Breastfeeding Interpersonal Communication, Mobile Phone Support, and Mass Media Messaging Increase Exclusive Breastfeeding at 6 and 24 Weeks Among Clients of Private Health Facilities in Lagos, Nigeria

Background: Although most health facilities in urban Nigeria are privately owned, interventions to promote optimal breastfeeding practices in private facilities have not previously been implemented. Objectives: We tested the impact of a breastfeeding promotion intervention on early initiation of breastfeeding and exclusive breastfeeding among clients of private facilities in Lagos, Nigeria. Methods: The intervention included training for health-care providers on the Baby-Friendly Hospital Initiative and breastfeeding counseling skills, provision of interpersonal communication and support to women at facilities and on WhatsApp, distribution of behavior change communication materials, and mobile phone and mass media messaging. We used logistic regression models adjusted for clustering to measure intervention impact in a cohort of women (n = 1200) at 10 intervention and 10 comparison facilities interviewed during their third trimester and at 6 and 24 weeks postpartum. Results: The intervention significantly increased the percentage of infants who were exclusively breastfed at 6 weeks (83% intervention; 76% comparison; P = 0.02) and 24 weeks (66% intervention; 52% comparison; P < 0.001), but had no impact on early initiation of breastfeeding (35% intervention; 33% comparison; P = 0.65). Among infants who were exclusively breastfed at 6 weeks, the odds of continued exclusive breastfeeding at 24 weeks were higher in the intervention arm than in the comparison arm (OR, 1.6; 95% CI: 1.2-2.1). Infants had increased odds of being exclusively breastfed at 6 weeks if their mothers discussed breastfeeding with a private health provider (OR, 2.3; 95% CI: 1.5-3.4), received text or WhatsApp messages about breastfeeding (OR, 1.7; 95% CI: 1.0-2.7), or heard breastfeeding radio spots (OR, 4.2; 95% CI: 1.2-14.7). Infants had increased odds of exclusive breastfeeding at 24 weeks if their mothers participated in a WhatsApp breastfeeding support group (OR, 1.5; 95% CI: 1.0-2.2). Conclusions: A breastfeeding intervention in private health facilities in Lagos increased exclusive breastfeeding. Implementation of breastfeeding interventions in private facilities could extend the reach of breastfeeding promotion programs in urban Nigeria. This study was conducted within the same geographic area as an ongoing, cluster-randomized impact evaluation of Alive & Thrive's overall IYCF program in Nigeria (registered at clinicaltrials.gov as {"type":"clinical-trial","attrs":{"text":"NCT02975063","term_id":"NCT02975063"}}NCT02975063). The research described here is a quasi-experimental longitudinal cohort study of women interviewed in their third trimester of pregnancy and at 6 and 24 weeks postpartum to measure breastfeeding intentions and practices (registered at clinicaltrials.gov as {"type":"clinical-trial","attrs":{"text":"NCT04835051","term_id":"NCT04835051"}}NCT04835051). The study was conducted in 20 private health facilities (10 intervention and 10 comparison facilities) in Lagos State, Nigeria. It followed the principles for implementation science in nutrition described by Tumilowicz et al. (38), including initiation and scoping, planning and design, implementation, and discussions of scale-up by the state Ministry of Health. The breastfeeding promotion intervention was implemented by Equitable Health Access Initiative in collaboration with Alive & Thrive from May 2019 to April 2020. Figure 1 shows the project's theory of change, including intervention inputs, processes, outputs, and expected outcomes and impacts. The intervention was designed to strengthen the capacity of private health-care providers to offer high-quality breastfeeding counseling. The intervention consisted of several components: 6 hours of initial and 2 hours of quarterly refresher training for facility managers and staff on implementation of BFHI and breastfeeding counseling skills; provision of interpersonal communication and counseling in person and on WhatsApp by health facility staff; distribution of BCC materials; delivery of mobile phone messages; and broadcasting of mass media messaging. More than 150 facility managers and health-care providers in the antenatal, delivery, immunization, and pediatric outpatient departments at the intervention facilities were trained on BFHI and breastfeeding counseling and support techniques. Facility managers received coaching to develop policies and procedures for implementing the 10 steps of BFHI. Women in the intervention group received in-person breastfeeding counseling and support at the facilities during antenatal, postnatal, and well- and sick-child visits. BCC materials at the facilities included posters and counseling cards, which were part of a state-wide mass media campaign and were provided to health-care providers in the intervention facilities to reinforce IYCF messages. Women received breastfeeding messages on foldable, pocket-sized cards provided to them at the facilities. Women were also invited to participate in WhatsApp