IRB Case: The ability of a biomedical IRB to effectively manage
Controls on novice researchers
The question that is posed in the presentation of this case is whether or not a biomedical IRB can provide adequate oversight of socio-behavioral research. Examining the posed question further, the issue is not really “oversight,” because that implies a level of ongoing supervision that is overstated based on the role and structure of IRB review. Much of what an IRB needs to do in reviewing a study for the protection of human subjects is predictive; ongoing oversight is not possible, despite the efforts to identify grave issues through data- safety monitoring boards (which would not apply in the case of much socio-behavioral research) and continuing review at regular, but infrequent intervals. In summary, an IRB does not have the ability to provide oversight. That said, there are risks that can be predicted, and risks that are unforeseen. If the question is re-asked as to whether or not an IRB that is used to review biomedical protocols can sufficiently review human subjects’ risks in studies of a more behavioral nature, then there are several ways of addressing the question.
Psychosocial risk can be present in any research protocol, whether anticipated as a potential adverse event, or occurring as an unforeseen event. It is an often underappreciated risk in biomedical research. As such, those on any IRB should always be alert to the risk of psychosocial issues and harm. In that sense, a biomedical IRB is certainly suited to review a protocol such as this one. Since this study recruited hospitalized patients, it is unlikely that a more behavioral-focused IRB could have done a more effective review. The IRB in question has decided to seek consultation in the future when reviewing studies of a more psychosocial nature. That is always a good idea, and I think this use of consultation would be preferable to “turfing out” such studies to a different IRB. Consultation and subsequent learning from unfortunate events such as this one is advantageous to all.
We are well aware that often “the devil is in the details” with research protocols in general, and perhaps very much so with this study. This study could have posed distress and risk whether it approached the research question with quantitative or qualitative methodology. I have reviewed studies that have incorporated qualitative lines of inquiry as well as quantitative questions and instruments that contain multiple assumptions. I have argued, many a time, about the unnecessary distress that questions that have built- in assumptions can cause to subjects, and how they can invalidate the meaning of findings. In this case, with a zealous and inexperienced investigator who may have her own beliefs about the value of spirituality and serious disease (she is a Divinity student, after all), this might be particularly the case.
An example of a quantitative question that illustrates assumption might read something like: “When I feel distraught about my health, I find that I can receive comfort from my belief in God,” with a Likert Scale that offers “all the time,” “more than half of the time,” “some of the time,” “less than half of the time,” and “never.” The embedded assumption in this question is that one has a belief in God. It does not provide an answer choice for someone to easily indicate no belief in God, and as such, it conveys a judgment that one who does not share this belief is “outside the norm,” a burden that someone who is struggling with poor health doesn’t need to have added to their already vulnerable state. This hypothetical, quantitative question needs a choice such as “does not apply to me.” In IRB deliberations, others have suggested in scenarios quite similar to this that the “never” category could be used. This does not provide for consistent data. On one hand, checking “never” might mean that one picked that answer because one does not believe in God and that answer is the best (but non-fitting) choice available. Or one might pick “never” for a completely different reason – that one has a very strong belief in God, but this belief does not provide any comfort in dealing with one’s disease. While this example shows how quantitative questions that carry assumptions can cause distress to subjects and invalidate findings, the same can be said for qualitative research. For example, inviting a subject to participate in qualitative focus groups on spirituality and terminal illness might subject someone to peer pressure, personal invalidation and extreme distress. With the presentation of a qualitative proposal, an IRB has less material available to review for bias and assumption that may harm subjects. Qualitative discussion and data collection operate outside the close assessment of the IRB, certainly well beyond any “oversight,” and thus, very thoughtful understanding of the theoretical nature of the protocol and any bias that the researcher brings is essential.
One of the telling issues in this case is that the researcher was rather vague in delineating a research question or specific aims. One must further wonder about her inexperience in that she had a planned cohort of 15 patients, rather than planning qualitative data collection until saturation. She is described as a staff nurse and a student (presumably for graduate studies, but not noted here) in a Divinity School. The nursing educational level of this student is not indicated, nor whether she was receiving supervision on this research project from a nurse or theological mentor with more educational preparation in conducting research. A further question that might be posed to the academic community at large is assessment of the educational benefit versus social and individual risk in requiring that students conduct independent research that they might not be qualified to execute. In nursing, research expectations are clearly delineated based on educational level. Baccalaureate level nurses are expected to be consumers of research; Master’s level nurses are expected to be critical evaluators of research; and Doctoral level nurses are expected to be involved in the design of research projects. Based on this guideline, if this nurse is prepared at the Baccalaureate level, and in a Master’s (or even undergraduate) Divinity program, the implementation and design of independent research is not supported by her nursing education and therefore the close supervision of an experienced research mentor would be critically important.
The case study points out that three of the non-scientific members of the IRB were not present at the time of initial review and approval. One hopes that minimal IRB requirements for at least one non-scientific member were followed. (Note: for biomedically-focused IRB’s, the definitions of scientific versus non-scientific membership is generally clear, but these same definitions are sometimes murky in IRBs that are focused on behavioral studies.) Regardless, all committee members should be encouraged to review protocols and submit comments, even when they cannot attend meetings, and especially if the areas are within their expertise. IRB’s can further encourage careful evaluation of protocols by all members by assuring that very clear protocol submissions are received before they are forwarded to members. Committee members who have to spend great amounts of time deciphering voluminous and arcane theoretical documents that obfuscate the issues necessary for IRB review are less likely to provide a quality review. Requests for clear summaries of the research plan can be required, with submissions returned for revision if this requirement is not met. This may be especially important in fostering the commentary of critical non-scientific members.
The “devil may be in the details” also in terms of how this research was described in the formal informed consent document. It is important that all who are reviewing and conducting research keep in mind that the very nature of research means that unanticipated risks can occur, and this should always be disclosed to participants in the informed consent process. Thoughtful IRB deliberation about unforeseen risks can help researchers modify protocols to reduce the risk of harm, but it is always present. This study points out that research can cause significant socio-behavioral harm, and as such provides important lessons for this committee. I think that the approach of the IRB in seeking consultation is appropriate and positive.