Lost in Translation: New Technologies are Closing the Language Gap in Healthcare

Mahamu Kanneh, a Liberian immigrant living in Gaithersburg, MD, was arrested in August 2004 on charges that he repeatedly raped and molested a seven-year-old girl over the course of a year. Kanneh’s case has failed to go to trial in the three years subsequent to his arrest and has now been dismissed by a Montgomery County, MD judge because the court could not find an interpreter fluent in Vai, Kanneh’s native language.

While this particular case remains in the appeals process, it doesn’t require a great leap of the imagination to conceive of other situations in which communication failures caused by lack of access to translation services can create immense difficulties. When the venue shifts from the courtroom to the exam room, insufficient access to interpreters can mean it’s private physicians or hospital systems that stand to face jail time or stiff financial penalties for noncompliance with federal regulations requiring the use of interpreters for individuals with limited English proficiency (LEP). Although this may seem harsh, one need only peruse the data and statistics databases of the Office of Minority Health or the Agency for Healthcare Research and Quality to understand how seriously an individual’s health can be

affected by a lack of understanding of the English language.

Supply and Demand

The issue at heart, however, is not one of English-speaking ability, but that of giving LEP patients information about their medical needs that they can understand in their native language. Typically, this is done through a medical interpreter, one who possesses command of a number of languages—from the vernacular to advanced medical jargon—who can facilitate communication accurately between patient and physician to ensure that proper care is given.

“In an ideal world, we’d get to the point that our healthcare workforce matches the linguistic makeup of our patient population, we’d have as thorough a distribution of bilingual healthcare providers as we have bilingual patients, and they’d all be located in the same place so no assistance anywhere would be needed,” says Don Schinske, Executive Director of the California Healthcare Interpreting Association, in conversation with MDNG. “But that’s not the case.”

A look at the numbers quickly reveals why that is so. There are far more patients than interpreters; according to the Bureau of Labor and Statistics (BLS), of the 31,000 interpreters and translators employed in the US in 2004, only 4,100 worked in the healthcare sector (usually in the hospital setting). There is little financial incentive to become a medical interpreter; salaried interpreters and translators had median hourly earnings of $16.28 in May 2004.

Statistically, LEP patients are not going to have ready access to medical interpreter services in either a hospital or private practice setting, but technological advances are beginning to change that reality. Physicians can expect to utilize a number of technologies, both on the horizon and currently available, to overcome language barriers and effectively treat their patients.

Two-Way Street

Machine translation has a history dating back to the early 1930s, but did not begin to really accelerate until the onset of the Cold War, when researchers involved in the Georgetown-IBM experiment utilized the IBM 701 computer to accurately translate Russian into English. Spurred by the experiment, the federal government began heavily investing in machine translation technology, hoping it would result in the production of fully automatic high-quality translation. Fast forward to 2007, and the federal government is again investing heavily in translation technology to aid its combat and intelligence efforts in the Middle East.

Because of the short supply of military linguists fluent in Arabic, and the risk to their lives on the battle field, the National Institute of Standards and Technology (NIST) is evaluating prototype two0way translation systems for the Defense Advance Research Projects Agency (DARPA), while the United States Joint Forces Command (USJFCOM) is testing IBM Research’s Multilingual Automatic Speech-to-Speech Translator (MASTOR) system with support from IBM’s Technology Collaboration Solutions group. Both research efforts aim to enable any soldier and the foreign national to whom he or she is speaking to converse in real time, with the intended meaning of words and phrases accurately conveyed to the listener. This system, no doubt, would be welcomed by more than a few physicians who have encountered a language barrier when trying to explain prescribed care to an LEP patient.

Unfortunately, the systems remain in the early stages of development and it may be quite a number of years before the technology trickles down to the public sector. Schinske predicts that if and when this occurs, healthcare won’t be the fi rst to embrace automatic translation software. “The legal issues alone would seem to suggest that healthcare would be the last adopter of something like that.” Indeed, IBM is currently exploring both public and private sector opportunities for

future implementation of MASTOR, healthcare included; however, no timetable has been put forward as to when the public will have access to such software.

On the Market

Something a little more accessible to physicians would be Integrated Wave Technologies’ Voice Response Translator, which can convert simple English commands into 16 languages or VoxTec’s Phraselator, which holds tens of thousands of pre-recorded phrases in various languages on a flash card. Both pieces of hardware are currently used by the military in Iraq, and VoxTec indicates that it is in limited use in the healthcare sector. However, technology like the Voice Response Translator and Phraselator is simply a slightly more advanced version of translation software that currently exists on the market for physician use in conjunction with Smartphones, PDAs, or computers. The usefulness of such software in facilitating communication during a patient visit can only be described as limited.

“Translation technology has completely revolutionized the industry and made the lives of translators much easier and much more efficient,” Nelva Lee, President of the Medical Interpreting and Translating Institute Online, told MDNG. “However, translation is not just language. There are also cultural aspects and idioms that you have to account for. If you have someone translating using just technology, it won’t be accurate.”

