“I avoid going to the doctor. In general I try to have as little interaction with the medical health system as possible.” — non-binary study participant

Gender 101

As a primer: one’s gender expression, gender identity, and gender assigned at birth may or may not be one and the same. Rather, people may have an infinite variety of gendered experience in the world.

The social construction of sex and gender is becoming more widely accepted, and people have lived beyond the binary across civilizations and millennia. Yet somehow, in 2020, almost every form field you’ll encounter in every interaction still looks like this:

This means that to our products, transgender non-conforming users don’t exist.

Across the U.S., more and more states are recognizing non-binary genders.

Mother Jones, Source: National Center for Transgender Equality

If our products don’t as well, we will continue to leave out entire populations of people.

What is the ideal member experience for gender collection in a healthcare context? Particularly for our TGNC members?

Healthcare is a dated industry filled with legacy systems and archaic technology. Oscar, a tech-forward health insurance company, was founded to disrupt that history and center the experience of the patient.

On the Growth team at Oscar, I set out to understand how we might provide the best possible user experience for gender collection in our products, for all our members. Here are our learnings.

Step 1: Do you need to collect gender at all?

A general guiding principle for gender inclusion allies in the UX design field is that if gender is not intrinsic to your experience, don’t collect it.

For instance, should a person’s banking experience be any different due to their gender?

Including gender in your product can also lead to significant product debt further down the line if your approach forgets inclusivity from the start. Something AirBnb had to remediate in its product after the fact.

Not accounting for diversity in your product can also lead to disjointed user experiences.

When there really is no use for it, sometimes it’s best to skip gender collection altogether.

Step 2: Understand its impact on your users

In some instances, say healthcare, gender can actually be incredibly important to one’s experience.

This experience with a seemingly simple field can be frustrating for TGNC people and also have significant impacts on their health. How can a doctor provide personalized care for TGNC individuals if they don’t know the patient’s gender identity? And how can a healthcare company properly serve its members, if their identity doesn’t exist in their data?

The lack of appropriate recognition of TGNC individuals within healthcare also has dramatic health consequences:

23% of transgender people avoided going to a doctor when sick or injured out of fear of discrimination in the past year.
33% of transgender people who saw a health care provider in the past year were harassed, denied care, or even assaulted.
2015 U.S. Transgender Survey

This leads to disenfranchisement with the healthcare system, resulting in negative health outcomes; compounded by factors of race, immigration status and class.

Furthermore, when legally changing one’s gender, the dreaded form field is compounded — where the name and gender on their identity card, birth certificate, prescriptions, doctors forms and health insurance forms all have to be updated. The process is often bureaucratic and dehumanizing.

If you’ve been following the news these days, the current administration is trying to undermine the Health Care Rights Law, “a lifesaving law that helps transgender people access the health care they need without discrimination from health care providers or insurers.” (https://protecttranshealth.org/)

The proposed legislation will introduce further confusion and friction for how TGNC people access care.

They are not edge cases. They are human beings.

As the majority of gender fields remain stuck in the binary in our products and healthcare forms, there’s an indication that the experiences of TGNC people have not been prioritized in the products and services we build. And it’s our role as designers to consider the experience of all our members, not just the majority.

“When you decide who you’re designing for, you’re making an implicit statement about who you’re not designing for. For years we referred to people who weren’t crucial to our products’ success as ‘edge cases’. We were marginalizing people. And we were making a decision that there were people in the world whose problems weren’t worth solving.
Facebook now claims to have two billion users. 1% of two billion people, which most products would consider an edge case, is twenty million people. Those are the people at the margins.
These are the trans people who get caught on the edges of ‘real names’ projects. These are the single moms who get caught on the edges of ‘both parents must sign’ permission slips. These are the elderly immigrants who show up to vote and can’t get ballots in their native tongues.
They are not edge cases. They are human beings, and we owe them our best work.”
- Ruined by Design, Mike Monteiro

If 10% of our 400,000 Oscar members are LGBTQ, that’s 40,000 members who may currently experience friction every time they encounter binary gender fields.

Data & disenfranchisement

Lack of inclusivity in your gender forms can also lead to inaccurate data, which can have significant impact on public health. The problem is industry wide in healthcare, as carriers and insurance companies use inconsistent and unclear field labels.

For instance, Healthcare.gov uses “sex” while CoveredCA uses “gender identity”. These labels often map to the same source for insurance carriers, assuming that our member’s gender identity, gender assigned at birth, and legal sex are always one and the same. But the differences in these data points can mean significantly different things for a person’s care.

