We originally started this blog page to speak with patients, but we're getting
many questions from fertility doctors who want to know how the personalized prognostics developed by Univfy can help their patients. I've decided to use this blog forum to address those doctors, but I welcome the public to read this series as well. Who knows, you might get a kick out of seeing how doctors talk to one another! Please feel free to send us a comment or question at email@example.com. You will not be identified in public unless you specifically ask me to do so.
My research team and I have worked hard to develop fertility predictive testing because we want more women to be able to realize their dreams of building a family. No doubt you've seen patients who waited too long to seek your help, simply because they weren't aware that time was such key factor in their fertility. On the flip side, you've seen women who probably could have had children give up after just one failed IVF cycle. Other women never seek help because they wrongly assume there is no hope for them.
Fertility predictive testing has the power to counter both of these scenarios: It can educate women about the importance of seeking help sooner rather than later, and it can provide more precise predictions of success for women who might otherwise assume their chances of having a baby are slim.
In reproductive medicine, we have a wealth of data with which to personalize outcomes prediction. Markers of ovarian reserve, embryo quality measures, reproductive history, and health data such as body mass index (BMI) and chronological age have all been individually associated with fertility treatment outcomes. However, unless these factors are built into a prediction model that can be objectively and quantitatively tested, the information they carry is not easily utilized because there are tremendous correlation and overlap of information among these predictors. For example, how do you compile the odds of increased or decreased live birth rates of each clinical factor into one composite number that would show the probability of having a baby in a way that is meaningful to your patient?
We have found that machine learning is a very powerful way to tackle such complex data sets in fertility predicitve testing. While we may not realize it, we're already using machine learning in our everyday lives: It powers most internet search engines, recommendations from Amazon and Netflix, and the "people you may know" feature on Facebook or LinkedIN. In other words, we are afforded more accuracy and personalization in our shopping experiences than most patients receive when making medical decisions.
To apply machine learning to infertility predictive testing to predict IVF outcomes, we first had to peel away some long-held assumptions in fertility statistics. We assumed instead that we did not know which of many clinical factors, such as age, BMI, clinical diagnosis, or sperm count are strong outcome predictors, and how they rank in terms of predictive value. In doing so, we opened the way to answering many complex questions.
In my next blog, I will remove some common misconceptions about predictors, and show how that paves the way to applying the boosted tree, a method that is well-established in machine learning.