As the nation confronts a reckoning over racial inequality, tech giants — with their vast budgets, institutional power, and growing focus on health — could be well-positioned to make a difference. They could invest in health tech initiatives to close racial gaps in local communities, conduct inclusive research, and create more affordable, accessible products.
But for the most part, that isn’t happening. In fact, while executives from Facebook, Apple, Alphabet, and Microsoft have all condemned racism in public statements, none of the statements made mention of how they plan to address racism in health or health care.
STAT asked spokespeople at five big tech companies working in health — Facebook, Apple, Amazon, Alphabet, and Microsoft — whether they have any health projects dedicated to addressing inequality. Only Alphabet provided concrete examples, including an effort by Google to map U.S. health disparities with the Morehouse School of Medicine and an initiative by Verily to enroll more women and people of color in clinical trials. Spokespeople for Facebook, Apple, Amazon, and Microsoft declined to comment on specific projects, pointing instead to their chief executives’ recent statements.
But by and large, those statements detail plans to address racism without tackling its impact on health.
Apple CEO Tim Cook announced last month that his company was committing $100 million to a new racial equity and justice initiative focused on education and criminal justice reform. Microsoft CEO Satya Nadella, meanwhile, penned a blog post citing the company’s ongoing work on criminal justice reform as evidence of his company’s commitment to addressing inequality.
A Facebook spokeswoman also pointed STAT to a list of the company’s existing health projects, such as a preventive health tool designed to nudge people toward health screenings.
Outside experts cautioned that those approaches run the risk of ignoring the disparities that place Black individuals and other people of color on unequal footing compared with white people. That means some of tech’s health ambitions could — perhaps unintentionally — worsen inequalities, rather than helping to reduce them, they said.
“The industry needs to take greater responsibility for how the tools it makes can contribute to racial inequality,” said Kate Crawford, co-founder and lead researcher at the AI Now Institute at New York University.
Facebook’s preventive health tool, for example, does not collect data on the race or ethnicity of users who engage with the tool. Some experts said that makes it impossible to know whether the tool is reaching members of underrepresented groups.
“It’s just like with Covid-19: We didn’t know there was an issue until we started collecting data on race and ethnicity,” said Adewole Adamson, a dermatologist who co-authored a 2018 paper which concluded that machine learning dermatology research based predominantly on white datasets can lead to worse outcomes for Black individuals and other people of color.
Other problems can arise when tech solutions are developed without a proper accounting of race and ethnicity, according to clinicians and startup executives.
For example, several smartwatches rely on green light to measure heart rate. But because skin with more melanin blocks green light’s shorter wavelength, the tools are less accurate for Black individuals and other people of color. Fitbits, Samsung watches, and several other brands use only green lights; a Fitbit spokesperson said it boosted the current in an effort to address the issue. In addition to green light, the Apple Watch also uses infrared sensors, which appear to make the technology more reliable when used on a variety of skin tones.
The potential for baked-in bias has cropped up in a number of other health tech tools and datasets. At-home genetic tests suffer from an underrepresentation of Black individuals and other people of color in their data, for example, while a commonly used hospital algorithm trained on a biased dataset was found to lead to worse outcomes for Black and Latinx people.
Experts said the problem, in part, is that such tools have been designed without a deep understanding of the ways in which structural racism shapes health. In the case of the hospital algorithm, for example, the system used cost to prioritize patients’ medical needs — which proved to be problematic, because health spending for Black patients was less than for white patients.
But, experts said, tech’s shortcomings on addressing racial disparities in health is also the product of disparities within tech companies themselves. The staff and leadership of many major tech companies remains overwhelmingly white — potentially impacting everything from product design to decisions about what types of health research are worth investment. Experts said that diversifying their own staffs will be a critical first step for tech companies, who need to do better at recruiting and retaining people from underrepresented backgrounds, including Black people and other people of color.
“If your decision-makers all look the same, you’re going to come up with some biased tools,” Adamson said.
Beyond diversifying internally, the tech companies can and should be taking other steps, such as investing in health tech leaders working in underrepresented communities, experts said.
“These are long-standing inequities, and there have been people working on these issues for years at a grassroots level. We need to give them a moment to drive the change,” said Kistein Monkhouse, founder and chief executive officer of digital health startup Patient Orator, which aims to improve health care outcomes through storytelling.
There is evidence that some companies are taking steps toward making their work more inclusive and accessible. Verily’s Covid-19 testing program includes a collaboration with the California Coronavirus Testing Task Force and Team Rubicon that’s aimed at increasing testing in underserved areas.
Employees from Google including Google Health, meanwhile, are involved in an effort to map U.S. health disparities with the Morehouse School of Medicine. And in 2018, Google Brain researchers published a study on machine learning and health equity aimed at helping clinicians and policymakers address some of the issues that Adamson highlighted.
Still, in a more recent example of the work the companies have ahead of them, Verily recently suspended spot bonuses for employees, opting instead to funnel that money toward social justice initiatives. Some staffers were displeased with the change, first reported by Business Insider, noting in a letter to management that the move “implies that these efforts are charity causes not worthy of their own investment.”
For their health efforts to be truly inclusive, tech companies will need to acknowledge the role systemic racism has played in health care, invest in health tech efforts designed at addressing disparities, and include people from underrepresented groups in their own ranks and in the design and rollout of their products.
“If tech is about making the world a better place, you can’t do that selectively,” Adamson said.
This article was published by STAT News on July 7, 2020. You can find the original article here.