“We All Join Hands”: Perceptions of the Kangaroo Method Among Female Relatives of Newborns in The Gambia

Family support is essential for kangaroo mother care (KMC), but there is limited research regarding perceptions of female relatives, and none published from West African contexts. In-depth interviews were conducted from July to August 2017 with a purposive sample of 11 female relatives of preterm neonates admitted to The Gambia’s referral hospital. Data were coded in NVivo 11, and thematic analysis was conducted applying an inductive framework. Female relatives were willing to support mothers by providing KMC and assisting with domestic chores and agricultural labor. Three themes were identified: (a) collective family responsibility for newborn care, with elder relatives being key decision makers, (b) balance between maintaining traditional practices and acceptance of KMC as a medical innovation, and (c) gendered expectations of women’s responsibilities postnatally. Female relatives are influential stakeholders and could play important roles in KMC programs, encourage community ownership, and contribute to improved outcomes for vulnerable newborns. Based on the concept that perceptions of newborns and their care are influenced by social phenomena, we aimed to construct accounts of participants’ experience, which were collected using in-depth interviews, observations, and reflexive field notes. The study formed part of the formative phase for a randomized controlled trial investigating KMC before stabilization in neonates weighing <2,000 g (Brotherton et al., 2020). The Gambia is the smallest country on mainland Africa, with the fourth highest population density and population of ~2.08 million in 2017 (The Gambia Bureau of Statistics, 2019). The predominant religion is Islam (90%), and the Mandinka are the largest ethnic group (35%) followed by Fula (25%), Wolof (15%), and Jola (10%; The Gambia Bureau of Statistics and ICF International, 2014). Polygamy is commonplace, and the most common household structure for all ethnic groups is patrilineal, multigenerational, and extended family groups residing within a compound of varying size, depending on the domestic cycle (Kea, 2013). Compounds typically include a husband, his wives, his married sons, and their wives and children, any unmarried children, widows or divorced sisters, and other extended family (Kea, 2013; Sear et al., 2002). The neonatal mortality rate in The Gambia reduced from 49/1,000 live births in 1990 to 26/1,000 live births in 2018 (UNIGME, 2019). Approximately 17% of Gambian neonates were born LBW in 2015 (UNIGME, 2019) and 12% born preterm (Chawanpaiboon et al., 2019), with complications of prematurity the second most common cause of neonatal death at the neonatal referral hospital (Okomo et al., 2015). The study took place at the Edward Francis Small Teaching Hospital (EFSTH) neonatal unit, the national neonatal referral unit and the only teaching hospital in The Gambia. Newborns were admitted to the study site from the EFSTH maternity unit (approximately 7,000 births per year) or from other hospitals or home with two thirds of admissions born outside EFSTH (Okomo et al., 2015). At the time of this study in 2017, KMC was not practiced at EFSTH nor widely in The Gambia (Ekholuenetale et al., 2020), although mothers were advised to provide skin-to-skin contact for 30 minutes immediately after feeding for clinically stable, small newborns. Small or sick newborns were cared for under radiant heaters or incubators, often with multiple occupancy, and with few nurses. Eligible participants were adult (>18 years) female relatives of newborns weighing ≤2,000 g who were admitted to the study site between April and July 2017. We used purposive sampling to identify participants by approaching mothers of currently hospitalized neonates or those discharged within the preceding month. They were contacted by the interviewers in person or by phone, and invitations were extended to their female relatives. Women who were willing to participate contacted the interviewers to arrange a convenient time, and transport expenses were provided. Because different generations and family lines may have different perspectives, we aimed to include maternal and paternal relatives from a range of generations. All participants interviewed were from different families and represented a different neonate. Sample size was based on the availability of participants within the study period. Semi-structured interviews were conducted over a 5-week period from July to August 2017 by the interviewers: a non-Gambian female midwife researcher and a multilingual Gambian female field worker. The Gambian interviewer enhanced the credibility of the interviewing team and was able to elucidate and interpret participants’ comments within the cultural context in which they were intended (Guba & Lincoln, 2005). Neither interviewer was involved in the clinical care of the participants or their newborn relatives. A semi-structured interview guide was used with open-ended questions concerning knowledge and perceptions of newborns, care of small newborns, and KMC (Supplementary File I). Written informed consent, including for audio-recording, was obtained in the participants’ preferred language, with impartial witnesses present for illiterate participants. Informed consent documents were in English, with verbal translation to local languages during the consent process, as per standard local consenting practice in view of the most common local languages having no formal written standard in routine use. Interviews were then conducted in Wolof or Mandinka, as preferred by the participant, in a private, nonclinical room at the hospital. A pictorial information sheet was used to assist the discussion (Supplementary File II). The interviews lasted between 30 and 40 minutes (average 37 minutes) and were recorded on an ICDPX 440 Sony digital recorder. The interviews were conducted by the same interviewers with the Gambian interviewer leading the interview and the non-Gambian interviewer present for observation of the interview process and reflexivity. The interviewers worked closely together to ensure understanding of the interview guide, and both were experienced in conducting interviews, including on similar topics. The interviewers were aware that as interviews were conducted in the hospital, participants possibly associated the study with the hospital and despite assurances of confidentiality and independence, this may have led to participants sharing what they thought the interviewers wanted to hear. To try and address this, we attempted to build rapport using a warmup session, and the semi-structured interview style allowed participants to lead portions of the interview. As only one interviewer conducted interviews, we were confident that internal validity of the questions was maintained between sessions. A pilot of two interviews was used to refine the interview guide and to ensure that the Gambian interviewer was familiar with the guide and able to readily translate into the spoken language. After each interview, the interviewers debriefed, which helped maintain reflexivity, improved interview technique, and challenged established assumptions during the analysis and writing. A field diary was kept to document the context and reflections from the interviews, informal conversations with hospital staff and insights into potential findings. Interviews were translated and transcribed into written English text by the same interviewers to ensure consistency and dependability (Tuckett, 2005). Three randomly selected transcripts underwent validation by an independent research nurse fluent in the local languages and English to monitor for accuracy of translation, and no major discrepancies were identified. The use of these research strategies contributed to the rigor of the data collection, especially the reliability and internal validity of data collected (Guba & Lincoln, 2005). All participants’ data were pseudonymized from the time of enrollment with unique study identification codes for confidentiality. All recordings were deleted from the recorder after transcription. Recordings and transcripts were securely stored on an access-restricted, central server at London School of Hygiene & Tropical Medicine (LSHTM). Ethical approval was obtained from the ethics committees at LSHTM (Ref. 12398) and The Gambia Government/Medical Research Council Joint Ethics Committee (Ref. 1535). Thematic analysis was conducted using an inductive framework (Braun & Clarke, 2006), allowing codes and themes to develop directly from the data. Due to time constraints, the full transcripts were read and coded by one researcher (the non-Gambian midwife interviewer), who then worked in a cell of qualitative researchers to map, reflect, and refine codes and interpretations of themes. This process was used to help strengthen the reliability of the coding (Guba & Lincoln, 2005). Transcripts were read twice with line-by-line coding on the third reading using NVivo 11 qualitative data analysis software (QSR International Pty Ltd.). The fourth reading focused on merging and reorganizing codes and examining unexpected findings and discrepancies. Codes were then collated into themes, which were refined through iterative analysis and thematic mapping. Themes evolved both directly from the data on a semantic level from explicit meanings and a latent level from interpretation of underlying patterns and ideas (Braun & Clarke, 2006). Quotes were selected to reflect the refined themes. This article was prepared in consultation with Standards for Reporting Qualitative Research (O’Brien et al., 2014).