The Kulera Contraceptive Counseling App

Unlike many domains in health, the provision of high-quality family planning (FP) services is not only measured by the achievement of good reproductive health (RH) outcomes but also considers the objective of helping clients maximize a complex set of preferences around future fertility, health, and well-being. In FP, the client’s role in her receipt of services is distinct from most other health contexts where providers play a leading role in decision-making. A high-quality FP program, therefore, would prioritize women and couples to have a right to “full, free, and informed choice” over FP methods and services. Informed choice has been prioritized by implementing partners, governments, and donors. For these reasons, FP programs dedicate significant resources to inform clients of available FP methods and services, and clients typically do not receive methods without a consultation with a provider.

Eliciting preferences during counseling has implications for behavior and improving client-provider communication. However, little is known about how the choice architecture for FP, the processes by which information is presented during counseling, shapes preferences and characterizes how women choose their preferred method. Because of the high value placed on informed choice, counselors may present as many as 15 methods and discuss as many as 10 features or attributes (e.g. effectiveness at preventing pregnancy, side effects, etc.) for each method. From a behavioral science lens, this information-intensive approach, which compels a client to interpret large amounts of data, may, in fact, reduce the effectiveness of counseling and increase the likelihood of choice overload, decision deferral, and discontinuation, which contribute to unintended pregnancies and poor long-term well-being. To promote informed choice, programs have begun to adopt user-centered design (UCD) approaches to FP/RH service provision. These approaches stress the client as the focal point of interaction and allow providers to elicit preferences. However, there is little evidence on the effectiveness of UCD in meeting women’s FP preferences, particularly in low- and middle-income countries (LMICs). Linking a woman's FP preferences to her behavior would have significant implications for improving service delivery and contraceptive continuation in LMICs, where 20% of users (44 million women) discontinue their method within a year.

We have developed a two-stage adaptive UCD counseling approach along with an Android-based counseling app that allows providers to efficiently screen and counsel clients within the app interface. We first elicit a client’s relative preferences for FP methods and method-specific attributes using a simple screening process in the app. Clients are then counseled on a subset of FP methods tailored to the client’s reported attribute and method preferences. By introducing a simplified, scalable, technology-driven approach to elicit preferences, we aim to promote information salience and improve FP/RH counseling and service quality. Our UCD-based app aims to increase provider capacity to deliver high-quality FP/RH services to all clients who demand them, which is the primary goal of the Malawi RH program.

We also assess the role of male involvement in counseling. Studies show that men's attitudes towards FP play a role in shaping women's RH behaviors; men are also important as FP/RH clients. However, rigorous evidence on the role of male involvement in FP counseling is limited, and findings on the role of men in FP decision-making have been mixed, particularly in LMICs. Importantly, existing studies of male involvement in FP have either: 1) studied couples who are jointly enrolled and counseled together, or 2) required women to be accompanied by their male partners to counseling. These couples, and particularly the male partners who participate, are not likely to be representative of average couples and male partners. In most FP programs, men are not required to be present for women to be counseled and often do not participate unless they are actively encouraged by providers or by women themselves. To test the impact of male involvement under programmatically representative conditions, we see how giving women the choice to invite (or not) their male partners contributes to FP/RH outcomes. 

UCD counseling strategies for FP/RH have received attention in the U.S. and other high-income countries. However, such approaches can be time-consuming to implement and difficult to scale to larger client bases, and evidence of UCD in FP in resource-constrained settings, where providers have less time to devote to counseling, is scarce.

Our target population is women and couples who demand FP services, and particularly clients who have an unmet need for FP/RH and contraceptive users who are dissatisfied with their method choice. In Malawi, contraceptive discontinuation for non-fertility related reasons is high (37%), particularly among younger and poor women, suggesting that there exist barriers to a client’s decision-making for choosing the “right / ideal” method that aligns with her preferences. Discontinuation has a significant impact on women and their families and has been identified as a key concern by the Ministry of Health (MOH). We directly seek to overcome these barriers. Our innovation also targets male partners who accompany women to counseling. We aim to understand the types of partners who are invited by women, how counseling is received by partners, and how partner participation shapes outcomes. Finally, our innovation targets FP providers who seek to improve efficiency and quality in counseling.

