Ethics statement

The study was reviewed by the KIT Royal Tropical Institute Research Ethics Committee and deemed to pose minimal risk to participants, as it involved anonymized interviews with professionals discussing programmatic experiences. Consequently, the committee granted an ethical waiver (Ref no S-235). Participants received an information letter outlining the purpose of the study, voluntary participation, and confidentiality measures, and provided written informed consent prior to the interviews. No financial or material incentives were provided for participation. Consent was reconfirmed verbally at the start of each interview.

All data were anonymized and stored on a secure server in compliance with the General Data Protection Regulation (GDPR) practices. Access to the server was limited to the lead and supportive researchers, and all transcript data will be deleted 6 months after publication as outlined in the consent form. The study was conducted in line with the ethical principles of the Declaration of Helsinki.

Study design

This study used an exploratory qualitative design to investigate how BFM can be operationalized to improve MDA programs for NTD control. Semi-structured key informant interviews (KII) were chosen as the primary data collection method to capture stakeholder experiences, operational structures, and perceptions of BFM use in sub-Saharan Africa. This design was well suited for exploring an under-researched area that required both conceptual and practical insights. KIIs provide an opportunity to collect information from a range of professionals with firsthand knowledge of the motivations, behaviors and perspectives of both the community and key stakeholders [34,35].

Participant selection and sample size

Participants were purposively selected based on their involvement in MDA program implementation, funding, monitoring and evaluation, or research, with a specific focus on those with experience in the ASCEND-supported countries where implementation of BFM was known to have taken place. Eligible participants held roles in national NTD programs, Ministries of Health, donor agencies, non-governmental organizations (NGOs), and research institutions. The sample included representatives who had contributed to work in a variety of anglophone, lusophone, and francophone countries across sub-Saharan Africa to ensure geographical and institutional diversity. As participants were recruited largely from within our NTD research network, and as prior exposure to BFM was necessary for participation in the study, the potential for selection bias was acknowledged. Efforts were made throughout the research process, including during creation of interview questions, facilitation of interviews, and analysis, to limit the influence of this bias in the study findings. These efforts included standardized interview guides, independent double-coding, and comparison of responses across stakeholder groups, among others. As the study was conducted within ASCEND supported contexts in sub-Saharan Africa, the findings are most directly applicable to preventive chemotherapy NTD programs operating under similar implementation structures, including community-based MDA delivery and routine drug coverage monitoring. While they may be relevant to other MDA platforms, transferability to non-preventive chemotherapy NTD programs or substantially different health system contexts should be interpreted cautiously.

Thirty individuals were invited via email to participate in interviews. Fourteen consented within the two-month data collection period. No trends in roles or geographic locations were noted among non-respondents. Interviews were conducted on a rolling basis as participants consented, until all 14 had been completed. Participants represented a range of roles: funders (n = 2), NGO implementing partner staff (n = 1), monitoring, evaluation and learning (MEL) officers (n = 2), national program staff (n = 6), and researchers (n = 3). The final sample captured diverse perspectives across implementation and policy levels, as shown in Table 1. No further recruitment was pursued after the fourteenth interview given the diversity achieved across stakeholder roles, limited data collection timeframe and recurrence of themes from earlier interviews.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Overview of key informants interviewed by organizational affiliation and role. Abbreviations: F = funder, IP = implementing partner, MS = monitoring, evaluation and learning staff, PI = national NTD program managers, R = professionals working in academia.

https://doi.org/10.1371/journal.pntd.0014373.t001

Sample size was guided by principles of thematic saturation [36] and information power [37] rather than numerical representativeness. Preliminary assessment of thematic recurrence occurred during the data collection period, followed by a more systematic evaluation of saturation was during coding and analysis. While interviews were not fully coded concurrently with data collection, ongoing review indicated a repetition of key themes. After completion of all interviews, comprehensive coding process confirmed that no substantively new codes or thematic categories were emerging across stakeholder groups. Consistent with established operationalizations of thematic saturation [38], recruitment ceased after 14 interviews, as additional interviews reinforced previously identified themes rather than generating new insights.

