I’m pleased to be joined by Christy Prada, the CEO of Future Fertility, a start-up developing AI-powered, non-invasive tools for personalized fertility care insights. We have an exciting conversation planned about the use of AI in the reproductive health space.
Before we dive in, I would love to hear a bit more about your background and your career journey that has led you to the work you’re doing at Future Fertility.
I was recruited to join Future Fertility while I was on maternity leave from Maple, a Canadian telemedicine company. I joined Maple as employee #3 and grew with them. I loved my time there, got to see and do so many different things in the Canadian healthcare space, and went on maternity leave thinking I’d be returning. I was approached by the Future Fertility board who was recruiting for the role, and I became obsessed with what the company was doing. I was fascinated by the industry and the innovative approach to solving a real clinical problem. As a new mom, it was also really appealing to work at a company that was focused on helping others achieve the joy that I have with my daughter. After many conversations with people at the company, different board members, and people in the industry, I just felt like I had to be a part of it. The clinical, technical, academic and scientific experience within the team was so impressive and I saw a ton of potential for growth. I'm so grateful to have such an amazing team to work with, we all bring something different to the organization and our unique blend of skills are what will take us to the next level.
What was exciting to you, other than the amazing team, about Future Fertility, that made you realize that you needed to be a part of it?
I knew I wanted to progress in my career to become a CEO at some point. I think choosing the time and the place to jump into that type of a role is really critical. You want to be sure it's the right environment and stage of growth for your experience, but also the right type of mission – something exciting enough to really focus you and push you. I felt like it checked all of those boxes for me, but it still comes back to the team. With the fertility industry being new for me, this meant that I could come in and really rely on the expertise there while also feeling like I had something to offer, bringing the experience of building and scaling the business at Maple. I felt I could complement my commercial and operational skills with the team expertise at Future Fertility. I didn't want to jump into the CEO role unless it made sense, both from an industry and a company perspective in an area that was exciting to me, but also in a way that my skills would translate well to what the company's needs were at that time.
It sounds like you learned a lot at your previous role at Maple. What else did you feel you took away from that role that helped you in your current role?
At Maple, while we were a startup, I think that we had the right amount of process and structure. When you're growing a company at this stage, everyone always talks about it being the wild, wild west, which is true, but there is a way to ensure things are still running effectively. While you need to have that space for agility, growth, and a little bit of chaos, you also need to have structure around it to allow people to collaborate effectively within their teams and across teams, allowing you to be responsive and agile. Having certain policies and processes in place are actually really critical for that, and I really learned and saw the value of that through my time at Maple.
In the reproductive health space, we’re seeing a huge wave of innovation around AI. How do you see AI helping to solve some of the key challenges or unmet needs in this space?
It's a good question. There are specifics in terms of how we're doing it, but I'll start by talking about the practice as a whole. What I've seen in my short time in this space is there are many different ways the fertility process can go right or wrong. There are so many different factors and variables that go into a successful pregnancy that I think it's really difficult for embryologists and clinicians to make sense of all of that data and how exactly each little factor plays a role. AI brings the ability to identify patterns across very large and diverse data sets. So the fertility space is a really perfect application for it, because of how many different factors are at play. I see AI as something that can help provide more information and insight about things that are not well understood in the process, allowing clinicians and embryologists to make better, more informed decisions for their patients.
What we're doing is bringing AI to focus on the eggs. There is no current standard of care on assessing the egg. There are measures for sperm, for hormone stimulation protocols, and to assess and grade embryos. We have measures for basically everything except for the egg right now. Looking at eggs with the human eye, even through a microscope, has not been successful in identifying differences or patterns that have an impact on success. Current decisions are made based on a patient’s mature oocyte count and generalized, age-based statistics relating to success rates. Technology, and machine learning specifically, creates the opportunity to make assessment more personalized. For example, not all 30-year-olds are going to have the exact same egg quality, nor will every cycle of eggs be the same for one specific patient. A more objective approach would allow you to measure: How good were a specific set of oocytes? Do they stack up with what we see from other 30-year-olds who had 10 eggs? What are your chances of success within those parameters? Does the quality look different than what we might expect within a group of that age range?
Regarding clinical actionability of results, how could a tool like this potentially change treatment or care decisions? Or even if you don’t directly impact care, how could it impact and improve the overall patient journey?
