Evaluation of the feasibility, acceptability, and impact of Group Antenatal Care at the health post level on continuation in antenatal care and facility based delivery in Ethiopia using a cluster randomized stepped-wedge design: Study protocol

Study design

This mixed-methods study will use a stepped-wedge cluster randomized controlled trial with randomization of the order of the start of the intervention done at the HC level (clusters) in the Amhara region of Ethiopia. The study will begin with a pilot to test the feasibility of the intervention conducted by HEWs at the HP level. If feasible, a randomized controlled trial will be conducted with three time periods: baseline period (T1), time period 2 (T2), and time period 3 (T3), each about six months in duration. The control period for the first half of clusters will occur at T1. These clusters will start G-ANC first at T2. These clusters will have a choice to continue G-ANC with a second cohort of women during T3 (Figure 1). The control period for the second half of the clusters will occur at T2. The second half of the clusters will start the intervention in T3. We plan to use a stepped-wedge study design because it allows for iterative learning (i.e., adaptive management) during the study to modify the intervention, which was novel compared to traditional cluster randomized control trials. The study was registered on September 23, 2021, on clinicaltrials.gov with registration ID: NCT05054491.

Figure 1. Time steps and clusters for G-ANC at the HP level.

Prior to the intervention, all women in the study HPs will receive the current standard of individual ANC. HPs primarily offer ANC and other preventive services but do not provide delivery care. At the time of the intervention, the HP in the selected clusters will provide G-ANC to one cohort of eight to 10 pregnant women. Group ANC participants will be followed prospectively from the time of study enrollment. All non-enrolled pregnant women in the intervention HP will continue to receive individual ANC. During the control period, care is provided in the same setting (HP) by the same personnel (HEW). The main difference is that the control group receives individual care, while the intervention group receives care in a group-based care model.

Study setting

The G-ANC study will be conducted in two zones of the Amhara regional state: West Gojjam Zone (South and North Achefer districts) and the South Gondar Zone (Farata, Dera, and Libokemem districts). Twelve HCs in the Amhara region will be selected as G-ANC study centers. G-ANC meetings will take place at the HP level only. Amhara region has an estimated population of 22.5 million (CSA/Ethiopia & ICF, 2017) and 12 administrative zones. The economy is primarily subsistence farming.

Participants

Twelve HCs will be selected in the Amhara region after meeting the eligibility criteria (Table 1) and categorized by 1) type of location (peri-urban or rural) and 2) client volume for the last year (low or high). Each of the six HCs will have three HPs, and these will be randomized in a sequence of receiving the intervention. Study participants will be pregnant women attending ANC, HEWs, and facility and regional managers working in study sites in Amhara region.

Intervention

During the intervention time period, pregnant women who come for their first ANC visit at the HP will receive standard individual care. At the end of this first ANC visit, eligible women who agree to participate in G-ANC will be added to the G-ANC cohort. Approximately eight to 10 women will be enrolled in a cohort. Cohort membership will be fixed (i.e., the same women and facilitators at each meeting). Two additional pregnant women who meet study criteria can be added to a G-ANC cohort if added to the cohort by the start of the second G-ANC meeting. Six monthly G-ANC meetings will be conducted for each cohort on a fixed day/time at each HP. Each cohort will meet for G-ANC meetings in a designated meeting space that is set up to ensure privacy during clinical consultations. The G-ANC meeting will be conducted by at least one HEW who is trained in the G-ANC service delivery model.

Each meeting will have the same general structure: self-assessment by the clients/peer-to-peer self-assessment initially taught and supervised by the HEW, such as measurement of weight and blood pressure using an automated cuff; individual assessments/clinical care by the HEW in a private space; discussions around pre-determined and appropriate gestational-age themes using illustrative cards and take-home booklets designed for G-ANC; and plan/reflect and peer-support activities that create group cohesion. Table 2 presents the G-ANC meeting schedule and content per meeting. Each meeting is expected to last 90–120 minutes and will be documented as an ANC contact. Social distancing in response to COVID-19 as per national guidelines will be enforced during meetings. Women will have the opportunity for additional private time with the HEW after the G-ANC meeting if desired or needed for health and/or safety. Women will be told that they can and should return at any time if they have questions or concerns. Phone/text reminders will be used to remind participants of G-ANC meetings as a strategy to improve ANC attendance.