Although some stakeholder categories included few participants, these individuals held roles with broad oversight or cross-country responsibilities, providing insight across multiple implementation contexts. Overall, participant selection prioritized analytic depth and information-rich informants with direct relevance to BFM implementation in ASCEND-supported settings.

Data collection and instruments

KIIs were conducted via Microsoft Teams from April 17^th until May 23^rd 2024. Each interview lasted between 45 and 60 minutes (mean: 54 minutes). The semi-structured interview guide was based on a review of the literature and programmatic context and piloted prior to implementation. Topics included previously used feedback mechanisms, professional experience with BFM, and perceptions on best practices of integration into MDA programs. The guide was adapted iteratively to explore emerging themes raised through conversations with key informants. Further themes from the semi-structured guide are presented in the Annex 1 Interview Guide, which was piloted prior to the KIIs.

Data analysis

Interview audio recordings were transcribed using Microsoft 365 Word and analyzed using NVivo 1.7.2. A hybrid thematic analysis approach, incorporating both inductive and deductive coding, was used to provide depth and breadth in the analysis and aimed to reduce the risk of confirmation bias [39]. The deductive component was informed by prior operational literature on BFM, including FCDO supported program reports, as well as established implementation frameworks such as the USAID NTD Sustainability Framework [40], which outlines key dimensions of NTD program delivery and sustainability. These sources provided sensitizing concepts that structured early coding while allowing themes to evolve inductively through iterative transcript review. Inductive coding involved detailed reading of the transcripts to derive concepts and themes based on the researcher’s interpretations. Overarching themes found via inductive coding were (1) operational benefits of routinely collected BFM data, (2) challenges in integrating and scaling BFM, (3) strategic role in reaching systematically missed populations, and (4) sustaining trust through BFM and closing the loop.

Initial coding was conducted by the first author. To support reflexivity and minimize bias, memos were kept during the process and code categories were regularly reviewed. Emerging themes were refined in consultation with the last author. For further quality assurance, a purposively selected subset of transcripts (n = 4), representing variation across stakeholder roles, was independently double coded by the first and second authors to assess consistency in code application. Coding discrepancies were discussed in analytic meetings and resolved by consensus, with refinement of code definitions where necessary. Thematic convergence across codes led to the development of key themes, which explored relationships between feedback mechanisms and the operational and strategic aspects of MDA delivery.

As this study drew on experiences from ASCEND-supported contexts, we were attentive to the potential influence of the research team’s professional proximity to the program. While familiarity with NTD implementation strengthened contextual understanding, analytic decisions were supported by structured coding procedures and collaborative theme review. Interviews were conducted by the second author with no prior involvement in ASCEND, and three of the five authors (second, third and fourth authors) were not directly engaged in ASCEND implementation, allowing for internal challenge and balanced interpretation.

Perceptions on how to best operationalize BFM for MDA

Informants acknowledged the many methods of collecting feedback surrounding MDA that were introduced during the ASCEND program in 2020, and that have been used in subsequent programs since its closure. There was significant variation in how BFM were implemented. This variability reflected the adaptability of feedback mechanisms to local contexts and the different methodologies surrounding MDA delivery across settings. Informants recounted various methods of collecting feedback, including the methods mentioned in Fig 1.

In some regions, feedback was systematically embedded into existing tools, like coverage evaluation surveys. These included the Supervisors Coverage Tool (SCT), a small-scale survey conducted immediately following MDA where feedback could be used to improve upon the current MDA campaign, and the Independent Coverage Survey (ICS), a large survey that targets a larger population but is conducted within 3--6 months of MDA.

Other informal methods of collecting feedback, such as community meetings or daily post-MDA debriefs (CDD meetings), played a larger role. Informants acknowledged additional initiatives where feedback was collected through social media, anonymous dropboxes, and non-anonymous daily MDA logs. Informants explained that while dropboxes would yield many comments which were at times not even relevant to MDA, critical information occasionally emerged through these channels which helped programs adjust elements of the MDA.

