OBJECTIVE: To develop and test the metrologic qualities of a questionnaire for a survey on determining factors in initiating hormone replacement therapy (HRT), using Ajzen's planned behaviour theory. DESIGN: Semistructured interviews to collect the behavioural beliefs needed to develop the questionnaire and the test-retest. SETTING: Family medicine practices and the community. PARTICIPANTS: In two family medicine practices, 42 women completed semistructured interviews that yielded the beliefs needed to design the questionnaire for the core survey. Thirty other women from the community took part in the test-retest 2 weeks later. MAIN OUTCOME MEASURES: The internal consistency of the constructs (tendency of answers within a group of questions) was determined using Cronbach's alpha coefficient. Stability of the global constructs (proportion of subjects who had an absolute difference in their scores of less than 1 between the two measurements) was determined using the adjusted agreement coefficient (adjusted Po). RESULTS: Findings for 12 of the 14 global constructs were above the acceptance threshold of 0.60. In one case alpha could not be measured because there was only one question for measuring it. On the test-retest, 12 of the 14 adjusted Pos had a reliability rating at or above 80%. CONCLUSION: The questionnaire developed using Ajzen's planned behaviour theory had the requisite metrologic qualities for integration into the core survey on determining factors for initiating HRT.
OBJECTIVE: To develop and test the metrologic qualities of a questionnaire for a survey on determining factors in initiating hormone replacement therapy (HRT), using Ajzen's planned behaviour theory. DESIGN: Semistructured interviews to collect the behavioural beliefs needed to develop the questionnaire and the test-retest. SETTING: Family medicine practices and the community. PARTICIPANTS: In two family medicine practices, 42 women completed semistructured interviews that yielded the beliefs needed to design the questionnaire for the core survey. Thirty other women from the community took part in the test-retest 2 weeks later. MAIN OUTCOME MEASURES: The internal consistency of the constructs (tendency of answers within a group of questions) was determined using Cronbach's alpha coefficient. Stability of the global constructs (proportion of subjects who had an absolute difference in their scores of less than 1 between the two measurements) was determined using the adjusted agreement coefficient (adjusted Po). RESULTS: Findings for 12 of the 14 global constructs were above the acceptance threshold of 0.60. In one case alpha could not be measured because there was only one question for measuring it. On the test-retest, 12 of the 14 adjusted Pos had a reliability rating at or above 80%. CONCLUSION: The questionnaire developed using Ajzen's planned behaviour theory had the requisite metrologic qualities for integration into the core survey on determining factors for initiating HRT.
Authors: Patrick M Archambault; France Légaré; André Lavoie; Marie-Pierre Gagnon; Jean Lapointe; Sylvie St-Jacques; Julien Poitras; Karine Aubin; Sylvain Croteau; Martin Pham-Dinh Journal: Implement Sci Date: 2010-06-11 Impact factor: 7.327