Design and Participants
The present research is a cross-sectional study that was conducted between Junes to November 2020 with the participation of breastfeeding women who had breastfed their babies during the onset of the coronavirus pandemic in the city of Qazvin.
Sample size estimation
The Green’s (27) rule of thumb was used to determine the sample for linear multi variable analysis. Considering 35 predictor variables (K), the sample size was calculated to be 330 persons.
Sampling procedure
In this study, we referred to five comprehensive health centers and specialized obstetrics and gynecology training hospital in the city of Qazvin that had the highest number of clients. We used a convenience sampling method. That is, by referring to the comprehensive health center and Kosar hospital in the city of Qazvin, breast feeding mothers with during the covid-19 pandemic who were willing to participate in the study were selected, and mothers who fed their children with only formula from birth, as well as those who did not answer the questionnaire completely (6 out of 330 persons) were excluded from the study.
Measures & Variables
In the present study, breastfeeding self-efficacy was the dependent variable, and the variables of spouse postpartum social support, perceived social support, anxiety and depression, and fear of Covid-19 were the main independent variables. In addition, demographic and reproductive characteristics and coronavirus infection status were studied as covariates. The tools used for this study, which have gone through the psychometric process of the Persian version and are valid and reliable, consisted of the following items:
A checklist of personal, family, social and reproductive characteristics: This questionnaire includes questions such as age, level of education and occupation of the woman and her spouse, mother’s understanding of the family’s economic status, number of pregnancies and number of live children, pregnancy rank, desire for the current pregnancy (from the perspective of the woman and her spouse), type of delivery, infant’s gender, skin-to-skin contact, previous breastfeeding experience, time of first breastfeeding, history of hospitalization of the baby and its cause, infant’s age, and the birth status of the baby (pre-term, term and post-term). This questionnaire was developed by researchers and professors.
Spouse Postpartum Social Support Questionnaire: This scale was introduced in 2017 by Dennis et al. It has 20 items with a 4-point Likert scale responses from Strongly Disagree to Strongly Agree. The final score is calculated as the total score of the items, which varies between 20 and 80. A higher score indicates more social support from the spouse. This questionnaire was developed based on the theoretical model of social relations and functional components of social support. The content validity was verified based on expert opinion. Psychometric evaluation, including internal stability analysis, exploratory factor analysis (EFA), and concurrent and predictive reliability and validity, showed that the psychometric properties of the English version are reasonable. The Cronbach’s alpha for this scale was 0.96 (28). Eslahi et al provided the Persian version and the psychometric properties of this scale. Confirmatory factor analysis and analysis based on Rasch linear model and concurrent validity, confirmed the acceptable validity of this tool. In addition, Cronbach’s alpha coefficient of 0.94 and the reasonable correlation between items, confirmed the acceptable reliability of this tool (29).
Multidimensional Perceived Social Support Scale (MSPSS) Questionnaire: The Multidimensional Perceived Social Support Scale by Zimet et al. was used to measure the perceived social support (30). This tool has 12 items based on the Likert scale from 1 (Highly Disagree) to 7 (Highly Agree) that measures the support that the person receives from the three sources of family, friends and important people in life. The range of scores obtained from this scale is 12 to 84. The Cronbach’s alpha of the total scale and also the subscales ranged from 0.85 to 0.91 and its reliability was reported through retesting as 0.72 to 0.85 (31). Psychometric properties of this questionnaire were examined by Salimi et al. Cronbach’s alpha coefficients of the three dimensions of social support received from family, friends and important people in life were 89%, 86% and 82%, respectively; which confirmed the reasonable reliability of the Persian version (32).
