Calculating odds for a successful in vitro fertilization
Nearly 12 percent of women between the ages of 15 and 44 have difficulty conceiving and turn to fertility services, according to the Centers for Disease Control and Prevention. But those services can be very expensive: A cycle of IVF, or in vitro fertilization, costs an average of $23,000, according to FertilityIQ.
Why so pricey? The procedure is complicated, and it can involve several rounds of hormone treatment, egg collection and embryo creation -- all in a very carefully controlled, sterile environment. And many women end up undergoing more than one cycle before they ultimately conceive.
Add in the fact that insurance rarely foots the bill (only 26 percent of companies with over 500 employees cover the procedure, according to a recent study by Mercer Health), and it's easy to see why the costs can be prohibitive, especially if a patient doesn't have an accurate picture of what her chances are for conception using IVF.
Enter Univfy, a company founded by Dr. Mylene Yao, an obstetrician-gynecologist who has developed models to better predict a woman's likelihood of conceiving using IVF. Yao likens it to a "FICO score for fertility" -- a way of assessing a patient's risk factors and determining her chances of having a baby. (FICO scores are assessments of consumer credit used to determine eligibility for loans, among other financial services.)
The company was started in 2009 and has partnered with IVF clinics across the U.S. to collect and analyze patient data, enabling it to develop predictive reports tailored to each clinic and each patient. According to Yao, 12 centers in the U.S. are currently using the reports, with 50 more in the pipeline to add the tool this year. The centers pay Univfy a subscription fee to use the services.
The company uses machine learning to map the likelihood of a patient conceiving over the course of three IVF cycles. It takes into account a number of factors, such as maternal age, body-mass index (BMI), hormone levels and the fertility of the patient's partner. Each report can be specific to the clinic and patient because each program is constantly updating its data for Univfy to provide a more accurate reading of what a given patient's chances of having a baby will be at that practice.
Yao said this not only helps patients by giving them a better understanding of what their own chances of conceiving really are (instead of being quoted a national average) but it also allows patients to qualify for a refund if they don't conceive.
"Without using machine learning, using traditional statistics, you would have to set up a threshold or cutoff -- different programs use different criteria: Your age has to be under 38, or your BMI has to be under 35, or your ovarian age has to be under a certain amount. So there would be all these cutoffs so that program could be viable. But then, only 10 percent to 15 percent of people will qualify," Yao said in an interview with CBS MoneyWatch.
"But using machine learning, it's not necessary to have cutoffs like that -- for each woman there's a probability," he added. "So based on the probability, we can also calibrate it and calibrate the refund levels. [For example], a woman who has high chance of success may be offered 80 percent refund, lower may be offered 50 percent refund."
Dr. Tamer Yalcinkaya, founder at the Carolinas Fertility Institute, has been using Univfy's report in his clinic for several years. He says many couples are concerned about spending so much money without knowing whether the treatment will take, and this report helps them understand their chances.
"First, I'm educating a patient, and for patients who are so inclined, a cost of a refund policy can be calculated, and the amount of refunds can be estimated according to an accurate prediction," Yalcinkaya said in an interview with CBS MoneyWatch. Some people will "want to buy some kind of insurance so that if they don't get pregnant, they'll get something back."
Yalcinkaya added that the Univfy program benefits both sides: "The practice can offer a refund program based on the real chances of that person getting pregnant ... and the patient pays a little extra. The risk is distributed in a fair way."
Dr. Bradford Bopp, a fertility specialist at Midwest Fertility in Carmel, Indiana, has been using Univfy's reporting tool for the last four years. He said the fertility process can be incredibly stressful and expensive, so having an independent way to evaluate the chance of a patient conceiving is helpful to both the doctor and the patient.
"The report can visually put out for them their individual chance for success and give them some data behind it to go, 'gosh I've got a 5 percent chance here,'" Bopp said in an interview with MoneyWatch. "It's done by statisticians from a completely independent place. It's a third-party, independent company that has no investment in me. They're just generating a report for me, and [the patients] can take it and do with it as they will."
"There are some couples that have no idea," he said, "and then we generate the report, and they have a better chance for success than they ever thought they would have. It gives them hope, and it gives them some measurable variables to make the decision that this is worth the investment."
