Factors associated with adoption of beneficial newborn care practices in rural Eastern Uganda: A cross-sectional study

Background: Beneficial newborn care practices can improve newborn survival. However, little is known about the factors that affect adoption of these practices. Methods: Cross-sectional study conducted among 1,616 mothers who had delivered in the past year in two health sub-districts (Luuka and Buyende) in Eastern Uganda. Data collection took place between November and December 2011. Data were collected on socio-demographic and economic characteristics, antenatal care visits, skilled delivery attendance, parity, distance to health facility and early newborn care knowledge and practices. Descriptive statistics were computed to determine the proportion of mothers who adopted beneficial newborn care practices (optimal thermal care; good feeding practices; weighing and immunizing the baby immediately after birth; and good cord care) during the neonatal period. We conducted multivariable logistic regression to assess the covariates of adoption of all beneficial newborn care practices. Analysis was done using STATA statistical software, version 12.1. Results: Of the 1,616 mothers enrolled, 622 (38.5 %) were aged 25-34; 1,472 (91.1 %) were married; 1,096 (67.8 %) had primary education; while 1,357 (84 %) were laborers or peasants. Utilization of all beneficial newborn care practices was 11.7 %; lower in Luuka (9.4 %, n = 797) than in Buyende health sub-district (13.9 %, n = 819; p = 0.005). Good cord care (83.6 % in Luuka; 95 % in Buyende) and immunization of newborn (80.7 % in Luuka; 82.5 % in Buyende) were the most prevalent newborn care practices reported by mothers. At the multivariable analysis, number of ANC visits (3-4 vs. 1-2: Adjusted (Adj.) Odds Ratio (OR) = 1.69, 95 % CI = 1.13, 2.52), skilled delivery (Adj. OR = 2.66, 95 % CI = 1.92, 3.69), socio-economic status (middle vs. low: Adj. OR = 1.57, 95 % CI = 1.09, 2.26) were positively associated with adoption of all beneficial newborn care practices among mothers. Conclusion: Adoption of all beneficial newborn care practices was low, although associated with higher ANC visits; middle-level socio-economic status and skilled delivery attendance. These findings suggest a need for interventions to improve quality ANC and skilled delivery attendance as well as targeting of women with low and high socio-economic status with newborn care health educational messages, improved work conditions for breastfeeding, and supportive policies at national level for uptake of newborn care practices. This was a cross-sectional household baseline study conducted among 1,616 mothers in two rural health sub-districts of Buyende and Luuka in Eastern Uganda. Both Buyende and Luuka health sub-districts are within Buyende and Luuka districts respectively, which are part of Busoga region contributing 10 % of the population of Uganda. Over 80 % of the population are peasants and live on less than US$1 a day. The crude birth rate in both districts (Buyende and Luuka) averages that of the country at 42 live births per 1,000 populations [3, 38]. Household interviews were conducted between November and December 2011. Mothers, who provided written informed consent, had given birth in the last 1 year, and had live babies, were exhaustively recruited. We excluded those who had stillbirths or whose babies died prior to interview to minimize the social consequences associated with asking mothers about babies that died immediately after birth. Buyende and Luuka are two of the control health sub-districts in Eastern Uganda where The Maternal Newborn Study (MANEST) was implemented. MANEST was a quasi-experimental 30 months study that started July 2011 and ended December 2013. The goal of the study was to learn how to integrate and scale-up interventions aimed at increasing access to institutional deliveries and care of complications through vouchers, and improving newborn care and uptake of Prevention of Mother to Child transmission of HIV (PMTCT) through home visits by community health workers, within the existing health system in Uganda. As part of the baseline assessment, to inform the final design of the intervention, data were collected among 1616 mothers in both health sub-districts who met the eligibility criteria i.e. had given birth in the last 1 year and had provided written informed consent and live babies. Data were collected using paper-based questionnaires by trained research assistants. The quality assurance of data was ensured through daily assessment via questionnaires filled-in by a supervisor; in cases of error or incompleteness of data, corrective measures were implemented immediately i.e. mothers were re-visited to ascertain correctness of the data, except for data they could not recall. The primary outcome was the proportion of women who reported that they adopted beneficial newborn care (NBC) practices. Beneficial NBC practices were grouped into five categories: (i) Optimal thermal care defined as: newborn after birth, was first dried, put skin-to-skin on mothers chest, wrapped in clean dry clothing and delayed bathe (after 24 h or more), (ii) Good cord care defined as: type of instrument used to cut the cord (such as a brand new razor