Effect of health facility linkage with community using postnatal card on postnatal home visit coverage and newborn care practices in rural Ethiopia: A controlled quasi-experimental study design

Background Postnatal home visit has the potential to improve maternal and newborn health, but it remains as a missed opportunity in many low-and middle-income countries. This study examines the effect of health extension worker administered postnatal card combined with health facility strengthening intervention on postnatal home visit coverage, newborn care practices, and knowledge of newborn danger signs in rural Ethiopia. Methods We employed quasi-experimental design using controlled before-and-after study in intervention and comparison districts of rural Tigray, northern Ethiopia. Training of health extension workers (HEWs) on postnatal home visit (PNHV), training of healthcare providers on maternal and newborn care, and capacity building of healthcare authorities on leadership, management and governance together with health system strengthening were the implemented interventions. Baseline (n = 705) and end line (n = 980) data were collected from mothers who delivered a year before the commencement of the actual data collection in the respective surveys. We used difference-in-differences (DiD) analysis to assess the effect of the intervention on PNHV coverage, essential newborn care practices and maternal knowledge of newborn danger signs. Results A total of 1685 (100%) mothers participated in this study. In all districts, more than 1/3rd of the mothers 633(37.57%) were in the age of 30–39 years. The difference-in-differences estimator showed an average of 23.5% increase in coverage of PNHVs within three days (DiD, p<0.001) and the provision of most postnatal contents significantly increased in the intervention district in the end line survey. The knowledge of at least three danger signs increased by 13.6% (p = 0.012).The DiD estimator showed an average of 27.6% increase to check the mothers for heavy bleeding (DiD, p = 0.011). This study also revealed that the checking of maternal blood pressure increased from 5.8% to 11.8% in the comparison districts and from 9.4% to 93.3% in the intervention district. The difference-in-differences estimator result showed a 9% difference in clean cord care practices among the participants (p = 0.025), 12.2% in skin to skin care (p = 0.022), and borderline significant increase in early initiation of breastfeeding (10.5%, p = 0.051). Conclusion We conclude that the intervention package was effective in improving the coverage of PNHV, increase in knowledge of newborn danger sign and essential newborn care practices. Hence, further strengthening the linkages between health facilities and community is imperative to improve the coverage of essential lifesaving maternal and newborn care services by HEWs at home. We employed a quasi-experimental study design using controlled before-and-after study in the intervention and comparison districts. The baseline survey was conducted in 2018 and the end line survey in 2020. We conducted the study in four districts of Southeastern zone of Tigray regional state. The selections of these districts were based on overall representation in relation to child and neonatal mortality indicators, and progress in implementation of the integrated management of childhood illness programs. One district was selected out of the four districts for the implementation of the intervention in consultation with the regional health bureau and the district health offices. The selection of the district was purposive and determined based on maternal and newborn service indicators for the districts and district health office interest in participation. It was also convenient to select the district by the researchers because of the easy access from Mekelle, the capital city of the region. The four districts surround the capital city (Mekelle) of Tigray region. The zone serves a population in excess of 567,700 inhabitants with the total households estimated at 129,031. With regard to the number of health professionals, there were 731 health care providers in the zone out of which 183 were HEWs [36]. At least one antenatal care coverage of the zone was, 97%, facility delivery was 89.2%. Out of all facility deliveries, only 18.2% of them had a minimum facility stay of 24 hours post-delivery. Postnatal home visit coverage by the HEWs within three days was 14.5% with only 0.71% mothers receiving the scheduled three postnatal home visits: within 24 hours, three days, and seven days [22]. Before the implementation of the intervention, community based survey was conducted to assess the coverage of postnatal home visits by the HEWs among 705 postnatal mothers in the proportionally selected 30 villages. The sample size was determined using a two-sided Z test of the difference between proportions with 80% statistical power, a 5% significance level. The outcome of interest used in the calculation of the sample size was the proportion of mothers who received PNHV within three days which was 14.5% from the baseline survey in northern Ethiopia. By assuming the effect size to be increased by 10% after a 1year