For the past several months, the Texas Department of State Health Services (DSHS) has kept track of coronavirus statistics across the state, including the number of COVID-19 related deaths.
The number of new deaths reported each day is often looked at as a key trend to show how the virus is spreading.
But big data, even in 2020, is not without its flaws.
More transparent COVID-19 data that is regularly released by the Houston Health Department includes not only the number of new deaths reported each day, but also the date for each death — as well as other important factors, such as each individual’s age, race, gender, and notes of other underlying health conditions.
When comparing the dates of death to the date the deaths are reported, there is a clear discrepancy.
For example, on June 15, the Houston Health Department reported seven new deaths. Only one was from June, and it had happened nine days before. Four were from May, and the other two dated back to April.
Scott Packard, the chief communications and public affairs officer for the local health department, told The Texan that the department reports the deaths on the day after they “receive and verify the information.”
Packard noted that delays in reporting are often due to “hospital, family, and funeral home paperwork processing.”
The department has stated that “when a hospitalized patient dies, the hospital should report it to the health department. When this does not occur, the health department may not learn of the death until the death certificate is filed, sometimes weeks later.”
Additionally, Packard said that “if a person dies outside of a hospital and does not have a positive COVID-19 test result, the medical examiner may order post-mortem COVID-19 testing if there is not an obvious cause of death.”
Lara Anton, press officer for DSHS, told The Texan about another potential cause for delays.
“Sometimes a death will occur in a jurisdiction other than where the person lived (i.e., they were in a hospital in a larger city), and the case information has to be shared with the correct jurisdiction,” said Anton.
Whatever the cause, the delays in reporting present an obvious problem: trends based on department reports of deaths will lead to low early numbers and inflated ones later on.
But since the Houston Health Department releases the date of death when they report new numbers, past dates can be updated to reflect more accurate trends.
In doing so, the number of daily deaths clearly climbed to a high in April and then declined in May with a smaller spike at the end of the month.
Of course, given the delays in reporting, deaths in June could prove to be higher than they currently seem.
But while the Houston Health Department releases information on the date of death, DSHS does not.
Anton said that the state health department’s COVID-19 dashboard “reflects newly reported fatalities and they are added to the count on the day that they are reported to DSHS.”
According to Anton, coronavirus deaths are “supposed to be reported to the local health department within 24 hours but it may be a few days between when it is reported to the local health department and when it is reflected on DSHS’s dashboard due to additional case investigation done after a death.”
DSHS releases data by county, so comparing state-reported numbers to Houston’s must be done by looking at Harris County — Houston County’s population of 23,000 is a hundred times smaller than Texas’ largest city of 2.3 million.
As of June 18, Harris County reports that in total there have been 305 deaths in the county, with 175 occurring in Houston.
DSHS appears to be one day behind the local health departments in reporting new deaths. For instance, Harris County reported 284 deaths on June 15 and DSHS data shows 284 deaths in the county on June 16.
The DSHS increase of nine new deaths in Harris County on June 16 aligns with the earlier mentioned seven new deaths reported by Houston on June 15.
But since the state health department does not include the date of death with the numbers they release, the trendline for COVID-19 deaths paints a starkly different picture than the data published by the Houston Health Department.
The steady climb of daily deaths throughout April is replaced with a trough and higher numbers in the first halves of May and June.
While the deaths in Harris County are only a fraction of the statewide deaths, they make up one of the largest fractions.
Of the 2,105 total deaths in the state, 298 are reported by DSHS to be from Harris County, and on June 16, just under a fifth of the new deaths were in Harris County.
Reporting those deaths as new when at least seven did not even occur within the previous week casts significant doubt on the reliability of the data as a means for the public and government officials to monitor trends in the spread of the virus.
Although it is possible that hospitals in other cities and counties have been able to report information on deaths more quickly than in Harris County, that can only be assumed when local health departments do not include the date of death in their reporting.
Information on the precise time of deaths might not come too soon, either.
“Epidemiologists will match reported deaths with death certificate data in the future but death certificate data is not available for months after a person dies,” said Anton. “This matching process is a standard practice with all infectious disease investigations and is usually done at the end of a year.”
Until then, the trend of deaths as reported by DSHS can only be trusted with a large degree of uncertainty.
The current number of coronavirus cases and hospitalizations in Texas is trending upward, but Governor Greg Abbott expressed optimism that the state is well prepared to handle any potential surges in the coming days.
Disclosure: Unlike almost every other media outlet, The Texan is not beholden to any special interests, does not apply for any type of state or federal funding, and relies exclusively on its readers for financial support. If you’d like to become one of the people we’re financially accountable to, click here to subscribe.
Daniel Friend is a reporter for The Texan. He participated in a Great Books program at Azusa Pacific University and graduated in 2019 with a degree in Political Science. He has studied C.S. Lewis’s science fiction trilogy and in his spare time you might find him writing his own novel partly inspired by the series.