Elena Yu's responses to questions after the end of Wednesday's session:

Question 1:

No single study of any minority population will allow us to make broad generalizations of the group. In the papers that we have published, we always made it a point to mention the limitations of our studies. I have not personally encountered the kinds of criticisms that you mentioned. However, I have heard colleagues and advanced graduate students at some leading schools of public health who make direct or indirect comments about minority health research in general as being "unscientific," "unreliable," or "lacking rigor." These are individuals who tend not to have any research or personal experience at all with ethnic minorities. I have also heard criticisms of "lack of generalizability" being applied to some minority studies where non-PPS (proportional-to-population size) sampling were used to generate the sample. In general, such comments stem from ignorance about the complex processes involved in doing good minority health studies. As we train more minority students to enter public health and major in epidemiology and biostatistics, there will be increasing awareness of the methodologic issues in minority health research. The heightened awareness is likely to lead to the development and acceptance of many innovative methods that have been used by scientists in recent years to study minority populations.

Question 2:

Yes. In general, nurses are excellent collaborators. They appreciate details and precision. Those trained in research understand the importance of ensuring reliability and validity of data. It is wonderful to work with them! I highly recommend including nurses as collaborators in health studies of any populations, especially minorities. One of my best collaborator in the Asian American Cancer Control Project is Katherine Kim, a Professor of Nursing at Grand Valley State University, Michigan. Working together has made us lifelong friends.

Question 3 (originally asked of Dr. Kalsbeek):

I know of several large-scale studies where a small number of minorities are included in the study. Because the research instrument is developed only in English, by design non-English speaking persons are excluded, either because they will simply refuse to be interviewed or because the interviewer would consider them as an "ineligible" person. If the study was planned to exclude these non-English-speaking populations, then such exclusion constitutes neither bias nor discrimination. On the other hand, if the study was designed to include them, then the lack of a non-English-language version of the study questionnaire constitutes "sloppy" science. But this is generally not seen as a discrimination issue.

An important issue related to your question is the ethics of obtaining consent in a population with limited education which happens to be also non-English-speaking. I have had the experience of doing surveys with such populations. We always prepared different translations of the instrument. However, written informed consent is difficult to obtain. It has the potential to be misunderstood.

For example, many of the Southeast Asians have been told by the service agencies they frequented as part of their orientation to life in the United States, that they should not sign any documents they cannot understand. They were often quite puzzled, and sometimes troubled, by the fact that we wanted them to sign their consent "just to talk." They see it as a "trap." We had to train our interviewers repeatedly over a much longer period that one would do in a mainstream project, on how to handle these kinds of situation.

The approach we take is: (1) First we advertise the study in the local newspapers so that the literate segment of the population understands the objective of our research; (2) When we approach the respondents in their homes, we explain the study objectives to all present at the household, and let those who understand repeat the explanations to those who did not understand until the targeted respondent agrees to be interviewed. (3) When a subject is illiterate, obtaining a signature becomes a meaningless exercise.

In those situations, we used the consensual verbal consent from family members as a basis for proceeding with the interview, reassuring the family and the respondent that the interview can be terminated at any point they so desire without negative consequences on them. In one unusual circumstance, where we had to list household members in order to apply a table of random numbers to select a person to interview, the parents got so worried they called police halfway through the interview, convinced that our interviewers must be communist spies in the U.S. who want to have their descendants killed. The interviewer phoned our office, we spoke with the police, and asked the interviewers to destroy all the interview data in front of the interviewees and hand them over to them. After that experience, we made it a point to inform the Police Department in every neighborhood where we contacted interviews. This was a good and unforgettable lesson for us.

6/14/2000, Minority_Health@unc.edu