Can AI improve healthcare in resource-limited settings without massive budgets or billion-parameter models?

Patricia Thaine sits down with Chenjerai Sisimayi, a clinical epidemiologist and data scientist based in Zimbabwe, and Tariq Khokhar, Head of Data for Science and Health at the Wellcome Trust. Together, they explore what it actually takes to build and deploy AI tools in healthcare systems across low and middle income countries.

The conversation cuts through the hype to focus on practical realities: How do you train a model when you only have 500 chest X-rays? What happens when you remove an AI tool from clinicians who've grown dependent on it? And why might a simple checklist outperform a sophisticated algorithm?

Chenjerai shares his work developing hybrid deep learning models for TB and COVID-19 classification using local data, and the challenges of integrating these tools into clinical workflows. Tariq brings perspective from large-scale trials, including the UK's EDITH mammography study, and raises critical questions about evaluation, unintended consequences, and what "good value for money" actually means in healthcare AI.

They also dig into often-overlooked infrastructure gaps: the lack of systematic language data collection for African languages, the promise of "small AI" approaches built on transfer learning, and why national AI strategies matter for getting these tools into practice.

Chenjerai Sisimayi is a Zimbabwe-based Clinical Epidemiologist and Data Scientist with two decades of experience working across Africa at the intersection of health, data, and innovation. He has supported governments, development partners, and private-sector organisations in applying analytics and technology to strengthen health systems and improve population outcomes.

Tariq Khokhar is Head of Data for Science and Health and Chief Data Scientist at the Wellcome Trust, a role he has held since 2019. He leads a team of experts who support innovations in data and AI that advance Wellcome’s strategies in discovery science, mental health, infectious disease and climate and health.  He brings deep expertise in data science, innovation, and public policy, with experience spanning industry, multilateral institutions, non-profits and philanthropy. Prior to Wellcome, he served as Chief Data Scientist and Managing Director at The Rockefeller Foundation and as Senior Data Scientist and Global Data Editor at the World Bank. Earlier in his career, he held roles including Open Data Evangelist at the World Bank and Technology Innovation Analyst at Development Initiatives. He also served as Chief Development Officer and Director at Aptivate, a non-profit focused on technology for international development. Tariq is currently a Trustee of UK Biobank and formerly Chair of the Board of Trustees at DataKind UK.

Selected Links and Publications

• Deployment of AI Models and a Mapper Algorithm to Enhance Clinical Decision-Making. ⁠https://doi.org/10.1109/i-COSTE63786.2024.11025118⁠

• AI-enabled Case Detection Model for Infectious Disease Outbreaks in Resource-Limited Settings. ⁠https://www.frontiersin.org/articles/10.3389/fams.2023.1133349/full⁠

• Assessing the Multi-Dimensional Risk of Stunting Amongst Children Under Five Years in Zimbabwe: An Application of Machine Learning and Advanced Econometric Techniques on Population Survey Data.

• ⁠https://wellcome.org/about-us/our-people/our-teams/data-science-and-health-team⁠ (Wellcome Trust)

• ⁠https://delft.care/cad4tb/⁠ (CAD4TB)

• ⁠https://pmc.ncbi.nlm.nih.gov/articles/PMC8459969/⁠ (an early CAD4TB evaluation)

• ⁠https://www.nihr.ac.uk/news/world-leading-ai-trial-tackle-breast-cancer-launched⁠ (EDITH AI Trail) 

• ⁠https://pubmed.ncbi.nlm.nih.gov/40816301/⁠ (Endoscopist deskilling effect with AI)

• ⁠https://lelapa.ai/⁠ (Lelapa AI) 

• ⁠https://lacunafund.org/⁠ (Lacuna Fund)

• ⁠https://www.newtonstree.ai/⁠ (Newtons Tree - middleware for deploying AI in health)

• ⁠https://www.bennett.ox.ac.uk/openprescribing/⁠  (Open Prescribing)

Who Owns Your Health Data?

The answer depends on where you live, and it's rarely the patient.

