Overall, it is hypothesized that facilities that can adapt the model to improve feasibility and maintain high intervention fidelity will have higher levels of satisfaction among pregnant women, leading to greater retention in ANC and increased institutional deliveries for better pregnancy outcomes among G-ANC participants. The most important intervention principles are group processes (i.e., the quality of highly participatory facilitation), group stability, and group size. Diluting these principles may adversely affect women’s comfort in the group, openness and freedom of the discussion, the validation of women’s experiences and needs, and the bonding and relationships between women and the providers.

We also hypothesized that: (i) G-ANC as the predominant ANC model can be implemented with high fidelity and deliver better quality ANC for women, particularly a more positive experience of care; (ii) facilities with smaller caseloads and more ANC staff will be able to transition and sustain G-ANC as the predominant ANC model; and (iii) facilities with fewer staff and higher caseloads, such as hospitals, will not be able to transition fully to G-ANC, leading to lower health care provider satisfaction with G-ANC and less likelihood that G-ANC at scale will be sustained over time. Figure 1 illustrates the G-ANC scale up theory of change.

This was a mixed-methods implementation research study that collected qualitative and quantitative data at baseline, during implementation, and at the end of the study period. The study team collected the required data and information using study instruments, process documentation, as well as the routine ANC register (MOH 405), a MoH tool used for routinely reporting clinical services for pregnant women. Process documentation also included observations of changes in the health facility (e.g., re-allocation of staff to ANC) and changes in the G-ANC model.

The study team abstracted pre-intervention quantitative data 12 months preceding G-ANC scale-up from facility ANC registers at study facilities, including any data available on previous G-ANC implementation. The same data were collected monthly from all facilities during the implementation period. The study team excluded personal identifiers and abstracted only aggregated descriptive information about each facility, such as total ANC clients. The study assessed fidelity through facility visits and meeting observations.

During implementation, a survey of women attending G-ANC was conducted. Every month the study team interviewed three pregnant women participating in G-ANC from each health facility, at the end of a G-ANC meeting. Women were allowed to participate in the survey only once. The study team conducted forty focus group discussions (FGDs) with pregnant women attending ANC at the mid-point (n=20) and at the end of the study (n=20).

Further, KIIs were conducted with health care providers engaged in G-ANC and ANC or health facility in-charges at the study’s mid-point (n=24) and end-line (n=24). At each point, they were purposively selected. Four health care providers were selected from each facility: two at midline and another two at the end line. Additionally, six county/sub-county health managers purposively selected participated in the KIIs. The study data and instruments are summarized in Table 2.

For iterative improvement, the study team conducted quarterly in-depth reviews to share data with facility teams and the sub-county and county health management teams (S/CHMTs), throughout the study to identify whether adjustments were needed in implementation.

The study was conducted in Machakos County, with a population of 1.4 million, and adjacent to the capital, Nairobi. There were 275 facilities that reported ANC data in 2020, per the KHIS/DHIS2. Ten health facilities in Machakos County participated in the G-ANC cRCT.

Purposive sampling in consultation with the CHMT was used to identify 12 health facilities for G-ANC scale-up. The selection was based on the geographical distribution (i.e., at least one facility in each of the eight sub-counties in Machakos) and number of clients who received ANC services per month. Five facilities were defined as high-volume (i.e., more than 70 ANC cases per month) and seven as low to moderate-volume (less than 70 ANC cases per month).

Data was collected from 12 health facilities in Machakos County ( Table 3). Five facilities were considered high-volume based on the ANC1 monthly caseloads (>70–80/month) and seven facilities were considered low-to-moderate volume based on monthly ANC1 caseloads. The low-to-moderate volume facilities were also considered as the predominant model facilities where G-ANC was offered as the predominant model for ANC, although women also had the option to have personalized ANC visits. At least one facility was selected from each of the eight sub-counties in Machakos County, as per the recommendation from the Machakos CHMT. The 12 facilities served 4.5% of the total ANC1 clients reported in 2020.

