The data presented here suggest that there is widespread understanding of the need for clean delivery practices to reduce the risk of infection to both mothers and their babies. This understanding expands beyond the health care setting, where it might be expected, and into the rural community. Despite this understanding, when looking at six key clean delivery practices that can be targeted for intervention – hand washing/use of gloves during delivery, delivering on a clean surface, sterile cord cutting, appropriate cord tying, proper cord care following delivery, and infant bathing and cleanliness – each appears to have room for improvement in this rural area in Northern Ghana. Notably, behaviors appeared to differ based upon delivery location – with facility-based deliveries and deliveries attended by skilled birth attendants much more likely to comply with clean delivery recommendations than home deliveries and those attended by traditional or untrained attendants.
In our study, the use of gloves during delivery and hand washing during and after delivery were mentioned infrequently, despite repeated discussion of the need for cleanliness to avoid infant infections. This may be due in part to the open-ended nature of the interview tool – respondents were not directly asked, “What about hand washing?” However, it is noteworthy that discussions prompted by “What needs to be clean during delivery?” rarely included discussion of hand washing. Respondents frequently mentioned the need for a clean delivery surface, including explicit discussion of avoiding delivering in the dirt. However, at least 3 of the 35 recently-delivered mothers in our sample delivered their infants on the way to the facility, including deliveries alongside the road. Cord cutting was done with a variety of tools, the most common of which were razor blades or scissors. Cord tying also utilized a variety of non-sterile materials, including string, rope, thread, twigs, and clamps. These data suggest that applying traditional salves to the cord – including shea butter, ground shea nuts, local herbs, local oil, or “red earth sand” – is still a common practice in this region of Ghana. The motivation behind doing so appears to be to prevent infection – “so that dirt will not enter it for it to get rotten.” Finally, these data suggest an appreciation for the need to bathe infants frequently and keep their surroundings clean in order to prevent infection.
Our findings also suggest that this community is undergoing a shift toward a greater percentage of facility-based deliveries. These findings mirror results published elsewhere: in the northern region where our study was conducted, 2003 data suggests that 29% of women delivered in a health facility , whereas 2009-2010 data suggest that number has risen to nearly 70% (data not shown). This mirrors trends seen on the national level . While this is good news from the perspective of the World Health Organization’s recommendation that all deliveries be attended by a skilled provider , it also suggests that facilities in this region need to be equipped to handle the increasing volume and complexity of patients. This may require additional staffing, re-training of existing staff, addition or renovation of physical infrastructure and close attention to quality improvement techniques. Perhaps most importantly, the technical skills of the providers need to be supplemented with an understanding of long-held traditional practices and beliefs. Community members and providers need to work together to ensure that birth traditions can be upheld in as clean a manner as possible in order to reduce cultural barriers to facility-based delivery.
These data also suggest that grandmothers are critical social gatekeepers, providing advice, guidance, and advocacy regarding how mothers and their babies ought to behave and be treated. While such a finding has been demonstrated in the context of breastfeeding [25–28], and while research in Ghana has supported the involvement of husbands as financial decision-makers , to date the literature has yet to demonstrate the importance of grandmothers in preventing neonatal infection. The data presented here suggest that future interventions and health promotion efforts are likely to be more successful if grandmothers are incorporated into the program planning and implementation phases.
Finally, these data suggest that health care providers and community members are not always in agreement with regard to maternal and child health practices. This suggests that not only do healthcare providers need to be educated about broad community perceptions, but they need to proactively ask about them with individual patients. Perhaps most importantly, providers need to be willing to discuss these issues openly and respectfully and work with patients and family members to find acceptable alternatives to traditional (or standard medical) practices. Providers also need to appreciate that unless acceptable solutions are devised, agreed-upon, and implemented collectively with a woman and her extended family, traditional practices may be resumed once she leaves the hospital. The results presented here suggest that researchers and policy makers need to engage health care providers and community members in working together to help plan interventions that maximize community participation.
