The doctors who dismiss PGT-P have almost never used it

You asked your fertility doctor about polygenic embryo screening, and they told you it's not ready. Maybe they said the science isn't there yet. Maybe they said professional societies recommend against it. You left the appointment feeling like the question was settled.

It isn't.

Here's what probably didn't come up in that conversation: a recent survey of 152 U.S. reproductive endocrinologists found that 89% of them have never offered PGT-P to a single patient. Not once. And yet 97% say they're familiar with the technology. That's a profession forming opinions about something it has almost no clinical experience with. When those same specialists were asked whether PGT-P should be allowed, they split 44% in favor to 45% against.

That's not a consensus. That's a coin flip.

The criticisms worth taking seriously

There are real scientific problems with how some polygenic screening has been done. Two in particular matter.

First: different scoring methods can rank the same set of embryos differently. One study found the top-ranked embryo only agreed across methods about 30% of the time. Which company you use changes the answer. That's a legitimate problem, and it means the scoring methodology matters enormously.

Second: most polygenic scores are built and tested by comparing unrelated people. But your embryos aren't unrelated people. They're siblings. They share the same parents, the same genetic background. The question that matters for embryo screening isn't "can this score tell apart strangers?" It's "can it tell apart siblings?" And for some traits, predictive power drops when you move to that harder test.

ASRM's December 2025 statement raised exactly these issues: insufficient within-family validation, inadequate cross-ancestry performance, risks of misleading patients. Those are the right questions.

But asking the questions isn't the same as answering them.

Someone actually ran the test

We tested 17 disease scores on sibling pairs. Not unrelated individuals. Siblings. The kind of comparison that actually matches how embryo screening works in practice.

16 of 17 scores showed no decrease in predictive performance within families. The scores that predict disease in the general population also predict disease among siblings. That means they're picking up genuine genetic signal, not statistical artifacts from population structure.

This is the test critics have been asking for. And when someone actually runs it, the scores hold up.

The logic builds on established genetic screening. We've tested embryos for BRCA variants (the breast cancer genes) since 2009. Genetic variants that predict disease in adults are present from conception. Polygenic screening looks at many variants simultaneously instead of one, which makes the validation challenge harder. But the empirical test now exists, and we publish our within-family validation results so anyone can evaluate the methodology directly.

How we build those scores matters too. We use a method called SBayesRC that separates genuine genetic signal from noise by modeling how genes actually affect disease, not just chasing statistical correlations. We calibrate across 8+ ancestry groups, because a score that only works for people of European descent isn't a finished product. These aren't things we do quietly. The data is published. Critics can check our work.

Where the actual hype lives

If the science is legitimate when done rigorously, where's the hype coming from?

From companies that skip the hard parts. Population-level validation numbers presented as though they apply to the embryo screening context. Relative risk reductions quoted without absolute context: a "50% reduction" sounds dramatic until you learn it means going from 1% risk to 0.5%. Scores built with older statistical methods when better ones exist. And no within-family validation at all, because the numbers look better when you don't test them against the harder standard.

Scott Alexander wrote one of the most thorough independent analyses of PGT-P and concluded that polygenic embryo screening is "obviously worth it" for families already doing IVF, while specifically flagging that the critical question is whether scores survive within-family validation. He described Herasight as having "the most impressive disease risk scores yet," precisely because we've done that work.

The question worth asking isn't "is PGT-P legit or hype?" It's "did this specific company validate their scores the way embryo screening actually works?" Did they test on siblings? What scoring method do they use? Do they report absolute risk, or do they dress up small numbers with relative comparisons?

What this means for your family

The profession is split. The evidence isn't.

When scores are validated within families, built with current methods, and reported honestly, polygenic screening works. It won't guarantee your child doesn't develop a disease. Nothing in medicine does. But it identifies which embryos carry higher or lower genetic risk for conditions like type 2 diabetes, heart disease, and Alzheimer's, and it gives you that information before transfer.

For families in which both parents have type 2 diabetes, the difference in risk between the highest and lowest-risk embryo can be as large as 23.5 percentage points. That's not a rounding error. That's a meaningfully different life.

Your RE may not have engaged with any of this data. Most haven't. That doesn't make them wrong to be cautious, but it does mean their caution isn't based on a careful evaluation of the current science. You deserve to look at the evidence yourself.

One couple with a strong family history of type 2 diabetes told our counselors that seeing the actual risk differences between their embryos turned an agonizing guess into a clear choice. That's what this data does. It doesn't make the decision for you. It makes the decision yours.

If you're curious about what screening could mean for your family's specific risk profile, you can explore the expected gains calculator. And if you'd like to talk through what the results would look like for your situation, reach out to our team.