Identifying and Prioritizing Barriers to TB Prevention and Care in High-Burden Countries: A Community-Engaged Approach Using Best-Worst Scaling
1University of California San Francisco, Center for Tuberculosis, San Francisco, United States of America
2Afrocab Treatment Access Partnership, Lusaka, Zambia
3APCASO, Bangkok, Thailand
4Access to Rights and Knowledge Foundation, Kohima, Nagaland, India
5Johns Hopkins Bloomberg School of Public Health, Center for TB Research, Baltimore, United States of America
*Corresponding author: Andrew.Kerkhoff@ucsf.eduAbstract
Objective
To comprehensively identify the multi-level barriers to tuberculosis prevention and care in high-burden settings, and prioritize them according to their perceived harm and modifiability.
Methods
We conducted a multi-phase stakeholder-engaged study, synthesized barriers from two scoping reviews, and then convened regional workshops in Hyderabad, India, and Nairobi, Kenya, prioritizing TB survivors, community advocates, and frontline healthcare workers (HCWs). Participants refined existing barriers and added new ones. They then completed a Best-Worst Scaling (BWS) exercise to assess the perceived impact (harm) of each barrier, and Likert-based questions to assess perceived modifiability. Hierarchical Bayes (HB) modeling was used to generate mean impact scores (MIS) for each barrier, and mean scores for barrier modifiability (range, 1-4) were calculated.
Findings
81 stakeholders from 28 countries participated, and 65 completed the survey (24% community representatives/advocates, 36% frontline HCWs; 17% were TB survivors). Stakeholder input expanded the barrier set from 31 to 39, with most newly added barriers at the health systems level. The BWS identified systems-level drug and supply challenges (MIS=5.8), patient/community-level financial factors (MIS=5.2), and inadequate provision of holistic care (MIS=4.9) as the highest-impact barriers; 7 of the top 10 barriers overlapped across regions. Asia participants placed 5 barriers, and Africa participants 10 in the high-impact/high-modifiability priority quadrant; four were shared across both: patient/community knowledge, HCW knowledge, HCW attitudes, and lack of community engagement mechanisms.
Conclusion
Incorporating the lived experience of TB-affected groups revealed substantial gaps in literature-derived TB care barriers. The high-impact, modifiable barriers identified provide actionable targets for programs in high-burden settings.
Article notes
Competing Interest Statement
The authors have declared no competing interest.
Introduction
Despite being preventable and curable, tuberculosis (TB) is the leading infectious cause of death globally, killing more than 1.2 million people annually.1 This, in part, reflects persistent gaps between the availability of effective interventions and their successful delivery and uptake.2–5 The TB care cascade involves multiple touchpoints, from prevention among those at risk for TB, to symptom recognition, care-seeking, and treatment initiation and completion for people with TB.6 At each step, barriers operating at multiple levels - patient and community, healthcare worker (HCW), and health system - can undermine individual-level prevention and treatment outcomes as well as community- and country-level progress in the fight against TB.
Complex, multilevel barriers to TB prevention and care are well documented,7–16 with numerous studies from diverse contexts describing factors ranging from limited disease knowledge, TB-related stigma, and gendered factors at the community level, to inadequate knowledge and high workload among HCWs, to fragmented services and insufficient resources at the health system level. However, knowing which barriers exist is not sufficient to improve demand for, access to, and retention in TB services, and the sheer number of documented barriers can make prioritization difficult. Improving care requires both prioritizing which barriers to address and matching them to strategies suited to the local context.
Unfortunately, several knowledge gaps regarding key barriers to TB prevention and care hinder implementation progress in high-burden settings. First, although there is extensive literature on TB barriers, it is unknown how well it reflects the lived realities of people affected by TB and frontline HCWs rather than researcher framing. Second, many relevant studies are more than a decade old and may not accurately reflect the changes in post-COVID-19 service delivery or the contraction in global health funding in 2025 and its harmful effects on TB programs. Third, TB barriers have not been comprehensively synthesized and systematically prioritized based on the harm they cause or the feasibility of modifying them. Without this prioritization, programs often repeat the same activities, expecting different results, and fail to address key barriers, depleting precious resources without achieving meaningful progress.
To address these gaps, we conducted a stakeholder-engaged process with two major objectives: (1) to comprehensively identify barriers to TB prevention and care by integrating the perspectives of TB survivors, affected communities, and frontline HCWs with an evidence synthesis; and (2) to prioritize barriers by integrating their perceived harm with the feasibility of modifying them, identifying the most actionable barriers. These findings are intended to inform a global digital tool that matches priority barriers across diverse settings with the most promising implementation strategies to improve demand, access, and engagement in care across the TB care cascade.