Jim St. Clair and Eric Sutherland join Patricia to unpack the legal gray areas, competing infrastructure models, and policy shifts shaping how health data flows (or doesn't).

Jim is a leader at the intersection of public health and technology, with over 25 years of experience in digital transformation. As former Executive Director of Linux Foundation Public Health, he expanded global collaborations in healthcare AI and patient identity management, partnering with WHO, UNDP, and national health agencies. He's shaped standards at HL7, IEEE, and ISO, and advises on AI workforce development, interoperability, and health IT modernization.

Eric is a Senior Health Economist leading the OECD's work in digital health, bringing together policy guidance for digital tools, integrated data, and responsible AI. He's accountable for evolving the OECD's Recommendation on Health Data Governance and supporting policy that balances data protection with timely access for patients, providers, researchers, and innovators.

They cover:

• Why patient data ownership may become inevitable, not optional
• The trade-offs between centralized, federated, and patient-intermediated architectures
• How the European Health Data Space is approaching cross-border interoperability
• What AI scribes and direct-to-consumer apps mean for data collection at the source
• The case for "data solidarity," balancing individual rights with public health needs

Join us in taking a practical look at what needs to change before patients can actually control their own records.

Ashley Girgis (Digital Health Program Manager, Terry Fox Research Institute) and Vinod Subramanian (Chief Operating & AI Officer, MedAdvisor Solutions) join host Patricia Thaine to explore the complex journey of healthcare data from research to real-world patient care.

Ashley shares insights from building Canada's $49M Digital Health and Discovery Platform, designed to break down data silos between hospitals, research institutions, and industry. Vinod brings 30 years of experience in healthcare, including pharmacy-driven patient engagement, discussing how platforms are transforming medication adherence and patient outcomes.

Together, they tackle critical questions: How do we ensure data represents diverse populations? What does "Privacy by Design" look like in practice? Why is patient consent the foundation of everything? And how can we bridge the gap between cutting-edge research and the pharmacy counter?

From OMOP data standards and FAIR principles to AI scribes and federated learning, they explore the technical infrastructure reshaping healthcare. But technology is just part of the story – they emphasize that trust, intentionality, and keeping patients at the center determine whether innovations succeed or fail.

Key topics:

• Building pan-Canadian data sharing infrastructure

• Privacy-first approaches in pharmaceutical technology

• The platform revolution in healthcare

• Bridging academic research and industry innovation

• Why healthcare lags in digital transformation compared to other industries

• Creating trust in AI-powered clinical tools

• The future of connected healthcare ecosystems

Whether you're building healthcare products, working with clinical data, or navigating regulatory frameworks, this conversation offers practical insights into the challenges and opportunities at the frontier of healthcare data.

*- About the Guests -*

Vinod Subramanian is the Chief Operating & AI Officer at MedAdvisor Solutions, leading AI strategy, operations, and innovation to elevate pharmacy-driven patient engagement. With prior leadership roles at Syapse and McKesson, he brings deep expertise in data, product, and cloud platforms. Vinod is passionate about using analytics, machine learning, and automation to simplify the medication journey and improve access to care. He blends technical depth with operational insight, focusing on solutions that empower pharmacies, enrich patient experience, and deliver meaningful health outcomes. LinkedIn handle: https://www.linkedin.com/in/vinod-subramanian X handle: @vindna

Ashley Girgis is a health informatics professional with a Master of Health Informatics from the University of Toronto. At Terry Fox Research Institute (DHDP.ca initiative) she is leading the development and execution of the $25M Digital Health Innovation Fund. Previously, Ashley evaluated LLM models, troubleshot data pipelines with engineering teams and supported EPIC EHR integration training and resource planning. She bridges technical and non-technical teams while keeping quality methodology and evaluation at the centre of her work (Visit her portfolio: https://girgisas.github.io/).

What if we could rebuild primary care from scratch – knowing what we know now about AI, robotics, and the lessons from eHealth Ontario?

