The study, conducted at two pediatric ERs and published in the Annals of Emergency Medicine, found that mistakes which could have "clinical consequences," like giving the wrong medication dose, were about twice as likely if there were no interpreters or if the translator was an amateur.

"The findings document that interpreter errors of potential clinical consequence are significantly more likely to occur when there is an 'ad hoc' or no interpreter, compared with a professional interpreter," said lead researcher Glenn Flores, of the University of Texas Southwestern Medical Center in Dallas.

An estimated 25 million U.S. residents have limited English proficiency - that is, they say they speak the language less than "very well."

By law, U.S. hospitals that receive federal funds have to offer some type of translation help for those patients. That can mean a professional interpreter who works for the hospital or telephone or video-based translation services.

Studies have found that patients like to have a translator available, and that it may improve care and cut costs, by avoiding unnecessary tests, for instance. But it hadn't been clear how well professional interpreters performed against amateurs, or no interpreter at all.

The current study was based on 57 families seen in either of two Massachusetts pediatric emergency rooms. All were primarily Spanish-speaking and 20 families had help from a professional interpreter.

Ten had no translation help and 27 had a non-professional interpreter. The reasons for the difference was unclear.

When trained interpreters were available, 12 percent of translation slips could have potential health risks to a child. But when the interpreter was somebody such as a family member or a bilingual member of the hospital staff, 20 to 22 percent of their translation errors were potentially risky.

In one example from the study, an amateur interpreter - a family friend - told the doctor that the child was not on any medications and had no drug allergies. But the friend had not actually asked the mother whether that was true.

A number of questions remain about professional interpreters, such as whether in-person interpreters are more effective than phone or video interpreters and what are the most cost-effective measures, Flores said.

Another issue is training, including the question of how much is enough.

Errors were least common when the interpreters had 100 hours of training or more. In these cases, only two percent of their translation slips had the potential for doing the child harm.

Few training programs for medical interpreters provide at least 100 hours of training, Flores said. Even when hospitals run their own programs, the hours involved vary widely.

"These findings suggest that requiring at least 100 hours of training for interpreters might have a major impact on reducing interpreter errors and their consequences in health care, while improving quality and patient safety," the researchers wrote. SOURCE: bit.ly/HDDLbj

(Reporting from New York by Amy Norton at Reuters Health; editing by Elaine Lies and Bob Tourtellotte)

• A Permanent Economic Emergency