Can Artificial Intelligence Make IVF Less Miserable?

AI won’t cure infertility, but it might help you plan better.

by Gillian Telling

If you’ve ever explored the option of IVF, there’s a good chance the insanely high sticker price shocked you at first. It costs an average of $15,000 to $25,000 per full round, which covers medication to stimulate the eggs, egg retrieval, embryo testing, and embryo transfer, among other costs. And with a success rate that ranges between 25% and 50% for women under 40, according to national averages, IVF is a little like going to the car dealership to buy a new car, dropping a bunch of cash, and crossing your fingers it will even start. Not knowing how many rounds of IVF it will take to have a baby, if it works at all, makes an already stressful experience even more taxing, and for many, financially out of reach.

Very few employers offer fertility benefits, though their numbers are growing, and fertility-focused lenders have cropped up to serve those looking to finance IVF. But another new player, Univfy, is using artificial intelligence to predict how many rounds an individual patient might need so they can plan accordingly. “When someone has cancer, you don’t tell them to have one round of chemo and see how that goes,” says Dr. Mylene Yao, M.D., an OB-GYN and Univfy’s founder. “The patient understands upfront that they might need to do it over and over until it works. The same is true for IVF.”

Everyone knows that the older you are, the harder it might be to have a baby. Beyond that, the probabilities get murky. “Age only accounts for 50% of the predictions,” Yao says. Univfy offers a report that breaks down the odds of IVF success in easy numbers: You might have a 65% chance of getting pregnant on your first cycle, an 85% chance on your second cycle, and so on.

“A lot of patients will get angry with the doctor. But when they have the data presented to them, it’s easier to swallow. They can’t get mad at a computer.”

The company’s AI considers factors Yao chose based on comparisons to live birth outcomes, including reproductive history, ovarian reserve numbers, clinical diagnoses from the female and male partner, and the length of time a person has been trying to conceive. The report also considers the success rate of an individual clinic, predictions that are more accurate and personalized than national averages. Yao says the tool is able to predict how likely it is that a patient will become pregnant over three rounds with 95% accuracy. (For comparison, in one recent peer-reviewed study, researchers developed an AI system that could predict IVF success with a 69% accuracy rate.) But the tool is also still evolving: As it runs more reports and collects more data, Univfy’s AI analyzes the factors that have contributed to the most accurate reports over time.

“We absolutely need this tool,” says Dr. Aimee Eyvazzadeh, M.D., a fertility expert in the Bay Area unaffiliated with Univfy. “It can be really hard for [doctors to tell patients] they only have a 3% chance of becoming pregnant because they don’t want to upset them.” A Univfy report can help take some of the emotion out of the injustice of infertility. “A lot of patients will get angry with the doctor, or they won’t believe you. But when they have the data presented to them, it’s easier to swallow,” she adds. “They can’t get mad at a computer.”

Yao’s interest in demystifying fertility began during her residency at McGill University when she kept encountering the same thing: unexplained fertility issues, which affect 15% to 30% of couples trying to get pregnant. “There were just so many people going through this,” she says. “And there are so many unanswered questions when it comes to each patient’s infertility. Even the doctors can’t really explain what’s causing it half the time.”

“The price and financial burden of IVF was way scarier than the needles and hormones for me.”

She decided to focus on infertility research, and began studying embryo development at Brigham and Women’s Hospital in Boston. “I thought, ‘That’s what I want to do: figure infertility out, and figure out how to best help these couples have a baby,’” she says. But it wasn’t until she was working as an assistant OB-GYN professor at Stanford University, where she led fertility and embryo genetics research funded by the National Institutes of Health, that she did what so many of her Silicon Valley colleagues were doing and turned to tech for answers. Univfy started as a research project, but eventually became Yao’s full-time job; the Series B company has raised $26 million, including investment from the women- and non-binary-focused v.c. firm Rethink Impact.

Univfy reports are available at no cost to patients at fertility centers in more than 50 locations across the U.S., Canada, EU, and U.K. Fertility centers pay Univfy a flat monthly fee to use the IVF success prediction model to run an unlimited number of reports for their patients. The reports were intended to prepare patients for the potentially long road of IVF, but they’re particularly useful when it comes to navigating the financial side.

“The price and financial burden of IVF was way scarier than the needles and hormones for me,” says Paige Chapin, a military wife in Virginia who was told she’d need to do IVF after she was diagnosed with severe endometriosis. The clinic she went to in Pensacola, Florida, offered Univfy reports, along with discounts if she paid for more treatments upfront.

“One thing I was totally näive to is that IVF is not a guarantee,” Chapin says. “I feel like a lot of people think that it is, and I definitely did. So it really helped mentally to have all the numbers laid out for me. When the report said I had a 45% chance of getting pregnant the first time, I was like, ‘Well, it’s not the best, but it’s not the worst.’ The second round jumped up to a 65% chance, and it was higher for the third round. It just really helped set expectations in a way that was manageable for me.”

She and her husband used the report to justify paying upfront for three full rounds of IVF at a steep discount. “I think it was about $16,000 or $17,000 per round at our clinic, but if you paid upfront for three cycles, it was $22,000,” she says. “So we went for that option. It was a no-brainer.”

Yao says most people who do at least three cycles of IVF become pregnant, but the biggest barrier to IVF success is the inability to pay to keep doing it indefinitely. At the same time, “Fertility is a situation where, to wait in order to save up money is not the best solution,” she adds. “That’s why we’re so excited about bringing this technology to providers so that they can, in turn, give the best kind of price offering to their patients.”

Many clinics now offer refund programs, in which a patient pays a lump sum upfront — typically covering three IVF rounds and all resulting frozen embryo transfers — and then, if the patient is unsuccessful by the end of those rounds, gets a portion of their money back. Yao says these programs, which are usually negotiated between the patient and the clinic directly, are typically only available to about 10% of patients, but when AI-powered predictions are involved, 50% to 80% of patients qualify.

In those cases, Yao says, “maybe the patient wants to consider adoption,” which can cost at least $20,000. “Or they may want to consider an egg donor, and they won’t have used up all of their resources.” And others just want to move on, she adds: “Some people say, ‘That’s enough. I don’t want to do more treatments. I don’t even wanna think about fertility ever again.’ But still, you get some money back to do other things in your life.”

With her first transfer, Chapin miscarried at six weeks. Her second transfer didn’t take, and her third miscarried at eight weeks. Her fourth transfer resulted in the birth of her now 1-year-old son, Owen. (All embryos used were retrieved during the first cycle.) “I was just so angry after that third miscarriage,” she continues. “I feel like if I hadn’t had the AI report telling me my chances were going to improve the more I tried, I wouldn’t have had it in me to do it again. It sort of allowed us to keep trying and then get our rainbow baby.”

Additional reporting by Greta Rainbow.