No matter how much we rely on it or want to think of it as an ideal solution, we can probably all agree that machine translation isn’t always accurate. Although we’ve made impressive technological progress in the past few years alone, machines still have difficulties with things like context, figurative language, idiomatic expressions, and new vocabulary.
Still, many hospitals and healthcare organizations rely on machine translation…with less than ideal results.
All US hospitals and healthcare organizations are legally required to provide interpreters and translators to patients who are not proficient in English, and many healthcare providers believe that machine translation is the solution. The benefits include instant translation, instead of possibly having to wait for an interpreter to arrive on-site or via an online platform, and little or even no cost (after all, apps like Google Translate are free).
Another advantage to machine translation is that while interpreters and translators can’t provide immediate translations of written documents, AI can.
But a recent article warns that while machine translations’ accuracy may have improved over the years, ‘bots still aren’t ready for the complex world of medical interpreting. This is especially true when it comes to less common languages.
For instance, journalist Nicole Wetsman cites a study published in early March which found that Google Translate had a 90% accuracy rate when translating medical instructions into a common language like Spanish, but only a 55% accuracy rate in a less common language like Armenian.
Even that 10% margin of inaccuracy is troubling when it comes to the medical field. Wetsman gives the example of a mistranslation in Chinese, a common language with a high AI translation accuracy rate. The English instructions for a patient to stop taking Coumadin, a blood thinner, was machine-translated into Chinese as “Do not take anymore soybean.”
It’s clear that a human interpreter or translator is still the ideal. But how can a healthcare organization or hospital provide human interpreters and translators if budgetary concerns are an issue? What about getting written translations quickly?
For US-based healthcare providers, a familiarity with state laws could help. Some states require a patient’s insurance to cover interpreter and translation costs. Others may have partnerships between organizations and translators. As federal laws change and evolve, it’s also a good idea to keep abreast of what may be covered or required on a national level, as well.
Another way to reduce cost is to rely on remote, rather than on-site interpreters. A few years ago, for instance, one US healthcare system saved a whopping $1.5 million annually by making this switch. Virtual interpreters combine the easy availability of remote work with the ability to visually observe patients, which can be helpful when working with them.
For faster written translations, healthcare organizations could plan ahead by saving template versions of prior translations of standard documents like medical instructions and discharge forms. In addition to this, healthcare organizations can work with a translation provider who’s able to fill their needs – for instance by offering around the clock or on-call translation services.
When it comes to healthcare and pharma translation, accuracy shouldn’t be a hoped-for result, but a must. Fortunately, there are ways for healthcare organizations to provide the best possible translation and interpretation services to patients. It may take some planning and readjustment, but the end result is a safer experience for patients, whatever language they speak.