Methods
Process overview
We conducted a multi-phase stakeholder-engaged process combining evidence synthesis with participatory refinement and preference elicitation. The process comprised two phases: (1) literature-based barrier identification from scoping reviews; and (2) regional stakeholder workshops for barrier refinement, expansion, and prioritization (Box 1).
Phase 1: Initial barrier identification
We identified an initial set of barriers through two scoping reviews: an umbrella review synthesizing existing systematic reviews on barriers to TB prevention and care across the care cascade,17 and a focused review on barriers specific to HCWs involved in TB service delivery.18 The final set comprised 31 barriers: 12 at the patient/community level, 6 at the HCW level, and 13 at the health system level (Table 1). We characterized each barrier by: (1) its definition; (2) the step(s) of the TB care cascade it could directly affect (prevention, care-seeking, screening/case-finding, diagnosis, treatment initiation, and treatment adherence), recognizing that some plausibly affect several steps; (3) its influence on TB outcomes (reduced engagement/retention or poor clinical quality); and (4) the level it primarily acts on: patient/community, HCW, or health system. We use “patient/community level” as a framework label encompassing people affected by TB, those at risk, and the wider community. Because HCW behaviors are substantially shaped by their working environment, we defaulted to a system-level classification when attribution was ambiguous.
Phase 2: Regional Stakeholder Workshops
Overview of workshops and participants
We convened two regional workshops, an Asia workshop in Hyderabad, India (September 22-23, 2025), and an Africa workshop in Nairobi, Kenya (October 14-15, 2025), each spanning two days. Day 1 focused on refining, expanding, and prioritizing barriers, while Day 2 focused on evaluating candidate implementation strategies to address the prioritized barriers (reported elsewhere). We deliberately prioritized participation from groups typically underrepresented in TB policy discussions: TB survivors, community advocates, and frontline HCWs. Participants discussed, refined, and expanded the barrier set based on their experience and expertise, and then completed a Best-Worst Scaling (BWS) exercise and modifiability assessment of the barriers. The workshops are described in detail in Box 1.
Best-Worst Scaling and Survey Design
The impact and modifiability assessment were embedded in a self-completed online survey covering up to 39 barriers; two were identified during the later Africa workshop and assessed only by African participants, while 37 were assessed in both workshops. We selected BWS (type 1) over alternatives (e.g., rankings or ratings) to assess the relative perceived impact of barriers, because it provides more robust discrimination among a large number of items while reducing cognitive burden.19,20 The BWS exercise was designed using Sawtooth Software (Discover) and had a near-balanced incomplete block design.21 Participants were shown 28 and 30 choice tasks in the Asia and Africa workshops, respectively, to ensure each barrier was seen three times, allowing for stable individual-level estimates. In each choice task, participants viewed four randomly drawn barriers and selected the one they perceived as causing the “most harm” to good or high-quality TB outcomes in their setting and the one causing the “least harm,” independent of perceived modifiability (Supplementary Figure 1a).
Participants also rated the perceived modifiability of each barrier on a 4-point Likert scale ranging from 1 (not modifiable) to 4 (easy to modify) (Supplementary Figure 1b), assessing modifiability independently of perceived harm. They were asked to consider barriers in the context of settings they know and work in, reflecting on typically available resources, staff, and funding; required approvals or partnerships; existing strengths that could be built upon; and whether changes could realistically be carried out and sustained.
Statistical analysis
BWS data were analyzed using a Hierarchical Bayes (HB) model in Sawtooth Software (Discover),22 to estimate individual-level preference weights for each barrier. These were rescaled to sum to 100 across all barriers to improve interpretation; we subsequently refer to these as “mean impact scores (MIS).” Mean scores (1-4) were also calculated for modifiability ratings. Estimates were determined overall and by region. To generate pooled overall estimates, BWS results for African participants were rescaled to the 37 common barriers assessed in both workshops. Regional estimates used the full barrier set for each workshop (37 for Asia and 39 for Africa), enabling Africa-specific barriers to be positioned within each region’s priority matrix while still allowing cross-regional comparisons. We combined BWS-derived impact scores with Likert-derived modifiability ratings to classify barriers into a 2×2 priority matrix for each region. A barrier was classified as a priority based on impact if its MIS exceeded the overall mean (indicating greater-than-average perceived harm) and as a priority based on modifiability if its mean Likert score was 3 or higher (indicating the barrier was perceived as at least somewhat modifiable).
