Most patients view test results before hearing from a physician

More than two-thirds of U.S. adults who had a medical test in the past year read the results in a patient portal before hearing from anyone on their care team, and roughly 1 in 15 of them said they did not really understand what they were looking at, according to a nationally representative study published Aug. 18 in JAMA Network Open.

The finding attaches national numbers to a shift physicians have been absorbing since the 21st Century Cures Act Final Rule required that electronic health information, including labwork and imaging, be released to patients as soon as it is available.

The first read of a result now routinely happens without a clinician present, and the question that follows arrives as a portal message rather than as a scheduled conversation. The study's second finding is the one with operational teeth: comprehension is not evenly distributed, and the strongest protective factor the researchers identified was patient-centered communication — patients feeling that clinicians explain things in a way they can understand and involve them in decisions.

Researchers led by Jemar R. Bather, Ph.D., assistant professor of biostatistics at NYU School of Global Public Health, analyzed responses from 4,982 adults in the 2024 Health Information National Trends Survey who had been offered online access to their medical records and reported having a test within the previous year. Weighted, the sample generalizes to nearly 175 million U.S. adults.

Of that group, 68.6% said they viewed immediately released results before hearing from a clinician. Among those early viewers, 6.6% rated their understanding of what the results showed and what they meant for their care as poor. Another 26.8% said they understood fairly well, 31.7% well and 34.9% very well.

"The Cures Act has restructured the traditional pathway between the release of a test result to a patient-clinician discussion," Bather said in an NYU release.

Which patients open results first?

Early viewing tracked with digital fluency more than with any demographic category. Each 1-point increase in the survey's digital literacy score was associated with a 24% higher prevalence of viewing results first, the largest single association in the model.

Adults 65 and older were 19% more likely than adults 18 to 34 to look first, women were 14% more likely than men, and households earning $100,000 or more were 13% more likely than households under $50,000. Each additional chronic condition raised the likelihood by 4%, and greater use of health-related social media was independently associated with early viewing.

What did not reach significance is worth noting for practices building outreach around this: after adjustment, the researchers found no meaningful differences in early viewing by race and ethnicity, educational attainment or urbanicity.

What predicts poor comprehension?

Geography and prior experience in the health care system did most of the work.

Compared with respondents in the Northeast, those in the Midwest had more than triple the prevalence of poor comprehension and those in the South roughly 2.5 times. Patients who reported experiencing discrimination in medical care had a 69% higher prevalence of poor comprehension.

The confidence intervals on the regional estimates are wide, and the authors applied no multiple-comparison correction, so some significant findings may reflect chance.

Two factors cut the other way. Each 1-point increase in patient-centered communication score was associated with half the prevalence of poor comprehension, and each 1-point increase in digital literacy with 37% less.

Study author José A. Pagán, Ph.D., professor and chair of the Department of Public Health Policy and Management at NYU School of Global Public Health, argued in the NYU statement that the fix belongs partly to the technology. Health systems hand patients their data without investing much in making that data interpretable alone, he said, and portal developers should consider using artificial intelligence to summarize results in plain language and open an easy path to follow-up questions.

The analysis carries the usual limitations of survey work. It is cross-sectional, so it cannot establish cause. Every measure is self-reported and subject to recall and social desirability bias, and the comprehension outcome was a single item bundling two distinct questions — what the result showed and what it meant. The survey also does not identify respondents' health systems, meaning the study cannot account for how differently organizations implemented immediate release.

Bather said the Cures Act's access goal is being met, but that a subset of patients who view results before being contacted is struggling to understand them, and that better patient-centered communication may help close that gap.