Research by Johns Hopkins epidemiologists has confirmed what some have long suspected about consent forms required of clinical trial volunteers: They use language far too difficult for most people to understand.

In a survey of 114 U.S. medical schools, researchers found the average consent form to be written at a 10th-grade reading level, while an estimated one in two American adults read at or below an eighth-grade level. The investigators recommend using language even simpler than that. The government-funded study appears in the Feb. 20 issue of The New England Journal of Medicine.

The study also found that sample text for these forms--provided to researchers by the medical schools' institutional review boards--generally fails to meet the IRBs' own standards for reading comprehension.

"While IRBs are charged with safeguarding people with low literacy, they may have an inadvertent role in creating unreadable informed-consent forms, as they write the templates that investigators use," says Michael K. Paasche-Orlow, lead author of the study and a postdoctoral fellow in bioethics and internal medicine.

However, the 52 medical schools that have been subjected to oversight by the U.S. Office for Human Research Protections had forms that were easier to read.

"This is the first bit of empirical evidence to show that federal oversight has any benefit," Paasche-Orlow says.

"Our study suggests that a fourth-to-sixth-grade reading level is a suitable target, because text at this level can best convey key concepts simply and directly," he says. "Alternative methods of obtaining informed consent, such as multimedia presentations, also may be beneficial. Every time subjects come in, you have to make sure they understand what's going on."

Paasche-Orlow and colleagues surveyed the Web sites of 114 U.S. medical schools (of a total 124) for IRB reading comprehension standards and consent form templates. They determined the reading levels of the forms using two standardized scales that determine the minimum grade level required to read and understand the text; the Flesch-Kinkaid score, for example, is determined by the average number of syllables per word and the average number of words per sentence. Data on the individual schools' research activities, local literacy rates and federal oversight were obtained from additional Web sites.

Sixty-one Web sites posted the schools' grade-level readability standards, which ranged from a fifth-grade reading level to a 10th-grade reading level. Forty-seven Web sites contained descriptive guidelines such as "simple lay language," and six contained no guidelines.

Only 8 percent of the 61 medical schools that posted grade-specific standards met them. The average readability score exceeded the stated standard by 2.8 grade levels. The ease of reading was not related to either the school's amount of research funding or the local literacy rates.

At Johns Hopkins, Paasche-Orlow plans to revise some of the forms, using focus groups of potential patients to develop and validate the improved language.

The study was supported by the National Heart, Lung and Blood Institute. Co-authors were Holly A. Taylor and Frederick L. Brancati.