Opinion polls can be valuable tools for physicians during the COVID-19 pandemic to gauge how well the public is assimilating public health messages and what patients’ major concerns are. For example, in a recent poll conducted by the Kaiser Family Foundation, 52% of respondents said they or a family member in their household had skipped or postponed medical or dental care because of the pandemic.
But polls aren’t infallible, and just as physicians need to understand how to read clinical studies in the context of the broader evidence base, so too they should understand how to interpret poll results. During a recent “Ethics Talk” videocast from the AMA Journal of Ethics® (@JournalofEthics), Mollyann Brodie, PhD, the foundation’s executive vice president and chief operating officer, outlined how to identify reliable polling and how to syncretize conflicting poll results.
Skepticism is healthy
“I think it is definitely harder and harder for … a casual observer of polls to be able to distinguish a really high quality, very representative one from [one that is] not so representative,” Brodie said. “What I would say to everyone … is you have to be a good consumer.”
She recommended following these guidelines.
Look for the hallmarks of transparency. “The more information that a poll provider … is willing to tell you—about how they selected their sample, how they defined their population, was it a probability based survey or a non-probability based panel, how did they weight the data, what were the questions, what order were the questions in—the more confidence you can have about what kind of conclusions you can draw from that data,” Brodie said.
Never draw strong conclusions based on responses to a single question. Strong conclusions usually require answers to multiple questions.
Consult multiple polls. “Sometimes the polls that get the most attention or the poll finding that gets the most attention … it makes news because it's an outlier,” Brodie said. “You want to question, ‘Well, why is that an outlier?’”
Be careful how much you read into differences. For example, a four-percentage point difference between populations might sometimes sound considerable, but the variation might be due to lack of precision of the measurement.
Keep in mind that the public isn’t always right. “The most important thing as public health professionals, as clinicians, as experts, is to remember that the data tells you a lot about the audience and about where people are,” Brodie said.
"It doesn't necessarily tell you what the right thing is to do, but it tells you what your challenge is going to be in terms of being a leader, in terms of being an expert, in terms of being an educator. You'll know what you're facing, and you might have to rethink your message and rethink how you're talking about an issue in order to meet the public where they are.”
Listen and claim CME credit
This episode of “Ethics Talk” is CME activity designated for 0.5 AMA PRA Category 1 Credit™. Check out previous episodes of the “Ethics Talk” podcast or subscribe to the series in iTunes or other services.
The AMA and the Centers for Disease Control and Prevention are closely monitoring the COVID-19 pandemic. Learn more at the AMA COVID-19 resource center. Also check out pandemic resources available from the AMA Code of Medical Ethics, JAMA Network™ and AMA Journal of Ethics, and consult the AMA’s physician guide to COVID-19.