I'll caveat my answers here, as I’m not a reproductive endocrinologist or an embryologist, but there are a few key ways these types of insights impact treatment decisions. First off, we know that about 70% of cycles for IVF will fail. In those cases where it's failed, without the ability to understand if the egg perhaps contributed to that failure, it can be difficult to determine the next course of action. I wouldn't say that this is the only tool that's necessary to impact treatment into the future, but it is an important data point. Physicians have to make decisions based on a variety of different data points and factor all of that into their decision making. If you knew the egg quality was great, then you could isolate and say that it was probably not the cause of the failure, and instead make different decisions around sperm selection, embryo grading or other protocols to consider. At least understanding which variable was not the issue allows them to make decisions about other variables in that process. Or, if you know egg quality is not good and will likely impact success, you may be able to recommend a donor egg sooner and prevent that patient from having to go through unnecessary multiple cycles. Obviously, that's the extreme, but it's a good example of how this could impact clinical decision making.
This then ties into the patient experience, because when you are trying to educate a patient on what happened and why, you're also trying to counsel them on next steps. One of the biggest pieces of feedback we’ve heard from physicians counseling patients is they don’t have something that explains egg quality and how that played a role. Having tangible tools to support in those counseling conversations is actually extremely valuable to physicians. It allows them to point to something objective and data driven, and it allows the patient to have something to take home to better understand their own health. We're big believers that an empowered patient is really key to this process. Whether the patient is or is not happy with their egg quality, at least they know. Giving them a tool to have that information is very empowering for them. There’s a lot of research and studies that show the value of having an engaged and informed patient as a part of their care journey. We're seeing a lot of patients in the IVF process looking to be more engaged, aware, and informed. They're doing research on the internet and better educating themselves. I think this type of patient-centric information is going to become the norm. People are going to expect to have all that information at their fingertips and to be able to better understand the decisions physicians are making and what plays into that.
Another great example of this is for egg freezing. As it stands today, in the majority of clinics, if you go in and get your eggs frozen, you just leave with a VISA bill and maybe some vague info about the number of mature eggs. With our products, you can have information about the quality of your eggs, take that home and feel a bit more connected to the process. But beyond this, with our product, you’ll also have a better sense of your expectations for success. There have been lots of studies that have shown that with the growth in egg freezing, people really don't understand how it works. Once the eggs are frozen, what is your personal likelihood of success? There's not a lot of good educational content on this generally, let alone in a personalized way. So patients can be empowered by giving them information to be able to look at, understand, and help manage those expectations for the future.
If we don't have these types of technologies and pools of educational resources that are clinically accurate and informative, people are going to go and find their information elsewhere. That’s not in everyone's best interest. There's a lot of good information out there, but there's also a lot of not-so-great information. Making sure they have the right information is important.
Exactly. There’s an entire black box of unexplained infertility, so at least piecing parts of the answer together can be quite helpful. Now, thinking about potential cross-sectional opportunities and future growth potential, how do you envision this technology to be paired with other information? What opportunities may exist in the future?
There are other applications beyond the egg freezing and IVF processes where we have current clinical use cases. AI is now being used to understand egg quality for drug development, to help understand the efficacy of drugs. There could also be opportunities to use information about egg quality as an indicator for other elements of health. I don't think that we're necessarily there yet. But there are so many different interrelated elements of the human body. It seems feasible that egg quality could be an indicator for other health issues, health conditions, or health status. Or maybe a proxy for it, in combination with other information to better diagnose, prevent, or treat other women's health issues.
We're still in early stages. There's such a low level of investment historically in women's health. I feel all these things are not well understood right now. As we devote more resources and energy, we're going to find more linkages. The good thing is there is more investment and focus happening in this space recently. I think that's going to continue to grow and where we'll start to see really interesting connections that this type of technology can support.
Other than the process of standardizing and improving oocyte selection, what other areas do you envision your organization or AI in general being able to assist with?
There are so many other elements along the fertility journey that AI can support with decision making. We see companies tracking follicle development and optimizing stimulation protocols. There are companies looking at sperm, embryo implantation and timing of implantation, and endometrial receptivity. Even further downstream and upstream, there's opportunity. I’ve seen interesting companies starting to think about other types of women's health issues, like endometriosis and PCOS. Further down the chain, thinking about pregnancy and miscarriage could also create more opportunities. If you extrapolate to both ends of that whole process, it creates so much opportunity to use AI to either better understand patterns in what good looks like or just to make sense of all the different data points and bring them together in a way that can help inform decision points.
Beyond that, there are lots of conversations about what can be done to standardize processes in the lab and the clinic, like workflow management tools, aiding in patient engagement and patient education, information sharing and streamlining how that information is delivered through the lab. I've seen tools that are helping with annotations in the lab. I think there's a lot that can be done to either parse through the data or make it easier for physicians and embryologists to really focus on the elements of their job that really require their skill set and expertise by removing some of the manual, administrative labor.
