Yale University.
Calendar. A-Z Index.
Yale Bioethics

IRB Case: An ethnographic study of homeless adolescents

Competence, sampling, and risk to subjects

Robert Heimer

From the description given, the study design and the intent of the qualitative research seem clear and sufficiently articulated to be viewed as a study of little risk for potential study participants; one exception, however, is the diary component. 

The decisions to waive written consent and parental assent are sensible.  In some of our studies that have included homeless youth, we encountered teenagers whose information we sought to capture.  We have described these individuals as emancipated minors in our IRB applications.  We have argued that since they are truly living outside of their nuclear family they are emancipated.  We have further argued that since the study protocol does not include any procedures or activities that place study participants at risk of investigator-caused injury, the benefit of parental advice and assent is nil. After all, the purpose of parental assent is to assure that individuals who cannot make informed choices because of their lack of adult perspective about risk, are given proper counsel by those they trust most, i.e., their parents.  But in this case, their parents are not trusted and in many cases involving homeless youth the parents are actually abusive.  Parents of children who have fled the parental home are often vindictive and might actually, for no reason other than spite, block their child’s participation in a study that the child wished to join.  We have used these arguments successfully to make the case for waiving parental assent. 

One decision regarding who should be approached to participate in the study is less sensible.  The IRB insisted that the investigator revise the exclusion criteria to exclude prospective participants considered to be possibly under the influence of drugs or alcohol.  I am unhappy about the use of the phrase “under the influence of drugs or alcohol.”  This seems too ambiguous.  If I have a drink or two, smoke a little marijuana, or even use a little of a stronger psychotropic substance, then I am “under the influence” but I am not too intoxicated to provide clear and valid answers to an interview.  The IRB should have been more precise and insisted that the investigator decline to interview anyone who appeared too intoxicated to provide meaningful information. 

I am concerned about one aspect of the study design that the IRB did not address.  The investigator planned to approach only groups of youths rather than individuals for a reason that may or may not be valid.  The choice, however, of approaching only groups introduces a sampling bias in that there may be significant differences between homeless youths who are more or less likely to be a part of a group.  If the interviewer is interested in obtaining a more thorough sense of the “environmental culture and organizational strategies” of these individuals, then individuals in both categories – the more gregarious and the more solitary -- should be included. 

We have done some diary data collection with a stigmatized and often homeless population - injection drug users.  We were interested in learning about the variety of injecting situations in which subjects found themselves during the course of a week, in order to better understand the contexts of injections that carried risk for HIV and hepatitis infection, for drug overdose, or for exposure to police detection and arrest.  As in the study proposed by Wallach, diaries were to be completed by only a small subsample of study participants.  Our chief concern from the standpoint of the subjects’ protection was that carrying a diary for a week in which they were to document their drug use might prove self-incriminating if confiscated by the police or found and read by a family member, co-worker, or social contact.  This seems to be a potential problem for this study also, and the IRB has offered no suggestions to minimize these risks.  To minimize them in our study (and to be sure we collected adequate data), we met with diarists daily, reviewed their entries, and took the pages of reviewed entries.  This minimized their exposure while simultaneously assuring us high quality data (see Stopka TJ, Springer KW, Khoshnood K, Shaw S, Singer M.  Writing about risk: use of daily diaries in understanding drug-user risk behaviors.  AIDS & Behavior. 8(1):73-85, 2004).  

The IRB considered the statutory requirements for reporting criminal activity, abuse and neglect, and addiction or mental illness.  We have conducted research in six states and in all of them the legal requirements for reporting were similar.  We were not obligated to report on use of illicit substances (which would include alcohol for minors), nor to report on past criminal activities.  To further protect our study subjects, we obtained a Certificate of Confidentiality for each study from the National Institutes of Health, to preclude the use of our research data in legal proceedings.  For criminal activity, the sole exception to blanket confidentiality that we have offered study participants is if they mention the intent to commit a violent crime in the future.  However, in nearly twenty years of research with injection drug users, we have never had to report this to the authorities.  On the other hand, we are obligated to report instances of child or elder abuse, which is relevant here since many homeless youths leave home as a consequence of abuse.

Next comment: Protect the researcher, include more subjects

Return to case description

Return to IRB cases main page