Connecting data silos using a social network approach to support AI for Women’s Health: Case study in fertility

In Jan 2023 Dr Michelle Perugini presented at both the Precision Medicine World Conference (PMWC23) in Silicon Valley and at JP Morgan Health Conference (Biotech Showcase) in San Francisco.  Below is the transcript of the presentation, the video of the presentation, and the presentation slides.

Slide 1

Hi Everyone, thank you for coming.

My name is Michelle Perugini, and I'm the Co-Founder and CEO of Presagen.

Today I would like to discuss social networks, as an approach which has been highly successful for media, with Facebook and Twitter, and for video, with YouTube.

I'd like to discuss how social networks can solve some of the biggest challenges in healthcare, particularly with a focus on women’s heath.

Slide 2

When we talk about women’s health, many still say its niche. Why is that?

Women are the primary healthcare decision makers in the household.

They do majority of the healthcare spending, and they represent half the worlds population.

That certainty does not sound niche to me. Possibly unique, and definitely under-served.

And therein lies a huge opportunity to create impact.

Slide 3

The women’s health market is estimated at $60bn annually.

But when you consider all the specific women’s health areas, like pregnancy and menopause and ovarian cancer, as well as other health conditions that affect both men and women, but disproportionately affect women, or affect women differently,

$60B seems like a gross underestimate.

Women’s health has been largely misunderstood and undervalued, and what we need now is a new targeted approach in a system that has been historically designed for men.

Slide 4

When you consider applications like precision medicine and artificial intelligence, their success really relies on DATA.

They rely on good quality, globally diverse, data, to ensure that technologies can support all women around the world, without bias.

Data will allow us to better understand the differences and treatment between men and women, between female demographic sub-groups,

and importantly, identify better ways to address largely ignored female specific issues.

Slide 5

However… women’s health has a huge data access and bias problem.

Women’s health issues are typically treated by small specialist clinics, rather than big hospitals.

This means that data is sparce and poor quality, as well as disaggregated and geographically distributed among many clinics globally.

It becomes a massive challenge then to connect those siloed and private datasets held by many different stakeholders.

Slide 6

At Presagen, we've solved the problem of connecting globally distributed and siloed data by using a social network model.

Slide 7

So, what on earth do social networks have to do with healthcare?

Slide 8

At a basic level, social networks are a decentralized network with a globally connected user-base.

The social element of the network allows collaboration and sharing amongst its users.

And the network grows via the network effect, which I'll explain in a minute.

Importantly, it democratizes and shifts power from large companies to the masses.

For example, with YouTube, there is a power shift from large studios to anyone with a camera that can create and share videos.

Slide 9

To explain this further, lets continue to use YouTube as an example.

Users who are the content creators can create and share videos for other users, who are the consumers.

Content creators are incentivized by royalties on advertisements.

Any user can be a content creator or a consumer, or both.

Slide 10

The great thing about social networks is the network effect.

Content creators create videos because of the incentives.

The more content creators on the network drives additional video content.

The more video content drives more consumers to join the network and watch that content.

And this in turn drives more demand for new content, which in turn drives more content creators to meet that demand, and so on, and the network grows.

Slide 11

At Presagen, we're using the same social network model to access global datasets.

Clinic users from around the world can share medical data, with patient consent, to support the development of AI healthcare products.

These clinics can be of any size, anywhere in the world which encourages diversity and allows anyone to participate.

Clinics are incentivized by royalties on the future sales of the AI healthcare product that is being developed.

And any clinics can be a data contributor, or a consumer, or both.

Because the AI is so specialized, we have a technical team that does all the hard work in the middle, which includes the AI build, clinical validation with support of the collaborators, regulatory approvals, and product distribution through the global network.

Slide 12

This model is also driven by the network effect.

Clinics become collaborating users due to the incentives, as well as being involved in innovating products and publications.

The data they contribute allows us to build additional products.

The more products we build encourages more consumers – or other clinics to utilize these products for their patients.

And the additional consumers drives more demand for new products.

So users drive which products they want and the expansion of the network into new verticals or industries.

And the more demand drives additional clinics to become collaborators, and so on,

and the network grows and organically expands across different healthcare verticals or industries of great value.

Slide 13

To make all this happen, there are significant healthcare data challenges that we needed to solve.

We've created a global decentralized cloud network that allows us to access these global datasets without ever having to move the data across borders or centralize the data into a single location or country, which many privacy laws do not allow.

The data remains protected because it remains local in each region.

Since we can’t centralized the data for AI training, we use our own unique federated AI training algorithm that allows us to train on data distributed all throughout the world without having to move or see that data.

And lastly, we solve the data quality problem with a powerful algorithm that can automatically and remotely detect errors in that data stored on our global network.

Slide 14

So now I would like to switch to talking about how we've used this social network approach to build our first product called Life Whisperer, which is aimed at helping IVF patient’s get pregnant sooner.

Slide 15

Life Whisperer has two AI applications in market, with authorization to sell into 48 countries.

The two AI applications assess embryo quality, which is a critical part of IVF treatment.

Life Whisperer Viability assesses the embryo to identify which embryo is likely to lead to a pregnancy.

Life Whisperer Genetics assesses that same embryo image to identify if the embryo is likely to be genetically normal.

We also have a third application in production, which is Life Whisperer Oocytes, which applies earlier in the IVF treatment, to assess the quality of eggs.

This new application is also applicable to patients wanting to freeze their eggs, in addition to IVF.

Slide 16

Our global collaborative network is growing, and is now at over 200 individual clinics globally,

representing 18 clinics networks that are collaborating with us, and over 100,000 patients across 13 countries.

As the network grows, it also helps us to be able to access data faster, and build products faster, as the network effect generates momentum.

Slide 17

Life Whisperer’s AI is truly globally scalable and unbiased, which ensures it is accessible and affordable to any clinic, and their patients, of any size, anywhere in the world.

Life Whisperer is currently being used in really advanced clinics in Europe and Canada, to small remote clinics in Trinidad and Tobago, all the way through to massive clinic networks like Indira IVF, which comprises over 100 individual clinics across India.

Soon we'll be looking to diversify and start connecting users and their data to our collaborative network in other areas of women’s health, starting with obstetrics and gynecology.

Slide 18

Thank you so much for joining me. If you’re interested in downloading this presentation you can go to Presagen.com. Thanks so much.