The G-ANC intervention will be implemented at the HPs that offer integrated point-of-care testing (POCT) to screen for anemia, syphilis, and asymptomatic bacteriuria. Availability of these rapid test kits improves the quality of ANC and reduces the need for women to travel to the HC for routine screening. The POCT approach was introduced by Enhancing Nutrition and Antenatal Infection Treatment (ENAT) (Mekonnen et al., 2023; Sawtell et al., 2023), a project initiated before this study as proof of concept for ANC service decentralization to the community/HP level. In ENAT, HPs started provision of the complete package of ANC services, including tests not provided at non-ENAT intervention HPs in the country or during standard care at the HP level nationally. The study provided test kits to all HPs for the duration of the study.

The G-ANC intervention and materials were adapted from Jhpiego’s G-ANC intervention model, which consisted of five monthly two-hour meetings with clinical assessments alongside structured group discussions and activities that were facilitated by nurses, midwives, or community HEWs (Grenier et al., 2019). It was modified in collaboration with the MOH for the Ethiopian context as a six-meeting model. The model is designed to increase early entry into ANC and retention in ANC in support of changes to the Ethiopian national ANC guidelines, which increased ANC contacts from four to eight contacts (Ministry of Health Ethiopia, 2022). Intervention materials in Amharic language include a pictorial self-assessment card women use at the beginning of each meeting to record weight, blood pressure, and any danger signs; meeting guide used by the facilitators to guide each meeting; five to six A4-size illustrated picture cards per meeting that are designed for nonliterate audiences (total 36 cards) and questions on the back of each card to foster discussion; an A4-sized booklet with small versions of all the picture cards that women can take home to share what they’ve learned with their partners and families; an implementation guide for HEWs to introduce G-ANC in their HP; a simple G-ANC register to aid the facilitator to track attendance and record other data; and three monitoring tools, including a meeting observation checklist to provide feedback and assess fidelity. The picture cards were pretested with HEWs before the pilot phase, and all the intervention materials will be revised as needed based on the pilot feedback. Each intervention HP will be provided with 12 chairs; a basic set of supplies and commodities, including POCT reagents to run activities; two weigh scales; three user-friendly Microlife PSA blood pressure devices for the self-assessments; and one privacy screen to divide the private clinical consultation area from the larger G-ANC meeting space.

All clients will continue to receive the components of standard care, and no standard therapies will be denied to the intervention group. Likewise, standard, individual ANC care will not be denied if requested at any time.

Training: During time period one (T1: control period), two HEWs will be trained on community mobilization and gestational age estimation (including knowledge and skills) at selected HP for two days. During the intervention period, two HEWs per HP from intervention HP will be trained on G-ANC facilitation, gestational age estimation, and community mobilization for five days. From control HPs, two HEWs per HP will be trained on community mobilization and gestational age estimation for two days. Midwives from intervention HCs who will provide on-site mentorship to HEWs during G-ANC implementation will be trained for four days on G-ANC-focused mentorship by the study team (i.e., senior health professionals from Jhpiego Ethiopia with specialization in maternal health). Midwives from control HCs will be oriented for one half day on community mobilization, screening procedures, and linking women who come to the HC for their first ANC visit to HPs under their catchment. Key stakeholders, including participating health facility managers and regional managers, will be oriented to G-ANC to build ownership and support. Midwives from HC will provide mentorship, supervision, on-the-job support, guidance, and feedback to HEW as they implement G-ANC in their HPs.

Outcomes

Primary outcome 1: Proportion of women with at least four ANC visits.

Primary outcome 2: Proportion of women who delivered in a health facility.

After exploring the balance by time period/intervention status and correlations, the effect of G-ANC will be estimated using generalized linear mixed effects models with binomial distribution and logit-link.

We will adjust for facility- and individual-level characteristics that show imbalance between the intervention versus control time periods and can be strongly correlated with the outcome scores. The primary analysis will not adjust for multiplicity and will be performed at 0.05 level of statistical significance.

Secondary outcomes: Satisfaction level of women with ANC services, proportion of women who reported self-efficacy or empowerment.

Analysis of secondary outcome measures will follow a plan similar to the analysis of the primary outcome. Subgroup analysis will be conducted for specific subgroups to determine if the interventions were different. The key subgroups of interest are marital status, education, and economic status.

Recruitment and consent

ANC clients will be recruited via four strategies: 1) pregnant women will be identified in the community as potential study participants through existing local community structures (such as the women’s development army leaders) so that the study can be explained to potential participants and they can be linked them to an HEW for their first individual ANC contact and to a data collector; 2) all new ANC attendees in the intervention HPs will be screened for eligibility during the first ANC contact with an ANC service provider; 3) women who come to the HC for their first ANC visit will receive their first individual ANC visit at the HC and then be linked to a catchment HP for eligibility assessment and enrollment if they consent to participate; and 4) eligible women will be identified by G-ANC-trained HEWs who make door-to-door visits in the community.