In addition, the sample size calculation for this study was based on detecting client satisfaction among women attending ANC. The sample calculated was the minimum number of ANC clients that would be representative of the total number of clients attending G-ANC to determine the overall level of satisfaction with G-ANC, based on the question “Among women attending G-ANC, what is the level of client satisfaction with G-ANC?” The following parameters were used in the calculation: 1) the estimated population of pregnant women attending ANC 4 visits in the 12 facilities; 2) a confidence interval of 95%; 3) a margin of error of 5% and 4) power of 80%. To estimate the population of pregnant women attending the 4 ANC visits in the 12 health facilities, we considered the following:

Participants were recruited from study health facilities: the health care providers, and women attending G-ANC. The women were informed about the survey by the G-ANC nurse-facilitator before the G-ANC meeting that day. Those who were interested were consented and interviewed by the research assistant. Those who took part in the survey were ineligible to take part in future surveys. County/sub-county health managers were recruited in coordination with CHMT for the KIIs. The recruitment is summarized in Table 5.

Participant selection was purposive and based on the criteria described in the Table 6 below.

Group ANC is an alternative ANC service delivery model that enrols pregnant women at their first ANC visit into groups (cohorts) with other women of similar gestational age (GA). These women are organized into cohorts and receive subsequent ANC during scheduled meetings together as a group.

The G-ANC eight-contact model includes the first individual visit (i.e., booking visit, intake visit, ANC1) plus seven G-ANC meetings ( Figure 2). Facilities in Machakos County have been implementing the Jhpiego G-ANC eight-contact model since 2019. The intervention includes facilitated, participatory discussions on topics most relevant to the stage of pregnancy. Three main components of G-ANC were retained: building peer support; conducting clinical assessment including self-assessments; and using a participatory facilitated learning approach. The six principles in the G-ANC model define the intervention and are used to measure fidelity to the intervention ( Grenier et al., 2020).

Scale-up was informed by co-design workshops led by the Machakos CHMT in January 2022 which included staff from the 12 facilities and Jhpiego technical staff. Participants mapped available resources for G-ANC scale-up in the 12 health facilities and discussed adaptations to the model to make it more feasible at scale, particularly in facilities that will adopt G-ANC as the predominant model. The G-ANC model and materials were then updated in January 2022. Materials include the following: meeting guides for facilitators; illustrated picture cards for facilitated group discussions (5–6 cards per meeting); self-assessment cards (1 per women); take-action booklets used by women during meetings; and an implementation guide that includes guidance on preparing and implementing G-ANC at scale. The G-ANC cohort tracker was used to record attendance for each group, and a G-ANC meeting observation checklist was used to provide feedback on fidelity. The study team supplied all the G-ANC related materials, equipment and supplies, as well as some ANC equipment (e.g., automated blood pressure devices, weighing scales). Facilities that lacked appropriate and adequate meeting space were supplied with a tent.

A series of trainings were conducted in February 2022 to: 1) standardize the county trainers and sub-county mentors on the revised materials and scale-up aims; 2) refresh previously-trained health care providers who have experience conducting G-ANC on the new materials and scale-up plans; and 3) train a fresh set of health care providers on G-ANC facilitation skills. The county trainers and mentors—mainly the county and sub-county Reproductive Health Coordinators—led the training activities, with support from Jhpiego. County trainers and mentors provided on-site mentoring and support as the 12 facilities scaled-up G-ANC. The Sub County Health Management Teams (SCHMTs) were responsible for G-ANC monitoring, including meeting observations, within their ongoing routine ANC/MNH supportive supervision responsibilities. As required, facilities mentored new staff on-the-job (OJT) to become competent G-ANC facilitators.

We assessed G-ANC implementation and scale-up in a routine setting where service provision and commodities were government administered, as opposed to prior studies where resources were externally supported. The intervention was fully implemented by staff at the study facilities who are health care providers employed by the government and routinely provide ANC services in accordance with standards of clinical care outlined in the national guidelines on ANC ( Ministry of Health Kenya, 2016).

G-ANC was integrated into the 12 health facilities and the county's existing healthcare systems. The 12 facilities introduced the 2022 model in March/April 2022 at the scale the facility staff felt was feasible. Staff at each facility adopted G-ANC and contextualized the model, considering the G-ANC principles and scaled up G-ANC within their ANC services. The sub-county mentors and county trainers supervised and provided continuous mentorship for G-ANC services.