We test two hypotheses: 1) targeted, client-centered counseling will enable women to effectively identify and realize their contraceptive preferences; and 2) the involvement of supportive male partners in counseling will allow women to translate their stated preferences into behavior.

Programs that have incorporated UCD approaches are informed by insights from behavioral economics and cognitive psychology. Choice architecture studies in marketing and management have noted that cognitive overload may lead to risk-averse behavior and impatience. Studies show that incorporating UCD may serve to “nudge” individuals to make better decisions – when faced with a smaller number of well-defined alternatives, individuals make more effective choices.

While UCD has received increased attention in the health sector, the inclusion of UCD approaches in FP is scarce, and evidence on the role of choice architecture on RH is limited. This is surprising given that FP decision-making is preference-sensitive. In the U.S., researchers have developed a tablet-based support tool, “My Birth Control” and found the tool improved outcomes without the need for extensive provider training, although longer-term outcomes of contraceptive concordance and continued use were not assessed. In LMICs, tools such as the Balanced Counseling Strategy (BCS), in which a decision tree-based algorithm identifies a client’s preferred FP methods, have been promoted as client-friendly approaches to counseling. An evaluation of BCS found that the approach was linked to increased postpartum FP use, especially among women who were supported by their partners. While interest in BCS and other strategies has grown, there is limited evidence on how these approaches impact choice and decision-making. Moreover, it is unclear how providers can implement these strategies, which require significant training and resources, to larger client bases. Therefore, we propose a simplified approach that tailors counseling to a woman's preferred attributes and methods to improve concordance, where women’s FP preferences can be more efficiently and effectively realized through their behavior.

Spousal preferences for FP are a key determinant of women's own access to and use of FP. Evidence on men’s role in FP decision-making remains limited and mixed in LMICs. In Jordan, couples counseling led to higher FP use, but this increase was not different from women who were individually counseled. In Malawi, a peer-delivered intervention found that male involvement increased FP use. In contrast, evidence from Zambia found that women who received a FP voucher together with their husbands were less likely to use FP compared to women who received the voucher alone. Our approach to male involvement is distinct in that women are given the choice to invite their partners to counseling. This approach differs from previous efforts in which couples are selected based on whether they jointly participate.

We conducted a small pilot study to test the impact of UCD counseling approaches on FP preferences and use in Malawi. In 2020, we recruited and randomized 782 women to one of 4 treatment arms: 1) a control, where women were counseled on 15 methods using a FP “Kulera” method flipchart that was developed by the Malawi MOH; 2) an arm in which women were encouraged to invite their male partners to counseling. Participants were counseled on all methods using the same approach and materials as the control; 3) an arm in which women were counseled using shortened flipcharts on up to 5 methods elicited from their stated baseline preferences; or 4) an arm in which women were encouraged to invite their male partners to counseling and were counseled using shortened flipcharts based on women’s stated baseline preferences.

After counseling, women were offered free transport to a clinic to receive free FP services for 1 month. A follow-up survey was conducted to assess if women’s stated preferences at counseling translated into behavior. While half of women changed their ideal method over the study period, only 1 in 4 changed their method use. Women who received targeted counseling were more likely to change their ideal method choice but less likely to use their ideal method at follow-up. This discordance highlights women’s inability or reluctance to act on their stated preferences in spite of increased access to FP. Follow-up interviews suggest that a longer service period may translate to concordance and uptake. Women who were encouraged to invite their partners reported higher concordance with their partner and switched methods, but were no more likely to be concordant in their use at follow-up. This implies that while these women could realize their preferences, they might have changed their minds due to their partners’ presence at counseling. Taken together, our pilot reinforces the need to assess the trade-offs that women face to realizing their true preferences for FP/RH.

