This story originally appeared on Spotlight PA.
Spotlight PA is an independent, non-partisan newsroom powered by The Philadelphia Inquirer in partnership with PennLive/The Patriot-News, TribLIVE/Pittsburgh Tribune-Review, and WITF Public Media. Sign up for our free newsletters.
In early February, more than a month into Pennsylvania’s rocky vaccine rollout, state officials were once again facing questions about missing data on race and ethnicity.
The Department of Health had failed to enforce its order requiring providers to collect and submit the information, crucial to understanding if shots were being equitably distributed. UPMC — one of the state’s biggest health systems — admitted it wasn’t doing so at the time.
Acting Health Secretary Alison Beam vowed to take action, warning providers that their vaccine allocations could be reduced or suspended if they failed to follow the order.
“It is essential that in the process of getting vaccine doses administered as quickly as possible, we are still able to collect vital data in terms of the number of doses administered and who is getting them,” Beam said. “We also need to collect race and ethnicity data to ensure that vulnerable communities and communities of color are being vaccinated.”
Time and again — from testing to reporting confirmed cases, and now vaccine administration — Pennsylvania health authorities have taken a reactive approach to collecting race and ethnicity data, rather than ensuring and enforcing collection from the start.
As a consequence, public officials from the state on down have been left to make crucial decisions along the way — such as where to site a clinic, where to send more testing, and where to send more personnel — without a complete picture of how the pandemic is affecting all people.
During the first six, critical months of the pandemic, Department of Health employees modestly improved the collection of race data during case investigations. Between April and September, state employees went from collecting it in just 40% of case investigations to roughly 62%, according to reports obtained by Spotlight PA.
Some local and county health departments in Pennsylvania, meanwhile, averaged an over 70% success rate during that entire time period, while York City managed to improve its collection from 15.5% in April to 72% a few months later.
Data is similarly lacking for testing, which, like vaccine distribution, is primarily in the hands of outside providers. A recent report the department is required to submit to the legislature showed race data was unknown for around 39% of positive tests since the start of the pandemic.
And while the Department of Health requires providers to submit this information — and has the power to punish those that don’t — officials have not issued any penalties or fines.
Race and ethnicity data is key to informing vaccine outreach efforts to communities disproportionately affected by the virus, local health officials told Spotlight PA. Black Pennsylvanians have died at a higher rate from COVID-19 than white ones, according to an analysis of U.S. Centers for Disease Control and Prevention data from APM Research Lab.
“It’s important to be as accurate as you can because that helps us to do targeted outreach and education to populations that may need to get more education about either the testing or the results,” said Barbara Kovacs, director of the York City Bureau of Health.
State officials have promised an “equitable” approach to vaccine distribution, even though data that could illuminate who is most at-risk for contracting the coronavirus is still incomplete.
In February, the state entered into an $11.6 million contract with Boston Consulting Group to handle, among other things, “data management” and “smart vaccine allocation,” according to a statement of work.
Consultant teams, some of which are paid a minimum of $110,000 each week, are tasked with developing criteria and a process to score vaccine providers based on elements including “equity considerations” so the department can make its weekly allocation decisions. The group is also charged with ensuring “equity is considered in all communications materials.”
Work on vaccine hesitancy and maximizing demand within communities that could benefit the most is not planned to start until March 21.
“This is not only a public relations issue,” Boston Consulting Group wrote in the document. “It is a very real health equity issue.”
Pennsylvania began publicly releasing race and ethnicity data on coronavirus cases and deaths in mid-April 2020, weeks after other states and municipalities.
What it revealed: Black Pennsylvanians at the time accounted for a disproportionate number of coronavirus cases across the state, though drawing strong conclusions was impossible because data was missing for almost two-thirds of cases.
But a data quality report from that time compiled by the Pennsylvania Department of Health — and obtained by Spotlight PA through a Right-to-Know request — shows that some local and county public health workers were able to gather this information in twice or three times as many cases as the state’s case investigators.
That’s a trend that continued through October, the last time the state ran the report.
Just six of the state’s 67 counties, as well as four local municipalities, have their own health departments. According to state data, around 54% of Pennsylvania’s population is not covered by one of those agencies.
The data quality reports split the Department of Health’s work into six regions, and case investigators in some areas had better success over the six-month period examined by Spotlight PA. While workers in the northwest and northcentral parts of the state collected race data in at least 76% of cases in September, workers in the southeast region did so in just 50% of cases.