Anxiety and depression questionnaire: In the present study, the Hospital Anxiety and Depression Scale (HADS) Questionnaire was used to assess the anxiety and depression. HADS was developed in 1983 by Zigmond and Snaith to diagnose anxiety and depression disorders in patients in non-psychiatric clinics. This scale has 14 questions in two subscales of anxiety and depression. Each item is rated based on a 4-point Likert scale from 0 to 3. The maximum score that can be obtained in each subscale is 21. Scores above 11 in each of the subscales mean having a significant psychological illness, scores from 8 to 10 indicate borderline cases, and scores from 0 to 7 are considered normal (33). Bjelland et al. (2002) examined the psychometric properties of HADS in a systematic review. They concluded that this tool is an appropriate tool for assessing anxiety and depression disorders for different groups, including somatic patients, psychotherapy patients, primary care unit clients and the general population (34). Psychometric properties of the Persian version were reviewed and confirmed by Montazeri et al. (2003) (35).
Fear of Covid-19 Questionnaire: This questionnaire was designed by Pakpour et al. In 2020. The tool consists of 7 questions. Participants express their level of agreement using a 5-point Likert scale. Answers include strongly disagree, disagree, neither agree or disagree, agree and strongly agree. The minimum possible score for each question is 1 and the maximum is 5. The total score is calculated by adding the scores of each question (from 7 to 35). A higher score indicates a greater fear of the Covid-19. Psychometric properties of this tool have been approved for the Persian version in the Iranian population (36).
Breastfeeding Self-Efficacy Questionnaire: The breastfeeding self-efficacy questionnaire has 14 items, and it begins with positively loaded questions with the prefix I can always; and it is rated based on a 5-point Likert scale from Always or I am completely sure (score 5) to Never or I am not completely sure (score 1). The lowest and highest scores are 14 and 70, respectively, such that the highest score indicates the highest level of breastfeeding self-efficacy (37, 38). The validity of this questionnaire has been reviewed and confirmed in a study by Varai et al. (16). The reliability of this questionnaire has been investigated in a study by Bastani et al. and a Cronbach’s alpha of 0.87 has been obtained (39).
Ethical Considerations
The implementation of this research has been approved by Qazvin University of Medical Sciences with the ethics code IR.QUMS.REC.1399.079. Initially, breastfeeding women who met the research inclusion criteria were contacted via telephone and the study objectives and working methods were explained. If they wished to participate in the research, the link to the online questionnaire would be sent to them. In order to observe ethical principles of the research, they were convinced that the information of individuals remained confidential and questionnaires were collected anonymously. The contents of the informed consent were listed on the first page of the online questionnaire and the continuation of the questionnaire completion process was considered as consent to participate in the study.
Statistical Analysis
In this study, 324 persons participated. The research data were analyzed using SPSS software, version 24. Mean and standard deviation were reported to describe continuous quantitative variables, and frequency and percentage were reported to describe qualitative variables. First, the normality of the distribution of breastfeeding self-efficacy scores was evaluated and confirmed using central distribution and dispersion indices, histogram graph, and Shapiro-Wilk Test.
In the first step, one-way analysis of variance (ANOVA) and Pearson correlation coefficient were used to examine the association between breastfeeding self-efficacy as a dependent variable and independent variables of the study including spouse postpartum social support, perceived social support, anxiety and depression, fear of Covid-19, Covid-19 infection status, and the demographic and obstetric characteristics. Considering a significance level of 0.05, the variables that had a significant correlation with breastfeeding self-efficacy were selected to be included in the multivariable linear regression model. These variables included spouse postpartum social support, breastfeeding intention, type of infant feeding, social support of the family, and depression.
Afterwards, the multivariable linear regression model was developed by considering the total score of breastfeeding self-efficacy as the dependent variable, and the above variables as independent variables. Multivariable regression model was based on stepwise method. Assumptions of using linear regression method including normal distribution of breastfeeding self-efficacy scores and lack of outlier data. In the initial model, given VIF>10 to examine the collinearity between independent variables, the family social support variable (VIF = 25.9) was removed from the model. After removing this variable, the problem of collinearity between the independent variables in the model was resolved.