This project is led by the Program on Women’s Empowerment Research (POWER) at the Boston University (BU) Global Development Policy (GDP) Center’s Human Capital Initiative (HCI). The project is aligned with the GDP Center’s mission to advance policy-oriented research that drives positive social change and well-being. The GDP Center is a university-wide center that supports large-scale research initiatives, including HCI. Since its inception, HCI has advanced interdisciplinary research on the role of human capital in development to inform policy solutions to global challenges including poverty, women’s empowerment, and sustainable economic growth.

Over the past three years, the POWER program has supported: 1) global research related to the implementation and analysis of programs that promote women’s empowerment and well-being in LMICs; 2) the dissemination of research and policy findings to local and global policy audiences; and 3) the expansion of data collection and policy engagement efforts. These activities have provided insight on the impacts of programs and interventions in women’s empowerment across a range of domains (FP/RH, education, women’s mobility, political and social participation). POWER comprises an interdisciplinary research team with expertise in the evaluation of women’s empowerment. The project team includes local experts to a) build local research capacity; b) foster ownership of research findings for local consumption; c) facilitate translation of research findings and engagement with local audiences; and d) foster longer-term collaborations with local partners. We have established long-term partnerships with scholars from the University of Malawi College of Medicine and Kamuzu College of Nursing. Partner organizations in FP/RH include the World Bank, the Malawi MOH, the Malawi Reproductive Health Directorate (RHD), and a number of NGO and private sector providers in FP/RH (PSI, Marie Stopes, FPAM, and others).

The project lead, Mahesh Karra, is Assistant Professor of Global Development Policy and Associate Director of the Human Capital Initiative at Boston University. His research utilizes experimental and non-experimental methods to assess the relationships between population, health, and development. Over the past 5 years, Karra has conducted impact evaluations in East Africa and South Asia. Prior to BU, Karra worked at the Population Reference Bureau (PRB) and Futures Group International, where he provided technical assistance to the USAID Health Policy Initiative and the Population and Poverty research network. Karra is a Research Associate at IPA Malawi and oversees several initiatives in Malawi, including two ongoing field experiments in FP/RH. This project will leverage the field sites, partnerships, and infrastructure Karra has identified. As Project Lead, Karra will provide administrative and technical support for the project duration. Karra will oversee the design, supervise field activities, and meet with institutional partners and contacts at the MOH. He will also work with the MOH, NGOs, and local providers to disseminate findings and identify scale-up opportunities. 

We aim to build on our pilot by examining longer-term outcomes of reproductive autonomy and well-being, particularly: 1) sustained contraceptive concordance, i.e. whether a woman’s stated FP preferences and fertility intentions align with her method use over the service period; 2) contraceptive continuation and method switching; 3) satisfaction with FP counseling; and 4) longer-term fertility, health, and well-being. Based on our pilot findings, we expect to observe: 1) a 30% sustained increase in contraceptive concordance; 2) increased switching behavior along with a 20% reduction in overall contraceptive discontinuation; 3) a 15% increase in male partner engagement in the FP counseling process; and 4) a 7 to 10% reductions in unintended pregnancy and unmet need for FP. Finally, our dissemination and outreach efforts, through engagement with communities, will allow us to present our findings at our project site, a catchment area of 40,000 households.

Funding from the MIT Solve Challenge, which will be supplemented with funds from other donors (the Gates Foundation, the Hewlett Foundation) and secured resources from the Global Development Policy Center at Boston University, will cover: 1) data collection and monitoring, including purchase of equipment; 2)  local personnel in Malawi; 3) additional testing, refinement, and implementation of the UCD counseling innovations, including development and refinement of the “Kulera” flipchart app, and post-counseling activities; 4) travel support; and 5) dissemination, capacity building, and outreach activities with stakeholders in Malawi.

