Language barriers and human subjects research.

Over at On Becoming a Domestic and Laboratory Goddess, Dr. Isis looks at challenges of opening up participation in human subjects research to potential subjects who are not fluent English speakers:

When one enters the university hospital here at MRU, there are a number of skilled and qualified translators that are available to help patients that can't dialogue in English to communicate with health care staff. They are able to sufficiently translate documents to allow a patient to provide some reasonable level of consent (my M.D. blog buddies can debate the quality of said consent). There is no infrastructure like this in research at most major research universities. Consent forms are written in English. Even if I could provide verbal translation for this man, it would not be ethical for him to sign a document in a language he cannot read himself and understand. Thus, it all more or less becomes a moot point. Beyond that, while I am a fluent Spanish speaker, I am not a qualified medical interpreter. I have no idea how to say "indwelling arterial catheter." Babelfish says it's, "catéter arterial dejado en un órgano." I know that can't be right, but who am I to question Babelfish?. Even if I were completely confident saying, "We are studying the physiological effect of [Dr. Isis's favorite stimulus] on vascular function" in Spanish, it's not the best use of my time as a researcher (unless one of you folks want to write me in for 10% effort, and then I will make sure I learn to say it).

So, the question is, is there an ethical issue here at all ... ? The National Institutes of Health mandate the inclusion of minorities in human research studies. In our area, members of the major minority groups often do not speak fluent English. However, the translation of study documents and the hiring of an interpreter to help with the consent is expensive and I have never known an investigator to include a translator in a budget when they could have a technician. Furthermore, if you are willing to translate a consent form into one language, what about all of the other languages that might be spoken in the area around where the study is being conducted, no matter how rare? Indeed, most people I know make the decision that the ability to understand and communicate in English at a 6th grade level is a criteria for participation.

Yet, if the ability to speak English is a criteria for participation, then we by default fail to include particular groups in research cohorts. We're back to research cohorts being comprised of middle-aged white men.

(Bold emphasis added.)

It looks to me like there is an ethical issue here. Plus, I think I see a scientific issue. Together, the two kinds of issues make me think that tackling the issue of translation should be a priority for researchers (and for the agencies funding their work).

Let's start with the ethics. The ethical principles that are supposed to guide research with human subjects (at least as understood in the U.S.) are described in the Belmont Report. These include beneficence, which requires maximizing the possible benefits and minimizing the possible harms of a particular intervention for the human subject, respect for persons, which includes valuing and taking steps to protect the autonomy of the human subject, and justice, which requires that the risks and benefits of the research -- and of the knowledge produced by it -- be shared equitably.

Respect for persons is the principle that seems most prominent in the situation as Dr. Isis presents it. Researchers have a clear duty to provide human subjects with information they can understand about the research -- about its potential risks and benefits, about the type of knowledge the research is intended to produce and who this knowledge might benefit, and about the rights of the human subjects during the research, including the right to halt participation at any time without jeopardizing their access to medical treatment. Even between researchers and human subjects who officially speak the same language, there are significant challenges to communicating clearly to human subjects (most of whom are not scientists, or even necessarily scientifically literate) in order to secure consent that is truly informed and free from coercion. When human subjects are not fluent in the researchers' home language, it's much harder to be sure that all the important information is being transmitted and to check that it has all been received accurately by its audience.

Communicating the crucial information badly can result in human subjects who are therefore hampered in the exercise of their autonomy. Arguably, one part of respect for persons is not putting persons in situations that can't reasonably be expected to navigate. This makes the requirement that human subjects of research be able to understand and communicate in English at a 6th grade level seem like an ethical safeguard, not just a practical constraint.

However, the principle of justice pulls in the other direction here.

We've discussed before how justice requires that an intervention developed using human subjects from a particular population needs, at the conclusion of research, to be available to that population. It would be an unfair distribution of risk and benefit if, say, a new drug developed through research with poor human subjects ended up being priced at a level that made it unavailable to poor patients who might benefit from it. Another aspect of the principle of justice, though, is the idea that access to participation in research studies is a good that should be distributed fairly. To the extent that a great deal of biomedical research with human subjects is intended to develop better interventions, the human subjects of this research may end up with better outcomes as a result of participating in the research than they would seeking the non-experimental interventions. As well, for some diseases or conditions where scientific understanding is at an early stage, experimental treatments may be the only treatments available. Thus, access to participation in a research study may be access to treatment, and being excluded from participation amounts to being excluded from treatment.

So, from the potential subject's point of view, access can be a benefit and exclusion can be a harm. Granted, the hope is that access to participation will include all the information one needs to weigh the costs and benefits of participation -- and this may require a skilled translator to ensure successful communication between researchers and subjects. Still, as the NIH requirements state:

It is the policy of the NIH that women and members of minority groups and their subpopulations must be included in all NIH-supported biomedical and behavioral research projects involving human subjects, unless a clear and compelling rationale and justification establishes to the satisfaction of the relevant Institute or Center Director that inclusion is inappropriate with respect to the health of the subjects or the purpose of the research. Exclusion under other circumstances may be made by the Director, NIH, upon the recommendation of an Institute/Center Director based on a compelling rationale and justification. Cost is not an acceptable reason for exclusion except when the study would duplicate data from other sources. Women of childbearing potential should not be routinely excluded from participation in clinical research. All NIH-supported biomedical and behavioral research involving human subjects is defined as clinical research. This policy applies to research subjects of all ages.

And this is where ethical considerations and scientific considerations come together. Especially in biomedical research, the aim of human subjects research is to develop safe and effective interventions that are broadly applicable -- treatments that will be effective for lots of different kinds of people, not just for a narrow subpopulation of people. Thus, conducting research on human subjects drawn from a single narrow subpopulation of people is less useful than conduction research with human subjects drawn from a variety of ages, ethnic groups, socioeconomic statuses, and so forth. A pool of human subjects that is more representative of the human population in all its variation is more likely to result in information applicable to that human population.

Now, the language one speaks need not be correlated with physiological function, genetic susceptibility to disease, or other features that are the object of biomedical study. Nonetheless, in American society one's home language and degree of fluency in English may well be correlated with other factors that are correlated with these "biological facts" -- factors such as employment status and health care access, diet, exercise, stress level, familial support systems, etc. To the extent that such factors may make a difference in the safety and effectiveness of an intervention, including subjects who are not fluent in English may reveal some of these complex relations. Ultimately, this should result in knowledge that can be applied more easily to a diverse patient population.

Including human subjects who are not fluent in English, though, requires good translation services; you don't get to skimp on respect for persons just to build a more representative sample population, nor even to increase justice. And good translators will cost money. So, given that the NIH requires the inclusion of women and members of minority groups in the human subject research studies they fund -- and given that cost is not an allowable reason for their exclusion, why aren't PIs including translators in their research proposals and budgets? If a translator is necessary to fulfill the requirements of the NIH guidelines, doesn't it make sense to include this expense in the research proposal?

Perhaps good translators are more scarce than good technicians, but if you need them to do the job properly, what's wrong with asking for them?