"I welcome the attention by the media and by many well-intentioned organizations to really point the spotlight at it, to push people to ask the hard questions about what's being done," said Dr. Christina Ghaly, the director of Los Angeles County Department of Health Services.
[Ads /]
As a result of an aggressive expansion of the statewide testing capacity, Ghaly and her team have managed to at least narrow some of these disparities.
Using data provided by the health care navigation company Castlight Health (the same data that Google Maps and Apple Maps use to surface COVID-19 testing sites), this analysis showed sites in the poorer urban centers of Los Angeles were estimated to have greater community need than sites in the relatively wealthier suburbs outside the city, as well as in Pasadena, though to a lesser degree.
Since California Gov. Gavin Newsom announced an increase in testing capacity back in late April, the state has also significantly ramped up testing in underserved communities, setting up close to 100 state-supported community sites in its Black and brown neighborhoods and lower-income neighborhoods.
That expansion appears to have helped, according to our analysis of testing data from the Los Angeles County Department of Public Health. Testing in Los Angeles County was initially far more unequal, but seems to be getting better.
At the end of April, mostly white cities and communities in Los Angeles County had, on average, more than 1,200 tests completed per 100,000 people. That was about 65% more completed tests than communities of color.
[Ads /]
After Los Angeles Mayor Eric Garcetti announced free COVID-19 testing for everyone in the county, the number of tests administered tripled.
A month later, the racial testing gap had shrunk to just about 25% between the most-white areas and the least white. About 6,200 tests per 100,000 people were completed in whiter areas and about 5,000 tests per 100,000 people in less white areas.
There was a similar pattern in wealthier and poorer areas. At the end of April, people were tested at a rate 60% higher in areas where the median household income was in the top quarter versus the bottom quarter. But by May 31, that gap had shrunk to less than 15%.
Ghaly said if she had to put her finger on one factor that has helped close this gap, it would be the "willingness of government at many levels."
"The Board of Supervisors has been incredibly supportive in using its funding available within the county to fund continued access at testing sites," she said.
[Ads /]
Since April, California has scaled up its statewide testing capacity from just 2,000 tests per day to more than 100,000 now, and has completed nearly 4.8 million tests since the pandemic began, the state's health department spokesperson told ABC News and FiveThirtyEight.
But what went into choosing those initial testing sites?
"It's a good question," Ghaly said. "I think at the beginning when there's not data to drive the decision, you end up citing things based on just logistical and feasibility factors. Individuals or cities or communities that are coming forward and offering insights for testing. Partners where there's already an established relationship, and there's an easy and quick ability to be able to set up the infrastructure for testing."
As cases and hospitalizations hit new records in the past couple weeks, however, the state is again struggling with testing capacity. The governor late last week issued a new advisory to hospitals and labs, asking them to prioritize testing turnaround for individuals who are most at risk of spreading the virus to others.
Ghaly said in LA County they are trying to learn from the data and "use it to do better in the future."
"This should be a constant process that any organization, but certainly government, goes through to look at and be critical...so that there can be progress."
Contributing: The nationwide data analysis and additional reporting for this story was contributed by a team, including Soo Rin Kim and Matthew Vann of ABC News and Laura Bronner of FiveThirtyEight. The full analysis is here and an explanation of how we did it can be found here.