That mirrors the experiences of county departments in the Philadelphia region. Departments in Bucks and Chester Counties collected race data in 39% and 46% of cases during that same month, respectively, while Montgomery County managed to do so in 55% of cases.
Philadelphia was not included in the state reports, and the city’s own data shows race data is missing in a quarter of cases.
While the reports provided a valuable look at which areas of the state were struggling to collect race data — something not available publicly — the department stopped running them in October. They “were unnecessary and redundant” since increased case counts during the fall spike were making it difficult to investigate all cases in a timely manner, said health department spokesperson Barry Ciccocioppo.
Ciccocioppo said the state’s case investigators can only contact so many people who test positive, and that some people are simply not willing to give this information.
“While balancing the case investigation and duties of collecting the case demographics, the case may only be willing to stay on the line for a given amount of time not giving the public health professional the opportunity to ask for this information,” Ciccocioppo said.
Testing data also remains an issue for the state, though one not directly in its control, as the vast majority of testing is done by outside providers like hospitals, pharmacies, and labs.
Race data is listed as unknown for close to 40% of those who have tested positive, a recent report from the health department to state lawmakers showed. And for the vast majority of positive cases — almost 70% — ethnicity information is not known.
Health department officials alerted those administering COVID-19 tests of the need to collect race and ethnicity data — as well as other patient information like address and date of birth — in April 2020. Still, “for a significant period of time, doctors were not filling out demographic data on the lab submission forms, which we require them to do,” Ciccocioppo said.
The department has stressed education, not penalties, to make improvements, Ciccocioppo said, pointing to a new online training on collecting demographic data.
As a result, “positive cases with race reported are now more than 60%, up from just 40% in spring 2020,” Ciccocioppo said. “We also are also able to report that positive cases with ethnicity reported are now more than 45%, up from just 30% in fall 2020.”
Still, a December alert sent by the department to providers — once again asking them to submit race and ethnicity data — shows progress has been slow.
“Among almost 5 million COVID-related reports submitted by labs and providers to [a state system] in the past 90 days, 42% were missing patient race, and 76% were missing ethnicity information,” the alert stated. “This information is necessary to understand health disparities across the commonwealth and to guide resource needs and allocation.”
As the state struggled last year, some county and local departments managed to collect demographic data in significantly more cases.
Health department officials in these areas told Spotlight PA that they dedicated additional staff resources to contact tracing and data collection efforts and followed up on missing information during case investigations.
Allegheny County averaged a 79% completion rate for race data from April to September and a 65% completion rate for ethnicity data. The rest of the southwest region, overseen by the state, reported an average 52% completion rate for race data from April to September and 45% for ethnicity.
Completion rates were even better in Erie County, where an average of 93% of cases included complete race data from April to September and an average of 75% of cases each month included ethnicity data.
And in York City, an average of 54% of case investigations completed between April and September included information on the contact’s race. York improved dramatically in its collection of demographic information over the period: In April, public health workers there collected race data for only around 16% of cases there, state reports show, increasing to 72% percent by September.
“We did work really hard to get that information,” said Kovacs, of York City. Like their counterparts at local health departments elsewhere in the state, York officials had to get strategic to collect demographic information not captured during coronavirus testing.
Erie added close to 60 staff members to work on contact tracing and trained contact tracers and case investigators to better collect demographic information, said Laura Luther, health equity supervisor at the Erie County Department of Health. The enhanced data was used to identify communities that would benefit from expanded testing.
“It’s incredibly important for us, pandemic or not, to make any decision data-driven,” Luther said. “For us to do any work we really want to ensure that the work is needed and really identify the data that’s available out there. We’re trying to collect and create that data ourselves to make sure that what we’re doing is actually what the community has identified.”
The Allegheny County health department started cross-referencing coronavirus case information with data already housed in a countywide system that streamlines information about people enrolled in county services, including aging resources, mental health and housing services, and public schools. If a person was included in that database, the county health department was able to add any missing demographic information to their coronavirus test data.
In York, having staff members who could communicate with Spanish-speaking callers was critical, Kovacs said. Between 35% and 40% of York residents speak Spanish, she said.
Still, there are challenges to this work.
“It’s hard to collect information,” Kovacs said. “When we do our call we ask people to self-identify what race and ethnicity they are. Some are willing to share, some aren’t. So we do have those unknowns, but I don’t think it’s a hard thing to ask people. Sometimes they don’t want to divulge that because they don’t trust us.”
Former Spotlight PA reporter Aneri Pattani contributed.