breastfeeding support groups, which were organized and managed by the staff member in each intervention health facility who was designated as the breastfeeding champion. The WhatsApp support groups provided women with an opportunity to meet as a group, receive information and support related to breastfeeding, and get answers to their breastfeeding questions. The women and their influential family members (male partners and the women's mothers or mothers-in-law) received text messages. Text messages by phone or WhatsApp were sent in bulk and were 1-way, but WhatsApp support groups offered 2-way communication through the texting function. Women and influential family members gave informed consent to receive the text messages. Mass media television spots on breastfeeding were played on LED screens at the intervention facilities. Theory of change for the Alive & Thrive private provider breastfeeding promotion study in Lagos, Nigeria. Abbreviations: BCC, behavior change communication; BFHI, Baby-Friendly Hospital Initiative; EHAI, Equitable Health Access Initiative; IYCF, infant and young child feeding. In addition to the specific intervention components delivered in intervention facilities during this study, Alive & Thrive's overall IYCF program was ongoing in Lagos State. Alive & Thrive's BCC strategy was based on behavioral science, systems strengthening, and social marketing (39). It included strategic use of data, interpersonal communication, social mobilization, mass media, and policy advocacy (40). Alive & Thrive did not provide support for breastfeeding promotion activities in comparison facilities during this study. Facilities in the comparison areas continued with their usual breastfeeding activities, such as individual breastfeeding counseling offered during provider visits, if at all, with no implementation of BFHI. Alive & Thrive's IYCF program included television and radio spots that were publicly broadcast throughout the state. In the areas where the intervention facilities were located, Alive & Thrive also supported IYCF community mobilization activities, which included sensitization and training of traditional and religious leaders on IYCF to improve awareness of optimal IYCF practices in their communities. In the intervention areas, selected doctors, nurses, midwives, community health extension workers, and traditional birth attendants in public and private health facilities were trained following comprehensive IYCF counseling training manuals to provide counseling to pregnant and breastfeeding mothers. The trained providers were expected to cascade the training to other health providers within their facilities. Data were collected during the third trimester of pregnancy and at 6 and 24 weeks postpartum. The main outcomes of this study were early initiation of breastfeeding and exclusive breastfeeding at 6 and 24 weeks, measured using the WHO IYCF questionnaire (41). Secondary outcomes included breastfeeding intentions, collected during the third-trimester survey, and breastfeeding knowledge, collected during the third-trimester, 6-week, and 24-week surveys. Breastfeeding intentions were measured using the breastfeeding duration component of the Infant Feeding Intentions scale (42). Breastfeeding knowledge questions were adapted from Alive & Thrive impact evaluations in Nigeria and other countries, and they used an open-response format (19–21, 36). The study was conducted in 20 private health facilities purposefully selected from the intervention and comparison areas that had been assigned as part of the overall Alive & Thrive Nigeria impact evaluation. For this study, we chose 10 facilities in the intervention area and 10 in the comparison area that provided maternity and pediatric services, such as antenatal care, postnatal care, and immunizations; were registered with the Association of General and Private Medical Practitioners of Nigeria and the Health Facility Monitoring and Accreditation Agency; had a monthly average of ≥40 antenatal clients and ≥20 deliveries over 3 months; and agreed to participate in the research. The selected facilities provided the research team with the current number of pregnant women in their third trimester. In the 3 lowest-volume facilities (i.e., those with <35 eligible clients) in each study arm, we attempted to enroll all pregnant women in their third trimester. For all other facilities, the number of women sampled was proportional to the facility size within each study arm. All eligible women attending antenatal care on the days of data collection were invited to participate until the target enrollment at each facility was achieved. Women who were attending antenatal care at the study facilities were approached by interviewers in the waiting room and invited to a private area for recruitment into the study. Women were eligible to enroll if they were ≥18 years old, in their third trimester of pregnancy, and current clients of a private health facility selected for the study. At 6 and 24 weeks postpartum, the women continued to be eligible if their infant was alive and neither the mother nor infant had a health condition that caused breastfeeding to be contraindicated. Women remained