摘要：Numerous questionnaires are available on Knowledge, Attitudes, and Practices (KAP) towards antibiotics' use by adults, but none of these questionnaires is fully validated. We undertook an exhaustive literature review to design a comprehensive KAP questionnaire concerning the personal use of antibiotics in Galicia, North Spain. The Item Content Validity Index (I-CVI) and modified Kappa statistic (K*), confirmed the content validity of the questions (0.78 ≤ I-CVI ≤ 1.00 and 0.78 ≤ K* ≤ 1.00). The S-CVI statistic showed the content validity of the scale (S-CVI/Ave: 0.95). Following face validity and pilot testing, the Test–Retest Reliability in a sample of 145 adults confirmed the reliability of the questions. We carried out Confirmatory Factor Analysis using cross loadings and modification indices to choose the most adequate model in data collected from 844 adults. We estimated the indicators of model fit and demonstrated that the selected model has a good to excellent fit, thus establishing the construct validity. The final version of the questionnaire was highly accepted by the general adult population as reflected by the response rate (95.85%) and the low percentage of unanswered questions (0.4–2.7%). Our fully validated questionnaire could prove useful for research as it permits generating high quality data and reducing measurement error.
其他摘要：Abstract Numerous questionnaires are available on Knowledge, Attitudes, and Practices (KAP) towards antibiotics' use by adults, but none of these questionnaires is fully validated. We undertook an exhaustive literature review to design a comprehensive KAP questionnaire concerning the personal use of antibiotics in Galicia, North Spain. The Item Content Validity Index (I-CVI) and modified Kappa statistic (K*), confirmed the content validity of the questions (0.78 ≤ I-CVI ≤ 1.00 and 0.78 ≤ K* ≤ 1.00). The S-CVI statistic showed the content validity of the scale (S-CVI/Ave: 0.95). Following face validity and pilot testing, the Test–Retest Reliability in a sample of 145 adults confirmed the reliability of the questions. We carried out Confirmatory Factor Analysis using cross loadings and modification indices to choose the most adequate model in data collected from 844 adults. We estimated the indicators of model fit and demonstrated that the selected model has a good to excellent fit, thus establishing the construct validity. The final version of the questionnaire was highly accepted by the general adult population as reflected by the response rate (95.85%) and the low percentage of unanswered questions (0.4–2.7%). Our fully validated questionnaire could prove useful for research as it permits generating high quality data and reducing measurement error.
