Left or right political leaning in the United States predicts both physician and patient beliefs about COVID-19 treatments, with the two groups perceiving information differently, according to a study to be published this week in the Proceedings of the National Academy of Sciences (PNAS).
A team led by University of Pittsburgh researchers surveyed a novel panel of 592 board-certified critical care physicians and a sample of 900 lay adults recruited from an online panel. A total of 410 physician responses and 882 lay responses were analyzed. Physicians were surveyed in three phases from April 2020 to April 2022, and lay adults were surveyed during the final physician survey.
"We focused on critical care physicians, because they are important decision makers in the treatment of severe COVID-19, and because their day-to-day judgments are less influenced by patient preferences, compared to other physicians," the authors wrote.
Physicians were asked to assess a clinical scenario about a seriously ill COVID-19 patient and decide which treatments they would administer. They also reported their beliefs about the quality of clinical evidence supporting use of the treatment and predicted the proportion of their peers who would make the same decision. Lay adults reported their beliefs about the therapies but didn't make treatment decisions.
All participants reported beliefs about the antimalarial drug hydroxychloroquine and the antiparasitic drug ivermectin in March and April 2022. They were also asked about their political ideology, consumption of news, and preferred cable news network.
Conservative participants most polarized
The study yielded strong evidence of political polarization on 8 of 10 physician outcomes and all 6 lay outcomes. For example, conservative physicians were about five times more likely than their liberal and moderate peers to say that they would treat a hypothetical COVID-19 patient with hydroxychloroquine, which has no scientific evidence to support its use for this indication.
Physicians' beliefs were, on average, less polarized than those of their lay counterparts. The difference, the authors said, was driven largely by agreement between liberal and moderate physicians, with conservative physicians showing polarization that was often similar to that of conservative laypeople.
"We know that there have been enormous disparities in prescriptions of hydroxychloroquine and ivermectin between Republican and Democratic counties," lead author Joel Levin, a PhD student, said in a University of Pittsburgh news release. "Our study suggests that these patterns are driven not just by patients' preferences, but also by what doctors think about these drugs."
Physicians' beliefs were, on average, less polarized than those of their lay counterparts.
Senior author Jeremy Kahn, MD, said that it was "disappointing but not entirely surprising to see that physicians are not able to completely put aside their ideological biases when they approach clinical decision making."
Conservative participants were more likely than others to prefer Fox News as their media source, which the researchers said suggests exposure to different information.
The authors also asked participants to assess the findings of a study showing that a drug didn't effectively treat COVID-19, identifying the drug as ivermectin for some but not for others. When participants were told it was ivermectin, which is not supported by clinical evidence, conservatives were more likely to say the study was less informative and methodologically rigorous and that the authors were probably biased.
'Physicians are human'
When physicians’ responses were analyzed separately, they didn't show a clear division along political lines, so the authors said they couldn't conclude that physicians evaluated evidence about ivermectin in a biased way, the authors said.
Kahn said that the study shows that, overall, people assess scientific data in a biased way. "Physicians are human, and they have biases that are driven by their political ideology," he said. "I think there's a notion among the public that decision makers and experts arrive at their viewpoints through analysis of empirical data and base their decisions on the best available evidence, but this study suggests that this is not the whole story."
The findings, the researchers said, show that education alone is not enough to minimize polarization.
"Our work highlights the limits of expertise and exposure to scientific evidence in mitigating polarization, underscoring the need for psychologically informed interventions to improve evidence dissemination on politicized topics," they concluded. "Such interventions might draw on bipartisan sources, appeal to shared values, or highlight implications that are desirable to liberals and conservatives alike."