Our approach tests how preference-based counseling may shape women’s FP/RH outcomes. Specifically, we study how a woman's FP decision-making is sensitive to: 1) targeted counseling that is adapted to her stated preferences; and 2) the presence of her male partner at counseling. By introducing digital tools for FP/RH, our innovation aims to increase efficiency in FP counseling, which may promote contraceptive concordance and RH outcomes. By taking a client-centered perspective, we build on approaches that seek to improve counseling, either through expanding access to counseling (e.g. through home visits) or improving quality at points of care (e.g. increased provider training). 

Our innovations are designed for potential scale-up (with scope for additional testing) from the onset. Both approaches we test require limited restructuring of existing counseling practices, minimal provider training, and almost no initial public investment. Counseling materials, such as the “Kulera” flipcharts, are already provided by the MOH, and the “digital flipchart” app can be distributed for free on Android-based mobile devices. We are confident that our innovation components are scalable and cost-saving relative to current practices.

In the next 3 to 5 years, we will expand our two UCD-based counseling approaches to an estimated 3,000 women and couples. Based on our pilot study, 52% of women reported concordance with their preferred method and actual method use. After one month of counseling, we observe an 8.8 percentage point (or 17%) increase in women’s contraceptive concordance. Based on this finding, we estimate that extending our timeline to one year and expanding our scope to include additional follow-up would increase and sustain contraceptive concordance by up to 30% from baseline. A second key goal is to examine how our UCD approach and app impact contraceptive discontinuation. In Malawi, 49% of women discontinue contraceptive use within a year for method-related reasons. Although our pilot was not designed to identify discontinuation, we propose that our innovation, as a result of improved concordance, may reduce discontinuation by up to 20% over one year. 

The lessons learned from our pilot and planned project will inform communities of the local FP/RH environment by illustrating the role of UCD counseling in identifying and shaping women’s and couples’ FP attitudes and behaviors. Our findings will also provide insight to communities, providers, and the Malawi MOH on the extent to which women’s preferences for FP/RH services are aligned and, through our innovation, being realized through their behavior. Our Android-based app will also empower providers to develop FP programs that promote effective intra-couple communication and that consider decision-making behaviors of women and their partners, with the aim of improving contraceptive concordance, reproductive autonomy, and other FP/RH outcomes. Through this project, we aim to experiment with features of the app to improve its effectiveness for both clients and providers. Finally, in contributing to women’s empowerment, our project may demonstrate that the benefits of improving FP/RH services are likely to extend beyond the health domain.

Our long-term plan is to have our UCD-based counseling approach and app adopted by FP/RH practitioners who provide contraceptive counseling to women and couples in Malawi and in other countries. Through our project, we will disseminate our findings on the effectiveness of the tailored counseling and the counseling app to providers through local partnerships with FHI360, BLM, FPAM, PSI Malawi, and the World Bank, as well as the Reproductive Health Directorate of the Ministry of Health in Malawi. We have worked with these organizations on previous projects and during the pilot phase, and we will continue to work with them during the design phase to ensure that the counseling process is adapted for a range of settings. On completion of the interventions, we will know the short- and long-term effects of the interventions, and we will collaborate with our local partners to consider how their existing counseling services and practices can be refined to more effectively help women and couples achieve their FP/RH goals. The flexibility with which our counseling process and app can be modified allows providers to more effectively adapt to demand and local constraints. The fact that the app can be accessed, modified, and utilized with any Android device (phone, tablet) also bodes well for its potential to be adopted by service providers in more resource-constrained settings.

As part of the evaluation approach, we will conduct a randomized controlled trial (RCT) that assesses the causal impact of UCD counseling approaches and tools on FP decision-making and behavior. Our key hypothesis is that women’s realized FP behaviors are contingent upon whether targeted counseling reflects their stated FP preferences. We also test the extent to which male involvement in counseling affects outcomes.