Reflecting on the outcomes of the various mechanisms for collecting feedback, informants most often acknowledged CDD meetings, where the CDDs would review feedback collected during the campaign and share among each other and with the MDA supervisors and program staff, as a standout opportunity for collecting BFM:

“During MDAs, there were opportunities for CDDs to collect informal feedback while distributing drugs or when they convened at the end of the day to share their experiences and challenges… CDDs were pivotal in both delivering MDA and collecting feedback. In many cases, their interactions with communities during [drug] distribution allowed them to pick up on subtle issues, like hesitancy due to misinformation or logistical problems in reaching certain households. However, follow-up was often limited, which reduced the impact of this feedback.” – PI6.

CDDs were praised for their capacity to identify pivotal local issues as members of the communities where drugs were distributed, but limitations were noted regarding whether feedback was consistently documented during the meetings. Some informants believed that the lack of procedure in documenting the feedback was a missed opportunity which may have inhibited follow-up to the feedback received.

While discussing the best methods for ensuring operationally useful feedback, informants felt that BFM collected in a timely manner, or received during the MDA campaigns and immediately passed on to supervisors, had the best chance of being acted upon. Among these methods of collecting timely feedback, CDD meetings, along with the SCT survey, were cited as approaches that could have the quickest impact, allowing information to flow directly from community members up to supervisors, who were present and often facilitating the meetings. An informant explained how this was useful in their program:

“Supervisors or program leads would often review feedback gathered by CDDs in daily meetings, and if issues came up, like a household needing different hours of outreach or community concerns about side effects, we could immediately adjust plans for the next day’s activities.” – PI6.

The coverage evaluation surveys, both the SCT and ICS, were regarded as a good way to systematically integrate collection of more large-scale feedback in a way that was perceived by informants as less anecdotal than CDD meetings. However, as the ICS operates on a longer time scale, often taking six months to a year from the time of the MDA to generate and analyze results, they would only be helpful in improving subsequent campaigns. Some informants also felt that the data from these surveys was important for supplementing what was discussed in meetings:

“Independent surveys brought an external lens to the challenges communities faced: like fears about side effects or logistical barriers to access. These surveys sometimes uncovered issues not visible in the routine data collected by CDDs or supervisors.” – PI1.

Informants felt that another potentially impactful opportunity for discussing feedback was supervisory meetings, which were often conducted after the MDA campaign had concluded. Here, supervisors and other staff members involved during the MDA convene to discuss their experiences during the MDA campaign. Informants felt that the systematic integration of BFM in the discussion agenda would be very beneficial given the supervisors’ decision-making power. Informants reflected, however, that like with ICS surveys, these changes would only be able to go into effect in subsequent campaigns.

In discussing how changes could be best operationalized, informants discussed the importance of having a systematic process, from integrating the collection of feedback through analysis, monitoring, and ultimately reviewing key feedback to make tactical adjustments to the MDA where necessary.

“The national NTD program is in charge of guiding all levels if there is any change you have to conduct... For example, they may come up with some feedback during data review meetings. They look at the challenges and experiences from the previous MDA, and this feedback helps to adjust for the next MDA. Whether it’s changing data tools, adjusting the template, or modifying the approach for supervision, the program reviews feedback to ensure improvements are made each year.” – MS1.

“… it’s essential to have a way to ensure that feedback isn’t just collected but also acted upon. The country teams were usually good at managing this, but sometimes the resources weren’t sufficient to close the loop effectively.” – F2.

Key informants also revealed that various feedback collected often went by different names depending on the context. For example, in some settings, it was referred to as “community monitoring” or “health worker insights,” emphasizing how terminology and practices varied based on local needs, program design, and stakeholder involvement. A programmatic informant discussed alternative framings for feedback mechanisms in their program:

“We definitely used oral, written, and digital engagement mechanisms, but they were often framed more broadly as tools for outreach or engagement rather than strictly [as] feedback [mechanisms].” – PI2.

Strategic role of BFM in improving drug coverage and reaching systematically missed populations

Informants were asked how feedback received could best be used to improve upon elements of MDA delivery. Most felt that such feedback could help programs tailor elements of their MDA strategies to better fit the local context, thereby facilitating the highest possible drug coverage among populations at risk.

