RESEARCH METHODOLOGY
5.9 DATA COLLECTION METHODS AND TOOLS
5.9.2 Structured Interviews: The structured questionnaire was identified as the most suitable instrument for collecting data on trial participants’ understanding, given that the main aim of
the study was to assess how the respondents understood the concepts under study and how they viewed or appreciated them (attitudes). Although the concepts under study are generally difficult to understand, questions were formulated to make them easy to understand. Both closed and open-ended questions were used. Open questions were used to invite the respondents to give their own opinions or views on trial procedures. The questionnaire
consisted of three sections: personal data, questions about understanding of research, and their opinions and suggestions on research in general. It was administered to all participants by one trained research assistant (See Appendix 2 and 4 for English and ChiChewa versions of the structured questionnaire).
The questionnaire was administered to a total of 203 participants from the microbicide study.
Seven women refused to participate citing constraints of time and fatigue. Four of the women
reported that they were taking advantage of the temporary absence of their partners to visit the clinics as they were participating in the microbicide trial without the approval of their partners.
Such women had to ensure that they were at home before the times that their partners came back. Three women indicated that were not interested in responding to the questionnaire as they had already responded to too many questionnaires during the past months.
5.9.2.1 The Development of the Structured Questionnaire
The questionnaire consisted of 100 response items and each interview took approximately 40- 45 minutes on average (see Appendices 2 and 4). For each of the three concepts under study, the questionnaire included some questions which covered understanding of procedures used in the study, the purpose of the procedure, and the implications of the procedure for each
participant personally. The dependent variables were measured using numeric scores. The questionnaire included several questions on understanding related to each of the three concepts under study. For each participant, a score was obtained for each concept and a composite score obtained after adding scores from each of the concepts under study. The study relied on the scoring method used by Joffe et al. (2001). This scoring method has been found to be reliable and valid. No weights were assigned to concepts or questions, as each one was viewed as equally important in facilitating understanding. Questionnaires used in previous related studies were consulted during the development of the questionnaire for this study (Ellis, Butow, Tattersall, Dunn & Houssami, 2001; Joffe et al., 2001; Kass, Maman & Atkinson, 2005;
Lindegger et al., 2006; Quieroz da Fonseca & Lie, 1995, 1999).
The questionnaire and the scoring system were reviewed by various experts including an epidemiologist, two biostatisticians, a seasoned immunologist and researcher, and three social scientists. All agreed that the results should be generalisable to the study population and the study should be replicable. The biostatisticians suggested some significant changes to the
layout of the questionnaire and recommended the use of tables to capture data on the concepts under study. They also assisted in the sample size calculation. One of the biostatisticians assisted by creating the database in SPSS and entering and cleaning all the data from the structured questionnaires. The social scientists provided useful insights into the use of words such as randomisation, placebo and double-blinding and the challenges that these may present.
They assisted in generating alternative phrases and ways that could be used in bringing out these concepts. For example, regarding placebo, they suggested that one could find out what the respondents knew by asking them about the number of different gels that were being provided to the participants and to list them. The epidemiologist mainly assisted generally in determining the procedures for this study.
Several measures were taken to ensure the reliability of the questionnaire. A limited number of Yes/No questions were included in the questionnaire in order to avoid responses based on guesswork. True/false questions were also used minimally for generating some data on general knowledge on the three concepts (Peterson, 2000). There were a number of questions using different formats under each concept to counter the effects of guess work. Some of the questions were not aimed at recall – but the use of disclosed information in making sense of the personal implications of the study procedures.
The main questionnaire was aimed at coming up with baseline data which would determine the design and implementation of an intervention. The questionnaire was aimed at examining the knowledge of the women concerning a real study in real life and not a simulation of a study as compared to that investigating understanding within the context of a simulated trial (Pace &
Emanuel, 2005). The questions also sought to assess actual understanding and not perceived understanding. English and ChiChewa versions of the structured questionnaire are included as Appendices 2 and 4. The English and ChiChewa versions of the informed consent forms which
were used are attached as Appendices 3 and 5 accordingly. After implementation of the
intervention aimed at improving understanding, a shorter questionnaire with about 50 questions was administered to participants in both the intervention and non-intervention groups. The English and ChiChewa versions of the evaluation forms are attached as Appendices 12 and 14.
The English and ChiChewa informed consent forms for the intervention phase are attached as Appendices 13 and 15.