For instance, “gender” can be interpreted as gender assigned at birth for population health interventions and risk models. Or “gender” assumed to be the member’s gender identity and pronouns when addressed by member support. And most of our products presume “gender” matches a person’s legal sex, or gender marker on identity documents for identity verification. When these data points don’t align, this introduces significant friction for our users, who often have to navigate uninformed and arduous channels to access care equal to cis-gender patients.

Step 3: Talk to the people it affects most

Some consumer facing products have done more to bring inclusivity into their experiences.

And some healthcare providers are starting to offer more inclusive forms to their patients.

A leading pattern in the field is the two-question method: a method used by some EHRs and academic researchers.

It yields near-zero missing data and two times the transgender spectrum response rate. It also uses language that is respectful of TGNC identities.

To ensure we were providing the best user experience possible, I tested a variety of approaches in a 400 person survey and with qualitative research.

I found that gender, particularly in a healthcare context, is highly personal. Participants were excited to see more inclusive option, but many feared the information would be used against them.

I put male on my enrollment form, but my body’s female. Doctors will presume I have a prostate, testicles.
Will this affect the care I receive? I didn’t want to put female either because that’s also not true.
trans male participant
“ I’m worried sometimes that if my doctor and medical providers don’t know my gender assigned at birth I may not receive care that I need. Trans women may need mammograms, but we also don’t have periods and may need prostate exams… I’d like to have something that reassures me that I won’t be denied care when answering such questions for my insurance company.”
— trans female participant

While “gender assigned at birth” was a new phrase for those outside the LGBTQ community and prompted some need for education, Option C, the two-question method was preferred by the majority of participants.

“More flexible but also clearly defines the question… I felt affirmed.”
— female, non-binary participant
“I would love to have my insurance ask me for my gender identity.”
— non-binary/other participant

Also of note was the data integrity results. With Option A, the binary approach, 5 participants’ selections for “sex” did not match their “gender assigned at birth.” Showing that “Sex” is subject to multiple interpretations and can lead to inaccurate data collection.

While in Option C, the Two-question method, 15 participants’ “current gender identity” did not match their “gender assigned at birth.” Without this method, those people would not have been recognized in our dataset.

Step 4: Provide options, context & transparency

When soliciting feedback from experts in the field and our users, some best practices became apparent.

I. Inform users how their information will be used

II. Multi-select dropdowns allow for a multiplicity of identities and recognizes the intersectionality of many experiences.

III. Typeaheads allow for infinite possibilities while retaining structured data.

IV. Other = othering

An “other” option feels like that identity doesn’t belong.

V. Include options upfront & together

Collecting additional gender options separate from the default feels othering; like those identities are less than those society formally recognizes.

VI. Pronouns & names aren’t “preferences.” They’re inherent to one’s identity.

Step 5: Advocate & iterate

To implement this work we formed a collaborative task force consisting of Product, Design, Engineering, Legal & Ops.

We worked to engage key stakeholders across the entire company on the importance of this work. It was resoundingly supported. But sometimes, professional side-eye became necessary in moments of education.

I further validated this work with additional user testing, and it was resoundingly appreciated, both by cis-gender and TGNC participants.

It makes me feel safe that they’re going to understand what to do with my body. I do care about that. That’s part of the uncomfortable-ness, I’m telling you I’m trans, but do you know what that means? I don’t want to have secrets with my health insurance.”
— trans male participant
“Very happy that these are being asked at all — healthcare will gender me correctly; not misgendered me on the phone.”
— non-binary participant
“I have a relative who is gender non-conforming so I can see this being important… I think it’s important to have on there.
— cis gender participant
“I appreciate that you value other genders. I would pick you over other companies because of that.
— cis gender participant

In health insurance, gender touches every aspect of our user experience: from members’ enrollment experiences and 3rd party data integrations. It will take at least a year to implement more inclusive gender collection methods into over 10 team’s roadmaps. But to us, it’ll all be worthwhile as we ensure product is welcoming to all our members.

We need a diverse workplace so that we can build products that reflect the real world.

I myself am not TGNC. But as a queer person, and as a designer, it’s my responsibility to think about the experiences of all my users.

Having TGNC colleagues I could collaborate with was intrinsic to the success of this work. And speaking to TGNC Oscar members at every step of the work was essential to our product development process. This work would not have been possible without those who graciously shared their vulnerable and personal stories.

Our products aren’t perfect yet, but with the contributions of those in this room, we’ll bring people out of the margins of edge cases.

References & resources