“At the moment that the information comes through in the timeline of MDA or mop-up, you are able to tailor your community mobilization strategy based on what you have heard. For example, if coverage was low in one area, supervisors and CDDs could immediately adjust their approach to address the reasons behind low compliance.” – PI1.

“During the five days of MDA, there was daily feedback with supervisors who would review data from the previous day and tailor strategies to address issues like refusals or low coverage.” – PI1.

Some informants cited examples of how the integration of novel BFM had already been programmatically useful, in some cases helping to find additional people who were missed during the same round of MDA:

“During mop-up activities, we found households that were not reached during the initial MDA despite being reported as covered. This was revealed through feedback during supervision, which then allowed CDDs to go back and administer the drugs.” – MS2.

Furthermore, some informants felt that use of BFM could help programs understand why sub-optimal coverage in specific populations within communities was occurring, potentially providing insights into why some individuals and groups were routinely left out of MDA. These included understanding community dynamics like gender and language barriers:

“There were other barriers, like needing permission from males for women to engage, or people not understanding due to language barriers. These are the things you uncover when you do root-cause analysis. The BFM really helped document this and reveal what we might not have seen otherwise.” – MS2.

Multiple informants referenced that feedback received helped programs to directly address drivers of non-participation in MDA, such as fear of adverse events:

“In [one country], an area reported low compliance because of rumors that MDA is killing people. This originated from an unrelated incident where a woman died during a previous MDA. Feedback mechanisms allowed us to identify this misconception and address it directly with the community.” – PI1.

Perceptions on challenges in integrating and scaling use of BFM

Several key informants flagged two critical issues in monitoring and evaluation of feedback that could inhibit operational changes. The first is that there is no platform where BFM data can be systematically reviewed and acted upon, as there is with other MDA data. For many countries that participated in ASCEND program, significant efforts have been made to streamline data entry and reporting to the District Health Information System 2 (DHIS2) platform. Informants explained that while there are many quantitative data fields on the MDA and its drug distribution performance, none exist for BFM:

“With DHIS2, the issue has always been that it captures quantitative data quite well, things like coverage, treatment numbers, but it doesn’t really have fields or a framework for qualitative feedback. This is where BFM insights are often lost, as there’s no standardized way to input those observations into DHIS2 or link them back to program adjustments. We’ve been advocating for updates to the system, but it’s a challenge.” – MS1.

The second issue is the lack of human resources for reviewing BFM data. Informants reported that there were no dedicated personnel investigating BFM results, which they believed limits the capacity for systematic monitoring and subsequent streamlining into the decision-making process.

Sustaining community trust through BFM and closing the loop

Informants believed that community engagement was important for MDA campaigns and NTD control beyond simply its operational benefits. They also felt BFM could support community engagement and lead to more sustainable program delivery.

“[engagement of marginalized groups beyond the easy to reach target population] was really quite high on the agenda at the time, especially for countries where they were kind of close to elimination and eradication of diseases. Certainly, to maintain that momentum, there was a real […] need to maintain engagement. And really […] keep people on board to reach that last mile.” – PI6.

“Community sensitization and outreach are critical. It’s not just about distributing drugs; it’s ensuring that the community understands the importance of taking them and trusts the safety of the medication. If the community is informed and engaged, the coverage rates improve, and reinfection cycles are reduced.” – PI1.

Informants further detailed how they felt collecting feedback could support the essential task of keeping communities engaged with the MDA program:

“BFM are not just a way to get feedback; it’s also a way to foster engagement. When communities see their input leading to real changes, it motivates them to stay involved.” – MS2.

At the time of the interviews, some informants felt that community engagement was especially important due to community fatigue, which they felt was threatening continued participation in programs. Informants described how fatigue was both an issue for CDDs and community members.

“Some communities show resistance over time, feeling fatigued by the repeated MDAs, particularly if they don’t see tangible results or benefits. This requires more focused engagement and communication to explain the importance of continuing the interventions.” – PI4.

“Fatigue among CDDs after many years in the program is a real issue. They sometimes manipulate data or report good coverage without actually reaching all households, which we catch during data audits or supervision.” – MS2.