Our findings complement those by Hill et al., who reported on clean delivery practices in Central Ghana.  They interviewed women who had recently delivered (30 IDI and 2 FGD), traditional birth attendants and grandmothers (20 IDI and 6 FGD), and husbands (12 IDI and 2 FGD), and analyzed the prevalence of clean delivery behaviors collected through a demographic surveillance system. In their study, they report that most women delivered on a covered surface, and had birth attendants who washed their hands, cut the cord with a new blade and tied it with a new thread. They also described as near universal the frequent application of products to the cord. Husbands were singled out as key in financial decision-making, thus the authors suggested incorporating them in home visits . There are also important contrasts to the work by Hill et al. Our findings suggest that in Northern Ghana there may be a greater use of non-sterile materials to tie the cord such as twigs, string, rope and thread, as well as the use of contaminating materials on the umbilical cord, such as red earth soil, as well as local herbs and oils. While husbands were singled out as prime drivers of newborn healthcare decision in central Ghana, we noted the prominent role grandmothers play in the north.
Our findings also supplement four other studies in Africa that have addressed prevention of neonatal infection through community-based practices [11, 30–32]. All four were intervention studies, yet Meegan et al. used their knowledge of community practices to inform and shape their intervention. The authors conducted an evaluation of the effect of a health-promotion program on neonatal tetanus among the Maasai in Kenya and Tanzania. While traditional cord care among the Maasai includes packing the umbilical stump with cow dung, the authors were able to work with local leaders to encourage substituting washing the stump with water or milk instead. This resulted in a dramatic drop in neonatal tetanus rates (0.75 per 1000 births in the intervention areas vs 82 per 1000 births in the control areas). This is an example of working with the community to provide an acceptable substitute for long-held traditional infant care beliefs, something that our data suggest will be critical if we are to address neonatal infection rates in rural areas of developing countries.
Our results also complement research conducted outside Africa. For example, Sreeramareddy et al. (2006)  found that in Nepal, only 16.2% of mothers who delivered at home used a clean home delivery kit, only 38.3% of the birth attendants had washed their hands prior to delivery, and nearly 94% of infants were given a bath shortly after birth . As seen in Table 1, our finding regarding non-sterile substances being applied to the cord is not uncommon in the developing world. Mustard oil was applied to the cord in 22.1% of deliveries in Nepal.
We believe the research reported here has several key strengths. First, this study represents 253 individual respondents who completed in-depth interviews or participated in focus group discussions in 2010. The study includes a diversity of perspectives, including women with newborn infants, grandmothers, compound heads, community leaders, formally-trained healthcare providers, and traditional healers. It also includes diversity within each of those groups. For example, among mothers, our sample includes women who delivered unassisted, assisted by a traditional birth attendant, or assisted by a skilled birth attendant. It includes women who delivered at home and in a variety of types of facilities, from the health center to the district hospital. It includes literate and illiterate women, women experiencing their first birth or one of many. And it includes women of both Nankani and Kasem ethnicity. This comprehensive approach helps ensure that our findings reflect a rich and variable portrait of newborn care in this region – including the influences of grandmothers, compound heads, community leaders, and health care providers.
Despite its strengths, there are limitations to this study. First, interviews were conducted by undergraduate- and graduate-student interviewers. It is possible that results might have been different if the community members perceived the interviewers to be more similar to themselves. It is also possible, on the other hand, that community members were less guarded among students than they might have been with local peers. Given the volume of information readily volunteered and the 20-year history of conducting interviews in the community we believe respondents were not inhibited by the student status of interviewers. Finally, the design of this study did not include an assessment of actual infections resulting from post-natal practices. Future experimental research is needed to assess the relationship between traditional and contemporary practices in rural northern Ghana in order to document actual neonatal infection rates. However, given the formative, hypothesis-generating nature of this research, we believe our findings provide useful information to researchers, clinicians, and program planners in Ghana and beyond.
The results presented here provide an important backdrop against which future interventions can be planned. Newborn-care interventions are not new – the NewHints Trial in central Ghana , Nepal’s Safe Delivery Incentive Programme (SDIP) , and the Pregnancy and Village Outreach Tibet (PAVOT) program  are just a few examples of programmatic attempts to improve the way infants are handled upon delivery. However, our results suggest that future interventions would benefit from thoughtful inclusion of grandmothers and other key community figures in addition to training traditional birth attendants and others who might attend home deliveries. Our results suggest that grandmothers play a very important role in infant care and must not be overlooked as important stakeholders with regard to infant care. In addition, our results uncover a notable disconnect between providers and community members – one that must be breached if future interventions are going to be successful.