in the study regardless of where they sought postnatal care or well-child services. We calculated the sample size for our 2 main outcome variables—early breastfeeding initiation and exclusive breastfeeding—using prevalence estimates from Alive & Thrive's 2017 population-based survey (36). We adopted the larger required sample size, which was for early breastfeeding initiation. To detect a difference of 8 percentage points (pp; from 45% to 53%) in the proportion of women who initiated breastfeeding within 1 hour of delivery in the intervention and comparison health facilities with 80% power, an α of 0.05, and an intraclass correlation of 0.005, we required 960 women across a total of 20 health facilities, which are considered clusters. We added 25% to the sample for attrition because the study took place in an urban area where we expected participants to be challenging to track. This resulted in a final sample size of 1200 women (600 per study arm). The evaluation was conducted by RTI International and Datametrics Associates Ltd. Data collectors were not blinded to the evaluation design because Alive & Thrive BCC materials were visible in intervention facilities. The data collection team was divided into groups of 3 to 4 interviewers and 1 supervisor. Interviewers and supervisors were trained for 5 days before the first round of data collection (i.e., during the women's third trimester). They participated in a 3-day refresher training before 6-week postpartum data collection and a 3-day refresher before 24-week postpartum data collection. Each training included a pilot of data collection procedures. Survey questionnaires were used to obtain data on women's breastfeeding intentions (third trimester) and practices (6 and 24 weeks), their breastfeeding knowledge, their exposure to the intervention, and demographic characteristics. Intervention exposures to interpersonal communication at the facility, mobile phone messaging and support, and mass media were measured at each survey wave. Demographic questions were taken from the Nigeria Demographic and Health Survey (43), and the household hunger scale was used to measure food security (44). Data collection tools were translated into Yoruba by 2 translators, who translated the tools separately, reviewed each other's translations, identified differences, and agreed on the best translation. The questionnaire was available in English and Yoruba. Interviewers administered the questionnaires and entered the responses into electronic tablets using Open Data Kit. Supervisors spot-checked 10% of the questionnaires and reviewed all completed questionnaires before uploading them to the server. The majority of the interviews were conducted face to face. Interviewers conducted approximately one-quarter of the 24-week interviews by phone because travel restrictions related to the coronavirus disease 2019 pandemic were put in place in Lagos in mid-March 2020. Interviewers already had the participants’ phone numbers, which they had been using to make appointments for the 6-week and 24-week interviews. Ethical approval for the research was obtained from the RTI International institutional review board and the Lagos State University Teaching Hospital Health Research and Ethics Committee, and permission was obtained for phone-based data collection prior to transitioning during the pandemic. The consent form was read aloud to participants, and they provided written consent at enrollment. All analyses were conducted in Stata MP Version 16.0 using survey commands to account for clustering effects within facilities. At each round of data collection, data were analyzed cross-sectionally using regression models to determine the intervention impact on breastfeeding knowledge, intentions, and practices and the association of specific intervention exposures with breastfeeding outcomes. Exposure variables for 6 and 24 weeks were constructed so that they were cumulative from the start of the intervention. Tests for differences were not conducted for variables with cell sizes less than 5, except for third-trimester analyses of maternal education and household hunger, which were performed using clustered χ² tests. Longitudinal models were fit to examine the relationship between breastfeeding intentions, measured during the third trimester, and exclusive breastfeeding practices at 6 and 24 weeks. To determine whether adjustments to regression models were needed to correct for unintentional bias that may have occurred from the nonrandomized design, the demographic characteristics of participants measured during the third trimester were used to calculate inverse probability weights. We found that applying inverse probability weights was not necessary to balance the study arms for key outcome variables; therefore, we conducted the analysis without adjustment. Women who missed their 6-week postpartum interview were allowed to rejoin the sample at 24 weeks. Characteristics of women interviewed at all 3 time points were compared with women interviewed at 1 or 2 time points. There were no differences between those with and without missing surveys at the P < 0.05 level with respect to key outcomes and demographic variables.