Dive Brief:
- Nearly one-third of American adults and 40% of adults under 34 report that they would be comfortable with an artificial intelligence-led primary care appointment, according to a new survey released by Outbreaks Near Me and SurveyMonkey.
- But the option isn’t their preference. Although survey respondents reported believing that AI in healthcare could reduce medical bias and improve diagnostic accuracy, over 80% of respondents would prefer seeing a human medical professional for prescribing pain medications, deciding when to go to the emergency room and other services.
- The latest survey suggests that, while AI hype may be on the upswing, entrenched patient attitudes and preferences for care could be slow to shift.
Dive Insight:
Advocates for AI have say harnessing the technology to assist in early detection, diagnosis, treatment plan development and patient communication has the potential to make healthcare more accurate and cost effective.
This latest report joins the growing body of research suggesting that, while patients have generally bought into the use case and agree that AI would improve healthcare, they are reluctant to have their own provider utilize AI. This winter, approximately 60% of Americans reported they would be uncomfortable with their providers using AI to diagnose them.
A research team at the Boston University’s Questrom School of Business ran a series of experiments testing patient reluctance in 2019. In one condition, students were told a doctor would diagnose them, and in the other, students were told a computer would assess them. Forty percent of students elected to see the doctor compared with just 26% in the AI condition.
Researchers have attributed the disconnect between professed AI comfort and low utilization with Americans’ love of individual experience. Researchers from the Boston team found that the more likely patients were to view themselves and their circumstances to be unique, the less likely they were to want to see an AI provider.
Part of the reluctance may also be due to patients’ mistrust of novelty. In February, 75% of respondents told Pew Research Center that they worried their providers were moving too quickly to implement AI without fully knowing the risks — a concern echoed by the WHO in May. The most pressing concerns, watchdogs warn, are misdiagnoses, cybersecurity threats and the risk of inadvertently magnifying medical bias.
According to the report out this week, some providers are among the most concerned about the risks of adopting AI. Forty-two percent of providers surveyed said the tech would help and hurt the field and a third said that AI will hurt more than it helps.
If AI evangelists can persuade those on the fence about the technology to utilize it going forward, AI could save the industry up to $360 billion annually. But that’s only if healthcare providers can help providers and patients overcome entrenched skepticism and preferences for human-to-human contact, which, as of this report, remains far off.