Research assistants will rotate among each of the intervention HPs during the recruitment period to explain the study to women who meet the inclusion criteria and agree to participate in the study.

The research assistants will obtain verbal consent from women in a private space in the HP using institutional review board (IRB)-approved recruitment and consent documents in the local language. Verbal consent will be sought instead of written consent as written consent is seen as a personal identifier in the community and verbal consent is preferred. This consenting method was approved by the ethics committee (see Ethics section). The potential G-ANC participants will be informed during the consenting process of the expected number of meetings (i.e., ANC visits) and their rights, risks, and benefits. The research assistant will explain the need to collect their names, phone numbers, and addresses, which will be used for study-related follow-up purposes only. Anyone declining participation in the study will be referred to individual ANC care during the intervention period. A baseline questionnaire will be administered to consenting study participants at the beginning of each step (time period).

For HEWs and health facility managers, study staff will explain the purpose based on the inclusion criteria. The HEWs will be oriented to the study including the randomization (they will not be blinded to the order of randomization of their HC and affiliated HPs) and their role in the study including screening first-time ANC clients for eligibility to participate. HEWs who will be facilitating G-ANC meetings will be informed that they may be asked to participate in an in-depth interview (IDI) at the end of the study related to implementation of group-based care and its effect on their job and satisfaction levels. Oral consent will be obtained in a private location at the health facility by research assistants.

For the regional health managers, study staff will contact the regional health bureau responsible for overseeing reproductive health programs in the Amhara region via phone or in-person and invite the regional health manager to participate in an IDI as per the inclusion criteria. An appointment will be scheduled to obtain informed consent and conduct the IDI, preferably at the regional health bureau office. Written consent will be obtained by the research assistant who will conduct the interview in a private location.

For the IDIs of women and HEWs involved in G-ANC, qualitative data will be collected at the end of the study after all G-ANC meetings have been conducted. Oral consent will be obtained from the HEWs and eligible women three to six months postpartum to be enrolled in the qualitative portion of the study.

Harms

The intervention in this study is for women to take part in G-ANC. Some of the components of the intervention are focused on nutrition, birth preparedness, breastfeeding, and postpartum care. Some risks, discomforts, and inconveniences may apply and will be minimized as outlined in Table 3.

Sample size

We used the existing data from Amhara region for two primary study outcomes (i.e., four or more ANC contacts [31.5%] and facility-based delivery [27.1%]) to calculate the number of clusters (i.e., HCs) to be randomized to be able to detect an odds ratio of 2 comparing post-intervention to pre-intervention proportion with 80% statistical power using stepped-wedge design with three time periods and two steps for switching from standard care to intervention (Figure 1). For the complete stepped-wedge design, the calculations indicate that a sample of 12 HCs with three time periods/two steps, six HCs switching from control to treatment at each step, and with an average of 30 women per cluster (i.e., three HPs with 10 women in each) per time period achieves 84% power to detect an odds ratio of 2 at 0.05 level of statistical significance. The underlying pre-intervention proportions and the post-intervention proportions associated with odds ratio of 2 are shown in Table 4.

The within-HC correlation (intraclass correlation coefficient [ICC]) is assumed to be 0.100. This level of correlation is consistent with the one reported by Ousman et al., where clustering was done at enumeration-area level with an ICC of 0.11 for ANC visits (Ousman et al., 2019). The total sample size for ANC clients is then 12 HCs x 3 HPs x 10 women/HP x two time periods = 720 women. An odds ratio of 2.0 translates to 50%–60% improvement in completion of four or more ANC contacts and 60%–70% improvement in facility-based delivery in the G-ANC arm compared to pre-intervention.

The study will be conducted in 12 HCs. The overall sample size is a total of 770 women. The pilot will enroll 50 women (i.e., 10 women per HP, five HPs), and the randomized control trials will include a maximum of 720 women (Table 5). Sample size for the qualitative data collection is indicated in Table 6.

Randomization

Health centers with catchment HPs that provide POCT for ANC services will be listed and randomized to begin the intervention at T2 (i.e., the first randomized set of facilities) or T3 (i.e., the second randomized set of facilities). These facilities will be randomized 1:1 before T1 to two different intervention start times (i.e. two sequences), stratified by location and client volume, using randomly permuted blocks of size 2 and 4. We will use centralized randomization as the allocation concealment mechanism. A remote and independent statistician in the Johns Hopkins Bloomberg School of Public Health (JHSPH) will generate and implement the allocation sequence, and communicate it to the study sites after confirming the eligibility of the clusters. Health extension workers will identify pregnant women and provide the list of women to Addis Continental Institute of Public Health (ACIPH). ACIPH field workers will take consent and enroll eligible women.

