University of Texas Southwestern Medical Center researchers think part of the answer to the crisis is in using predictive modeling to analyze and quantify the effectiveness of mitigation strategies against the coronavirus.
There was a time when scientists couldn’t track hurricanes or predict their severity. Then increasingly sophisticated technology enabled meteorologists to make data-based conclusions about the conditions that produce violent weather and track the paths of storms. Lives were saved.
The coronavirus poses a similar challenge: the inability to know the virus’s next outbreak until dozens, if not thousands, of people are at risk. UT Southwestern Medical Center researchers think part of the answer is in predictive modeling as a tool to analyze and quantify the effectiveness of mitigation strategies against the coronavirus.
UT Southwestern researchers are about to put that theory to the test, making publicly available on the institution’s website what they have learned from predictive modeling about the virus and how our behaviors affect its spread. The goal is to keep the region, especially policymakers, ahead of the virus.
As businesses reopen, North Texans need innovative, critical insight about what the future might hold. And right now, North Texans have more questions than answers.
How is the virus spreading? Do we need to tighten social distancing guidelines? Are Texans adhering to or ignoring basic safety guidelines like washing hands, wearing masks and distancing. Can certain businesses and large entertainment venues reopen safely? Are we doing our best to keep each other safe?
The answers to these questions can be the difference between a safe reopening or one with a spike in infections. And that’s where UT Southwestern enters this picture.
The medical center won’t track people or the disease, at least not directly. Instead, researchers will crunch data about viral infection rates, personal behavior and other relevant factors to predict worst-case, best-case and most likely scenarios based on the best available data and scientific assumptions. If successful, researchers might be able to determine the effectiveness of intervention strategies, and, like meteorologists who predict weather patterns, provide early warnings about emerging patterns of infectious transmissions in North Texas.
That’s a missing part of the effort to increase public safety. Diagnostic testing and contact tracing are designed to break the chain of transmission by identifying people who should be isolated after having had contact with someone who has tested positive. But modeling can take this to another level — to better identify dangerous conditions and trends before they cause a spike in new infections.
“We might be able to predict the next hot spot,” said Trish Perl, chief of infectious diseases at UT Southwestern. And that early warning might help policymakers to know how to increase resources, such as where to locate a testing van, instead of “flying blind,” Perl said. And this might just touch the surface of possibilities. “There are still a lot of applications of modeling that can be explored in the context of public health and infectious diseases,” she said.
As a tool for analyzing patterns of infectious diseases, predictive modeling is a relatively new work in progress. It is better known as a tool that companies and governments have used to project the outcome of military clashes, the winners of sporting events, forecast corporate earnings among other applications. Much of the modeling of the coronavirus has focused on forecasting demands on the health system. And often that meant turning to patterns from previous outbreaks and observations to determine a mitigation strategy.
But this coronavirus is novel — less contagious than measles but more deadly because no vaccine exists and the incubation period of this coronavirus can be up to two weeks. An infected person can have multiple contacts before they show symptoms, and each of those individuals could infect dozens more. So on the advice of medical experts, public officials encouraged social distancing, issued stay-at-home orders and closely monitored hospitalization trends — the best known metrics to prevent infections from overwhelming health care systems.
UT Southwestern researchers noticed a major gap in the analysis of coronavirus data. On a macro level, state and federal statistics focused on taking inventory of infections, deaths and recoveries, as well as the availability of ICU beds, personal protective equipment and ventilators.
However, the coronavirus’ impact in a region, county, city or neighborhood isn’t always apparent in the broad data. By looking at granular, nuanced data and projecting forward, UT Southwestern hopes its assessments will help policymakers devise more effective responses.
UT Southwestern reminds us that predictive models aren’t crystal balls and can’t game out every possible scenario. Ultimately, the utility of modeling depends on the quality of data and the accuracy of a wide range of variables and assumptions. And that is difficult since much is not known about the virus, such as the distance the virus spreads from a sneeze or cough, or how long can it live on various surfaces or linger in the air.
UT Southwestern is on the right mission at the right time. As we seek a more normal existence, predictive modeling could help businesses and political leaders protect us. The best foundation for North Texans to plan the next steps forward will come from better understanding the impact of specific policies and actions, answers that UT Southwestern’s work could provide.
©2020 The Dallas Morning News. Distributed by Tribune Content Agency, LLC.
Looking for the latest gov tech news as it happens? Subscribe to GT newsletters.