Baseline quantitative data were collected for the period March 2021 to February 2022 from the 12 study facilities. Monthly data abstraction and quarterly surveys were conducted from the inception of the project in March 2022 until November 2023. Midline data were collected in April 2023, and endline data in December 2023 using KIIs and FGDs. Additionally, a quantitative sub-study was incorporated, focusing on exploring the determinants influencing the choice of place of delivery among women participating in G-ANC from the eight health centers of the 12 health study facilities.

The study team worked with experienced research assistants who had a minimum of a diploma certificate in a relevant health discipline including social sciences or equivalent. All research assistants were extensively trained before any data collection. Training was on the study objectives, collection of data through assessment tools, surveys, FGDs, KIIs, and in-depth interviews (IDIs), and moderating the sessions. To ensure language and translation accuracy, the research assistants were educated on the exact wording of each specific question. Additionally, all research assistants were trained on how to obtain informed consent from participants as well as responsible/ethical research conduct.

ICRHK was the local research partner and was responsible for leading data collection. For quantitative data, data quality checks for completeness and accuracy were done at different levels. ICRHK data collectors verified data before submitting it, and the data manager at ICRHK monitored data on an ongoing basis. For qualitative data, the study team closely monitored the data collection process, involved multiple analysts to check interpretation, and monitored personal values and ideas that may bias qualitative findings. Participant validation of the data was achieved through county/sub-county and facility quarterly review and feedback sessions in Machakos County.

The data was managed according to ICRHK, Jhpiego, and John Hopkins Biostatistics Centre (JHBC) data policies. Soft-copy data was de-identified of personal identifiers such as name, age, and residence. Data was stored on encrypted drives, and only authorized study staff had access to it.

Qualitative data was transcribed from audio to text. Findings were summarized and structured under key thematic areas related to the logical framework's outputs and outcomes. Analysis was done using the Dedoose Version 9.0.17, cloud application for managing, analyzing, and presenting qualitative and mixed method research data (2021). Los Angeles, CA: Socio-cultural Research Consultants. LLC www.dedoose.com. A free one month trial with all standard Dedoose features with no payment information required can be found on https://www.dedoose.com/signup.

Quantitative data will be analyzed and presented in the form of descriptive and inferential tables. Following this analysis, quantitative findings will be triangulated with qualitative results to compare and contrast findings, deepen the analysis of interesting or contradictory findings, discuss challenges, raise issues that needed further investigation, and suggest implications for the program implementation.

Ethical approval was obtained on July 13, 2022 from the African Medical and Research Foundation AMREF ESRC (ESRC P1219/2022) in Kenya and from the National Commission for Science, Technology and Innovation (License No: NACOSTI/P/22/19555) on the 22 ^nd August 2022. Permission to conduct the research was obtained from the Government of Machakos County, Department of Health and Emergency Services in the office of the County Director of Medical Services (Ref No. MKS/DHES/RSCH/VOLI/233) on the 29 ^th August 2022. In addition, IRB approval was obtained from the Johns Hopkins Bloomberg School of Public Health (JHSPH) Institutional Review Board (IRB No. 21094) on the 21 ^st June 2022.

Written informed consents were obtained from all participants as per the United States (US), international, and local regulations, and ICRHK’s Informed Consent Process. The informed consent described the study objectives, procedures, voluntary participation, and the risks and benefits to the study subjects. Consent forms for the mothers were translated to the Kiswahili language. Consent forms for KIIs were not translated since health care providers are educated and work in the English language. Participation in the study was entirely voluntary. All eligible pregnant women, health care providers, and county/sub-county health managers who met the inclusion criteria were enrolled in the study. Confidentiality was observed throughout the study period.

Data collection is complete, and analysis underway. Baseline data was collected 29 August-2 September 2022 by trained research assistants for the period from March 2021 to Feb 2022 (considered the baseline period) in the 12 study health facilities. Implementation period ran from March 2022 to November 2023. Trained research assistants collected midline data between March-April 2023 and endline data in December 2023. The final study results will be ready for publication in September 2024.