From contact tracing to communicating with the public, the COVID-19 pandemic presents ethical challenges to data professionals. Dr. Nick Tsinoremas, IDSC Founding Director, moderated a July 23, 2020 online discussion on “Data Ethics during COVID-19” with Brian Lee, Senior Advisor, Informatics, Centers for Disease Control and Prevention (CDC), Office of the Chief Information Officer; Kenneth Goodman, PhD, Founder and Director, UM Institute for Bioethics and Health Policy, and Director, Data Ethics and Society, IDSC; and Soyeon Ahn, PhD, Chair and Professor, Department of Educational and Psychological Studies (EPS) in the School of Education & Human Development.
“Good science depends on producing, understanding, analyzing and communicating data, and then vetting it with other scientists,” said Nick Tsinoremas, PhD, IDSC Foudning Director. “With COVID-19, the timeline is very compressed, creating both practical and ethical challenges for data scientists at every step of the process.”
More than 200 faculty, students and professionals attended the virtual discussion, which was hosted by IDSC, the UM Institute for Bioethics and Health Policy, and UM Ethics Programs. Michael Mannino, Ph.D., director of programs, IDSC, organized the event.
“With COVID-19, the more information we have, the more lives we can save,” said Goodman. “But we are not doing a very good job of gathering essential data, or analyzing it in a way that engenders trust with the public.”
Collecting and Analyzing Data
Tsinoremas said ethical issues come into play when collecting COVID-19 data, such as different reporting systems used by cities, counties and states.
“It is very challenging to have multiple sources of data from populations with different attributes,” said Ahn, who leads the UM Research, Evaluation, and Measurement program. “Another issue is that the people collecting data may not be applying the same standards or understand the consequences of their processes. That increases the difficulty of harmonizing all this information.”
From the CDC’s perspective, Lee said that variety of data feeds is one of the public health organization’s biggest challenges. “Our colleagues are working on validating the collection processes and data fields so we can have ‘apples to apples’ comparisons,” he said. “This certainly is not as simple as analyzing one set of data collected over a decade.”
Lee said the CDC’s professionals are addressing the issue of determining what data is most important and useful for an intended audience. “We want to produce information that can be easily understood by the public, while providing information about the lineage of the data for public health professionals,” he said. “We are working with stakeholder groups to determine which information should be given the most priority.”
Once the data has been gathered, writing the analytical software or preparing a database needs to be done with veracity, transparency, and accountability, said Goodman. That’s particularly important as COVID-19 as different types of data become available.
Ahn said that analysis of evolving COVID-19 information may require bringing in professionals with different data skill sets, such as what to add as variables and how to process the content.
Lee said much of the CDC’s work goes into designing and developing transparent data sets. “We want anyone in the world to be able to see how the datasets have been produced, linked and cleaned,” he said. “As sources change or fields are added, we need to have a clear audit trail – not just to confirm the validity but to gain the extra meanings from the data.”
Goodman said the CDC has been a “trusted broker” for providing information during the pandemic. “The CDC has gotten things right, although it has been vastly under-resourced,” he said. “Our public health departments are also doing a good job with their local data.”
Back in 1854, John Snow tried to understand a cholera outbreak in London by tracing individual cases. He put dots on a map where people were getting sick and was able to identify contaminated wells in the neighborhood. “Contact tracing can be a great public health tool,” said Goodman.
But ethical issues also come into play said Tsinoremas, adding “How do we balance the desire for personal privacy with protecting the public?”
One approach is for public health agencies to set clear ground rules on how COVID-19 data will be collected and used, said Goodman. “Trust is vital for managing the process,” he said. “However, many people feel their mobile devices are under surveillance for marketing purposes. The big tech companies need to address this issue in order to help mitigate worries about a robust contact tracing process.”
Communicating With The Public
Another daunting ethical issue for public health professionals is communicating COVID-19 findings to the public, said Ahn. “We need to present the data so it can easily be understood and interpreted.” For example, a data chart should not be compressed horizontally or vertically to highlight a COVID-19 finding or trend.
Tsinoremas pointed out that many people are looking for simple, binary answers to questions like, “Should I get a COVID-19 test or not?”
Responding to that question, Goodman said it is very difficult to communicate scientific findings when they are expressed as probabilities. It’s even more difficult in the contentious U.S. political climate. As he said, “We have done a terrible job in conveying the importance of simply wearing a mask. It is against that background that we need to think about how we can best communicate the importance of public health measures to the public.”
by Richard Westlund