A key metric used by Gov. Larry Hogan as an indicator of how the state is doing in its fight against the COVID-19 pandemic is coming under scrutiny.
Lawmakers say they want to better understand how the Maryland Department of Health calculates the much-reported positivity rate. They also want to know why it is so much lower than other sources, such as Johns Hopkins, that use state-reported numbers in their own calculations.
Acting Maryland Deputy Health Secretary Jinlene Chan acknowledged the state uses a different method to arrive at the rate but said officials have used a consistent approach since the the first cases were reported in Maryland.
“There is consistency,” said Del. Shane Pendergrass. “You consistently don’t add up on a daily basis to the number of tests, and in part that can explain why the Hopkins data is different.”
The state health department publicly reports online the positivity rate daily along with the number of positive and negative tests and the total number of tests.
Getting those numbers to fit together so that calculations can be independently verified is often a frustrating exercise. Often, the number of positive and negative test results do not add up to the daily number of tests reported
Pendergrass said neither she nor her husband, a mathematician, has been able to figure out how the Health Department arrives at the number it reports daily.
Maryland uses the number of total positive cases reported on a given day and divides that by the total number tests reported by that day — regardless of when the test was given. Pendergrass said the numbers on the site — positive and negative tests — do not equal the total number of tests.
“It seems cryptic,” said Pendergrass.
Chan agreed that the number of positive and negative cases should equal total tests reported, even though Pendergrass said “they’re not.”
“I’ll go back and do some of the math. There may be some data reconciliation issues as well and I believe there is some element of looking, there may be some tests that are not, I don’t know if they are indeterminate or there may be other matters,” Chan said.
The positivity rate — the percentage of people who test positive for the virus in a given period — is one of a number of key metrics used to track how well a county or state is doing in keeping the virus in check and preventing health care systems from being overwhelmed.
Sen. Clarence Lam, D-Howard and Baltimore counties, called on Chan and the department to release the specific methodology and formula used to determine the rates it publicly reports.
“There are a lot of discrepancies and questions here,” said Lam, who is also a physician and researcher at the Johns Hopkins Bloomberg School of Public Health.
The CDC recommends a positivity rate of below 5% for two weeks before states should begin easing restrictions.
And while Chan and others stress that no one metric is perfect, Hogan and health officials rely on the positivity rate and tout that figure when it drops. In the last week, Hogan himself has noted the record low numbers.
An increase above 5% could raise concerns and, if high enough, could put Hogan in a politically difficult position where he might have to reimpose tighter restrictions at a time when there is “pandemic fatigue” and the local and national economies struggle to regain their footing.
With a statewide reopening on pause, those decisions to impose additional restrictions fall not on the governor but on county and city leaders of jurisdictions whose numbers might be worse than the state.
On Wednesday, the state reported the seven-day rolling rate was 3.61%, down from more than 5% on June 24. In the last week, the state has reported three new low rates. Hopkins reported the rate at 4.78% on June 24.
Meanwhile, Johns Hopkins reported the positivity rate for Wednesday at 5.2%. Hopkins also uses a seven-day average but filters out multiple tests by an individual.
Covidactnow.org, a site noted by a number of local health officers, reported on Wednesday that Maryland’s seven-day average was 5.3%. It’s website lays out the specific formula that divides the total number of positive test results over seven days by the total number of positive and negative results for the same period.
Senate President Bill Ferguson said he once used the state website as his primary source but now uses Covidactnow.org. The site pulls its data from the state for its own calculations.
Ferguson told Chan the department needs to clear up why the positivity rates are different on the state site.
“Clarity on this question I think is really crucial as we move forward.”
Chan said a lack of uniformity on how terms are defined and reported out could be the reason that different sites, including Johns Hopkins, Covidactnow.org and Harvard, are reporting different rates, though she said many of the sites base their calculations on statistics reported out by the health department.
“One of the things we would like is really a common definition that we could then use to compare apples to apples. I think that there are a number of different sites showing different numbers but again, it does not make them wrong but it doesn’t necessarily make ours,” Chan said trailing off. “We picked a methodology that we believe is the most correct and what is the most stable for us.”