right market fit and attract additional investment funding. 

Our core technology provides advanced analytics, tailored solutions for specific use-cases, and cutting-edge AI capabilities provided as a service. 

Analytics and Visualization: Our platform enables users to visualize data, uncovering behavioral and clinical trends while assessing performance and risk.

Use-case Specific Tracking: We identify and track minimum de-identified data essential for specific purposes such as patient management, disease care, e-learning, training, and supply-chain efficiency improvements.

Cohort Segmentation: Users can segment their stakeholders (healthcare facilities, personnel, patients, drugs, tests, etc.) into relevant groups for targeted analysis.

Customizable Dashboards: Our platform allows users to customize dashboards, facilitating easy comparison of different cohorts, extracting meaningful insights, and guiding evidence-based decisions, particularly focusing on dynamic clinical and behavioral traits.

Reinforcement Learning and AI as a Service: We offer a platform featuring Reinforcement Learning and AI capabilities, including:

• Easily Deployable Models: Machine learning models are readily deployable to anticipate user behavior, needs, and preferences.

• Access to State-of-the-Art Algorithms: Users can access cutting-edge algorithms through their platform profiles and deploy them effortlessly without coding. These include recommendation, user, and demand models.

• Model Performance Tracking: Full transparency of running models and a comprehensive verification suite enable users to understand model tradeoffs and performance, facilitating continuous improvement.

• Constant Improvement: Regular updates introduce new traits, enhancing recommendation accuracy, predictions, and uncertainty management for improved outcomes.

Adaptive Interventions: We enable users to define and deploy adaptive interventions quickly, crafting personalized content and message interventions using various RL-based algorithms tailored to individual needs.

Experimentation: Our platform facilitates quick experimentation with various designs, reducing preparation time from months to minutes, and maximizing statistical power while minimizing risk.

Impact Measurement and Tracking: Experiments and interventions are monitored in near real-time, with thorough statistical analyses measuring and tracking their impact.

Large Language Models (LLM): Our platform features an LLM-based assistant enabling users to query data and send interventions with just a sentence, empowering patients and providers with access to data and standards of care.

Speech-to-Text Hands-Free Assistant: MedScribe, our LLM-based product, seamlessly integrates into electronic health records and digital tools, enhancing patient-provider interactions with data integration and personalized summaries, fostering patient engagement and health management.

Full time staff 10

Part time staff 1

Contractors 3

Advisors 4

Although Causal Foundry was registered as a company in November 2022. The founding team has been working on this project since July 2020 as part of the Foundation Benshi, a not-for-profit organization funded by the Bill and Melinda Gates Foundation to research how behavioral interventions could be used to improve global health outcomes. 

Over the last 12 months Causal Foundry has purchased the IP and migrated the staff from the foundation in order to ensure that the project will have a sustainable business model to avoid overdependence on grant funding. 

Causal Foundry is deeply committed to diversity, equity, and inclusion, evident in its leadership structure and workforce composition. With a female CEO at the helm and a 60% female management team, the company exemplifies gender diversity and inclusivity in leadership roles. 

This female-led approach is complemented by the company's sponsorship of working visas, ensuring talent transcends geographical boundaries, and providing opportunities for individuals to contribute regardless of their country of origin. The team of 18 people includes representation of  12 different nationalities, including team members based in the communities we wish to serve who provide insight into specific local needs. 

By embracing diversity as a cornerstone of its culture, Causal Foundry not only drives innovation within its team but also sets an example for promoting positive societal change.

Causal Foundry provides value to the populations it serves by increasing access to information and engaging patients and providers to improve health care outcomes. We work with partners' existing digital tools and create new ones to achieve these goals.

Patients: this group are the ultimate beneficiaries of our efforts to improve global health. Through our interventions, patients receive personalized support, access to understandable health information, and enhanced engagement with their healthcare providers. This leads to better health outcomes, improved disease management, and, ultimately, a higher quality of life for individuals worldwide.

Healthcare providers: Value is delivered to healthcare organizations in two ways: first, by analyzing data to gain insight into patient behaviors. Second, by creating interventions to nudge users to desired behaviors, e.g., prompting a caregiver to take the blood pressure reading of a patient at risk of complications. We provide value to these organizations by increasing efficiencies and supporting them to create more impact with fewer resources. 

Companies involved in the pharmaceutical supply chain: Causal Foundry can help companies predict and manage demand, avoiding stockouts and increasing profitability by analyzing demand and purchasing behavior. This activity delivers a measurable financial return and is a service we can charge for. 

Pharmaceutical companies: the platform can also be used in clinical trials to boost engagement and feedback. This is especially relevant with populations where follow up is complicated due to personnel limitations or large geographical areas.  This would also be a paid-for service. 

Grant-giving organizations: Measuring impact and results can be challenging for philanthropic and government organizations. The platform can provide value for organizations such as the Bill and Melinda Gates Foundation by measuring and driving engagement in interventions carried out by their grantees. 

Our plan to become financially sustainable is based on a hybrid of grant financing and SaaS. We aim for the company to start turning over a profit from sales of access to the dashboard within the next five years. 

We identify that our platform can be used by various actors in the healthcare ecosystem. Several organizations will gain a financial benefit directly from the use of the platform. 

A seed grant from BMGF assures short-term financial security, and funding for specific projects from BMGF and Google.org supports it. We will continue to apply for grants to enable us to work with partners with whom we can drive the biggest impact on global health.

The plan for financial sustainability relies on developing paid-for services. The first of these is the e-commerce use case which is well developed. We are already gathering evidence of the financial upside of the recommendation system and for healthcare distributors and retailers. We can create a tiered monthly fee for operators in this space, dependent on market size. 

We have agreed to financial terms with a Nigerian start-up partner to pay a fixed fee per month for their integration on the platform. Partnerships, such as this one, within the fast-growing innovation ecosystem of Africa will integrate into the platform with increasing ease over time. The purpose is twofold, to cover the running costs of the platform while contributing to our impact goals. 

An unexplored area for income will be the use case for pharmaceutical companies to increase engagement in clinical trials. In addition to driving income for Causal Foundry, this will promote equity by increasing the potential for diversity in clinical trials, engaging patients in rural and hard-to-reach areas of the global south.