Research Design and Study Population
A cross-sectional study was carried out among Malaysian Umrah pilgrims attending a weekly Umrah orientation course organized by private Umrah tour companies from March to June 2018. This study was the second stage of a large study. In the first stage, we conducted exploratory factor analysis of the measurement tool.
Sample Size and Sampling Method
A total of 200 Umrah pilgrims were recruited through a multistage sampling method for the 72 items in the KAP questionnaire for prevention of RTI during Hajj. The sample size for this study was based on a simulation study as recommended by Hair (2010) for CFA. Therefore the sample size for this study was fixed at n = 200 when the anticipated domains were seven or less and items commonality was less than 0.5 and no under identified domains.
Measurement Tool
A self-administered questionnaire for the measurement of the knowledge, attitude and practice of Malaysian Hajj and Umrah pilgrims towards the prevention of RTIs. All the domains, as well as the sub-domains, have been developed and exploratory factor analysis (EFA) was done in the previous part of this same project [17]. The knowledge domain was based on a literature review and a qualitative study to include aetiology, transmission, signs and symptoms, risk factor and prevention measures. The attitude domain was factored on barriers to prevention and self-motivation while the practice domain was developed based healthy lifestyle and protective practices. The final questionnaire consisted of 72 questions and items.
Data Collection Procedures
All data were collected from June 2018 to August 2018. A self-administered questionnaires were distributed to the Umrah pilgrims before their weekly course that met the inclusion criteria. Pilgrims are that aged 18 and above, able to write and speak in Bahasa Malay and are willing to participate are considered to have fulfilled the inclusion criteria. Participants were briefed on the purpose of the study, the procedures, and the confidentiality of their responses. Informed consent was obtained from the participants that are willing to be part of the study prior to the administration of the questionnaire. The pilgrims were also instructed to give their honest responses when answering the questionnaire. The completed questionnaire was immediately retrieved from the participants at the end of the day’s orientation. The time to complete the questionnaire was approximately 10 to 15 minutes.
Data Management and Preliminary Analysis
All data were entered and checked for missing data using SPSS software version 24 and then transferred to R version 3.5.0 for Item Response Theory (IRT) and Confirmatory factor analysis (CFA) analysis. Data analysis was done using R version 3.5.0 in the R Studio environment.
Item Response Theory (IRT)
Considering the unidimensionality of the items consisting of dichotomous responses of the knowledge, the section was analyzed by two-parameter logistic item response theory (2-PL IRT) analysis, using the ltm package version 1.0.0 6.
Confirmatory factor analysis (CFA)
Confirmatory factor analysis (CFA) was conducted to confirm the factorial structure of the KAP questionnaire identified in the EFA published in the other part of this study. The attitude and practice domains were analysed as recommended by lavaan package version 0.5–22.[18] Several indices indicated a good model fit for the construct, they include: the ratio of chi-square to degree of freedom (χ2/df) < 5.0, root mean square error of approximation (RMSEA) ≤0.08, comparative fit index (CFI) >0.9, Tucker Lewis Index (TLI) >0.9, and p >0.05 for the chi-square test [19]. For composite reliability, semTools package version 0.4-14 5-6 was used to determine the Raykov’s rho [19, 20]. Hair et al (2009) suggested that model fitness can be decided by at least a minimum of three different indices. A good relationship between items and respective factors are shown by a standardized factor loading greater than 0.5 as well as a p-value of less than 0.05 and it therefore further proves the validity of the construct. A composite reliability of the domains was calculated with a value of 0.7 and above was considered acceptable [21, 22].