This study uses an interrupted time series methodology to demonstrate the impact of implementing the BFHI package among an initially EBF-averse patient population. The absolute rate of EBF at hospital discharge increased from 2.4–49.0% from the first month to the final month of data recorded, and the above analysis conveys the improved trends in EBF practices over time. This analysis showed overall increases in the adoption of EBF during both the “Intervention” and “Follow-Up” periods, as well as a statistically significant change in the monthly increase in EBF adoption when compared to baseline. Moreover, these observed behavioral trends in the adoption of EBF practices persisted up to two years following the initial intervention period.
Implementation of the full BFHI package is a complex intervention. Each of the WHO/UNICEF-validated “10 Steps” involves multiple discrete actions and changes that require buy-in from hospital administration, healthcare staff, and patients. Using Powell et al.’s framework from the implementation science literature, this process included implementation strategies related to planning, education, restructuring, and quality management (29). A timeline of when specific interventions were implemented is available in Additional File 1; however, many of these changes were gradual, and the effect of the BFHI package was likely cumulative over time.
From the CMC team’s perspective, a key factor in the successful implementation of BFHI and the resulting increase in EBF rates was the engagement of a core team of care providers. This multidisciplinary team included midwives, nurse educators, pediatricians, and obstetricians. They served as “champions” for the BFHI process and led the process of policy revision and implementation of new procedures. The CMC team cited pediatricians as very influential in women’s decisions to breastfeed, a finding that is supported by several previous studies in Lebanon (7, 11). Hospital administration and management were also engaged in the BFHI process, allowing for institution-wide reforms and unified messaging. Positive feedback from future mothers and their families regarding CMC’s BFHI improvements and the surrounding promotional campaign has provided the momentum for the process to continue.
Several challenges arose during BFHI implementation, many of which reflect cultural barriers to breastfeeding that persist in Lebanon. For practices such as skin-to-skin contact after birth, adoption of such practices by women occurred quickly, demonstrating that successful behavior change is possible. However, many CMC patients, family members, and even occasionally outpatient healthcare providers still expressed doubt that breastmilk is nutritionally sufficient for newborns. Some mothers also resisted the rooming-in policy, expressing the need to rest while their infant is cared for in a nursery. The CMC team addressed these challenges by adopting patient-centered communication techniques and focusing on the benefits of breastfeeding for both mother and newborn. Women were empowered to make their own decisions and were supported accordingly. Continuing educational outreach, particularly on the sufficiency of breastmilk for infant nutrition, will be important to address these challenges and further increase EBF rates in the future.
Another challenge in implementing this initiative is reflective of the fact that CMC employs a combination of full-time clinicians in addition to part-time providers from the community. For those part-time providers (who primarily practice outside the hospital but bring their patients to CMC for delivery), it has been more difficult to become engaged in the Phase 2 activities aimed at changing clinical practices, and accordingly, EBF adoption has been slower in this group. Further targeted outreach to these providers is planned.
Finally, an ongoing audit of the long-term breastfeeding practices of new mothers in the outpatient setting (i.e., following discharge from the hospital) is still actively being developed. Although our study and others show that BFHI can be successful in increasing EBF in early infancy, maintaining EBF through the recommended six months of age remains a challenge (7, 30). For example, in their BFHI trial in Saudi Arabia, Mosher et al. showed a significant decrease in EBF rates at six months among both the intervention (BFHI hospital) and control (non-BFHI hospital) groups (21). Systematic reviews have shown the importance of ongoing postpartum support to mothers for EBF continuation (18, 31), corresponding to Step 10 in the BFHI package. In Lebanon, a previous study has shown the positive impact of a telephone support hotline on EBF rates (8), an encouraging result for the hotline set up by CMC as a part of this study. A recently-completed randomized controlled trial at two other hospitals in Lebanon has shown the effectiveness of a comprehensive package of professional and peer postpartum support on EBF knowledge and practice (30, 32). Looking forward, CMC will evaluate such initiatives for their utility in maintaining breastfeeding practices after new mothers have left the inpatient setting.
Strengths and Limitations
This study has a number of important strengths. Although data were analyzed retrospectively, they were collected continuously throughout the initiative, which allows for evaluating trends over time rather than mere pre/post comparisons. Interrupted time series is a strong quasi-randomized study design when the use of a control group is not feasible (33, 34). In addition, this study’s use of patient data allowed for an objective measure (as recorded by healthcare providers) of EBF practice, compared to mothers’ self-report in post-discharge surveys used in other studies (21). The EBF data have been strictly audited as a part of CMC’s partnership with JHI and validated by the quality department at CMC. Finally, the study sample size was large, with 2,002 live births included.
It is also important to acknowledge several limitations of this study. It was conducted retrospectively and in a single center, meaning that no control group was present; however, the interrupted time series analysis is a methodologically appropriate means for addressing this limitation (33). In addition, because de-identified hospital reporting data were used, we were not able to measure demographic information of mothers or infants and account for any demographic shifts over time. This is an important avenue for further research, which could include prospective enrollment of participants and collection of individual-level data, in order to analyze trends among specific subgroups of mothers. Finally, our outcome measure in the evaluation of this program is the EBF rate at discharge (typically 2–4 days after birth), which, although clearly associated with EBF rates at one and six months of age (7), is admittedly a surrogate marker for such practices. Future studies at CMC should include post-discharge follow-up through six months of age, to allow for comparison with national EBF trends.