Page 103 - IJPS-7-2

International Journal of
Population Studies Knowledge, attitudes, practices, and COVID-19 in Lebanon

variables. In the second logistic regression, attitude was Table 1. Sociodemographic characteristics of the study
taken as the dependent variable, and knowledge and sample
sociodemographic characteristics as the independent Sociodemographic characteristics Frequency Percentage (%)
variables. In the third logistic regression, practice was
selected as the dependent variable, and knowledge and Total 405 100.0
sociodemographic characteristics as the independent Age 405 28.38 (12.02) a
variables. The fourth model considered the practice score Sex
as the dependent variable and the knowledge and attitude Male 82 20.2
as the independent variables. Variables that showed p <0.2 Female 323 79.8
in the bivariate analysis were included in the multivariable Marital status
models to decrease potential confounders. Afterward,
another set of three logistic regressions, taking the same Currently not married 276 68.1
dependent and independent variables, was conducted after Currently married 129 31.9
stratification by sex. In all cases, models’ equations for Education level
significant variables were presented. Non-university level 42 10.4
University level 363 89.6
3. Results Employment status

3.1. Sample description Employed, medical field 65 16.1
The mean age was 28.38 ± 12.02 years, and most participants Employed, non-medical 114 28.1
were female (79.8%) and single (68.1%). Only 10.1% had Unemployed 226 55.8
a family member who was infected by COVID-19, and Family member in the medical field
only 27.9% had a family history of COVID-19. Table 1 Yes 114 28.1
summarizes all sociodemographic characteristics. No 291 71.9

3.2. Principal component analyses of the KAP scales Living place
Rural 108 26.7
Using the principal component analysis (PCA), the
COVID-19 knowledge scale items produced eight factors Urban 297 73.3
that had an eigenvalue over 1. The first component Religion
(knowledge of preventive measures) explained the most Christian 71 17.5
variance by 15.89%, followed by the second component Muslim 254 62.7
(knowledge of the transmission mode of the virus), 7.97%. Other 80 19.7
The total variance was 58.38%. The Cronbach’s alpha value Tested positive for COVID-19
of the COVID-19 knowledge scale was 0.626 (Table S1). Yes 41 10.1
The COVID-19 attitude scale items produced two factors No 364 89.9
that had an eigenvalue over 1. The total variance was
56.55%, with a Cronbach’s alpha value of 0.615 (Table S2). Ever tested for COVID-19
Regarding the COVID-19 practice scale, it could yield four Yes 164 40.5
components with eigenvalues greater than one. The first No 241 59.5
component (protective measures) accounted for 37.77% of Having a family member who has
the total variance, while the second component (practice been infected with COVID-19
of preventive measures) contributed to 12.53%. The Yes 113 27.9
Cronbach’s alpha value of the COVID-19 practice scale No 277 68.4
was 0.899 (Table S3). Cronbach’s alpha values of the FCV- I do not know 15 3.7
19S and CAS scales were 0.856 and 0.846, respectively. Household crowding index 405 1.14 (0.55) a

a
3.3. Bivariate analyses Note: Mean (SD) of age and household crowding index
The mean scores of knowledge, attitudes, and practices were
16.29 ± 2.86, 16.5 ± 1.8, and 100.78 ± 12.11, respectively. with COVID-19 patients were significantly associated
No significant association was found between the KAP with higher knowledge scores (Table 2). However, a higher
scores and sex (Figure 1). Married status, university degree, household crowding index and higher anxiety were related
Christian religion, older age, and direct or indirect contact to lower knowledge scores. Older age, being married, being

Volume 7 Issue 2 (2021) 97 https://doi.org/10.36922/ijps.v7i2.342