Our RCT design follows our pilot design. Following recruitment and a baseline survey, 3,000 women will be randomized into one of 4 groups: 1) a control in which women receive counseling using the default “Kulera” 15-method flipchart; 2) a treatment group in which women receive targeted counseling using a tailored “Kulera” flipchart implemented on an Android app; 3) a treatment group in which women are encouraged to invite their male partners to counseling and receive standard counseling; and 4) a “joint” treatment group in which women are encouraged to invite their male partners and will receive targeted counseling using a tailored flipchart on the Android app. Women will be individually randomized into groups prior to counseling. We will randomize women such that assignment is balanced according to neighborhood, FP use, and stated ideal method.

Following counseling, women will have free transport and FP services at our partner clinic for a 1-year period. We will conduct follow-up surveys with women at the clinic, by phone, and through home visits over this period.

Our design builds on the pilot in these key ways:

1. By extending the service and follow-up period, we can assess longer-term impacts on women’s FP preferences, (dis)continuation, fertility outcomes, and health.

2. We test new and innovative digital approaches to UCD counseling through an Android-based app that we develop. The app streamlines counseling by replacing print flipcharts and by guiding providers through the counseling process, which will improve efficiency and counseling quality. 

3. By expanding our sample size, we have statistical power for additional inference. We can examine interaction effects between our targeted counseling and partner invitation arms. In our pilot, many invited partners were unable to attend a long counseling session. Since our targeted approach improves efficiency, we propose that shorter sessions and an invitation would significantly encourage male participation and improve women’s outcomes. While our pilot provided proof of concept for our innovations, we were only powered to detect impact of each innovation separately. Power calculations indicate that we need 3 times the sample size to disentangle interaction effects from main effects and to detect impacts on longer-term outcomes. Specifically, we need to recruit 2,768 women to detect a 4 percentage point (a 22%) reduction in contraceptive discontinuation over one year with 80% power. Recruiting 3,000 women will allow for a 5-7% attrition rate, which is within range of the pilot and prior studies in Malawi.

Our key beneficiaries (direct clients) for this project are our sample of 3,000 women and their partners. We build on our pilot by examining longer-term outcomes of reproductive autonomy and well-being, particularly: 1) sustained contraceptive concordance, i.e. whether a woman’s stated FP preferences and fertility intentions align with her method use over the service period; 2) contraceptive continuation and method switching; 3) satisfaction with FP counseling; and 4) longer-term fertility, health, and well-being. Based on our pilot findings, we expect to observe: 1) a 30% sustained increase in contraceptive concordance; 2) increased switching behavior along with a 20% reduction in overall contraceptive discontinuation; 3) a 15% increase in male partner engagement in the FP counseling process; and 4) a 7 to 10% reductions in unintended pregnancy and unmet need for FP. Finally, our dissemination and outreach efforts, through engagement with communities, will allow us to present our findings at our project site, a catchment area of 40,000 households. 

We will collect a range of data to measure and evaluate outcomes of interest. Data sources include: 1) survey data from participants at baseline, during counseling, at follow-up at the FP clinic, through phone surveys, and home visits; 2) process evaluation data collected through the app by our field teams to evaluate the implementation effort, particularly the counseling process and post-counseling activities; 3) administrative data from clinics and providers to assess changes to the service environment over time; and 4) qualitative data (IDIs, FGDs) collected from stakeholders, NGOs, and other organizations that operate in FP/RH in Malawi. These data will be collated using a range of tools (quantitative surveys, interview guides, app data, field notes) by our field teams through direct participant interaction (visits, phone interviews) and indirect methods (analysis of administrative health data, etc.).

We test two hypotheses: 1) targeted, client-centered counseling will enable women to effectively identify and realize their contraceptive preferences; and 2) the involvement of supportive male partners in counseling will allow women to translate their stated preferences into behavior.