Blinding: Study staff and those assessing outcomes will not be blinded to knowing the study time period to which facility is randomized. The participants will also not be blinded about the type of care they will be receiving (G-ANC or routine/conventional care).

Study implementation

Group ANC at the HP level is a new intervention, therefore the study will be rolled out using a phased approach.

Pilot phase: Prior to study activation, we will pilot G-ANC in up to five purposively selected HPs in Amhara region that will not be included in the study. This pilot phase will allow us to adapt G-ANC materials to adjust for the literacy levels of the HEW and conditions at the HP. After the materials have been adapted, up to two HEWs per HP (total 10 HEWs) will be trained to facilitate G-ANC. In the pilot phase, we will enroll a maximum of 50 women (up to 10 per HP x up to 5 HPs). The pilot phase will determine feasibility, i.e., whether HEWs are able to facilitate the G-ANC meetings, there are enough eligible ANC clients at the HP level, we are able to enroll enough women at the HP who are of a similar gestational age, and women come back to the HP for the G-ANC meetings. After completing at least two G-ANC meetings in the pilot, we will determine the feasibility of moving to Phase 2 of the study. Training modules, data collection tools, and intervention delivery models will be refined, with the added engagement of midwives and health staff, based on pilot study feedback.

Trial phase: Phase 2 includes T1, T2, and T3. During T1, pre-intervention data will be collected from the first set of 18 HPs by interviewing pregnant women during their pregnancy and again postpartum. The women in T1 will not receive the intervention. During T2, the G-ANC intervention will be implemented in the first set of 18 HPs (first six clusters), and pre-intervention data collection will be conducted in the second set of 18 HPs (i.e., second six clusters). During T3, the G-ANC intervention will be implemented in the second set of 18 HPs. If the first set of 18 HPs chooses to continue the G-ANC intervention with a second cohort, an additional six months of data collection will occur during T3. Participant timeline is depicted in Figure 2.

Figure 2. G-ANC study timeline with pilot and trial phases.

Once study participants are enrolled, the participant timeline is about 8.5 months: enrollment with a baseline questionnaire, follow-up for about six months during ANC, and endline survey up to six weeks post-delivery. Participants in the qualitative research will have a single contact for the IDIs.

The overall duration of the entire study is 25.5 months. The implementation process will be documented in these sites, with some corrective measures expected to be taken after each block, depending on implementation process documentation, findings from pause and reflect sessions, and qualitative study findings before commencing the next block. A transition plan will also be developed by Jhpiego to hand over implementation to the government.

Participant retention

Implementation of a paper-based participant tracking system will be used to accurately assess participant retention in the study. The G-ANC registration and participating registration book will be used to track ANC attendance. This will provide the HEWs with a listing of all participants who have missed scheduled G-ANC meetings.

Data collection

Both quantitative and qualitative data collection approaches will be used in this study. Data will be collected in the pilot phase and then during study implementation (Phase 2).

Once enrollment begins, all new ANC attendees in the selected HPs will be screened for eligibility during the first ANC contact with a health care provider. The screening and enrollment questionnaire will be completed. Basic demographic data will be collected from all eligible consented women (all who meet the inclusion criteria). All data will be de-identified for those who decline study participation. Baseline and endline data will be collected from all eligible and voluntary women participating in all study time periods. During the first period, data will be collected from 18 control HPs. In the second period (T2), data will be collected from 18 intervention and 18 control HPs. In the third period, (T3) data will be collected from 18 intervention HPs.

Data quality will be monitored through the use of trained research assistants and supervisors, data quality supervision, and use of an electronic data collection program to minimize data entry errors.

The study team will select respondents for the IDIs at the end of the study based on preliminary findings from the quantitative survey. IDIs with women will be conducted at the end of T2 and T3, after completion of all G-ANC meetings (Table 7). Specific topics/domains to explore with G-ANC participants include satisfaction with care, any perceived changes in self-efficacy or empowerment, health literacy, improved communication and/or trust with providers and the health system as a whole, factors in uptake/non-uptake of postpartum family planning, and the overall impression of the G-ANC model. In this study, we will use standardized pretested tools to assess the levels of empowerment, self-efficacy, and satisfaction. For empowerment and self-efficacy, we will use eight questions related to their current pregnancy and calculate a composite score to understand the woman's level of empowerment. We will ask 11 questions about their satisfaction with the ANC received during their last pregnancy and with the G-ANC for their current pregnancy.