blade, surgical blade or sterilized pair of scissors), type of material used to tie the cord (clean thread), and no medicinal substance (local or not local) put on the cord), (iii) Good feeding practices defined as: initiating breastfeeding within the first 1 h after birth and exclusively breastfeeding in the first month of life, (iv) Weighing the baby immediately after birth, and (v) immunization (if the baby was given oral polio vaccine (OPV) and/or BCG after birth). These NBC practices were further combined into an index of all beneficial NBC practices, which was dichotomized as (“Yes = 1”, if the mother practiced all the beneficial newborn care practices and “No = 0”, if the mother practiced none or just a few). Age distribution was checked for normality and found to be skewed (to the right). We then categorized age as follows: ≤24 (less than or equal to twenty-four years), 25–34 (Twenty-five to thirty-four years) and 35+ (Thirty-five years and above). Parity (number of pregnancies carried beyond 28 weeks) of mother was grouped into 1, 2-4 and 5+, while Trimester at first ANC was categorized according to weeks of gestation when the mother had her first ANC visit as follows: trimester 1 < 13 weeks, trimester 2 = 14–26 weeks and trimester 3 = 27–40 weeks. Number of ANC visits was categorized into 1-2, 3-4, and 5+. Distance to health facility where mother delivered was categorized into (< or =5 km, >5 km and not known). The other variables i.e. Marital Status, education level, occupation, husband’s education, skilled delivery (delivery by midwife, doctor clinical officer or nurse at a facility), delivery mode, ANC visit, were left intact. To generate household socio-economic status (SES), we considered the following variables: floor material, roof material, wall material, fuel used for cooking, source of light and other household possessions (i.e. radio, type of bed, table refrigerator, television set, sound cassette player, and telephone), agricultural land, and farm animals (chicken, goats, cows, pigs, sheep). These variables were screened for relevance and reliability using Cronbach’s alpha (which was found to be 0.628) and acceptable [39]. The final list of variables included agricultural land, type of floor material, type of roof material, wall material, fuel used for cooking, and source of light. We performed Principal Component Analysis (PCA), scored the first principal component, and used it to generate an asset index. The asset index was then used to group all households into wealth quartiles; i.e., <25 % = Lowest, 25–50 % = Low, 50–75 % = Middle and >75 % = High socio-economic status) [40]. We merged the ‘lowest’ and ‘low’ quartiles into “low” because lowest had very few values while “middle” and “high” were left intact. This resulted in three socio-economic status levels, namely: low, middle and high, as presented in the Tables. We computed descriptive statistics to determine the proportion of mothers who adopted beneficial newborn care practices separately for each health sub-district and conducted bivariate analyses using Pearson chi2 test to assess the association between adoption of beneficial newborn care practice and mothers socio-demographic and other characteristics. Before inclusion in the model, we assessed for collinearity of the explanatory variables and there were none. All variables with a p-value less than 0.1 (p < 0.1) at the bivariate analysis were included in the multivariable analysis. We then conducted multivariable logistic regression analyses using the likelihood ratio test to assess the covariates of a mother adopting all beneficial newborn care practices after adjusting for number of ANC visits, skilled delivery, husband’s education status, education; occupation, delivery mode, trimester at first ANC visit, socio-economic status, and health sub-district. Missing values accounting for 2 % in the final model were excluded from the analysis. A p-value less than 0.05 (p < 0.05) was considered significant at the multivariable analysis. While we intended to run separate multivariable regression models for each health sub-district, we were not able to do this due to the limited number of women reporting adoption of all the beneficial newborn care practices in each health sub-district. In order to account for the differences in adoption of beneficial newborn care practices between the two health sub-districts, we controlled for health sub-district of residence in the adjusted analysis. The Hosmer and Lemeshow’s goodness-of-fit test was used to assess how the final multivariable model fit the data and was found to be 0.896, 8 d.f, p = 0.35, which was satisfactory. We estimate that this study had a post-hoc statistical power of 81 % to detect an odds ratio of 0.64 as significant at an alpha-level of 0.05 when comparing adoption of all beneficial newborn care practices between the two health sub-districts. Data were analyzed using STATA version 12.1. Makerere University School of Public Health Institutional Review Board approved the study. Written Informed consent was sought from study participants after reading to them and adequately explaining to them the aim of the study. Participants were informed of their right to withdraw from the study at any stage of the interview.