Life or Death: Fostering an Open Health Data Market

Liberty. Equality. Data. Podcast #10 with Dr. Jaan Altosaar

In this episode, we talk with health data and machine learning expert Dr. Jaan Altosaar, a Postdoctoral Officer of Research at Columbia University Irving Medical Center focused on machine learning methods for health and science. Jaan has a background working with health data from Princeton, DeepMind and Google Brain.

We discuss the complexities of the current medical data systems, how to gather data yourself and what are the implications getting your own personal data, and how you can utilize your data for your benefit, such as genome sequencing. We discuss open source’s role in the development of the data markets and how governance of AI/ML models could evolve to fit a growing need for more data utilization in a way that works for entire ecosystems of individuals.

You can listen to this episode on Spotify, Apple Podcasts, and Google Podcasts.

You can find the transcript of the podcast here.

In case you don’t have the opportunity to listen to the entire podcast, here are the key takeaways from the conversation with Jaan:

1/ Collective dissonances are all around us. How can we utilize data, technology, and science in order to solve the most pressing issues in health?

Jaan began by explaining his background and how our individual and collective memories influence how we perceive the world. Jaan gave a number of illustrations of how cognitive dissonances — the way how we perceive conflicting information — could be used to solve problems that exist in the world today: especially how using technologies and data could help create new incentives to build new tools and solutions on health and medical space and beyond.

Jaan explained that his passion for working in the overlapping domains of health, data, science, and technology is oriented around the ambition to make the world a better place:

“What devices should we employ to change the world in the ways that we want to? Health is a nice area to work in because it is all about life and death.” — Dr. Jaan Altosaar

In the first part of the Podcast, Jaan and Markus discussed extensively how data and AI could be used to improve decision-making processes in medical situations. Here are some practical examples:

• Intensive care units (“ICUs”): how can we build an algorithm that determines the rank of priority which of the patients gets treated first?
• Cesarean delivery: What factors determine which mothers have to undergo C-section procedures?
• Accessing your genome data: what are the current challenges that lie deeply embedded in the healthcare system that curtail the ability to access and utilize patient medical data?

2/ Unlocking the opportunities with data in the health and wellness ecosystem: reversing incentives, empowering individuals

Jaan elaborated on his personal experiences working with hospitals in the health service industry; he provided practical illustrations of how various segments of the health and medical services industry suffer from systemic shortcomings and inefficiencies. To some degree, part of the problem is because there is much data, and system-wide ways to manipulate the system: for instance, a person (patient) may have a medical record indicating that he has both type I and type II diabetes (which is physiologically impossible), but that helps hospital get additional income.

”Somewhere in the system things have been set up in such a way, that there is this emergent property that seems like really evil, and nobody can quite know how to ascribe blame.” — Dr. Jaan Altosaar

Liberty. Equality. Data. Podcast host Markus noted that in many instances, such problems are actually driven by the complexity of the system itself, and the most critical actor — the individual patient — is left outside. Paternalistic approaches that individuals can not handle their medical and related health data have led us to quite a denigrating status quo. All of this makes the bottom-up empowerment of individual patients even more exciting.

“I think there is no option but to give people a little bit more choice. It is a global user experience problem where we know for a fact that we will never have enough psychiatrists or therapists, doctors, or nurses. … The only solution that I know is to build some technology and do some research on how that technology is built.” — Dr. Jaan Altosaar

3/ Open-Source Solutions have enormous potential

Jaan and Markus discussed the potential of open-source tools and a community of developers who could build new types of applications of individuals’ own data. If the open-source has been a foundation for so many enterprise applications, why can’t such solutions be brought to individuals?

“I already know that open source solutions work. The really hard part is how do we collectively abdicate as much responsibility as possible? The moment somebody becomes responsible, their fighter system kicks in: they are concerned about the liability. What if the AI system that I built kills somebody? I don’t want to be responsible for that.”— Dr. Jaan Altosaar

4/ Decentralization of decision making: what role will user-held, user-generated data play in this process?

Jaan and Markus shared their excitement and insights about the repeating patterns of decentralization. With crypto and decentralized governance ideas taking the main stage nowadays, much hope and actual work is being done to rethink the currently existing, centralized legacy systems of governance.

Jaan gave one illuminating example of thinking how bottom-up approach to designing governance approaches could be applied in addressing contemporary problems related to the use of data and AI: allocating liability for decisions made by algorhithms that power self-driving cars should be based not on some theoretical cost-benefit assessment, but on the answers that could be collected from people who would be affected by a likely accident caused by such a self-driving vehicle.

5/ The future is about giving individuals a choice to make their own decisions: decisions about their health, their wealth, and their freedom

As we look into the future, it is crucial to ensure that agency and autonomy are given to individuals themselves. We are moving into a world where people will have ownership and control over their own data. In the coming years, new tools will empower individuals with their own data: think of an AI doctor that detects changes in your behavior and nudges you to make meaningful adjustments to your routines. Many such use-cases already exist at an enterprise level.

Jaan ended with a cautionary tale and suggested taking an attitude of irreverence — we should start by questioning whether we need certain things in the first place. If we do, we can determine what interventions would be most desirable to individuals and be careful not blindly follow paths superimposed by legacy enterprises:

“Yes, we know that enterprise-first approach works; however, I think that we need to be really careful not to let the their mindset to get the best of us.” — Dr. Jaan Altosaar

You can listen to this episode on Spotify, Apple Podcasts, and Google Podcasts.

You can find the transcript of the podcast here.

Connect With Us and Stay in Touch

Prifina is building resources for developers to help create new apps that run on top of user-held data. No back-end is needed. Individual users can connect their data sources to their personal data cloud and get everyday value from their data.

• Follow us on Twitter, Medium, or LinkedIn.
• Listen to our podcast.
• Join our Facebook group Liberty. Equality. Data. where we share notes about Prifina’s progress.
• You can also explore our GitHub channel and join us at Slack.