Aug 30, 2026 · 34 min · 12 segments
AI is changing healthcare fast, but behind the headlines, it’s standards like FHIR that ensure data is safe, consistent, and ready for real-world use. In this episode, they explore how robust digital…
Josh MandelGuest
Peter BirchHost
Michael LawleyGuest
And hey, Michael, you know, thinking about those scenarios, and particularly in Australia too, and I think one or both of you did touch on this point of how the community, in inverted commas, is going to shape what this looks like.
But, you know, healthcare industry overall is a big blob of people, and then there's those in the wonderful HL7 community, however you wanted to define that.
But then, of course, in Australia, we've got the SPARCT program that's accelerating the development and use of standards around fire and interoperability.
Particularly for Sparked, with all of these things happening, this context around AI and what it means for standards and interoperability, is that happening to Sparked and it needs to respond? Or has it got an opportunity to really shape how it's applied in Australia? What are your initial thoughts on the role of Sparked in all of this, Michael?
How much can Spark shape things? I think if you look at Spark as a community rather than as a standards group, I think that's its role.
So it's bringing together a diverse range of people, not just the techies, but the techies, clinicians, and increasingly we're seeing some patient involvement, some patient voices, and it can be the home for these conversations and for influencing And understanding what people see as the risks, I think, you know, for someone like myself, white male, well off, the risks are much different to someone who's got other socioeconomic challenges, come from a minority group and so forth.
Those risks may not be visible and we have problem of models tend to the average and the norm and so they have a tendency to eliminate or exclude risk.
variety, even while they have the opportunity to accommodate a much broader representative sample.
So I think as a community, that's really the opportunity of responding to AIs and hosting the conversations, understanding what the risks are and helping influence where the vendors go with some of their solutions and effectively act as a bit of a guide rail.
What's your take, Josh, on the role that community can play, bringing together really diverse kind of perspectives from within the healthcare space and professions, much like Spark is doing here in Australia? You're a little bit in a different position, but definitely in terms of use of data in a different position.
Yeah, I mean, maybe one thing I can do is try to reason by analogy a little bit.
I'll say in the US, I have helped lead projects in the context of a fire accelerator called the Argonaut Project, which was maybe like the first of these accelerators that got started 10 years ago.
And by far the most effective way that we could bring the community together and the places where we've had the biggest impact was when there was a clear regulatory mandate that was sort of setting the roadmap in the context of the US, what was our EHR certification program going to look like? And that regulatory mandate left room for industry to figure out how it should all work.
It can be hard to take an industry group and kind of spur new interesting activities without some looming threat or deadline.
But If you apply the right kinds of pressure and direction, it can be much faster and much more powerful to have that community figure out the solutions, brainstorm what might be possible, come back with ideas versus trying to hand a set of standards down through the regulations.
And so each year in Argonaut, we try to pick half a dozen projects where we think we can make meaningful progress in a year.
They are often fairly roughly aligned with these kinds of national priorities, but sometimes they just come from the actual technical roadmaps and customer demands from the participants in that community.
So there's a whole host of different community groups who are focused on things like how to test AI systems or how to represent the attributes of a model in something like a standardized model card.
And at least the way that I look at it, it's better to try to figure out where are the areas that are under-supported? What are the places where if Spark doesn't take on a project, it's unlikely that the ecosystem is going to move in this direction? And really try to use that as the core prioritizing function.
And hey, Michael, you know, thinking about those scenarios, and particularly in Australia too, and I think one or both of you did touch on this point of how the community, in inverted commas, is going to shape what this looks like.
But, you know, healthcare industry overall is a big blob of people, and then there's those in the wonderful HL7 community, however you wanted to define that.
But then, of course, in Australia, we've got the SPARCT program that's accelerating the development and use of standards around fire and interoperability.
Particularly for Sparked, with all of these things happening, this context around AI and what it means for standards and interoperability, is that happening to Sparked and it needs to respond? Or has it got an opportunity to really shape how it's applied in Australia? What are your initial thoughts on the role of Sparked in all of this, Michael?
How much can Spark shape things? I think if you look at Spark as a community rather than as a standards group, I think that's its role.
So it's bringing together a diverse range of people, not just the techies, but the techies, clinicians, and increasingly we're seeing some patient involvement, some patient voices, and it can be the home for these conversations and for influencing And understanding what people see as the risks, I think, you know, for someone like myself, white male, well off, the risks are much different to someone who's got other socioeconomic challenges, come from a minority group and so forth.
Those risks may not be visible and we have problem of models tend to the average and the norm and so they have a tendency to eliminate or exclude risk.
variety, even while they have the opportunity to accommodate a much broader representative sample.
So I think as a community, that's really the opportunity of responding to AIs and hosting the conversations, understanding what the risks are and helping influence where the vendors go with some of their solutions and effectively act as a bit of a guide rail.
What's your take, Josh, on the role that community can play, bringing together really diverse kind of perspectives from within the healthcare space and professions, much like Spark is doing here in Australia? You're a little bit in a different position, but definitely in terms of use of data in a different position.
Yeah, I mean, maybe one thing I can do is try to reason by analogy a little bit.
I'll say in the US, I have helped lead projects in the context of a fire accelerator called the Argonaut Project, which was maybe like the first of these accelerators that got started 10 years ago.
And by far the most effective way that we could bring the community together and the places where we've had the biggest impact was when there was a clear regulatory mandate that was sort of setting the roadmap in the context of the US, what was our EHR certification program going to look like? And that regulatory mandate left room for industry to figure out how it should all work.
It can be hard to take an industry group and kind of spur new interesting activities without some looming threat or deadline.
But If you apply the right kinds of pressure and direction, it can be much faster and much more powerful to have that community figure out the solutions, brainstorm what might be possible, come back with ideas versus trying to hand a set of standards down through the regulations.
And so each year in Argonaut, we try to pick half a dozen projects where we think we can make meaningful progress in a year.
They are often fairly roughly aligned with these kinds of national priorities, but sometimes they just come from the actual technical roadmaps and customer demands from the participants in that community.
So there's a whole host of different community groups who are focused on things like how to test AI systems or how to represent the attributes of a model in something like a standardized model card.
And at least the way that I look at it, it's better to try to figure out where are the areas that are under-supported? What are the places where if Spark doesn't take on a project, it's unlikely that the ecosystem is going to move in this direction? And really try to use that as the core prioritizing function.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
