What is data driven decision making? Two professors explain it

KELOLAND.com Original

FILE – In this Wednesday, March 11, 2020 file photo, a technician prepares COVID-19 coronavirus patient samples for testing at a laboratory in New York’s Long Island. Wide scale testing is a critical part of tracking and containing infectious diseases. But the U.S. effort has been plagued by a series of missteps, including accuracy problems with the test kits the CDC sent to other labs and bureaucratic hurdles that slowed the entrance of large, private sector labs. (AP Photo/John Minchillo)

SIOUX FALLS, S.D. (KELO) — It’s been said often lately: Local and state public leaders and health care officials say decisions made about COVID-19 must be data and science driven.

But what exactly does that mean?

“Data science, statisticians, and mathematicians team with other professionals.  They use data and mathematical and statistical models to make data-driven recommendations to health care and public policy leaders and decision-makers, who then base their decision on this,” said Kurt Cogswell, a math professor at South Dakota State University.

Where does that data come from? In short, it seems like from everywhere.

There’s data from COVID-19 patients and testing, not just from South Dakota but from around the world. There’s data about the number of intensive care unit beds from South Dakota and the U.S. There’s even data from cell phones related to traffic patterns.

But take one set of data from one particular area as an example.

“Much of the data that is being used is coming from the medical records collected from those that are being tested and treated,” said Shawn Chiappetta, a math professor at the University of Sioux Falls.

“Looking at the COVID-19 infection rates, one can take the daily known infections as points on a curve and use techniques to find an approximate function, or rule, that mirrors the data points creating the model,” Chiappetta said.

“With the model in hand, we can then use it to predict what might happen at a future time, ” Shawn Chiappetta

Predictions about what happens in the future can result in public policies such as banning gathering of 10 or more people or shelter in place declarations.

Before a public policy is established, data and public policy options are combined. Known mathematical and statistical models of disease spread are to determine the best public policy and health care decisions by predicting several factors with COVID-19, Cogswell said.

Those factors include predictions about the infection rate over time and how much will certain public policies reduce rates, when will peak rates overwhelm hospital staff and beds, the total number of people infected and whether or not they’ve been tested or show symptoms and best ways to modify pubic behavior, he said.

The city of Sioux Falls, for example, established a ban of gatherings of 250 or more people on city-owned property earlier this month. As more data became available, the city adjusted rules. Mayor Paul TenHaken has frequently said the decisions are driven by data and science.

TenHaken had a news conference today talk about additional COVID-19 measures the city will take. A video from the news conference is included below.

There’s lots of data to be used.

“In the old days, in the 1960s and 1970s, we felt like we had to narrow the data,” Cogswell said.

Cogswell teaches the rapidly developing field of data science. There’s been a change in the thinking of narrowing the data with the development of the data science field, he said.

“…we want all the data we can get our hands on,” Kurt Cogswell

In 2020, there’s better methodology, mathematical calculations and computer software to use more data than ever before and apply it responsibly to issues such as COVID-19, Cogswell said.

“We live in a time where we are ‘data rich,'” Chiappettta said. “In regards to COVID-19, the fact that we have many other sources of data can help fill in the gaps and identify trends that we might not otherwise factor into the models. “

Today’s data includes traffic pattern information that can be gleaned from smartphones and data gathered from social media, Cogswell said.

While those who analyze and use data will find that information useful, there are some hitches, Chiappetta said.

“Currently, there are many questions being debated about those protections using cellphone and traffic patterns for obvious reasons,” Chiappetta said. “It is a tricky time as the data can be useful to give a more complete picture and see directions of possible growth in the infections in an effort to be proactive, but we must simultaneously contend with the privacy issues that are raised with these data sources.”

Data changes frequently and that can change how public officials and health care workers respond, the professors said.

Cogswell said COVID-19 predictions and responses are similar to meteorologists predicting the weather. Sunday’s forecast of an upcoming Thursday storm will use the latest and most complete data available on Sunday, he said. By Tuesday, the meteorologist will have more data to more accurately predict Thursday’s storm.

Specifically, new numbers on infection rates with COVID-19 are coming in daily.

“If we continue thinking about the infection rates, upon getting new data, we rerun the approximation techniques, say each day, to get updated functions that give us adjustments in the curve,” Chiappetta said.

Cogswell said it’s an important time for those working in data science. He’s been telling his students that now, more than ever, they need to really learn because they will have an impact on what happens in the future, Cogswell said.

Data-driven decisions won’t end when COVID-19 ends, but will continue with the decision made during the aftermath, Cogswell said.


Copyright 2020 Nexstar Broadcasting, Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.

Continuing The Conversation

Trending Stories

Don't Miss!

More Don't Miss

Your Guide To
Coronavirus

KELOLAND News is covering the Coronavirus outbreak. We have created a guide to everything you need to know to prepare. We also have the latest stories from across the globe feeding into this page.


 

More Contests