A1xLLcqAgt0qc2RyMz0y
Veteran
COVID-19 hospital data is a hot mess after feds take control
https://arstechnica.com/science/2020/07/covid-19-hospital-data-is-a-hot-mess-after-feds-take-control
With weird discrepancies and fluctuations, COVID trackers say the data is less useful.
As COVID-19 hospitalizations in the US approach the highest levels seen in the pandemic so far, national efforts to track patients and hospital resources remain in shambles after the federal government abruptly seized control of data collection earlier this month.
The Trump administration issued a directive to hospitals and states July 10, instructing them to stop submitting their daily COVID-19 hospital data to the US Centers for Disease Control and Prevention—which has historically handled such public health data—and instead submit it to a new database in the hands of the Department of Health and Human Services. The change was ostensibly made to streamline federal data collection, which is critical for assessing the state of the pandemic and distributing needed resources, such as personal protective equipment and remdesivir, an antiviral drug shown to shorten COVID-19 recovery times.
Watchdogs and public health experts were immediately aghast by the switch to the HHS database, fearing the data would be manipulated for political reasons or hidden from public view all together. However, the real threat so far has been the administrative chaos. The switch took effect July 15, giving hospitals and states just days to adjust to the new data collection and submission process.
As such, hospitals have been struggling with the new data reporting, which involves reporting more types of data than the CDC’s previous system. Generally, the data includes stats on admissions, discharges, beds and ventilators in use and in reserve, as well as information on patients.
For some hospitals, that data has to be harvested from various sources, such as electronic medical records, lab reports, pharmacy data, and administrative sources. Some larger hospital systems have been working to write new scripts to automate new data mining, while others are relying on staff to compile the data manually into excel spreadsheets, which can take multiple hours each day, according to a report by Healthcare IT News. The task has been particularly onerous for small, rural hospitals and hospitals that are already strained by a crush of COVID-19 patients.
Technical snags
Once the data is collected, hospitals have several options for offering it up to the federal government. They can submit the data directly to the HHS system (called TeleTracking) via an online portal, authorize an IT vendor to submit it to the HHS for them, publish it on their website in a standardized format, or have state officials submit it on their behalf.
Many of these options have proven difficult as well. Some hospitals that have historically reported data directly to their state governments found that their states aren’t yet authorized to submit hospital data to the new HHS database on their behalf. This has left some hospitals, such as those in New Mexico, with the burden of submitting data to both the state and the HHS. For the hospitals who try to submit on their own, some have scrambled to get all the necessary data collected only to face technical problems inputting data into the portal.
Such was the case for some hospitals in Georgia. "All of this is taking the very valuable and precious resources” away from the fight against COVID-19, Anna Adams, vice president of government relations at the Georgia Hospital Association, told Healthcare IT News.
Amid all the administrative and technical hurdles, the national data on hospitalizations has become a hot mess. The COVID Tracking Project—which collects data on a variety of COVID-19 pandemic metrics—wrote in a blog post July 28 that US hospitalization data is no longer reliable.
The blog noted that between July 20 and July 26, federal totals of currently hospitalized patients has been, on average, 24-percent higher than the totals reported by states. On a state-by-state level, some states are reporting fewer cases than the HHS, some are reporting more, and some federal data has significant day-to-day fluctuations not seen before the reporting transition.
Dark data
This may be due to a variety of factors, including double-reporting by hospitals, or hospitals only reporting to the HHS and not their states now. Some numbers of COVID-19 patients may be different because of dueling definitions states and the HHS use to define COVID-19 patients. For instance, some states may not report suspected or probable cases, or those that tested positive for COVID-19 after being admitted to a hospital for something else.
In a July 30 update, the tracking project noted the continued problems, concluding: “Taken together, the gaps and uncertainties in the previously stable hospitalization data mean that this crucial indicator has become much less useful for understanding the true severity of COVID-19 outbreaks.”
Likewise, Dave Dillon, vice president of media and public relations at the Missouri Hospital Association, expressed frustration at the timing of this data collection switch.
“It's worth mentioning that as we moved toward this change we were approaching the number that would have met or exceeded the maximum hospitalization we'd seen during the virus,” Dillon told Healthcare IT News. "We went dark at the same time we were getting close to what our previous peak was. Moving from a known platform that all of the individuals could easily manipulate … has harmed our ability to have that situational awareness.”
According to The COVID Tracking Project, hospitalizations reached a peak of 59,885 on July 23, just shy of the high of 59,940 hospitalizations on April 15. The project reports that the number of hospitalizations has since declined but that the numbers they are reporting are likely undercounts.
Meanwhile, an investigation by NPR noted that there were irregularities in the process used by the Trump administration to grant TeleTracking Technologies the $10.2 million contract to set up the federal database. In particular, the CEO of the Pittsburgh-based company has links to the Trump Organization. Congressional investigators are now looking into the matter.