“Our voices matter”: A before-after assessment of the effect of a community-participatory intervention to promote uptake of maternal and child health services in Kwale, Kenya

Background: Community-participatory approaches are important for effective maternal and child health interventions. A community-participatory intervention (the Dialogue Model) was implemented in Kwale County, Kenya to enhance uptake of select maternal and child health services among women of reproductive age. Methods: Community volunteers were trained to facilitate Dialogue Model sessions in community units associated with intervention health facilities in Matuga, Kwale. Selection of intervention facilities was purposive based on those that had an active community unit in existence. For each facility, uptake of family planning, antenatal care and facility-based delivery as reported in the District Health Information System (DHIS)-2 was compared pre- (October 2012 – September 2013) versus post- (January – December 2016) intervention implementation using a paired sample t-test. Results: Between October 2013 and December 2015, a total of 570 Dialogue Model sessions were held in 12 community units associated with 10 intervention facilities. The median [interquartile range (IQR)] number of sessions per month per facility was 2 (1-3). Overall, these facilities reported 15, 2 and 74% increase in uptake of family planning, antenatal care and facility-based deliveries, respectively. This was statistically significant for family planning pre- (Mean (M) = 1014; Standard deviation (SD) = 381) versus post- (M = 1163; SD = 400); t (18) = – 0.603, P = 0.04) as well as facility-based deliveries pre- (M = 185; SD = 216) versus post- (M = 323; SD = 384); t (18) = – 0.698, P = 0.03). Conclusions: A structured, community-participatory intervention enhanced uptake of family planning services and facility-based deliveries in a rural Kenyan setting. This approach is useful in addressing demand-side factors by providing communities with a stake in influencing their health outcomes. Community units (CUs) are established as part of the Community Health Strategy of Kenya’s Ministry of Health (MoH). Each CU comprises of ~ 1000 households and is aligned to official administrative sub-units (sub-locations) comprising of several villages. Each CU is served by ~ 50 community health volunteers (CHVs) i.e. each CHV serves ~ 20 households and is supervised by a community health extension worker (CHEW) who is typically an HCW from the primary care facility to which the CU is linked. At the time of implementing the current study, the County Government of Kwale had adopted the MoH’s Community Health Strategy and prioritized setting up of CUs for high-volume facilities serving large catchment populations. The DM intervention was implemented between October 2013 and December 2015 nested within the framework of the MOMI project that was funded by the European Commission Seventh Framework Programme (Grant Agreement #265448). This project was implemented in 10/20 (50%) facilities in Matuga sub-county and their associated CUs (intervention facilities) and included interventions at multiple levels including the county health administration, health facility as well as community. The intervention facilities were selected purposively as they were the only ones that had active CUs at the time i.e. CUs with a clearly-mapped geographic scope and CHVs selected and trained as per the MoH’s guidelines. As a result of the Kenyan government’s policy of free maternity services enacted in early 2013, the bulk of rural dispensaries in Kenya established maternity delivery units [38, 39]. These units enabled pregnant mothers to access delivery services at primary care level. Complicated deliveries are typically referred to more specialized levels for advanced care. Delivery units at lower levels are typically manned by a nurse-midwife and consist of 1–2 delivery beds. Additionally, all pregnant women in Kenya receive ANC follow up at primary care level including any recommended prophylaxis and supplementation. The DM sessions followed a series of standardized steps as outlined in the study-specific procedures developed a priori to guide the organization and conduct of each session (Additional file 1). The procedures required that local CHVs mobilize participants from their communities to attend sessions disaggregated by age and gender. These CHVs also selected a date and venue for the session and informed the local administrator (chief/village elder) as well as an HCW from the local facility who would be present during the session to clarify any health-related issues. Since DM sessions were meant to be held at the convenience of community members, no specific number was planned from the onset. The CHVs were encouraged to convene sessions as regularly as practicable aiming to conduct at least one session per month in their community. During the session, a local community member, typically a CHV chosen to suit the age and gender of the session’s participants and who had prior training on effective conduct of a DM session, would act as session moderator. Prior training for moderators focused on encouraging use of open-ended and probing questions, conducting the session using techniques that affirmed each participant’s contribution and promoting reflective listening with paraphrasing of each participant’s contributions. Moderators were also trained to remain neutral and ensure that they maintained group control so that that some participants do not dominate while encouraging silent ones to engage in the discussion. Each DM session was initiated