IRB Case: An ethnographic study of homeless adolescents
Protect the researcher, include more subjects
The IRB’s decision to approve the research presented in the case is ethically sound. The investigator is proposing to conduct a study of considerable value while at the same time doing so in a relatively risk-free manner. Several questionable concerns, however, arose throughout the review process, and each is considered below.
In its discussion concerning the ethical and regulatory acceptability of waiving informed consent and assent, the committee correctly determined that the intended population, i.e., homeless adolescents, created too strong a burden for the investigator to obtain either physical consent via photo identification, or parental consent from a legal guardian. As a result, the decision to waive written consent in favor of verbal consent and the use of an information sheet is legitimate. Given the title of the research project, it is curious that the discussion concerning why the investigator initially proposed using subjects over the age of 17 was abandoned so quickly. It may be understood that the waiver of consent will indeed create the opportunity for the investigator to include adolescent minors in his study, but this should have been spelled out more explicitly and agreed to beforehand. Furthermore, it is unclear whether an age ceiling for the subject population was ever espoused.
One of the most concerning questions brought forth by the committee was whether the investigator had provided enough safeguards for himself. While it was not stated explicitly, the implication was that such safeguards could be required for eventual study approval. Institutional Review Boards, whether biomedical or socio-behavioral in scope, are given the responsibility of reviewing research in order to ensure the adequate protection of human subjects. At no point in any of the Codes of Federal regulations are review committees guided to use the protection of researchers as a prerequisite for study approval.1 From a purely ethical standpoint, it makes perfect sense that a group of concerned individuals would voice opinions about the possible lack of safety for a graduate student researcher who proposes to conduct his investigation in a potentially dangerous environment. It is important, however, to remember that this concern should be expressed qua individuals concerned for the safety of another, rather than under the official auspices of the Review Board. A legitimate compromise could, for example, involve the relaying of the concern to the investigator via an informal email or telephone call, or perhaps even as an attachment to the protocol determination. But it absolutely should not play a role in the rendering of the determination itself, particularly in a case as seemingly benign as has been presented here.
The IRB’s decision to request that the investigator confirm the reporting requirements for the state in which the research was to take place was both ethically and prudentially justified, but presents a practical issue. If the investigator will have no real means for obtaining the actual ages and identities of the subjects in the study, it would be difficult (if not impossible) to determine whether a potentially reportable scenario involves a subject who meets the criteria for obligatory reporting. From one perspective this may not pose much of a concern, as the case study states that the investigator would be under no legal obligation to report suspected cases of child abuse or neglect. But again, from an ethical standpoint, it is less clear that the investigator would be justified in acting as the purely unbiased and non-invested ethnographer were he to encounter an instance of possible abuse. Legality may not be the crux of the issue here.
Also at issue is the IRB’s decision to revise the exclusion criteria to exclude all prospective participants whom the investigator deemed under the influence of drugs or alcohol, or whose behavior was “sufficiently ambiguous to prevent the researcher from obtaining a reasonable level of comfort for his safety.” The latter harkens to the above discussion concerning the legitimacy of IRBs basing determinations on investigator safety. The former is striking because it seems to be at variance with the investigator’s goal to obtain the types of subjects he wishes to interview. If the key subject population is “youth (who) exhibit a variety of emotional and physical problems including depression, substance abuse, and violence,” it is puzzling that the Board would eliminate subjects who exhibit behaviors—however harmless—that appear to fit the inclusion criteria so well.
A final comment concerns the lack of compensation for subjects participating in the study. There is very little discussion about the reason(s) for the absence of compensation, beyond the statement that “there were no plans to compensate study participants in any way.” This is troubling primarily because of the complexity of the groups involved. On the one hand there is the investigator and the IRB, both of whom have an overt ethical obligation to conduct research in such a way so as not to coerce subjects into participating. So the decision to refrain from compensating the subjects appears prima facie to be ethically acceptable. On the other hand, there is the study population, which consists of an intensely vulnerable group: adolescents with probable histories of drug abuse and violence who have been rendered homeless. The question arises as to whether or not this type of group, simply by virtue of its vulnerability, may be the exception to the rule to refrain from payment. Might there be a compromise beyond the investigator’s preparation to “provide participants with publicly available information on how and where to access help for drug and alcohol abuse, housing, etc…upon request by the participant?” Perhaps an offering of coupons or gift certificates for food would help to avoid the normal skepticism that actual money would simply be spent on nefarious goods. Some type of compensation seems ethically acceptable, if not advisable.