Without a public health model that focuses on high-risk populations, ‘we will continue to see Covid transmission throughout the city.’


We all learned in middle-school math that averages can lie. The classic example: Bill Gates walks into a crowded bar, and the average wealth of the patrons goes up by several million dollars. Robert Reich likes to say that the average height of him and Shaquille O’Neal is six-foot-2. 

Yet that’s the fuzzy math being deployed to measure how the city is doing in its battle against Covid-19. The city’s seven-day average case rate for the week of Sept. 7 per was 6.8 per 100,000 residents. That’s one of the metrics the state looks at for moving a county through tiers. 

That rate is high – it would put San Francisco in the state’s second-highest risk level.  But break that average down by race, and the numbers tell a catastrophic story for some residents of the city. 

The city’s seven-day average case rate for the same week in September was 2.3 for white residents, 4.71 for Asian Americans, 6.5 for Blacks and 16.4 for Latinx — more than double the city’s “substantial” risk rating by state metrics,  and well beyond the  “widespread” risk category for case rates above 7.   

In a transmissible virus, researchers said, it’s dangerous to allow the impact on segments of the population to be diluted by tossing them in with everyone else. Those harder-to-reach but persistent cases become all the more important to stop the spread of the virus. 

The San Francisco Department of Public Health acknowledges the disparity. “As we continue to invest in more testing resources, those resources will be focused in the southeastern part of the city so we reach the people most at risk,” Dr. Grant Colfax, the director of the department, told Mission Local last week.

That is good news, and Colfax reassured Mission Local about the city’s intent to do more after Mission Local wrote a story last week describing the city’s low testing rates in the communities most impacted. But – as we pointed out in that story – DPH has known about the disparity since at least April, when UCSF and the Latino Task Force did a study in the Mission District, but only recently started to test in the most severely impacted communities. 

The numbers the city publishes to keep the city informed on how we are doing paint an overly optimistic picture. But perhaps worse is that the city’s strategy to rid San Francisco of Covid-19 appears to be less urgent in the communities that it knows are high risk.  That model leaves everyone vulnerable.  

Efforts to get an explanation of the city’s public health model regarding Covid-19 went unanswered — as did repeated requests for more data on testing. 

One researcher who works closely with the city and asked for anonymity said it was simply incomprehensible.  

So, we asked other epidemiologists outside of San Francisco about the data and the strategy. 

Tomi Akinyemiju, a professor in Population Health Studies at Duke University specializing in epidemiologic methods, said that releasing aggregated data tends to obscure data and render it difficult to understand. Instead, she said, demographic factors should be acknowledged.

“When we look back in five to 10 years on the pandemic, one of the top failures we will see is a lack of transparency, especially regarding data,” Akinyemiju said, adding that the problem isn’t unique to San Francisco. 

Specifying data on the ethnic, racial and geographic profiles of residents being tested is not a difficult ask. 

“We know how to collect this data, so it is baffling to me why we are not doing it,” she added. “It is hard to see how this isn’t impacted by politics.”

Test where the cases appear

San Francisco has had a remarkably low death rate, but here too, the racial disparities are apparent: Of the 99 Covid-19 deaths, Asians comprise 31 percent of the deaths, compared to their 34 percent of the population; Latinx, 27 percent, compared to their 15 percent of the population; Whites, 17 percent, compared to their 40 percent of the population; and Blacks, 8 percent, compared to their 5 percent of the population.

Because the death rate has been so low, the city has been able to focus on cases, but testing has not been aimed at the most impacted populations.

Population and Covid-19 Cases by race and ethnicity. Source DPH. Data prepared by Hayden Manseau.

“If you are finding more cases among Latinos, you would be doing more testing among Latinos,” said Ali Mokdad, a professor of Health Metric Sciences and epidemiology at the University of Washington’s  Institute for Health Metrics and Evaluation. “Something isn’t right.” 

Mokdad said it clearly indicates the city’s contact tracing is not working well in the Latino community, because if it was, testing would be higher there.

The city reports that contact tracers are reaching 79 percent of the Covid-positive cases, and 84 percent of their contacts. But again, we don’t know if those numbers represent what is happening in the most impacted communities, because the city does not disclose  that data – data that would make it easier to monitor the city’s progress. 

Last month, the city said it was hiring 30 new bilingual contact tracers, but the city has not gotten back to us on how many have since been hired. 

Meira Epplein, a professor in Population Health Sciences and epidemiology at Duke University, agreed that the strategy should be to test much more in the high-risk communities. And this has been an issue across the country. 

We should be testing as many people as possible, but we should be testing the people who are at risk of the worst outcomes, which are Latinx and African American communities,” she said.

Overall, as we reported earlier, 9 percent of the city’s nearly half million tests have been conducted on Latinx residents. Presumably this has gone up in the last few months, but we don’t know for sure because the Department of Public Health has declined to provide the data. 

“I have no idea why the city isn’t reporting tests by ethnicity,” Epplein said. 

“It makes no sense.”  

DPH has not explained its reasoning. 

Early on, the city could blame the lack of a federal strategy

At the outset of the pandemic, epidemiologists said, cities across the country were distracted and hampered by basic supply issues. Scarce resources tied to worldwide demand and a scattershot approach to medical supplies made testing problematic for every city and San Francisco was no exception. 

Resources are still limited but, as the Health Department likes to point out, San Francisco now tests more than most cities. Most of those tests are done at the city’s SoMa and Embarcadero testing sites, where the positivity rates are at 1 to 2 percent, according to sources. Again, DPH would not provide those numbers. 

