Bridging the Gap: The Last Mile of Healthcare with Bilal Mateen, Digital Square @ PATH

05/02/24 • 57 min

Episode 3: How do we meaningfully bridge policy and real impact at the last mile of healthcare.

Shubs Upadhyay interviews Bilal Mateen, Executive Director of Digital Square at PATH about digital health challenges and successes in underserved communities. They explore the importance of safety, regulatory considerations, and the need for inclusive data sets and data infrastructure. The discussion also covers the role of community health workers and the impact of AI and large language models in healthcare.

Addendum : [As of October 2024 Bilal is now Chief AI Officer at PATH]

Key Topics:

Digital Public Goods: Exploring how digital solutions can be accessible public goods.
Health Data Poverty: Discussing how this issue affects global health equity.
The importance of strong data infrastructure
Regulatory challenges : Bilal explores some of the work that still needs to be done
AI in Healthcare: Insights on the use of AI and large language models to improve healthcare outcomes.

Chapters:

00:00 Introduction and Background
05:06 Digital Public Goods
07:04 How Health Data Poverty Plays Out
08:59 Reaching the Last Mile of Healthcare
15:47 AI and Large Language Models in Healthcare
29:00 Investing in Data Science Ecosystems and Regulatory Frameworks
32:59 More Global Representation in Regulation
37:03 Considering Local Nuances in AI Deployment
39:12 Divergent Approaches to Regulating LLMs
45:28 Regulation of LLMs as Medical Devices
48:40 Recommendations to Innovators about Healthcare Regulation

Links:

AMIE paper by Alan Karthikesalingam et al
Digitalsquare.org
Audere Website
Viamo Canada Call Centre
Siontis et al Paper on Diagnostic test comparisons

Episode 3: How do we meaningfully bridge policy and real impact at the last mile of healthcare.

Addendum : [As of October 2024 Bilal is now Chief AI Officer at PATH]

Key Topics:

Digital Public Goods: Exploring how digital solutions can be accessible public goods.
Health Data Poverty: Discussing how this issue affects global health equity.
The importance of strong data infrastructure
Regulatory challenges : Bilal explores some of the work that still needs to be done
AI in Healthcare: Insights on the use of AI and large language models to improve healthcare outcomes.

Chapters:

00:00 Introduction and Background
05:06 Digital Public Goods
07:04 How Health Data Poverty Plays Out
08:59 Reaching the Last Mile of Healthcare
15:47 AI and Large Language Models in Healthcare
29:00 Investing in Data Science Ecosystems and Regulatory Frameworks
32:59 More Global Representation in Regulation
37:03 Considering Local Nuances in AI Deployment
39:12 Divergent Approaches to Regulating LLMs
45:28 Regulation of LLMs as Medical Devices
48:40 Recommendations to Innovators about Healthcare Regulation

Links:

AMIE paper by Alan Karthikesalingam et al
Digitalsquare.org
Audere Website
Viamo Canada Call Centre
Siontis et al Paper on Diagnostic test comparisons

Previous Episode

Health Data Poverty Part 2 with Prof Alexandre Filho

Episode 2: Practical Solutions to Health Data Poverty

In this episode, we continue our exploration of health data poverty, showcasing a team that's making a real difference based on the issues discussed in Part 1 with Dr. Xiao Liu.

We're thrilled to have spoken with Prof. Alexandre Filho, a Professor of Machine Learning in Sao Paulo, Brazil. He shared insights on how his team has been maximizing the impact of data-driven technology in Brazil. Prof. Filho's remark, "The world is becoming more like Brazil," sets the stage for a discussion on what the global community can learn from Brazil's approach to data diversity, especially in serving underserved communities.

Key Topics:

Global Lessons from Local Successes: How Brazil's high-quality, diverse data sets serve as a model for the world.
Overcoming Data Challenges: Tackling issues with data availability and the performance limitations of algorithms developed in affluent settings when deployed in rural areas.
Leveraging Local Data: Details on the team's work, including a paper on neonatal mortality prediction using routinely collected data.
Direct Benefits to Data Providers: Ensuring that those from whom data is collected see benefits from its use.
Balancing Impact and Scalability: The trade-offs between highly tuned, local solutions and the need for scalable models that perform well across broader contexts.
Benchmarking for Better Health: Extended efforts in the ITU/WHO focus group on AI for health, evaluating the performance of models across different LMIC settings.
Advancing Local Impact: The use of transfer learning to enhance model performance and impact locally.
Prof. Filho’s Recommendations: Key advice for innovators and implementers in the EU, US, and UK.

Guest Bio:
Alexandre Chiavegatto Filho is an Associate Professor of Machine Learning in Healthcare at the Department of Epidemiology, School of Public Health, University of São Paulo. He directs the Laboratory of Big Data and Predictive Analysis in Health (Labdaps), which includes a team of 30 researchers focused on developing AI algorithms to improve healthcare decisions.

Find the team's work on Google Scholar

Next Episode

Creating impact with AI in isolated communities

Episode 4: Creating real impact with AI in isolated communities

In this episode, Dino, a pioneer in the digital health landscape, shares his insights on how digital solutions and AI are revolutionizing the healthcare industry. He discusses the critical challenges such as overcrowded healthcare facilities and the urgent need for support for healthcare professionals. Dino’s extensive work at Audere focuses on utilizing digital tools to improve healthcare delivery and outcomes, particularly in underserved communities.

Key Topics:

Digital Solutions for Overcrowded Healthcare Facilities: Dino explores how digital health can alleviate pressure on healthcare systems.
AI Tools in Healthcare: The role of AI and language models in enhancing diagnostic and conversational capabilities in healthcare settings.
Building Trust and Reliability: The importance of trust in technology, especially large language models, and how it's achieved through rigorous data handling and prompt engineering.
Partnerships and Local Collaboration: How forming strong local partnerships contributes to the successful implementation of technology solutions in healthcare.
Regulatory Support and Technology Verification: The necessity of regulatory backing in ensuring the safety and efficacy of new technologies.

Impactful Insights:

Improving access to rapid diagnostic testing through digital tools and AI.
Enhancing patient-provider conversations with advanced language models to make healthcare more accessible and efficient.
The significance of trust, reliability, and local partnerships in implementing successful digital health solutions.
Utilization of specific data sets and prompt engineering to enhance the accuracy and relevance of responses from large language models.
The role of specialization and niche focus in addressing particular healthcare and technology challenges.

Links