Programs that have incorporated UCD approaches are informed by insights from behavioral economics and cognitive psychology. Choice architecture studies in marketing and management have noted that cognitive overload may lead to risk-averse behavior and impatience. Studies show that incorporating UCD may serve to “nudge” individuals to make better decisions – when faced with a smaller number of well-defined alternatives, individuals make more effective choices.

While UCD has received increased attention in the health sector, the inclusion of UCD approaches in FP is scarce, and evidence on the role of choice architecture on RH is limited. This is surprising given that FP decision-making is preference-sensitive. In the U.S., researchers have developed a tablet-based support tool, “My Birth Control” and found the tool improved outcomes without the need for extensive provider training, although longer-term outcomes of contraceptive concordance and continued use were not assessed. In LMICs, tools such as the Balanced Counseling Strategy (BCS), in which a decision tree-based algorithm identifies a client’s preferred FP methods, have been promoted as client-friendly approaches to counseling. An evaluation of BCS found that the approach was linked to increased postpartum FP use, especially among women who were supported by their partners. While interest in BCS and other strategies has grown, there is limited evidence on how these approaches impact choice and decision-making. Moreover, it is unclear how providers can implement these strategies, which require significant training and resources, to larger client bases. Therefore, we propose a simplified approach that tailors counseling to a woman's preferred attributes and methods to improve concordance, where women’s FP preferences can be effectively realized through their behavior.

Spousal preferences for FP are a key determinant of women's own access to and use of FP. Evidence on men’s role in FP decision-making remains limited and mixed in LMICs. In Jordan, couples counseling led to higher FP use, but this increase was not different from women who were individually counseled. In Malawi, a peer-delivered intervention found that male involvement increased FP use. In contrast, evidence from Zambia found that women who received a FP voucher together with their husbands were less likely to use FP compared to women who received the voucher alone. Our approach to male involvement is distinct in that women are given the choice to invite their partners to counseling. This approach differs from studies in which couples are selected based on whether they jointly participate.

As part of our pilot study, we developed an integrated database management system to conduct all study activities and monitor data collection. Our data system is developed in CommCare, which is an open-source mobile health (mHealth) platform designed for data collection, client management, decision support, and behavior change communication. The mobile applications we build in CommCare are used by client-facing enumerators and counselors in visits and include audio, image, and video prompts, which are ideal for administering our counseling interventions. CommCare supports J2ME feature phones, Android phones, and Android tablets and can capture photos and GPS readings. CommCare also supports multi-languages and non-Roman character scripts as well as the integration of multimedia (image, audio, and video). CommCare also allow applications to run offline and collected data can be transmitted to CommCareHQ when wireless (GPRS) or Internet (WI-FI) connectivity becomes available. CommCare has been used in a wide range of settings, including on other USAID-funded FP/RH projects.

All monitoring and evaluation data (baseline, post-counseling follow-up) collected during the pilot was administered by our research team using the CommCare system. As part of the management system, a prototype counseling app was developed for counselors to facilitate administration of the counseling sessions. The app streamlined the pre-counseling screening process and the post-counseling follow-up and guided counselors to:

