Turns out it was a fool’s errand—the BCRAT wasn’t designed for people like me. Or Munn, for that matter.

In the demographics section, the assessment asks, “What is the patient’s race/ethnicity?” There were six options in the dropdown menu, in this order: “White,” “African American,” “Hispana/Latina,” “Asian American,” “American Indian or Alaskan Native,” and “Unknown.” None of these options alone would do—and I could only pick one.

I guarantee that if you ask someone who’s bi- or multiracial what it’s like to fill out a form, they’ll have a story for you. I, the daughter of a Black father and a white mother, have my share. Over the decades my options have expanded to such exotic categories as “Other,” “Biracial,” “Multiracial,” “Mixed Race,” “Some Other Race,” or “Two or More Races.” Sometimes I have the luxury of picking two.

It’s disheartening, but the lack of an appropriate racial/ethnic category on the BCRAT, however, may have serious consequences.

Cancer epidemiologist: Lack of data on nonwhite women mars risk estimates

The lack of a single bi/multiracial option makes sense; I assume my Black/white risk would differ from that of a woman with, say, an Asian/white background like Munn. But in the age of AI—not to mention a burgeoning multiracial population in the U.S.—surely the algorithm could let me choose two and calculate my risk?

NCI acknowledges the BCRAT’s limitations, namely a lack of diverse racial and ethnic data, which can skew results. The tool may underestimate risk for Black women with previous biopsies and Hispanic women born outside the U.S., and offer inaccurate risk estimates for American Indian and Alaskan Native women.

“The more data you can give an algorithm or a prediction tool or an AI, the better it’s going to be in giving you an answer, in this case, a risk estimate,” Robin Hines, Ph.D., a cancer epidemiologist at the University of Central Florida College of Medicine, tells Fortune. “If I have 280,000 women, I’m going to have much less error in my estimate than I will have if I have just over 3,000 women.”

Hines is referring to the 280,000 white women in the Breast Cancer Detection Demonstration Project. Their data, along with information from the NCI Surveillance, Epidemiology, and End Results (SEER) Program, were used to create the BCRAT. Risk estimates for nonwhites are based on SEER and other study data for far fewer women in these racial and ethnic groups:

• Black: 3,244
• Hispanic: 2,497
• Asian and Pacific Islander: 1,563

Dr. Mitchell Gail, now a senior investigator in the Biostatistics Branch of the NCI Division of Cancer Epidemiology and Genetics, developed the calculator in 1989, before the National Institutes of Health Revitalization Act of 1993 called for “an increase in the number of women and individuals from disadvantaged backgrounds (including racial and ethnic minorities) in the fields of biomedical and behavioral research.” Also known as the Gail Model, the BCRAT algorithm was last revised in 2017 and may be updated periodically as new data become available, according to the NCI.

Olivia Munn’s OB/GYN wants you to know your breast cancer risk

I took the BCRAT twice. When I input “African American,” the tool generated a lifetime risk of 9.8%, compared to an average risk of 10.1% among 35-year-old Black women in the U.S. When I input “White,” my risk jumped to 12.4%, compared to an average of 12.6%. I don’t know the nuances of the algorithm, but my biracial risk is probably more complicated than simply averaging the two scores.

How, then, did Munn’s OB/GYN, Dr. Thaïs Aliabadi, calculate the biracial actress’s risk? Aliabadi, who’s preparing to launch a calculator on the website of her new podcast, SHE MD, declined to specify which assessment tool(s)—the BCRAT is one of many—she used for Munn. But in her push to get every woman to understand her risk of breast cancer, Aliabadi tells Fortune the lack of racially and ethnically diverse tools is an issue.

“We have good tools to use for the general population but, obviously, we need to have more studies and focus more on different ethnicities,” she says. “Hopefully, by collecting a lot of data—and artificial intelligence—we’ll be able to put all this data together and come out with a better test.”

President Joe Biden on Monday signed the Executive Order on Advancing Women’s Health Research and Innovation, which concedes “we know far too little about women’s health across women’s lifespans, and those gaps are even more prominent for women of color, older women, and women with disabilities.” It calls for the investment of federal dollars in closing that gap.

For now, Aliabadi’s biggest concern is that not enough women—regardless of their race or ethnicity—are even aware of their ballpark risk, myself included. That I’m 35 and had never heard of a breast cancer risk assessment until Munn’s Instagram announcement is “exactly the problem,” she says. While I’m dissatisfied that most existing tools don’t, or can’t, account for biracial women, I do feel empowered with at least an idea of my risk. And though my estimates are relatively low, they’ve prompted questions I want to ask my doctor.

“I want to change that and make that standard of care for all women—not just high-risk women, all women,” Aliabadi tells Fortune. “It takes two minutes. Look at how much time we spend doing other things in life. What’s two minutes to calculate your lifetime risk of breast cancer?”