implementation of the intervention with un equal cluster size evaluation, the design effect was 1.81 which was calculated by considering the intra-class correlation (roh) = 0.05 and coefficient of variation (CV) = 0.12 [37–39]. With a non-response rate of 10%, 980 mothers were sampled for the end line survey. A total of 1685 mothers were included both for the intervention and comparison districts. A multi-stage sampling technique was applied to select the study participants. In the first stage we selected 30 villages to realize the total sample size. All households’ of mothers having delivered in the 12 months preceding the baseline and end line were listed and registered by the HEWs in communication with Women development group (WDG) leaders in the selected villages. We sampled a total of 373 mothers in the intervention and 1312 mothers in the comparison districts by using simple random sampling techniques. The study was approved by the Institutional Review Board of Mekelle University, College of Health Sciences (No.1437/2018). Verbal consent was obtained from the study participants after explaining the objectives of the study and the use of verbal consent was approved by the ethics committee. Privacy and confidentiality of the respondents were maintained during data collection and analysis. Mothers and newborns in the comparison group received the routine care services. We performed a multi-stage sampling strategy to select communities as sampling units in proportion to their population size. Households meeting the eligibility criteria were randomly selected for the interviews from each sampling unit. Baseline and end line data were collected in 2018 and 2020, respectively with an interviewer-administered structured questionnaire that was adapted from Ethiopia demography health survey (EDHS) and the last 10 km (L10K) survey [40, 41]. The tool contains items regarding socio-demographic, status towards model household, community based participations like pregnant women forum, participation in WDG, community health insurance membership, availability of HEW’s cell phone at home, time taken to visit the household, ANC attendance (both facility and home), place of delivery, birth notification, attendants at birth, postnatal visits, contents of PNC provided, maternal knowledge on postnatal danger signs and essential newborn care practices. It was initially prepared in English and then translated into the local language (Tigrigna) and translated back to English by language experts. The questionnaire was pre-tested prior to the commencement of actual data collection outside the study districts. A total of 20 field workers (BSc and above in nursing, and midwifery) were recruited for the data collection and 2 days training was given for the data collectors. Four field supervisors had also participated in the data collection. The primary outcome measures were: PNHV by HEWs and essential newborn care practices. The secondary outcome measures were postnatal care contents and maternal knowledge on newborn danger signs. By considering the quantitative (low coverage of scheduled PNHV) [22] and qualitative finding (poor attention of healthcare authorities, lack of effective supervision, poor functional linkages, inadequate logistics and supplies, poor community participation and support), an intervention was designed by the authors in consultation with the regional health bureau and the district health offices. Hence, to improve those maternal and newborn health services, Mekelle University School of public health with support from Tigray regional health bureau and Tigray KMC project implemented an intervention from August 10, 2019– August 20, 2020. Healthcare providers in the intervention district were trained on essential newborn care practice and postpartum care at facility and to link the mothers with and for the provision of PNHV by the HEWs using postnatal card. The postnatal card contains both maternal and newborn postnatal care contents documented in a separate section to be addressed by health extension workers (HEWs) on consequent visits that are recommended by the world health organization (WHO) (within 24 hours, 3rd day, 7th day and 42nd days) (S1 File). All responsible healthcare providers in the maternal, newborn and child health (MNCH) clinic were expected to counsel the mothers about the postnatal danger signs and inform them to have awareness about the presence of the service at home before the mother discharged from the facility and hence, linked them with HEWs. Evidences also demonstrated that there is a positive association between any recalling mechanisms for an appointment like SMS and healthcare utilization [11, 42, 43]. We provided training for all HEWs (39) to conduct home visits during pregnancy and postnatal period. In addition, the intervention included strengthening the health facilities, mainly through training of midwiferies, nurses and health officers who were assigned in the MNCH clinic about the provision of essential postnatal contents within 24 hours at facility and linkage with HEWs (a total of 36). Six days training was provided for the facility directors and supervisors, regional MNCH expert, district health