In the ASCEND BFM approach, “closing the loop”, or informing community members of the changes that had been made in the past from feedback, was the final stage of the BFM process. When asked about perceptions regarding closing the loop, informants agreed that it was a necessary part of the process.

“The value of the feedback mechanism is to build that trust and value people’s voices and give them a stake in what’s happening. In doing that, they may be more willing to engage, participate, and encourage others to take up services. So, if you don’t close the loop—if you gather information, act on it, and fail to tell people what you’ve done—you’ve broken the(ir) trust, and it doesn’t work.” – R2.

As with the methods of collection and operationalization of BFM, there were mixed opinions about how the loop should be effectively closed. Informants agreed that ‘closing the loop’ should be communicated through trusted and appropriate channels for the specific communities.

“Feedback should be delivered back to the community in ways that make sense to them—like through trusted messengers, community meetings, posters, or local radio.” – R2.

“We spent money on that campaign, on radio, on cars that are working in the neighborhoods... and they say we don’t hear radio a lot... Here the boxes or the cars in the neighborhood, we trust them. So, we know that next time... we invest in cars and this people that are going around.” – MS1.

The MEL specialist commented that one community did not listen to the radio but was rather accustomed to loudspeakers mounted on local cars, which resulted in wasted resources from programs on radio messages that failed to reach community members. This example of communication channels, used to close the loop, must be trusted, appropriate for the local context, and accessible to community members.

Strengths and limitations

A key strength of this study lies in the potential for transferability of its findings across NTD programs implementing MDA. Additionally, the inclusion of perspectives from stakeholders across diverse professional roles and geographic settings further strengthens the relevance of these findings to similar MDA delivery environments.

This study has two limitations. First, recruiting participants from within our existing NTD network may have introduced the possibility of selection bias. Prior exposure to BFM was a prerequisite for participation, which may have influenced the responses and skewed discussion toward conceptual alignment or optimism regarding BFM potential. However, the interviewers sought to mitigate this by attempting to maintain objectivity throughout the research process, including during question formulation, interviewing, and analysis. Second, these findings reflect the perceptions of key informants rather than direct observations or community perspectives, which may constrain insights into how BFM are perceived at the community level.

Recommendations

As MDA is an opportunity not just to distribute drugs but engage with the community, informants recommended that BFM be integrated into existing MDA structures and used consistently to gather timely and actionable insights. For maximum impact, feedback should be analyzed by dedicated staff and included in digital monitoring platforms. The authors recommend that this should be a task for M&E staff who are also looking at drug coverage, to make appropriate associations between community perspectives and drug coverage. Ensuring publication in peer reviewed journals, not only in project reports and other forms of gray literature, can improve visibility and bolster the evidence base to support the use of BFM by national programs. Crucially, community members should be informed about how their feedback is used to ensure transparency, foster program ownership, and strengthen participation.

While these recommendations reflect informant perspectives on strengthening BFM integration, their implementation must be considered within the realities of resource constraints, workforce capacity, and competing program priorities in many endemic settings. Phased integration leveraging existing tools and M&E personnel may improve feasibility, particularly in lower-capacity contexts.

We recommend conducting formal experimental trials that seek to provide empirical evidence of the capacity for BFM to improve drug coverage, which will be fundamental to convince NTD program leadership in both international and national health programs of its value. Such trials should include comparison arms between programs implementing BFM and those that do not, using clearly defined approaches based on successful methods identified in this and similar studies, and be conducted over multiple years to assess sustained adherence. Future research should incorporate direct community perspectives alongside professional stakeholder insights, for example through focus group discussions, participatory workshops, or mixed-method designs that combine qualitative inquiry with coverage data. Integrating community voices will likely be integral to validate, refine, or challenge professional interpretations of BFM effectiveness and to ensure that feedback mechanisms remain responsive to those they are intended to serve.

We furthermore recommend that the WHO work to standardize terminology for BFM, methods of data collection and key indicators, and integrate these components into preventive chemotherapy distribution guidance, guidelines for coverage evaluation surveys and broader M&E frameworks. Future empirical studies should inform the development of standardized BFM questions, reporting metrics and practical guidance on effectively closing the loop with communities.