using a dialogue stimulator/starter, in our case, an informational picture booklet. The purpose of this starter was to stimulate initial discussions focusing on the issues targeted for deliberation. Specifically, the issues discussed during the sessions revolved around promoting uptake and utilization of FP, ANC and facility-based deliveries, including discussions around barriers and facilitators to uptake and how to effectively deal with these as a community and individuals. The informational picture booklet was simple, specific, culturally sensitive, posing a single problem without providing a solution and adapted to the audience’s age and gender. The session moderator then posed a series of questions that aimed to identify and define the issues and confirm relevance to session participants. For example, “What did you see in the pictures? Did you identify a health problem? What was the problem? Does this problem occur in this community?” Participants then proceeded to provide individual testimonies of actual experiences with the issues identified. This step was also meant to get session participants to start talking and enabled them to define the issues under consideration from their own perspective and to emotionally own the problem as well as begin to reflect on any needed improvement. The next step in the session was meant to identify current actions to addressing the issues identified and the extent to which they could achieve desired results. The question posed was “Why does the issue identified persist despite current efforts?” This step was meant to promote an analysis of the causes of the issue and develop consensus that the current situation could be improved. This step was also meant to identify new actions/options necessary to solve the issue from the perspective of the community. Through brainstorming, a list of actions was generated and appraised in terms of effectiveness and feasibility. The final step involved generating commitment by participants to consider and list the consequences of taking or not taking the recommended actions. The question posed was “What do you think will be the results of carrying out the recommended action?” Having confirmed the importance and urgency of actions to be taken, session participants then proceeded to prepare an action plan detailing what will be done, by whom, when and with what resources. For each session, a facilitator, typically another CHV, kept a record of issues that were discussed and the agreed upon action plan. They also completed a session event log and shared this with study investigators who provided regular supportive supervision in conjunction with county/sub-county health administrators. Community health volunteers were provided with training on how to effectively conduct a DM session. This was an adaptation of the CHV training curriculum offered by the Kenyan MoH and incorporated aspects of the standardized DM procedures. The CHVs did not receive any monetary payment for their services. Instead, they were reimbursed for travel and meals when they attended trainings. They were also trained on how to organize themselves into informal community self-help groups for income generation. The trainings lasted a week at a time and were meant to improve the capacity of CHVs to effectively conduct their roles as well as to promote an avenue for continued self-sustenance. The local CHEW supervised CHVs’ activities and each provided monthly written reports of their activities. The 12 CUs associated with the 10 intervention facilities where DM sessions were held were sampled purposively as they were the only ones in Matuga sub-county at the time of implementing the MOMI project that were active. Depending on geographic scope, each CU covered several villages. Villages where DM sessions were held were selected at the convenience of the CHVs organizing the meeting. Participants during the sessions were also sampled purposively according to the required age and gender. Separate sessions were held by age and gender to ensure cultural appropriateness and promote effective discussions. The total number of participants per DM session was restricted to 40 and each lasted up to 30 min. In order to obtain buy-in, a series of meetings was held with community gatekeepers (religious leaders and local administrators) in collaboration with county/sub-county health management teams and other stakeholders prior to and during intervention implementation. Ethical approval for the study was obtained from the Ethics Review Committee of the University of Nairobi and Kenyatta National Hospital (P151/03/2014). A research permit was also obtained from the National Commission for Science, Technology and Innovation (#4703). Participants in the DM sessions provided group, oral informed consent. Data on the number of DM sessions held per month was logged into a Microsoft Excel (2010) spreadsheet (Microsoft Inc. Seattle, WA, USA). Continuous data on the outcomes of interest were then abstracted per facility from the District Health Information System (DHIS)-2. Outcomes of interest included uptake of FP, ANC and facility-based deliveries with specific DHIS-2 indicators abstracted being: 1) number of women of reproductive age (WRA) receiving FP commodities, 2) number of new ANC attendees and 3) total number of deliveries. These were compared pre- (October 2012 – September 2013) versus post- (January – December 2016) intervention implementation using a paired sample t-test. All statistical analyses were conducted in Microsoft Excel (Microsoft Inc. Seattle, WA, USA) and all statistical tests were evaluated using an α-value of 0.05.