The principle of justice articulated in The Belmont Report requires equitable selection of human research subjects. Equitable in this context means that the risks and benefits of the study are distributed fairly. Fairness has two components: 1) avoiding exploitation of the vulnerable (e.g. preying upon a poor, uneducated population) and 2) avoiding the unjustified exclusion of any population ( whether out of bigotry, laziness or convenience).
Recruitment strategies invariably shape the selection of research subjects and the extent to which a pool of participants really represents a cross-section of society. Institutional Review Boards (IRBs) are charged with evaluating whether study recruitment plans and materials used to obtain informed consent are easily understood and free of misleading information. This is relatively straightforward when researchers, IRB members, and study subjects all speak the same language. But when studies are done in geographical areas that include numerous cultural and language communities, it can be quite tricky.
One of the barriers that prevents people from enrolling in (or even knowing about) studies is a lack of awareness and planning by researchers to address language differences. The human research protection regulations at 45 CFR Part 46.116 require that informed consent information must be provided to research participants (or their representatives) in language understandable to them. IRBs are supposed to be vigilant about this and require investigators to obtain translated Informed Consent Documents (ICDs) for use with non-English speaking research subjects. But researchers commonly balk at this expectation, saying it’s unreasonable. (A disproportionate number of objections have been raised to me thusly, “And what am I supposed to do if someone shows up speaking only Swahili?!”)
The reasons researchers may object to proactively planning for language differences are various. Some object in principle to being told they need to include all comers who otherwise qualify for the study. Some worry about the considerable expense of ICD translation because they failed to include those costs as a line item in the budget they negotiated with a sponsor to do a study. Others express concern about how they can communicate with non-English speaking subjects once enrolled.
All of these obstacles can be overcome. And, in fact, both The Common Rule and the Civil Rights Act require avoiding bias, which–in the case of research studies–means planning for recruitment, consent, and ongoing communication with subjects in languages other than English. For researchers in geographical areas with heterogeneous populations, planning for interaction with native speakers of Spanish, Mandarin, Cantonese, Korean, and other common non-English languages just makes good sense. (For an example of good planning, see UCLA’s IRB website, which provides translated consent materials in over 30 languages, including Braille.)
To be clear, planning does not mean grabbing a file clerk with a Spanish surname and asking her to translate a complicated document containing medical terminology. I have seen hundreds of translated ICDs full of errors ranging from the confusing or incomplete to the downright absurd (imagine a form that says to be eligible for enrollment, a subject must have 18 anuses). Planning means knowing the major language groups in your geographic area, consulting with the IRB about availability of translation and interpretation services, and negotiating with the sponsoring agency or company to include these legitimate costs in the study budget.
To the extent that most clinical studies are done in healthcare environments that already have translation and interpretation services available for patients, these resources should be utilized for the responsible and fair performance of human research. Equitable recruitment can be more challenging in non-medical settings because so many social/behavioral studies are not funded by sponsors who can pay the added costs of translation and are not performed in hospitals with language support services. But IRBs understand this and they also are equipped to approve justified exclusions when they make sense. For example, when a survey instrument has not been validated scientifically in a language other than English, it would not be desirable to hastily translate the survey in order to enroll subjects with limited English proficiency. In that case, including non-English speakers would actually be problematic because their data would not be useable; so exposing them to risks or inconveniences would not be warranted.
At the end of the day, the justice principle requires fairness. We must be equitable with regard to access to research and we must ensure all research subjects truly understand the known risks and anticipated benefits of any study in which they are invited to participate.