Most Americans still back the use of personal data for public health research, despite the turmoil of 2020 and beyond

A survey conducted in February 2020 found that most Americans strongly supported the use of personally identifiable data for the purpose of preventing the spread of disease, improving safety or fulfilling other standard public health goals. But did their opinions change during the following weeks as social cohesion was eroded by a real-life pandemic, racism protests and political divisiveness?

To find out, the multidisciplinary team from Texas A&M University that sent the first survey on data privacy attitudes and preferences followed up with another survey in November 2020.

“Concern over data privacy has grown in this new era of big data and artificial intelligence, and beginning in 2022 it became a major focus of federal and state legislatures,” said the study’s lead researcher Cason Schmit, JD, assistant professor of health policy at Texas A&M and associate director of the university’s Population Informatics Lab. “Often, new privacy laws do not include provisions that permit the use of data for public health, despite the fact that the U.S. public strongly prefers using identifiable data for socially beneficial uses, like promoting population health, over more commonplace data uses, like profit driven commercial data uses.”

The study, funded by an X-Grant from Texas A&M, was published in The Milbank Quarterly. Others on the research team—all associated with Texas A&M—were Brian Larson, JD, associate professor with the School of Law; Thomas Tanabe, JD; Mahin Ramezani, MS, research data scientist with the Texas A&M Transportation Institute; Qi Zheng, PhD, professor of biostatistics; and Hye-Chung Kum, professor of health policy and management and director of the Population Informatics Lab.

The two surveys had a total of 1,373 responses: 504 in February and 869 in November. The sampling targets for age, gender, race, income, education and census region were generally met for both surveys, and health insurance coverage rates mirrored those published by the U.S. Census Bureau.

Over one-third of all respondents reported having at least one chronic health condition, nearly half had used a health provider within the past year, and around one-fifth had visited an emergency department in the past year. Scores for the Concern For Information Privacy instrument showed no significant changes between February and November 2020. Two subgroups received an in-depth analysis: African Americans and Democrats versus Republicans.

“The main takeaway is that public support for using data for public health purposes remained strong, despite the tumultuous year,” Larson said. “We found just a few statistically significant differences.”

Respondents were most accepting of research by universities/researchers and nonprofits and were least accepting of research by governments and businesses. The use of personally identifiable data was most supported for public health and research purposes, followed by identifying criminal activity, marketing/recruitment and profit-driven activities.

Although the February sample indicated similar general preferences across subgroups, the November sample found much less consensus by African Americans, Democrats and Republicans.

African American participants showed the strongest negative shift in relative preference for using data for research (48.4 percent to −5.8 percent) and promoting public health purposes (35.2 percent to -0.7 percent). Conversely, their relative preference of using data for profit-driven (−54.3 percent to −2.6 percent) and marketing activities (22.6 percent to -10.1 percent) improved. Their relative preference for businesses using data significantly declined (17.9 percent to −13.2 percent), whereas the relative preference for the government using data increased significantly (−27.8 percent to −4.8 percent). Their relative preferences for using government sources of data had a statistically significant increase (−45 percent to 1.3 percent), whereas using health and education data sources both significantly decreased (in health, from 7 percent to −13.3 percent and in education, 13.8 percent to −6.8 percent).

“The steep drop in comfort with public health and research data uses among African Americans during 2020 shows just how fragile trust can be,” Schmit said.

In the November 2020 survey responses, some African Americans expressed distrust of government and business. As one African American noted, citing the Tuskegee Experiment, “I will trust the government having my info when rich and poor are experimented on, until then I will remain skeptical.”

Despite some differences, there were strong similarities between Democrats and Republicans in the November survey. Both expressed the highest levels of comfort with researchers (32.5 percent for Democrats and 28.9 percent for Republicans) and nonprofits (26.5 percent and 16.4 percent) using data. Both expressed the highest comfort using data to promote population health (44.7 percent and 39.7 percent) and scientific research (43.6 percent and 33.5 percent). Both expressed the most comfort with the use of identifiable education records (11.6 percent and 5.9 percent). Additionally, both Democrats (-48.7 percent) and Republicans (-41.6 percent) expressed the most concern with profit-driven data uses.

“These similarities between Democrat and Republican constituents are incredibly promising for any new national privacy legislation because it demonstrates that some values are shared across party lines,” Schmit said.

“The bottom line is that elected officials must understand their constituents’ policy preferences to balance individual privacy with public benefit, especially given our findings of bipartisan backing even in a politically polarized year,” Kum said. “Our findings support including public health and research data-use allowances in new privacy legislation.”