Interesting – it seems like because this space is in its early stages, there's a lot of fragmentation in the different solutions that solve different problems. It seems like there's some consolidation happening at the care level for IVF in other countries. What are your thoughts on the likelihood or potential utility of consolidating these solutions under one standardized workflow? For example, instead of just assessing egg quality, you could also assess sperm quality and do a genomic test for endometrial receptivity all upfront. What are the pros and cons of that versus more of a stepwise approach to address different unmet needs?
There are a few things to unpack there. You're absolutely right that over time, there's likely to be consolidation in the space. Because it's so early right now, there are lots of startups going deep in different areas. In many cases, it takes years of research, tons of data and validation to get a model right, and companies are often starting by investing heavily in one area. As such, we're a leader in this space and I'm sure our colleagues in other parts of that journey would feel the same. So, I think the maturity, space, and size of these companies just means we need to break the problem apart and not focus on too many things. You want to make sure you have enough resources to do it properly. All of us are thinking about what other products we can do and what other solutions we can bring forward. Where do you learn from and leverage what you've done now, and how can that translate? As the industry grows, you'll start to see more companies become multi-focused companies.
On the clinic side, there is value in having multiple tools available in a streamlined way. That probably means we’ll need consolidation in the space or a willingness for companies to partner and work together to integrate. There's the ability to integrate across AI companies, but also with other ancillary tools and supports. Companies doing cryopreservation and storage have different lab equipment, like time-lapse incubators, cameras, and lasers. There's a lot of opportunity to create that cohesive experience. It doesn't necessarily always mean it has to be the same provider doing everything as long as you're able to integrate into that lab workflow. It means that a lab could have multiple different technologies in place. It’s up to all of us to be thinking about how to integrate processes in a way that's as seamless as possible.
It seems like there’s a lot of excitement from the founder and investor side. What does the adoption of AI look like from the provider and clinic perspective? Or even the patient perspective?
We’re seeing really great uptake. For instance, we recently moved into Latin America and had a ton of interest. We’re onboarding clinics at a rapid pace there, and clinics are implementing the technology and using it right away. I think because we've been able to streamline the process for implementing it in the lab, it’s easy for the embryologist to use, and because we’ve validated the clinical applications of this in multiple regions globally, we’ve reduced the barriers to adoption. We've made it easy for the doctors to adopt and integrate into practice and created lots of tools to help them understand how to use it for counseling. The feedback we get on our reports from patients and clinicians is really amazing. Obviously the clinics we're working with are early adopters in the space, but we are continuing to see the number grow, and the adoption curve is continuing to shift. I think that the number of naysayers or folks that are hesitant is shrinking, and we are excited about the momentum this technology is receiving globally.
Are you primarily seeing adoption from private practices or academic centers? Is there any sort of behavior in terms of the adoption curve by type of institution?
Honestly, it's a mix. There are academic centers looking to do research studies and understand the new, next big thing, and there are smaller private centres looking to implement new technology to compete with larger players. Most of the big clinic networks in Europe have their own internal research boards and scientific divisions that want to validate on their own, in their clinics. That's also true in the US. So we're working with a lot of these clinics to implement in their labs to not only validate, but also better understand how it can be used in clinical practice, and what else can we do to better understand the role that the egg plays in that journey. We're doing different studies with clinics globally to further prove out and expand clinical use cases of the tech, while also trialing use with their patients. It's been really refreshing to see the breadth of stakeholders who want to do research and publish, not just embryologists but also clinicians and academics. We are benefiting from research on both the lab side as well as the physician side, while also improving adoption by working with key opinion leaders globally.
It sounds like there's uptake in multiple regions globally. Are there any geographic areas that have had significantly greater uptake? And if so, do you have any hypothesis to explain why?
Europe tends to be early adopters of new approaches in the fertility world. Spain is well known for being a global center of excellence around reproductive care. I think naturally that means the rest of Europe is also focused on that as well. There are large clinic networks in the UK and Spain that have been doing research and working with us for quite some time. I think they were early adopters of the technology, because of that reputation for being academic thought leaders, but we also have had some really great partners in other regions. One of our founding partners is in Canada and we continue to do a ton of work with the lab there. We have a couple of clinics that we're working with in the US who have been extremely forthcoming with sharing data and doing studies with us. So we've been really lucky and grateful to have partners in different places all over the world, because it means not only are we working with lots of groups, but having a diverse set of data points, which means that our AI continues to get better and continues to be able to be applied in different clinical settings.
What are some of the challenges you're currently facing or expect to face that could potentially moderate the adoption of AI in this space?
