Background: The theory of planned behavior (TPB) has been used to explain breastfeeding behaviors in Western cultures. Theoretically-based investigations in other groups are sparse.
Objective: To evaluate cross-cultural application of TPB-based models for breastfeeding duration among new mothers in Hong Kong.
Method: First-time breastfeeding mothers (N = 209) with healthy newborns provided self-reports of TPB predictor variables during postpartum hospitalization and information about breastfeeding experiences at 1, 3, 6, 9, and 12 months postdelivery or until they weaned. Three predictive models were proposed: (a) a strict interpretation of the TPB with two added proximal predictors of breastfeeding duration; (b) a replication with modification of the TPB-based model for more fully employed breastfeeding mothers from a previous study (Duckett et al., 1998); and (c) a model that posited perceived control (PC) as a mediating factor linking TPB motivational variables for breastfeeding with breastfeeding intentions and behavior. LISREL was used for the structural equation modeling analyses.
Results: Explained variance in PC and duration was high in all models. Overall fit of the strict TPB model was poor (GOFI = 0.85). The TPB for breastfeeding employed women and the PC-mediated models fit equally well (GOFI = 0.94; 0.95) and residuals were small (RMSR = 0.07). All hypothesized paths in the PC-mediated model were significant (p < .05); explained variance was 0.40 for perceived control and 0.36 for breastfeeding duration.
Discussion: Models were interpreted in light of the TPB, previous findings, the social context for breastfeeding in Hong Kong, and statistical model-building. Cross-cultural measurement issues and the need for prospective designs are continuing challenges in breastfeeding research.
Publications ranging from early Chinese medical texts (Gartner & Stone, 1994) to current documents from the World Health Organization (WHO) (2001b) have recommended breastfeeding as the infant feeding method that best provides essential nutrition and health benefits. The health benefits of breastfeeding are dose-dependent, with infants who breastfeed longer and more fully demonstrating greater reductions in disease (Lawrence, 1997). Based on consistent research findings in industrialized and developing countries, both the American Academy of Pediatrics (1997) and WHO recommend exclusive breastfeeding during the first 6 months of life, with continued breastfeeding for another 6 months or more while solid foods are added to the diet.
Despite these evidence-based recommendations, global breastfeeding initiation rates and patterns of feeding are variable (WHO, 2001b). Among Asian industrialized centers, Singapore has the highest breastfeeding initiation rate (61%) (Fok, 1997). In Hong Kong, breastfeeding initiation rates have increased from 19% (in 1992) to 41% (in 1998) according to Chee and Horstmanshof (1996), who also reported that exclusive breastfeeding is rare and duration is short. Most mothers breastfeed only a few weeks (Chan, Nelson, Leung, & Li, 2000). Current breastfeeding promotion activities have been moderately successful in increasing initiation rates, but breastfeeding duration has not been affected (2000).
Although a vast body of research exists concerning the demographic variables influencing women’s infant feeding decisions (e.g., prenatal education, maternal educational level, support, age, delivery type) little is known about how these factors affect women’s infant feeding choices in Hong Kong. As a result of public-funded healthcare, all pregnant women have access to early prenatal care and childbirth classes that include one session on breastfeeding. Chan and colleagues (2000) reported that women who breastfed for 3 months or longer completed at least a secondary education and worked in professional or paraprofessional jobs. Unlike researchers in North America, no differences in age or parity between breastfeeding and formula-feeding mothers were found. However, the sample was small (N = 44) and not randomly selected. Variables such as rooming-in and avoiding supplementation have been associated with longer breastfeeding duration elsewhere in the world but have not been studied in Hong Kong.
Maternal motivation and family situations influence the breastfeeding decisions of Hong Kong women (Fok, 1997). Chinese mothers often spend the first month after delivery convalescing at home, away from societal obligations (Fok, 1996). This is referred to as “doing the month.” Traditionally during this time, the new mother is encouraged to rest in bed and to take special foods that will enhance recovery from the birthing experience and increase milk production (Fok, 1997;Holroyd, Fung, Lam, & Sin, 1997). Chee and Horstmanshof (1996) described this practice as creating a “period of social isolation, when others offer to care for the infant” so that the mother can rest (p. 8). Use of a confinement maid or a nanny to care for the infant is common among middle-income women.
Decisions about breastfeeding may be influenced by maternal position within the family and roles of other family members. Participants in an ethnographic study (N = 7) of current postpartum practices in Hong Kong reported that they adhered to the traditional dietary and activity restrictions if encouraged to do so by their spouse or older relatives (Holroyd et al., 1997). The mother-in-law, traditionally the most powerful family member, exerted a special influence on infant feeding decisions. Fok (1996) interviewed more than 4,000 Chinese women in Singapore during a 5-year period and found that mothers who wanted to breastfeed would not do so if any significant other felt they should not. This maternal behavior is attributed to a characteristic Chinese desire to “save face” (a complex etiquette for social interactions) by agreeing with those of a higher social status (i.e., husband, mother-in-law; p. 16).
In Hong Kong, 75% of childbearing-aged women work full-time (5.5 days a week); little part-time work is available (Hong Kong Census and Statistics Department, 2001). Government-mandated maternity leave is 10 weeks with a minimum of 2 weeks taken before the woman’s expected date of confinement (Hong Kong Labour Relations Promotion Unit, 2002). Compliance with maternity leave regulations, however, varies with employers. Work environments rarely provide support for lactating women (Tarrant, Dodgson, & Tsang, 2002). Of 235 Hong Kong breastfeeding mothers surveyed about their reasons for early weaning, 67% reported that their maternity leave (usually 6 weeks for study participants) was insufficient time to establish nursing (Chee & Horstmanshof, 1996). Chan et al. (2000) found that women (44%) intending to exclusively breastfeed for 3 months or more had completely weaned their infants by 6 weeks and other (13%) were supplementing breastfeeding with formula.
Joan E. Dodgson, PhD, MPH, RN, IBCLC, is Assistant Professor, Duke University School of Nursing, Durham, North Carolina.
Susan J. Henly, PhD, RN, is Associate Professor, School of Nursing, University of Minnesota, Minneapolis.
Laura Duckett, PhD, MPH, RN, is Associate Professor and Director of Research, School of Nursing, University of Minnesota, Minneapolis.
Marie Tarrant, MN, RN, is Assistant Professor, Department of Nursing Studies, The University of Hong Kong, SAR, PRC, and PhD Candidate, School of Nursing and Public Health, Edith Cowan University, Perth, Australia.
Accepted for publication February 13, 2003.
The authors thank Shirley F. Tsang, MBA, Xiu-Hong Peng, MPhil, Wing K. Choi, Hang C. Poon, Yee W. Chiu, and Ka. L. Lee, Research Assistants in the Department of Nursing Studies at the University of Hong Kong, who collected data.
Funded by The Research Committee, University of Hong Kong, Hong Kong SAR People’s Republic of China. Dr. Dodgson was an Assistant Professor of Nursing at Hong Kong University while conducting this study.
Corresponding author: Joan E. Dodgson, PhD, MPH, RN, IBCLC, Duke University School of Nursing, 811 9th Street, Suite 200, Durham, North Carolina 27705 (e-mail: firstname.lastname@example.org).