In-depth interviews will be conducted by experienced research assistants with a master’s degree in public health with HEWs trained in G-ANC and who facilitated G-ANC meetings in intervention sites and health managers. Specific topics/domains using open-ended guiding tools to explore with providers will include attitudes towards providing G-ANC; perceived changes in their communication and relationship with patients (if any); changes in empowerment, self-efficacy, and satisfaction related to their perceived ability to do their job well; changes in workloads; sustainability of group care; suggested changes to the model/logistics required to offer G-ANC; and perceived effects of G-ANC on colleagues and clients. IDIs with health facility managers and regional health managers will be conducted at the end of T2 and T3 in the intervention areas.

Data analysis

Quantitative data analysis. Data analysis and the reporting of results for this study will be conducted in accordance with norms for analyzing cluster randomized trials as described in the Consolidated Standards of Reporting Trials guidelines with considerations for the step-wedge extension (Hemming et al., 2018; Hemming et al., 2019). If there is variation within the clusters, the statistical analysis will be adjusted based on the pollution levels. The primary analysis will be conducted using intention-to-treat analysis. After exploring balance by time period/intervention status and correlations, the effect of G-ANC will be estimated using generalized linear mixed effects models with binomial distribution and logit-link. The model will include G-ANC as the intervention indicator, time period and their interaction as the primary predictors, and HC as the random intercept. We will adjust for facility- and individual-level characteristics that show imbalance between the intervention vs. control time periods and can be strongly correlated with the outcome scores. The primary outcomes will be compared across time periods both within the intervention and control groups. Comparisons will be made within a given time period (e.g., T1, T2, and T3) for both the intervention and control conditions. The amount of missing data will be assessed for each variable and overall, for the sample. If more than 5% of data on covariates are missing, multiple imputation procedures will be used assuming data are missing at random (Sterne et al., 2009).

Qualitative data analysis. Following each audio recording, either the moderator or an official transcriber familiar with local language and English will transcribe from the audio recording. Analysis will be thematic content analysis with similar themes and subthemes. Coding of textual passages will be done using open code 4.03 Umeå: Umeå University, an open-source tool known for qualitative analysis.

Data management

Following the computer system validation, the data manager will proceed to build, test, validate the master plan, and document testing of the study database using the SurveyCTO platform. SurveyCTO will automatically capture quantitative data and synchronize with the server. SurveyCTO will be programmed so that automated quality checks and data classification systems will be run. Data will be maintained in a way that preserves their accuracy, integrity, and legibility with user access restrictions and data retrievable by designated personnel only. A security system that prevents unauthorized access to the data will be maintained. The database management system will be password protected, with each member of the research team responsible for data management having their own password. After data collection, ACIPH and JHSPH will de-identify the data. Quality control reports will be sent to the study team bi-weekly indicating missing, overdue data, and outstanding queries during the baseline and endline survey time periods. Throughout the fieldwork, field supervisors and the research team will observe the research assistants conducting interviews and carry out field editing. By checking the research assistants’ work regularly, the field supervisor can ensure that the quality of the data collection remains high throughout the data collection period. The de-identified dataset will be made available on FigShare or another open access site.

Ethics

Study research assistants will be trained in a multi-day training workshop, covering the important points of the approved study protocol, including privacy and confidentiality. We will review and discuss all relevant content in the JHSPH Ethics Field Training Guide. We will review data collection procedures, which will ensure ethical conduct of research and data integrity. Ethical approval was obtained from the ACIPH institutional ethical review board, Addis Ababa, Ethiopia (ACIPH/IRB/006/2021) and the JHSPH IRB, Baltimore, Maryland, United States of America (IRB14448). ACIPH and JHSPH IRB amendments will be obtained for important protocol modifications, if any.

Dissemination of study results

The study team plans to disseminate findings among national and sub-national stakeholders through in-country dissemination events and globally through peer-reviewed journal manuscripts and international conference presentations. About five publications, including conference abstracts and journal manuscripts, are planned. Study findings will be presented to national and sub-national health officials to determine if and how G-ANC at the HP level will be integrated into policy and included as a strategy for service delivery. The study will contribute to the body of knowledge that will inform decision-makers locally and globally on whether G-ANC is a feasible service delivery model at the HP level that is more acceptable and effective than individual ANC.

Study status

The study as of September 2024, the study i active. Quantitative baseline and endline surveys of T1 and T2 have been conducted. Currently, preparations are underway for T3, but implementation of this step may be delayed until security improves in the region.