Factors associated with father–infant bonding during the COVID-19 pandemic: an internet-based cross-sectional study in Japan – Scientific Reports

Study design

This cross-sectional study used data from the Japan COVID-19 and Society Internet Survey (JACSIS), a nationwide cross-sectional survey conducted in 2021³⁴. Additionally, we used data from pregnancy and maternity surveys, which include data on fathers whose partners were currently pregnant or postpartum¹³. This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist³⁵ (Supplementary Information).

Sample

The study samples for this survey were retrieved from the pooled panels of the Internet research company Rakuten Insight, Inc., Hong Kong, which holds approximately 2.3 million panelists who voluntarily registered in exchange for small incentive points when completing questionnaires^13,34. Figure 1 presents a flowchart of the data extraction process. We excluded those whose youngest child had not yet been born (n = 560). The following were also considered invalid responses and excluded from the sample: respondents who did not select a specific option when asked to do so in a dummy question (n = 125); respondents who reported less than 22 weeks of delivery (n = 28); respondents who reported data on the apparent imbalance between weeks of delivery and birth weight (n = 33); respondents who reported abusing all seven substances (alcohol, sleeping medications, opioids, sniffing paint thinner, legal-high drugs, marijuana, and cocaine/heroin); and respondents who selected all 16 medical histories on the list (n = 15). To identify father–infant bonding during the COVID-19 pandemic, we included data after the first COVID-19 cases were reported in Japan, which were the partners of those who gave birth between January 16, 2020, and August 18, 2021 (n = 1055). No missing data were observed.

Variable measurement

Outcome

The Mother-to-Infant Bonding Scale (MIBS) was used to measure father-to-infant bonding. The Japanese version of Mother-to-Infant Bonding Scale (MIBS-J) was developed by Yoshida et al. (2012) based on the Mother-infant Bonding Questionnaire by Kumar³⁶. This scale was originally developed to measure bonding between mothers and children but has also been validated for fathers¹⁰. The scale consists of 10 items, such as “I feel loving towards my baby”, which are answered on a four-point scale of “almost always strongly agree”, “sometimes strongly agree”, “sometimes slightly agree” and “never agree¹⁰”⁸. The maximum score was 30 points, with higher scores corresponding to weaker bonding¹⁰.

Kitamura et al.³⁷ reported that the MIBS-J was structured into two factors, “Anger and Rejection” (items 2, 3, 5 and 7) and “Lack of Affection” (items 1, 6, 8, and 10) by exploratory factor analysis. As these are different aspects of the fathers’ bonding^23,37, MIBS-J scores in this study were calculated separately for Anger and Rejection and Lack of Affection in this study.

Covariates

Previous studies have found that the father’s presence at birth^11,12, participation in paternity classes^11,12,13,14, number of children⁸, parental leave status¹³, and developmental delay in the child³⁸ affected father–infant bonding. As paternal attributes, medical history including depression²⁹ and the Edinburgh Postnatal Depression Scale^10,29,39 were reported as factors associated with father–infant bonding. The Edinburgh Postnatal Depression Scale is a 10-item self-report scale for screening postpartum depression³⁹ in mothers and fathers, with a cut-off score of 8/9 and has been found to identify probable postnatal depression⁴⁰. Partner relationships^8,9,14 and partner violence²⁹, anxiety⁷, and fathers’ adverse childhood experiences^20,41,42 have also been shown to be associated with father–infant bonding.

Predictors

We describe the predictors conceptually separately from the covariates to convey the purpose of this study more clearly. The following were used as predictors: desired pregnancy, child abuse-related behavior, and variables such as emergency cesarean section, premature birth, low birth weight and Neonatal Intensive Care Unit (NICU) admission, which were considered to potentially impede early contact between the infant and father. Family adaptability, partnership, growth, affection, and resolve (Family APGAR) also included as predictors. Family APGAR is a five-item scale used to determine self-reported family dysfunction, with higher scores indicating highly functional families⁴³. The variables were related to the recent COVID-19 pandemic, history of COVID-19 infection, anxiety about COVID-19 infection, and the impact of the COVID-19 pandemic on childcare and work, including teleworking.

Statistical analysis

The study participants were divided into two groups, depending on their partners’ parity because these differences have been reported as effect modifiers of father–infant bonding⁸. Descriptive statistics for all variables included in this study were mean standard deviation (SD) and sample size (n, %). For comparisons between the two groups, the student’s t-test was used for continuous variables and a chi-square test was used for categorical variables. The predictors and covariates were entered into the linear (Gauss–Markov-type) regression model as follows. In the first step, age and education level were entered using the forced entry method to examine associations with factors from the MIBS-J as demographic variables. Then, in Step 2, the above variables, which have been shown to be associated with father–infant bonding in the previous studies, were added to the Step 1 model with forward–backward stepwise variable selection approach while keeping the variables in the Step 1. The input/removal criteria during the stepwise iterations were measured based on the distribution of F-statistics with 5% for the input and 10% for the removal, respectively. Two-sided p value of < 0.05 was considered statistically significant. Multicollinearity was tested using variance inflation factor (VIF), and no multicollinearity was observed. Statistical analysis was conducted using IBM SPSS statistics Version 23.

Ethical approval

This study was approved by the Bioethics Review Committee of Osaka International Cancer Institute, Japan (approval number: 20084). Informed consent was obtained electronically from all participants prior to the survey. Data were collected anonymously. This study was performed in accordance with the ethical guidelines for medical and health research involving human subjects enforced by the Japanese Ministry of Health, Labour, and Welfare, and the Declaration of Helsinki.