Status: Complete / Role: Lead UX Researcher
The momentum of social unrest during the summer of 2020 re: race, in combination with the evident impact of COVID-19 on underrepresented communities fueled this project – taking a deep look at our own product and practices and assess what opportunities we could pursue in response to that.
Qualitative interviews allowed us to identify the motivations and barriers of participants participating in research.
Taking a lead from the motivations - led to more academic research, specifically for race-based research and taking more steps to publicize its existence to our users. A study called the Black Representation in Genomics Research (BRGR) study was submitted to our IRB and approved.
On the flip side, we took a lead from the barriers – specifically about the misunderstanding of company/research practices and lack of transparency and visibility about the work being done.
The Product and Problem
The Health and Ancestry Kit, 23andMe's bread and butter, relies on reference panels from other 23andMe customers to improve its accuracy and granularity. Additional product features, like polygenic-risk scores (PRS) and whether or not someone might think cilantro tastes like soap are also impacted by this because within the product, customers can answer research questions that might lead to new data trends.
This type of analysis benefits greatly when there are more individuals of diverse backgrounds since there's more variance between data points; both the product and the database are very Euro-centric as a result of skewed data points and are generally more applicable for European/White populations.
Participation in research was relatively low in comparison for underrepresented communities in our research program and we knew that it was something to address.
Due to our current Eurocentric-skewed database, the overall accuracy, ethnic/racial granularity, and relevance of our product, failed to satisfy both new and existing customers.
-
What were some of the motivations and barriers of the participants who did participate in research?
-
How can the barriers be mitigated/addressed?
-
What opportunities could we identify to build upon the intrinsic motivations of existing participants?
Research Process
Preparation and Planning
We didn't want to reinvent the wheel in this space because there was already so much existing literature and data on research engagement in URM communities. Using already accomplished-work provided us a benchmark to start thinking about potential frameworks and/or ideas that this research could feed into.
Being on the Research Ethics team meant that we knew about the historic racial injustices towards individuals from said communities, particularly African American men, and how that perspective towards medicine and science continues to exist. Knowing that, we factored a level of sensitivity and care around the questions we asked to mitigate bias and any potential uncomfortable moments in the sessions we hosted.
A few highlights we learned from existing literature:
-
Generations passed the 'discomfort towards medicine and science' to their offspring, indicating a level of familial influence.
-
Individuals from URM communities faced various factors, like access to transportation, demand of time, and legal status, that prevented them from engaging with potential research initiatives, indicating that research participation is (and will remain) a privilege.
-
Relationship-building with URM communities continues to be the most effective strategy in engaging URM populations. Although challenging, building trust is a priority and lacking trust results in a lack of unsustainable engagement and progress.
Methodology and Research Execution
The potential sensitive nature of this topic presented two issues to consider -
-
With it being sensitive, there would be a lot of anecdotal data that would be incredibly valuable and powerful to understand for stakeholder buy-in and potential recommendations.
-
With it being sensitive, there might be a lot of discomfort and/or resistance to being honest about the particular subject that would prevent any opportunity for us to have next steps.
Interviewing felt like the right choice because of its ability to let me build rapport, probe into areas of interest and encourage active participation in concept ideation, but to also read body language (through Zoom...) if it was getting a little uncomfortable for them.
Recruitment was done through a recruiting platform and screening questions targeted ethnic/racial classification. All participants were 23andMe customers since this particular platform allowed us to create a cohort of them in our recruitment efforts. Consent status to participate in research was not required since we wanted perspectives of both those who consented (to understand why they participated) and who didn't consent (to understand why they didn't participate).
Analysis and Recap
It was very noticeable that there was a lot of misconception and confusion about what the research program was since a majority of the participants didn't even realize that we had an ongoing research program. Participants also sometimes interpreted the content wrong (which was on us) and corrective feedback intended to clarify the assumption during the interview reassured users about how their data was being used.
Next Steps
These findings -
- motivated our Consumer Insights team to launch the "Black Health Initiative," a large-scale qual and quant project that focused on understanding the African-American/Black experience in healthcare.
- established an initiative to develop recurring communication materials (monthly newsletter and blog posts) for research participants to provide updates on the research.
- supported the decision to continue conducting in-depth interviews with other demographic groups
Prior to my departure from the company, I finished interviewing a cohort of South Asian individuals and planned on sharing broadly. The goal was to run analyses across each cohort to see if there were overlapping themes across populations that pointed to potential initiatives.
Limitations
These recommendations are meant to holistically address any concerns about the 23andMe research program to increase transparency and visibility into the work being done. Post launch, we would monitor the enrollment rates of those from URM communities for any change.