Discussion

Chapter 16: Methodology

Selection of Institutions

Five areas of the country were selected as sites for the SIS: Ann Arbor, Baltimore/Washington, Dallas/San Antonio, Raleigh/Durham, and Seattle/Tacoma. These sites were selected because they include institutions that receive some of the highest amounts of federal dollars for human subjects research and because we were trying to balance our sample with respect to geographic region, rural/urban settings, and expected ethnic mix. At each of these five sites, a university hospital, a VA hospital, and a community hospital were selected. If other federal government or military hospitals were present at a site, the most highly funded of these institutions were included. A total of nineteen institutions were selected, as presented in table 1. Interviews were conducted at sixteen of the nineteen institutions selected. At the University of Washington Health Services Center and the Seattle Veterans Affairs Medical Center, the institutional review board (IRB) process could not be completed within the time constraints of the SIS. Baylor University Medical Center declined to participate in the study.

Recruitment of Patients

At each of the participating institutions, patients were recruited from the waiting rooms of three outpatient departments: medical oncology, radiation oncology, and cardiology. On the days that patients were seen in each of the departments, a clinic staff member informed patients arriving at the clinic that a study was being conducted to examine attitudes and beliefs about participation in medical research. The staff member also asked patients if they were willing to have a study interviewer approach them in the waiting room to see if they were willing to be interviewed. Interviewers approached a systematic sample of these patients and, following a brief description of the SIS, asked individuals to participate.

Each patient who agreed to participate in the Brief Survey completed a written consent form that authorized the SIS staff to consult one or more of the following sources to ascertain whether the patient was or had been a participant in a research project: doctors, investigators, research nurses, a research office, a research database, and their medical/research records. The survey, composed mostly of multiple-choice questions, was designed to take roughly five to ten minutes to administer. Patients completing the Brief Survey received $5.00 for their time and effort devoted to the study. All patients who indicated in the Brief Survey that they believed they were medical research participants were asked if they were willing to participate in an In-Depth Interview (roughly forty-five minutes). These interviews raised many of the topics from the Brief Survey. A sample of those who agreed to this further participation were contacted to arrange for an interview at a time and place convenient for them. Patients completing the In-Depth Interview received $25.00 to compensate them for their time and effort devoted to the study as well as to pay for any expenses related to participation in the study, such as transportation and parking.

A target of 150 Brief Surveys (50 each in medical oncology, radiation oncology, and cardiology) was set for each institution. A target of 100 total In-Depth Interviews was set for patients selected from all institutions. Both the Brief Survey instrument and the In-Depth Interview guide appear in a supplemental volume to this report. Electronic files containing the final data from the Brief Survey and transcripts from the In-Depth Survey are maintained along with other records of the Advisory Committee.

Data Collection and Analysis

Brief Survey

The Brief Survey instrument was refined based upon focus groups of patients conducted at two institutions not participating in the SIS: the University of North Carolina at Chapel Hill and Georgetown University. The instrument was then further refined based upon pilot testing at these same institutions. The instrument consisted predominantly of questions with multiple-choice answers addressing:

1. General attitudes toward medical research.
2. Any perceived differences in understanding among the following terms: medical research, medical study, clinical trial, clinical investigation, and medical experiment.
3. Beliefs about research participation.
4. Reasons for either participating in research or not participating in research (when applicable).
5. Demographic and other background information (such as race, sex, age, and employment/insurance status).

All survey forms were labeled with an identification number for each patient, rather than with patients' names. Data were entered into a computerized database and analyzed using standard statistical methods.

In-Depth Interview

An In-Depth Interview guide was developed based on the focus groups and pilot testing at the University of North Carolina at Chapel Hill and Georgetown University. The In-Depth Interview contained open-ended questions that allowed participants to speak more extensively about the issues addressed in the Brief Survey. For example, patients were asked to describe their attitudes about research generally, their own experience as research participants, how they arrived at their decision to participate, and the informed consent process for their particular project. All of the interviews were audiotaped and transcribed. All cassette tapes and transcripts were labeled with an identification number for each patient and never with patients' names. All transcripts were read in their entirety by Advisory Committee staff members, and then data were coded and analyzed using text analysis software.

Determination of Research Participation

To assess how well patients' reports of their participation in research matched their documented enrollment in research projects at the participating institutions, a mechanism for determining research participation was developed for each institution. In each instance, we sought documentation of participation in research from sources such as patients' medical or research records. This information was supplemented by information from investigators and research nurses.

A second level of review was conducted in those cases in which there was an apparent discrepancy between a patient's own description of having been, or not having been, a research subject and the documents or other sources of this information. A physician or a research nurse on the Advisory Committee staff reviewed the patient's interview, the patient's medical and research records, institutional databases, and other sources of information at the local institution for evidence that could either resolve or verify the discrepancy.

Expert Panel Assessments of the Research Projects

To identify some of the basic characteristics of the research projects in which the patients we interviewed were or had been participating, we convened an expert panel and asked this panel to make some preliminary judgments based on the information provided in the consent forms of these research projects. The panel consisted of eight physicians: specialists in oncology, radiation oncology, cardiology, nuclear medicine, and radiology, as well as general internists.

We attempted to secure a copy of an unsigned consent form for every research project in which a respondent in the SIS was a documented participant and that had been conducted at one of the study institutions. Although 336 consent forms were requested, only 236 were received in time to be reviewed by the expert panel.

Each consent form received was reviewed by the expert panel, which met for one day. After agreeing how the forms would be evaluated, the panel broke into four teams, each consisting of two physicians, one who had content area expertise in the project being reviewed and another who did not. If a team could not reach consensus on the evaluation of a particular consent form, it was brought to the larger group for review. If a consent form was received after this meeting of the panel it was sent to the panelists for review. The expert panel characterized the research projects on three dimensions: (1) type of research (therapeutic, diagnostic, or other); (2) degree of sickness of the population (expressed as a high burden for those with diseases such as AIDS, a medium burden for those with conditions such as hypertension, or a low burden for those who generally were healthy), and (3) incremental risk assumed by those who participated in the project compared with those who were not participating in the research project (measured as minimal or more than minimal incremental risk).