• Administer pre-counseling recruitment protocols and document consent;
• Elicit a client’s most updated preferences for contraception and methods using a screening survey. The identification of a client’s preferences for contraceptive methods, and particularly method-specific attributes, was implemented as a two-stage algorithm:
  • A client was first asked by the counselor to identify all possible features (method-specific attributes) that were most important to her when choosing a contraceptive method. Figure 5 presents the list of method-specific attributes that a client was prompted to consider.
  • Following elicitation of the client’s method-specific attributes, the counselor asked the client to assign a relative ranking to her top three most valued method-specific attributes (e.g. does she prefer that a method have a lower incidence of side effects over a method that is more effective at preventing pregnancy? By how much?). To facilitate the ranking process, clients were given 20 counters and a board that presented the attributes listed in Figure 5. Clients were asked to place counters in piles on top of the attributes that they had identified to be the most important to them – the number of counters assigned to each attribute would reflect the weight of importance of that attribute relative to other identified attributes. These revealed weights were used to identify a ranking of a client’s preference over method-specific attributes. Based on her elicited ranking of method attributes, the counselor, through the app, would identify the attribute that the woman revealed to be most important to her (e.g. methods with low incidence of side effects).
• Counsel the client on a tailored subset of methods based on her elicited preferences. Upon confirming the client’s method attribute preference ranking in the app, the app would identify a colored tailored flipchart that only included the subset of methods that corresponded to women’s stated preferences. Counselors were instructed to use the identified flipchart to counsel the client on a subset of up to five contraceptive methods. Figure 4 presents the various flipcharts, and Table 5 presents the details on the attribute-method correspondence mapping for all pre-designed tailored flipcharts. Particular emphasis was placed on making the subset of methods presented salient, where women were primed to consider the relative ranking of a method across their stated preferred attribute.
• Conduct post-counseling protocols, which included informing clients of the availability of free transport and reimbursement for all FP/RH services at our partner clinic.

Since this app was developed within the same system as our baseline and follow-up surveys, baseline data on women’s contraceptive preferences was integrated into the app to dynamically tailor the counseling process within the app in real time. Moreover, data from counseling was extracted in real time at follow-up, at which point clients were re-interviewed about their experiences with counseling. 

Over the past 5 years, we have partnered with IPA, a U.S.-based non-profit research organization. IPA comprises a group of leading academic researchers in development economics, behavioral economics, and psychology. IPA Malawi has provided technical and research assistance in health and development to governments and NGOs in Malawi, including the MOH. We will work with the IPA team to implement the project and will continue to work with IPA if the project is scaled. IPA Malawi will oversee field activities, including: 1) hiring and training of field staff; 2) data collection and monitoring; 3) UCD counseling and app rollout; and 4) assisting with dissemination to stakeholders.

We will partner with Dimagi, a social enterprise based in Cambridge, MA, and will use their open-source software suite, CommCare, to develop our instruments, collect data, and monitor data quality. We will use CommCare to develop the “Kulera” counseling app, which will adapt the tailored flipcharts into an Android-based interface. The app will be tested as part of our impact evaluation, and we will collaborate with Dimagi to tailor the app platform to larger client bases. We have worked with Dimagi on the pilot study and on other evaluations over the last 8 years.

Finally, we will work with the Malawi MOH and the Malawi RHD on dissemination and will regularly meet with MOH and RHD representatives. The project team will work with the MOH and other organizations, including Population Services International (PSI) and the FP Association of Malawi (FPAM), to ensure that the project and app design is appropriate for the setting and can be effectively measured for scale-up. Pilot funding was granted by the Hewlett Foundation. Through POWER, the Hewlett Foundation has committed additional support for longer-term data collection, testing of the counseling approaches, and refinement of our app prototype.

It is our aim to work with local partners and service providers beyond the scope of this project to ensure that the app and our UCD approach, as well as the lessons learned from the project, help to inform existing counseling practices. Over the next 5 years, we will continue to present the digitized tablet-based app and counseling process in meetings with the MOH, RHD, and a number of NGOs as well as with community-based health workers and service providers in Malawi. Through these engagements, we will refine and update the counseling algorithms and app interface, and we will also train service providers on how they can tailor the app to their best practices. In our current project, clients are recruited through a door-to-door strategy, which more closely reflects community-based approaches to FP/RH service provision. Following the project, we will work closely with the MOH and with service providers to adapt the counseling app and process for clinical settings, where provider availability to counsel clients may be more limited. Given that the app is developed on a free, open-source platform, service providers can tailor the interface to their needs at a minimal cost. Moreover, we hope to train providers on how they can use the data generated from the app to develop counseling processes that can be more effectively tailored to and respond to clients’ changing preferences.