There are a few key challenges that will impact the adoption of AI in this space. First off, you have to make clear, unique value propositions to the lab, to clinicians, and to patients and show them how you’re solving a problem and making it easy for them to use it. There's some nuance in what it means for each of those groups. Having a clear understanding and clear messaging for each is a really important element.
I'm also a big believer that it’s not enough to just have an interesting or awesome piece of tech or AI. It needs to really solve a true clinical problem and fill that gap. So having that type of principle has always been important for us and continues to be a driving force as we innovate and bring in new product ideas. We want to make sure it's solving a real clinical problem. In terms of the adoption or implementation side of things and how to get groups on board, having good engagement, training, and education is key. We need tools and resources to support clinicians’ and embryologists’ understanding of what the insights from the tool mean. Getting that feedback and continually having ongoing conversations and engagement with those groups to ensure that helps. And then there’s the integration and making sure it's easy to use – that we're embedding ourselves into the workflow, and it's not creating additional work. If anything, it should hopefully be reducing the amount of work. Adoption will almost be a non-starter if you can't find a way to integrate it into workflows.
In terms of AI overall, not just in this space, there's always concerns with how datasets are developed, how algorithms are developed and trained, and whether there's any biasing. When you're looking at a technology, what are some qualities that you think are indicative of a good data set? How can you ensure that you're building a good data set?
I love that question. There are a lot of things happening out there just generally in the world around AI. There’s buzz lately around ChatGPT and all the different things that you can use it for. With that comes an influx of companies doing AI and some will be good and some will be not as good. We actually just recently put out a blog series to help educate and inform on what good looks like. It starts with a very clear definition of the clinical problem you're solving and knowing which elements are playing into that. This step is very critical – you have to be very specific and in the weeds about what it is you're trying to solve for. There are also elements around the data that are important. You want to make sure you have enough data, but it also has to be representative of different patient populations and different outcomes. This is an essential starting point for building a good data set, in addition to many other factors that will have a big impact on potential success and usability.
Something else that’s interesting about this space is that the majority of fertility and IVF care in the US is either out of pocket, or covered by some employers, which I think naturally creates accessibility issues that elicit the concern of whether you’re training your data on certain socioeconomic populations and leaving out others. Maybe in the future, access to this type of care may change, but these datasets may have been trained using specific populations and therefore the algorithm could not be applicable to all populations. What do you think about this potential challenge?
It's a great point. I think it really kind of double clicks on the bias component of building a data set and why it’s so important. Population health is huge and varies drastically, not only just country to country, but also across states, and even towns. With different socioeconomic factors at play, there’are lots of challenges and root drivers of the social determinants of health that can impact not just access, but also other health challenges that different regions have. The reality is that the model is going to only be as good as its data. It’s a key part of our strategy to receive data from many different parts of the world. We may find that certain regions in the US where we've collected data happen to skew more towards certain populations. With the data we get from clinics across different parts of Canada, India, Europe, Latin America...we at least start to get some more diversity. I think the only way to solve that is to keep driving forward and ensure your model is refreshing, updating, and incorporating new data sets and is reflective of that. So it will always grow and build and bring in those drivers and outcomes from other regions. Of course it doesn't completely solve this problem, because until you have everybody in the model you're not fully removing that bias. But I think it’s the best that can be done at this stage given some of the challenges with access, particularly in places like the US and Canada where cost is so high and there's limited public dollars to support it.
It’s interesting, because when you think of different patient populations, often we think racial or genetic, but sometimes we need to go further than that. There’s epigenetics too, which hugely influences reproductive health, and different environmental factors like diet, lifestyle, and stress can tie to socioeconomic factors. But the research has been lacking, there’s not a big data set that shows these population level epigenetics, so I think it’s really great that big data tools are trying to crack at that a little bit more. I’m excited to see where that goes and how that may change our understanding of reproductive health moving forward.
I think we're only just scratching the surface and unfortunately there’s a lack of public dollars available. I do find from my work in the Canadian health system, where there are a lot of challenges and nuances with getting primary health care data for instance – this type of an environment is a lot more closed-loop. Fertility doctors work with embryologists and it's actually a really amazing rich data set that's fairly well contained and defined when compared to other areas like primary care. It just becomes a little bit easier to dig into and get a starting point. I think that's why we've seen so much explosion in the last few years of AI in the fertility space, because the data is well available, well documented, and there's a clear understanding of how it can be used. It’s a good starting point, but I agree with you there’s a long way to go. It'll get harder the further out we go and the more data we try to encapsulate, but there's a lot of value in doing so.
That wraps up my questions for today. It was great speaking with you and exciting to learn more about the work you’ve been doing to help contribute to precision medicine in this space. Thank you so much for your time!