Postpartum maternal bonding scale: Development and validation in a low- and middle- income country setting

Introduction: The World Health Organization’s Nurturing Care Framework recommends promoting secure postpartum maternal-infant bonding practices through responsive caregiving for healthy child development. Various instruments exist to assess maternal-infant bonding, but they differ in theoretical foundations and constructs, limiting their broad application and comparability. Notably, there is a lack of bonding instruments developed for low- and middle-income countries (LMICs), where children under five are most at risk of not reaching their developmental potential. This paper describes the development and psychometric validation of a conceptually grounded postpartum maternal bonding scale in an LMIC context and highlights its potential applications in similar settings. Methods: Based on a literature review of bonding concepts and measurement processes, we developed a postpartum maternal bonding scale using a cultural adaptation model for psychometric instruments for children and adolescents. This involved identifying and reviewing existing bonding-related tools, generating items, iterative rounds of expert reviews, and pretesting with postpartum women. We then conducted a final survey with a large sample of women at 42 days postpartum to establish the scale’s psychometric properties. The study was conducted in the Thatta and Sujawal districts of Sindh, Pakistan. Results: An initial pool of 44 items was developed following a literature review and interviews with postpartum women. After multiple rounds of expert review and cognitive pretesting, a 30-item tool was selected for field testing. Using data from 310 postpartum women, we examined the tool’s structure through exploratory (EFA) and confirmatory factor analysis (CFA), leading to a refined 12-item tool. The EFA revealed three factors related to Emotional, Cognitive, and Behavioural bonding. Taking the four highest loading items from each domain, we performed CFA using three models: a first-order model with the three domains, a second-order model, and a bifactor model, which included an overall bonding construct. The bifactor model showed the best fit (comparative fit index = 0.951; root mean square error of approximation = 0.066; standardized root mean square residual = 0.045). This indicates that both an overall bonding construct and specific domains can be measured separately. Pairwise domain correlations were all below 0.67, and internal reliability statistics ranged from 0.63-0.72 (Cronbach’s Alpha) and 0.64-0.77 (global omega). Regression analysis showed associations between bonding scores and factors such as cesarean delivery (reduced behavioural bonding score for mothers having caesarean: -0.94, 95% Confidence Interval -1.86 to -0.01, p-value 0.047), maternal disability (reduced overall bonding score for mothers with severe disability -1.54, 95% CI -3.12 to 0.03, p-value 0.054), and probable postpartum depression (reduced overall bonding score in mothers with probable PPD -1.57, 95% CI -2.70 to -0.45, p-value 0.006). Conclusion: The 12-item postpartum maternal bonding scale (PMBS) is a conceptually grounded instrument. It is a brief, easy-to-administer tool with potential cross-cultural use in low- and middle-income settings after cultural adaptation.