The support and level of interest that we have received from the Ministry of Health, and particularly the Reproductive Health Directorate office at the MOH, was critical to the success of the pilot study and will continue to be critical to the success of the proposed project. We have worked with both the MOH and RHD during the study design phase to ensure the interventions are appropriate and have the potential to be taken to scale in Malawi. To this end, our dissemination efforts and collaboration with the MOH and RHD are engrained in our study design and counseling interventions from the outset – the fact that our counselors are trained by an MOH-approved master trainer speaks to the strength of our collaboration on the intervention design. Over the past two years, we have presented our pilot study findings in policy meetings with the MOH and RHD as well as in research meetings to stakeholders at the International Food Policy Research Institute (IFPRI) and World Bank in Malawi. We have discussed the policy relevance and implications of our findings, particularly the effectiveness and cost effectiveness of our various intervention components in improving outcomes at scale. We have also discussed the role of data collection, particularly process evaluation data, in contributing to a better understanding of program impact.

Our commitments to diversity, equity, and inclusion have been ingrained into our efforts from the onset. In adopting a user-centered approach, we have made it an objective to ensure that clients from all backgrounds can be engaged. Through collaborative efforts and engagement with local stakeholders, particularly the Malawi RHD and local FP service providers, we were able to ensure that our counseling messages, app design, and materials were appropriate for the local context. For the pilot study, we hired a master trainer and family planning counselor from the RHD to train counselors who became part of the intervention team. The RHD and PSI Malawi provided counseling materials (demo kits, contraceptive methods flip charts, training inserts). We recruited counselors, recent Kamuzu College of Nursing graduates, through collaborations with the University of Malawi College of Medicine. We also consulted with practitioners and local agencies to help us develop information materials written in local vernacular, with a recognition of local cultural norms, and that were accessible to all clients who were approached, including clients from low socioeconomic backgrounds, minority ethnolingustic groups, and clients who were illiterate or had lower educational attainment. Finally, we ensured that our intervention components were respectful of male partners and other family members, and encouraged couples to engage in family planning without excluding the male partner.

Our outreach and dissemination strategy has included the production and presentation of briefs, reports, process evaluation assessments, summaries, and blog posts accessible to non-academic audiences, including policymakers, practitioners, and program implementers. We have contributed to the production of these and other materials in English and local languages, and we have hosted workshops, trainings, research seminars, and meetings with high-level public officials at the MOH and RHD, policymakers in women’s health in Malawi, NGOs and policy institutions, universities and research institutions, and local communities where we have worked. We will continue to engage with all of our stakeholders for the purpose of informing the policy discourse through capacity building and advocacy.

The app itself is developed in CommCare on a free, open-source platform and is therefore not costly to update and maintain; it is also part of our team’s longer-term strategy to conduct trainings on the app with service providers and public health personnel who manage information systems for FP/RH - we aim for our app to be adopted by these service providers and integrated into their current practices. These workshops and trainings will promote local ownership of the app and of the counseling process and will allow for more effective targeting. With this said, there may be some costs associated with managing and adequately storing any client data that is collected through the app on a secure server. Currently, costs for data collection and management through the app for research purposes consist of a small monthly subscription fee to maintain server space. In addition, materials for counseling, including contraceptive demo kits, physical copies of the Kulera flipchart, inserts, and brochures, can all be procured from the MOH and public sector either for free or for a small cost. Taken together, the total costs related to implementing the tailored counseling process and the app are, in fact, low. 

Support for the pilot study and preliminary testing of our UCD counseling approaches in Malawi was provided by the Program for Women's Empowerment Research (POWER) at the Boston University Global Development Policy Center (GDP Center) through a research grant from the William and Flora Hewlett Foundation ($100,000 from 2018-2020). Supplemental funding for the project to cover travel, student researcher support, and refinement of the counseling app content was provided by the Human Capital Initiative at the GDP Center and the Rafik B. Hariri Institute of Computing at Boston University ($25,000).