Lost in Translation: Ethics of Translation Services in Health Care

Sydney Shelby ’25 is pursuing a double major in biology and public health with a minor in philosophy, and was a 2023-24 health care ethics intern at the Markkula Center for Applied Ethics. Views are her own.

Imagine jumping out of a plane. Though you were initially scared to even get in the plane in the first place, you mustered up the courage to still do it, and you are now halfway between the sky and the ground. As you go to deploy your parachute you realize that the instructions to do so are written in a language that you do not understand. As the ground creeps closer and closer you become unsure of how you will overcome the language barrier to open the parachute and save your life. 

This is the essence of the experience faced by the 29.6 million people with limited English proficiency ( LEP) in America as they fight to engage with the health care system. 

What is Currently Being Done?

Title VI, "Improving Access to Services for Persons with Limited English Proficiency," from the Civil Rights Act of 1964, is a core pillar of oral interpretation and written translation services. This is protected by anti-discrimination laws regarding national origin. 

In 2000, Bill Clinton signed Executive Order 13166, which specifically mandated that all health care institutions that receive federal funding must provide interpreters for LEP patients. However, Clinton’s executive order had one fatal flaw: it didn’t specify who could do the interpreting. Health and Human Services (HHS) does not require that hospital staff serving as interpreters possess national certification. However, the fact that an individual has above-average familiarity with speaking or understanding a language does not suffice to make that individual a qualified interpreter. Executive Order 13166 was not sufficient for establishing protections for LEP patients, and new legislation is needed to ensure this population’s health and safety.

The Ethics of the Ambiguity in Executive Order 13166

Without clarification, health care institutions have capitalized on remaining in compliance with Executive Order 13166 without hiring licensed medical interpreters. These are the two major avenues:

1. Health care institutions will encourage health care staff who can speak the same language as LEP patients to interpret for patients.
2. In emergent situations, patients will need to rely on bilingual family members to act as interpreter.

Both of these avenues violate the bioethical principles of autonomy and justice.

In response to avenue 1: Though there are increasingly more doctors and medical professionals coming from diverse backgrounds, this does not make them qualified to interpret patients, nor make it ethical. In a survey by the Center for Medicare and Medicaid Services, around 45% of the practices they surveyed used bilingual staff as their primary method of communicating with LEP patients. 

There is a major power imbalance present when a doctor is explaining a treatment that they already believe should be used for the patient. When a physician or staff member is then not speaking in their primary language, they may translate words incorrectly and/or accurately interpret a patient’s story. Most importantly, they may use words that unintentionally (or intentionally) coerce the patient to pursue a specific treatment plan without them fully understanding and/or knowing that they had other options. This severely underwrites the patient’s right to autonomy and fully informed consent. 

Patients are not able to make an informed decision if options are not presented in a way that makes them feel like options. This can leave patients feeling dissatisfied with or unaware about the risks associated with their treatment. This only further exacerbates existing mistrust that many of these populations have with the medical field. Physicians need to continue to be educated on the cultural and societal factors that could impact their patient interaction, but leaving the interpreting up to the physician leaves room for error and coercion. 

In response to avenue 2: Though there is a level of comfort associated with having family members interpret, there are concerns that can arise from bringing a non-neutral third party into the conversation about a patient’s care. Family members may give their own version of events, and their emphasis may skew the consultation. It can also be difficult to check the veracity of the interpretation, leading to a misdiagnosis. Similarly, a family member may find it hard to share difficult or bad news and they may have emotional or cultural reasons to distort your message. Additionally, the usage of minors as translators can cause harm as well. This can range from the child being taken out of school to act as an interpreter to being traumatized by their exposure to a complex medical scenario. 

Minority communities should not have to inequitably bear the challenges associated with having to come up with an interpreter. Not only is this an issue of justice, but it is an issue of autonomy and informed consent. Family members are likely unable to be completely neutral when providing a patient with options related to their care, creating another situation in which patients don’t have the ability to make a fully informed decision.

Translator Deficits and Possible Interventions

The easiest solution would be to make it easier for health care institutions to employ translators and interpreters when needed. It has already been proven that the utilization of translators and interpreters overall reduces health care costs by ensuring patient satisfaction and quality of care. However, it would also help to further ensure financial coverage for interpreters through Medicaid, as that can vary state to state. So, the largest barrier that remains for health care institutions to easily access this tool for LEP patients is the ever-present interpreter shortage in the field of health care. For example, in 2023 it was estimated that there were only 388 certified Mandarin medical interpreters and 177 Cantonese nationwide for more than 900,000 who speak the Chinese language. This is clearly not a sufficient number to ensure the quality of care for every one of those 900,000 patients. Additionally, it could be easier (though still challenging) to find translation services for widely spoken non-English languages like Spanish, Chinese, Vietnamese, but it can be very difficult to find translators for languages that are not as common among U.S. residents. Still, these patients deserve a high quality of care no less than anyone else. In cases like these, the utility of the integration of interpretation technologies into health care systems is highlighted. However, this integration of technologies is also not without ethical setbacks. 

In considering the integration of translation technologies like telehealth or Artificial Intelligence (AI) translation, it is important to weigh the potential pitfalls. Tele-health services allow, and will continue to do so, for interpreters to be streamed into any appointment without having to physically be there in person. Yet, studies have found that patients with limited English proficiency have documented disparities in care access, satisfaction, utilization, and quality when it comes to telehealth services. Even if telehealth does improve, it does not create more translators. 

The recent rise of artificial intelligence provides an interesting opportunity to account for this deficit. Theoretically, having access to a computer that could translate any language you wanted it to would substantially help with the interpreter shortage. However, the way that AI is trained can lend itself to sizable inaccuracies when it comes to translation. This is especially true when you consider the ability to train AI on languages that are not as common using machine learning techniques. The question becomes how ethically permissible is it to implement systems that will knowingly lead to disparities for particular language populations? Additionally, when medical translators are trained, they receive training on cultural competency, not just translation. The best translation relies not only on understanding the language, but in understanding the cultural contexts for which it is placed in. For example, the verbiage used in Spanish dialects can be highly dependent on the region where the language was learned. 

Data on minority populations, languages, and cultures is not always the most sizable, especially in the field of health care. If AI systems were trained with a small or inaccurate data set, it would subject these LEP patients to a lower standard of care. Morally, we have an obligation to equitably maintain a quality of care for all. Populations with existing health care disparities cannot be subject to compounded barriers to care. 

Conclusion

In conclusion, it cannot be overstated how important it is that LEP patient populations have the ability to obtain good translative services. It is important to end by stating that though quality interpretive services are the goal, it does not eradicate the need for health care providers to be trained in cultural competency for their patient population. It is only the combination of reliable translation with culturally informed care that will be sufficient in addressing the long-standing inequities present in LEP populations today.