This idea of the necessity of human contact pervades any conversation dealing with translation technology in the healthcare setting, and for obvious reasons. Although there are undoubtedly some individuals who believe a MASTOR-like system will be the cure-all for the world’s medical interpreting needs, most discussion regarding this topic concludes that human intervention, in some form or another, is necessary in a patient—physician encounter.

“Often, a doctor will be dealing in situations of a high range of difficulty in which they’ll want a person in there with them, if not a couple of people,” predicts Schinske. “Cases of surgical informed consent and mental health, for example. So much of psychiatric diagnosis and treatment is about language; there’s no blood test, just talk.” Schinske goes on to discuss examples from pediatric oncology and situations such as domestic violence in which physicians “play the role of social worker” to illustrate his position further. “I bet we’ll see, over the next 10 years, technology taking on a lot of different roles, but in certain types of situations, you’ll always have a need for an almost all-human encounter.”

For this reason, professionals in the medical interpreting field hold telephone and video services in high esteem. The technology addresses many of the problems associated with machine translation, as well geography, a reality that, according to Schinske, is a historic inefficiency that has plagued the medical interpreting field. “I hate to call the use of video an emerging technology because it’s long been used in other settings, but for the use of language interpretation in healthcare, it’s something relatively new,” Schinske explains. “You wheel a monitor into a room, and in 10 or 15 seconds, you’re operating fully with an interpreter. It’s a minor miracle, because you can’t bring someone up the stairs or across the hall to interpret in that length of time, let alone across the state.”

Such a service is the crux of the Healthcare Interpreter Network (HCIN), a consortium of three public California hospitals—Contra Costa Health Services, San Joaquin General Hospital, and San Mateo Medical Center—that utilize shared voice and video over Internet Protocol resources. The pilot program began in August 2005, and as of August 2006, the network was routing approximately 3,000 video conference and phone calls per month in five languages: Cambodian, Hindi, Hmong, Spanish, and Tongan.

Room for Interpretation

The system allows interpreting resources to be allocated more efficiently through expansion of the interpreter base; the ability to categorize calls based on special requests, such as a male or female interpreter; and a feature that, when all interpreters of a particular language are occupied, automatically routes those calls to an audio-only commercial interpretation service. These abilities drive down costs, which have already been lowered by distributing them across the HCIN.

“When I first heard about the HCIN, I thought it sounded like a good idea. But I underestimated the impact this technology could have. At San Mateo Medical Center, this has been one of the single best steps we have taken to improve the quality and safety of our care,” says Nancy Steiger, RN, MS, CEO of San Mateo Medical Center.

Although the creation of such a network among small private practices would undoubtedly be benefi cial, until the practice becomes more widespread and the feasibility can be accurately assessed, phone translation services like Language Line Services, Network Omni, and CyraCom—that test, train, and staff medical interpreters who specialize in round-the-clock medical interpretation and healthcare document translation—are excellent alternatives. CyraCom’s services, for example, are currently being utilized by the Ronald McDonald Houses of Dallas and Fort Worth and by Cook Children’s Health CareSystem in Fort Worth to accommodate the area’s rising LEP patient population.

“The phone has made it so much easier. It’s decreased the stress level of families staying here as well as our staff ’s,” says Myke Holt, Executive Director at Ronald McDonald House of Fort Worth. “Before we got the phone, we had to rely on someone from the hospital staff to come and interpret—which was inconvenient for both of us and sometimes took too long—or we had to rely on a bilingual family staying here, which was hit or miss.” A phone-based interpretation service, announced recently by Verizon in conjunction with Language Line Services, has the potential to put high-quality interpretation services into the hands of patients as well as physicians. The Verizon Prepaid Interpreter Service (VPIS) program allows individuals to purchase prepaid access cards that put them in contact with interpreters speaking any of 170 languages; certified medical interpreters are available to individuals in what Verizon has determined to be the 22 most spoken languages.

For quicker communication, cards are also available for individuals who only need English-to-Spanish translation; however, the service would most likely have to be introduced to an LEP patient by an English-speaking physician who would place the order and then have an interpreter explain to the patient how to access the service. VPIS is unfortunately quite expensive for the contact that it affords—$35 for 10 minutes of access to an interpreter; $50 for 15 minutes; and $100 for 33 minutes—and there is no word yet as to whether use of the service would be covered by insurance.

At the end of the day, however, the fact of the matter remains that, despite the relatively high cost of a service like VPIS, opportunities are expanding for LEP patients to utilize medical interpreter services. Because it is highly unlikely that the ranks of human interpreters will ever grow to accommodate the full load of interpreting needs in healthcare, this expansion is in large part occurring because of technological innovation. Despite his lack of faith in technology to create the magical automatic translation algorithm, Schinske expects translation technology to engender far-reaching change in the field of medical interpreting. “As for the level to which technology is being used now within healthcare language services, we’re just seeing the beginning of it. It’s all up from here.”

Bradley Schmidt is a freelance writer and former MDNG editor.