Table 3 presents the results in a summative manner. The ‘+’ indicates that the relationship is significant and positive (in the expected direction); the ‘– ‘indicates a statistically significant relationship but in the opposite direction. A blank cell denotes that the relationship was not statistically significant.
A variable that is statistically significant in both analyses is taken as an indicator of change possibly attributable to the intervention. Based on this decision rule, the intervention possibly had a role in improving potential and actual access to antenatal care and health facility delivery services for mothers as well as positively impacting quality of care provision.(See Table 2)
It was no surprise that the intervention had an impact on transport expenditure, as focused provision of transport vouchers was a key element of the intervention. Beneficiaries of the voucher scheme were poorer women identified in intervention communities, who were reached by a community health worker with information on birth preparedness and received a transport voucher in their 3rd trimester of pregnancy. The voucher scheme covered return transportation to the health facilities for the mother and one companion, which was provided by boda-boda (motorcycle taxi) drivers affiliated to the programme, thus reducing financial barriers to care. In addition, by making a reliable and fast means of transportation available for free to women, the intervention also addressed the second delay in the three delays model, i.e. delay in accessing care, which can further contribute to improving maternal and neonatal health.(9)
In terms of realized access, the intervention seemed to have facilitated deliveries in health facilities. Given that the intervention activities included demand creating activities such as
re-orienting traditional birth attendants into birth companions, health education sessions aimed at changing views on skilled birth attendance, and community scorecard dialogues aimed at promoting social accountability, our finding is eminently plausible. However, it is wise to keep in mind changes in the contextual environment that could have also played a role. A major environmental factor was the abolition of user fees for maternity care, enacted by the government in 2016. Literature suggests that this led to the increase of facility based deliveries from 44% (in 2012/13) to 62% (in 2016).(10). Our study design did not allow us to disentangle these factors. Taken together, the voucher and the user fee elimination are key to enhancing social justice, especially when placing a priority on the most vulnerable, and makes eminent economic sense(11, 12)
The intervention package also seemed to have a positive impact on quality of care, as measured by quality of ANC and PNC received by mothers; however, our results indicate that it did not have the same impact on actual access for the baby, where we found a negative impact on whether postnatal care was provided to the baby within 48 hours of delivery. This could be plausible for a few reasons.
First, it is possible that the MANI intervention with its goal of health systems strengthening (e.g. providing EmONC training and refreshers, TWG focusing on service delivery, application of organizational capacity assessment tools, etc.) could have had a greater focus and emphasis (especially in implementation) on the mothers compared to their babies, thus resulting a in a differential impact. .
Second, literature indicates that while women are generally able to report accurately on aspects of postnatal care received, indicators related to new-born care received by the child are subject to greater recall bias than those related to postnatal care received by the mother. (13) A recent review of population-based survey data on maternal and new-born care in 20 Sub-Saharan Africa countries showed that approximately two-thirds (65%) of women and their babies had received some form of postnatal care, only 3% reported receiving all seven interventions included in the analysis(14)
Third, it is important to keep in mind the context of Bungoma County, where several projects of the MNH Programme’s County Innovation Challenge Fund (CICF) were implemented at the same time as the MANI HSS project. Most notably, two projects focusing on neonatal care by Save the Children International (SCI) and by Mount Kenya University (MKU) were active in both intervention and control sub-counties. In addition, the Liverpool School of Tropical Medicine (LSTM) was implementing its training programme ‘Making it Happen’ (MiH) in the region. In this dynamic environment, marginal end-user changes that could be attributed to the MANI HSS project would be difficult to detect, especially with our design.