ADPH in Ongoing Battle to Make Sure COVID-19 Test Reports Are Clear, Timely

A healthcare worker administers a swab test for COVID-19. Source: Wiki Commons, U.S. Marine Corps photo by Cpl. Sarah Marshall

Elected officials, business leaders, school systems and news media — among others — keep a close eye on the numbers of Alabama COVID-19 cases every day, with a wary watch on trends the data might indicate.

When those numbers suddenly spike on a single day, as they sometimes have in the past couple of weeks, many people may scratch their heads in confusion or fear.

The explanation in almost all cases is one of bureaucracy. In not-so-technical terms, it’s a “data dump.”

The great majority of results of tests given by hospitals and county health departments, and those processed in larger quantities by laboratories, go through the system swiftly and almost untouched by human hands. The automated process feeds the number of tests processed and the count of positive results into a computer, which then sends that data to the Alabama Department of Public Health in Montgomery. ADPH then compiles the totals and feeds them to its COVID-19 website and dashboard, updated at about 10 a.m. each day if everything goes right.

But with the onset of the pandemic, there are many medical establishments that have conducted public COVID-19 testing but weren’t aware that they were required to report all test results to the state. The majority of these are small operations such as storefront urgent-care clinics, especially independent ones that are locally owned and have just one or two locations. For those healthcare providers, their day-to-day practice is taking care of back pain, strained muscles, sports injuries and flu shots. Those providers have never dealt with a mass outbreak of a communicable disease.

“As new or different entities are coming on board, or perhaps places that have never had a reason to previously report notifiable diseases because they didn’t do a lot of communicable disease-type work — there are still entities that are unaware of the requirement to report,” said Dr. Karen Landers, assistant health officer for ADPH.

Finding providers that haven’t been reporting cases can be tricky, but it happens in many ways. Often, it’s linked to a report that ADPH receives about a new case that lists an entity the department didn’t know about. “Through our ongoing education, some entities realize that they should be reporting when they haven’t previously,” Landers said.

When the staff at ADPH discovers a test result coming from a provider that isn’t in the department’s automated system, officials contact that provider to find out how many other tests it has conducted, both positive and negative.

“We ask for a historic file, and part of that file is any backlog of any reports they have. It is an ongoing process, as we have many, many entities reporting to us,” Landers said, adding that labs in other states that process tests for health care providers in Alabama are part of that group.

Backlogged Cases Are Added to the Totals

Those backlogged tests must then be put into computerized form, even something as simple as an Excel spreadsheet that can create a data file ADPH can use.

In most cases, such small operators have only a handful of unreported test results, and ADPH can simply slip them into the normal daily flow with little effect on trends. But if a provider has been testing for weeks, or even months, that can potentially mean hundreds of cases to add to the database — or in the latest instance, thousands — on a single day.

On October 23-24, ADPH reported 3,747 positive tests that dated back several weeks. In a period when the state agency was typically reporting around 1,100 new positives a day, the big jump was noticed by news media that report the daily results, especially those that also compute moving averages of the new positive results reported in the previous 7-day and 14-day periods, a statistical process commonly used to smooth out data which fluctuates greatly over time.

CDC’s COVID-19 diagnostic panel. Source: Wiki Commons,

Birmingham Watch started using moving averages in weekly analyses of the mounting number of COVID-19 reports (as well as deaths attributed to the virus) as early as mid-April. In the case of the October data dump, we subtracted the additional cases reported as backlogged, so that our moving averages would reflect current data and not cases that occurred well before the dates in our current averages.

When data dumps involve a larger number of cases, ADPH posts a notice on its COVID-19 online dashboard to explain the anomaly. Such notices are shown on “Tab Zero,” linked with a row of numbered tabs at the bottom of the main dashboard page. That link does not appear when there are no recent notices, but as of Tuesday, the link showed a notice that 846 positive cases from northwest Alabama dating from June through August had been reported, along with 90 more from Covington County from August through October.

‘Data Dumps’ Fuel Suspicions

In any other medical event, such statistical anomalies could easily be explained. But in the current political climate and amid this year’s presidential elections, nearly every aspect of the pandemic is scrutinized by supporters of both President Donald Trump and his opponent, former Vice President Joe Biden. Sudden spikes in any COVID-19 related numbers become fodder for partisan broadcast commentators and social media warriors, looking for any reason to support an agenda or distrust what government agencies may report about the pandemic.

“That’s why we publish the notices on Tab Zero, and also show the current risk indicator on Tab 7A,” Landers said. (The backlogged reports do not affect the current risk indicator.)

A similar problem doesn’t exist with data about hospitalizations and deaths that result in whole or part from COVID-19 because all Alabama hospitals and medical examiners are part of the ADPH reporting process.

Data dumps may continue to be a problem as new testing methods come online, particularly those that can be administered with only minimal training. One such new test, which Landers calls a “credit card test” because it uses a cardboard base about the size of a credit card, uses a nasal swab taken as with the standard PCR test. The sample is then put on a piece of cardboard, and a reagent chemical is applied to the sample, with the result visible within minutes. It’s similar to tests for strep throat or over-the-counter pregnancy tests.

“Many of the urgent care providers have not previously been aware of reporting responsibilities because it’s not something they did most of the time. A lot of the urgent cares do the rapid point-of-care testing, and that presents its own set of problems because it’s not hooked up with an electronic interface to transmit data into the Alabama Department of Public Health system,” Landers said.

As this new test and others like it are developed and made widely available, more testing could be conducted by providers outside the normal automated reporting process, creating the potential for more data dumps as those new providers are discovered by ADPH.

“It’s going to be an ongoing problem,” Landers said. “It will be an ongoing educational issue as long as new entities continue to come on board.”