Patricia Thaine sits down with Adam Cole (Senior Solutions Architect, Smile CDR) and George Smitherman (former Ontario Deputy Premier and Minister of Health) to explore the gap between healthcare's $100 billion budget in Ontario and its struggling digital transformation.

They dig into the real blockers: how 17 hours of weekly admin burden is pushing physicians away from practice ownership, why your EMR can't talk to the local hospital despite decades of "interoperability" promises, and what happens when venture capital starts eyeing those 10,000 primary care clinics.

The conversation spans from George's insider view of eHealth's ambitious failures to Adam's vision of AI handling 80% of routine cases – while tackling the thorny questions around privacy walls, innovation sandboxes, and whether grandma's Alexa companion is just the beginning of ambient AI monitoring.
From the minister who wrote Ontario's privacy legislation to the architect building tomorrow's clinical systems – they're mapping the path from good intentions to actual implementation.
About the Guests:

Adam Cole is a veteran healthcare technology leader with over two decades of experience driving digital innovation across enterprise and startup ecosystems. Currently Senior Solutions Architect and Head of Sales Engineering at Smile CDR, Adam has held senior roles with organizations like McKesson, the Canadian Clinical Trial Network, and HealthChain. A Canadian IT Professional of the Year and Smithsonian Laureate, he combines deep technical expertise with a passion for ethical, sustainable, and ROI-driven innovation in digital health.

Hon. George Smitherman is a trailblazing public servant, entrepreneur, and advocate with over three decades of leadership across government, healthcare, and social impact sectors. As Ontario's first openly gay MPP and former Deputy Premier, he led major portfolios including Health, Energy, and Infrastructure, overseeing transformative policies and major healthcare expansions. Since leaving politics, George has founded organizations driving innovation in workforce development, aging, and sustainability, continuing to champion pragmatic solutions to healthcare recruitment and community well-being.

Healthcare moves at the speed of trust – and right now, trust is the bottleneck.

Patricia Thaine sits down with Michael Lynch (Product Manager, Providence Health) and Sam Schifman (Innovation Architect, Vantiq) to talk through what it actually takes to build AI infrastructure that works in healthcare.

They cover the messy reality: sparse data, 40+ ways to encode "negative" in a single COVID report, why physicians need blood pressure context (not just raw numbers), and how agents might coordinate patient discharge without anyone hunting down a physical therapist.

The conversation moves from HL7's AI transparency standards and FHIR interoperability to MCP-based agent communication, federated learning, and the shift from deterministic code to systems you teach rather than program.

If you're dealing with unstructured clinical data, building tools clinicians will actually use, or trying to scale AI without breaking trust – this one's for you.

- About the Guests -

Michael Lynch is a product manager with nearly two decades of experience building and launching healthcare solutions that truly make an impact. His focus is on leveraging AI and machine learning to create smart tools that reach millions of patients and help clinical teams deliver their best work. He has led major transformations, founded a startup, and stays hands-on with the latest advancements by teaching and contributing to open-source projects. While pursuing his PhD in computer vision for medical imaging, he focused on creating datasets to teach machines human-like intelligence – a foundation that continues to drive his approach to AI innovation today.

Sam Schifman is a seasoned software architect with deep expertise across industries including healthcare, finance, education, and cybersecurity. He entered Healthcare IT in 2019 as Chief Architect at Diameter Health, where he focused on unlocking clinical data for government agencies, insurers, and Health Information Exchanges. A longtime advocate for responsible AI, Sam has worked with HL7 on AI standards, explored NLP and GenAI in clinical contexts, and frequently shares his insights as a speaker and guest lecturer at institutions like Harvard and conferences such as HIMSS. Now at Vantiq, he's focused on how Adaptive Automation / Agentic AI will empower the future of healthcare and improve outcomes for all.

- Links -

Michael's Aggregate Intellect Talk: Opportunities for Agent-Based Applications in Healthcare

Michael's LinkedIn: linkedin.com/in/michaellynchphd

Sam's YouTube: youtube.com/@Schifman_health

Sam's LinkedIn: linkedin.com/in/samschifman