The Research Brief is a short take about interesting academic work.
Table of Contents
The big idea
Most doctors use language that is too complex for their patients to understand, but some have the unique ability to tailor their language to meet their patients’ communication needs and overcome the confusion that is so common in health care. These are the key findings of our new study recently published in Science Advances.
This language-matching strategy – what we call “precision communication” – appears to be especially helpful for the one in three Americans who have low health literacy. Prior studies have shown that individuals with low health literacy have worse comprehension of medical information and instructions and poorer health outcomes compared with those with adequate health literacy.
To conduct our research, we analyzed hundreds of thousands of secure email messages between doctors and patients with diabetes. Using sophisticated computational linguistics techniques, our research team discovered that only about 40% of patients with low health literacy have a doctor who adapts the complexity of their language to match the language their patient uses. We also found that even fewer patients are cared for by doctors who are consistently attuned to the kind of language that their patients use – whether it be low or high health literacy – and then adapt their communication accordingly.
We found that patients fortunate enough to be under the care of doctors who practice this form of precision communication were better able to understand and act on their doctors’ advice and instructions. Patients whose doctors don’t match their language to their patients’ health literacy are more likely to be confused and may get sicker. The benefit of this approach was so strong as to eliminate the usual gaps in understanding between patients of low and high health literacy.
Why it matters
Perhaps no health care experience is more universal than being sick and not understanding one’s doctor. Not only is this a frustratingly common – and often dangerous – experience, it’s also a massive and costly public health problem. Despite the outsized impact of this problem, few clinical studies have examined the issue, and no study has used artificial intelligence methods or has been large enough to draw firm conclusions.
Doctors and patients are relying more on secure messaging, a digital communication innovation that has expanded during the COVID-19 pandemic. Our findings suggest that most doctors can and should adjust how they listen and respond to patients to achieve more effective communication.
Patients who find themselves confused should ask their doctor to restate their explanations and advice in more approachable ways. And our study suggests that health systems should carefully consider the ways that they can best support doctors and patients to achieve shared meaning. This includes how they train clinicians and how they allocate and reimburse for the time, personnel and technologies that can promote communication.
What still isn’t known
While prior research has shown that understanding one’s health condition and its treatment is key to getting healthier, we do not know how beneficial this form of precision communication is to achieving better health outcomes. We also can’t yet determine whether doctors’ written communications reflect how they communicate verbally – in person – although the results of the patient survey that we used in this study suggest overlaps between doctors’ written and spoken communication.
What’s next
We are designing studies to examine whether language-matching improves health outcomes such as blood sugar or blood pressure control. We have also developed and are testing whether an automated feedback system embedded in the electronic health record can enable precision communication in email exchanges. The system rapidly analyzes patients’ email messages and alerts the physician if the complexity of their email response is too high.
Dean Schillinger received funding from NIH/National Library of Medicine grant R01LM012355 and NIH/NIDDK grant 2P30-DK092924
Nicholas Duran receives funding from the National Science Foundation (NSF 1660894). The current work is funded by the National Institutes of Health (NIH R01 LM012355).