The sites are available by appointment only and easily accessible if you have a car. Both are also accessible by public transportation, but that requires extra time — and, while SoMa is situated at the edge of some impacted communities, neither site is particularly convenient to where the vast majority of the most impacted populations live. 

Data map from the SF Covid Tracker shows cases per 10,000 residents dramatically higher in the southeastern part of the city. Source DPH.

Stefan Baral, a Johns Hopkins University epidemiologist, said that testing sites around the country are “set up for those who can take a few hours off from work compared to the people who are more shift workers.” This testing approach disadvantages people who are carrying the greatest burden during lockdown, he added. 

The limitations to that approach became immediately clear on Monday morning at a pop-up testing site at the 16th Street BART station. A woman who gave her name as Alba, who works nearby at the Food Hall at 16th and Valencia streets, immediately ran over. Could she come by on her break?  

Diane Jones, a retired SFGH nurse working at the site, reassured her that no appointment was necessary. But that site, run by UCSF and the Latino Task Force, would be up for only three days. 

The other walk-up site in the Mission is a once-a-week pop-up at 701 Alabama St., known as the Hub, run by the Latino Task Force. It serves 250 people every Thursday. The city’s sites at SoMa and Embarcadero, meanwhile, can test some 3,000 people a day. In comparison, during the week of August 31, the city performed about 1000 tests at eight pop-up sites in high-risk communities.

The lack of focus on low-risk communities has been an issue from the start, epidemiologists said. 

Covid strategies across the country, said Baral of Johns Hopkins, “focus on the protection of the wealthy.” 

Six months into the crisis, he said, we understand a lot more, “and yet we have the same intervention for everyone. When you have something that is driving (infections) and a singular strategy, that just increases disparity.” 

And the disparity is already there: Not only are low-wage workers going to work to keep society running and their own households fed, they return home to crowded living conditions and unwittingly pass the virus on. 

Baral said that the best predictor of someone’s vulnerability to Covid is square footage per household. The more room you have, the less likely you are to go out to work. The less room you have, the more likely you are to work outside the home and return with an infection. 

Household density around the city could have guided the siting of testing locations. The city’s average family household is 3.3 individuals. Twelve of San Francisco’s 40 neighborhoods have larger-than-average households. Tthe Excelsior and Bayview-Hunters Point have the largest household sizes, both at 4.2.  

Interestingly, in terms of Covid-19 cases, Bayview is the city’s most impacted neighborhood with 374 cases per 10,000 people, and 12 of the city’s 99 deaths. The Excelsior is the fifth-most impacted neighborhood, with 225 cases per 10,000 residents and fewer than 10 deaths, according to DPH. 

Of the dozen densely packed neighborhoods, only two are outside the southeast: Treasure Island (average family household: 3.5) and the Presidio (average family household: 3.6), according to the U.S. Census and a report on neighborhood profiles.  

And yet, the city’s two biggest free testing sites are in SoMa (with an average family household of 2.9) and the Embarcadero (with an average family household of 2.5). South of Market ranks ninth in impacted neighborhoods while the Embarcadero — included in the financial district — ranks 19th.

The surge in testing and what it buys San Francisco

Since the early supply issues were remedied sometime in the late spring, San Francisco began testing in earnest and has been hailed as a Covid-testing marvel. 

In fact, since this summer, DPH has been testing so many folks – some 4,000 a day –  and mostly in low-risk populations, that the positivity rate is 2 percent. And even if it is logging in 6.8 cases per 100,000 over a seven-day period, doing all of those tests can create magic for a case rate that the state considers at “substantial” risk for spread.   

Thanks to its stellar testing rates, the state awards bonus points so that San Francisco’s cases per 100,000 have been adjusted from over six – again, a rate that is at the second highest level of concern or at substantial risk –  to 3.9,  a level that state metrics establish as a moderate risk. 

Those bonus points for loads of tests and a positivity rate of only 2 percent essentially move the city closer to reopening.  No matter that epidemiologists say that much of that prodigious testing is focused on the wrong population – a lower-risk population that ensures the city a lower positivity rate.    

The intent behind the city’s sluggish move into the communities where positivity rates are higher is unclear. Motivations are hard to discern from the way in which the city uses data – and keeps much of it under wraps. 

Re-Opening

As long as San Francisco  remains in a partial shutdown, residents in much of the city, where incidents are low, can feel like Covid-19 infection is highly unlikely. And it is. But as more people move around the city as schools, theaters, restaurants and museums open, transmission possibilities increase, epidemiologists said.   

Mokdad of the University of Washington said that if he were a shot-caller in San Francisco, he would move daily, easily accessible testing sites to the impacted communities and run an additional test in those communities to see how many people have been exposed – another indicator that would say if more testing is needed – and broadcast a public health message that is tailored to that community. 

To be sure, the San Francisco Health Department does offer food and a hotel room after someone tests positive, and Latinx residents represent 46 percent of the 2,245 people who have isolated in city hotel rooms – a number commensurate with their rates of infection. 

In July and August, the city also spent $2 million via Supervisor Hillary Ronen’s Right to Recover legislation to replace two weeks of wages so that Covid-positive residents without other resources could get some income. But that money is gone and is no longer being offered. 

Despite the solid citywide average, pandemic fatigue and a sagging economy, San Francisco cannot let its guard down, epidemiologists say. 

“If we are seeing an overrepresentation of cases in one group, the logical next step is to tailor interventions to match that need,” said Akinyemiju.

Akinyemiju envisions what will happen without such tailored interventions: “Cases will continue to spread in those groups, the healthcare system will be taxed because many individuals in the most impacted groups don’t have access to healthcare, and we will continue to see Covid transmission throughout the city.”

Keep coming back and if you haven’t already, support Mission Local’s reporting.