office director, and MNCH expert about leadership, management and governance (15 participants) with emphasis to postnatal care services. It also included mobilization of necessary postnatal care supplies for all the health facilities in the intervention district. Following training, HEWs were instructed to create a register of pregnant women in their catchment areas and update the list every month through home visits and discussions with WDG leaders to identify current pregnancies and mobilize the community to maximize the demand for postnatal care at home. WDG leaders were expected to conduct pregnant women forum and report to HEWs on monthly basis. HEWs were expected to make at least 3 postnatal home visits for mothers and newborns (on days 1, 3 and 7). Mothers were expected to deliver and stay at least 24 hours in health facility. After having received the essential postnatal contents for their newborn and themselves, they were supposed to discharge having postnatal card labeled with postnatal contents for the mother and newborns. For those mothers who delivered in facilities other than the intervention district and at home, they were expected to receive their 1st PNHV from HEWs within 24 hours after delivery. The HEWs were instructed to conduct PNHV using postnatal card and were expected to provide a package of essential lifesaving contents for the mother and/newborns. The implementation of the study was undertaken with the support of Tigray regional health bureau, Mekelle University, and the KMC project. Supportive supervision was conducted based on the recommended schedule i.e. every month [44] by the supervisors from the respective health facilities in accordance with the existing district health service structure. We also planned quarterly meetings with HEWs, district management bodies and regional health bureau MNCH representatives. However, only a one day performance review and refresher training meeting were organized after 4 months of implementation. Implementation status was followed through the supportive supervision of the district supervisors and the research team. During the filed visit the supervisors observed and checked the provision of essential postnatal care both for the mother and newborns within 24 hours in health facilities. They also checked whether the postnatal card is filled and completed with the necessary contents that are supposed to deliver at health facilities within 24 hours. In the community, supervisors checked whether the postnatal card was filled with the recommended schedule. They also interviewed with some of the mothers about the provision of PNHV. As part of the monitoring system, all health posts were provided with uniform postnatal registration books which were developed by the research team. The outcome evaluation of the intervention was conducted through household surveys in the end line in between August and September 2020. We used the WHO recommendation of four postnatal care visits at day 1, i.e., within 24 hours of birth, day 3 (second visit), between days 7–14 (third visit) and week 6 (fourth visit). The coverage of postnatal home visits was defined as the percentage of women and/or newborns that were visited at home within 3 days after delivery. We measured newborn care practices with five items (initiation of breast feeding within one hour, skin to skin contact care, clean cord care/applied nothing (harmful) to the cord, bathing the newborn ≥24 hours after birth, and provision of colostrum to the newborn). Maternal knowledge on newborn danger signs was measured with 11 items. Participants were asked to mention spontaneously the key danger signs of newborns (Red eye/pus draining from the eye, cord bleeding/pus draining from the cord, jaundice, low body weight/preterm, low body temperature, fever, fast breathing, shortness of breathing, poor feeding, movement when stimulated/no movement even when stimulated, and convulsion). Those mothers who had mentioned more than or equal to three newborn danger signs were considered as having good knowledge about newborn danger signs [45]. The data was entered in to SPSS version 23 and exported to Stata software (version 14.0) for the analyses. Descriptive statistics was performed by computing frequencies across the intervention and comparison districts. We employed difference-in-differences (DiD) analysis to assess the contribution of the intervention package towards PNHV coverage, newborn care practices, knowledge of postnatal danger signs and postnatal care contents provided. The DiD analysis is based on comparing the percentage differences in the intervention district (before and after the intervention) to differences in the comparison districts and assumes that trends in both groups are the same in the absence of the intervention. Pearson Chi square tests and t- tests were used to compare differences between intervention and comparison districts/ we conducted test of homogeneity in the districts from the baseline to the end line surveys. We created a model that included dummy variables for the treatment and time variables: 0 for the comparison district, 1 for the treatment district); and (0 for the base line, 1 for the end line surveys).