Ethics statement
The University of California, San Francisco Institutional Review Board determined this study did not meet the threshold for human subjects research (#26-45819).
Results
Workshop participant characteristics
Overall, 81 stakeholders from 28 countries, including 23 designated by WHO as high-burden for TB, HIV-associated TB, or MDR/RR-TB, participated across the two workshops (43 in Asia, 38 in Africa; 47 [58%] in person; Supplementary Table 1). Of the 65 who completed the survey-based barrier assessment (32 in Asia, 33 in Africa), 25% were community representatives, 35% frontline HCWs, 20% policy/decision makers, and 15% researchers; 17% reported being a TB survivor, 57% reported 10+ years of paid or unpaid work in the TB field, 43% reported involvement in national TB program or policy decision-making, and 36% worked primarily in urban settings, 6% in rural settings, and 58% in both (Supplementary Table 2).
Barrier impact, modifiability, and prioritization
Across the 37 common barriers, MIS ranged from 0.8 to 5.8 (mean 2.7, standard deviation 1.3). Those perceived as causing the greatest harm to TB outcomes were drugs and supplies (systems-level; MIS=5.8), financial factors (patient/community-level; MIS=5.2), lack of holistic prevention and care (systems-level; MIS=4.9), financing and resource allocation (systems-level; MIS=4.7), and lack of community engagement mechanisms (systems-level; MIS=4.7; Table 2). Health system-level barriers comprised 6 of the top 10 highest-impact barriers. Despite small differences in the barrier sets evaluated (37 common barriers, plus 2 Africa-specific barriers), there was substantial concordance in prioritization across regions, with seven of the top 10 highest-impact barriers being consistent across Asia and Africa workshops (Supplementary Table 2). Of the two additional barriers evaluated only during the Africa workshop, multisectoral coordination and cross-border and regional coordination ranked 8th and 16th (of 39), respectively.
Overall, 21 (57%) of the 37 common barriers had a mean modifiability score of 3 or higher (indicating they were perceived as at least somewhat modifiable), while no barrier had a mean score of 2 or lower (indicating perception as very difficult or not modifiable) (Table 2). There was substantial concordance in barrier modifiability ratings across regions (Supplementary Table 2).
Priority matrices combining the perceived impact (i.e., harm) and modifiability of each barrier are shown by region in Figure 1. Five barriers in Asia and 10 in Africa fell into the high-impact/high-modifiability quadrant, representing immediate priorities for implementation efforts. Four barriers appeared in the priority quadrant across both regions: patient/community knowledge, HCW knowledge, HCW attitudes, and systems-level lack of community engagement mechanisms. Beyond these common priorities, Asia’s immediate priority quadrant included trust in healthcare (patient/community-level), while Africa’s included stigma (patient/community-level), symptom perception (patient/community-level), HCW communication skills, holistic prevention & care (systems-level), intervention acceptability (systems-level), and drugs & supplies (systems-level). Although drugs & supplies was one of the two highest-impact barriers in both regions, it fell just outside the immediate priority quadrant in Asia due to slightly lower perceived modifiability. Among the 10 highest-impact barriers in each region, 7 in Asia fell into the strategic opportunities quadrant (including the top 3), compared to 4 in Africa (none of the top 3).
Discussion
This stakeholder-engaged process yielded a more complete and practice-informed understanding of the barriers that undermine demand, access, and retention along the TB care cascade in high-burden settings. By centering the voices of TB survivors, community advocates, and frontline HCWs (voices that rarely shape global TB priority-setting),23,24 we identified many barriers missing from prior reviews and developed a pragmatic framework for prioritizing barriers to address in resource-constrained settings. The resulting priorities reflect how TB care is experienced and delivered on the ground, rather than how barriers appear in the published literature, and provide targets for high-impact strategies to improve TB outcomes.
Participants across Asia and Africa converged on the barriers perceived as most harmful to good TB outcomes (7 of the top 10 shared across regions), despite substantial contextual differences. Challenges related to drug and supply reliability, financing and resource constraints, and the absence of holistic, person-centered care surfaced as dominant barriers. The workshops occurred against the backdrop of substantial disruptions to global TB funding in 2025,25,26 which may have influenced stakeholders’ perceptions. However, the cross-regional consistency of these barriers suggests that certain system-level constraints remain universally limiting regardless of regional structures or epidemiology. Stakeholder-driven refinement also highlighted barriers previously underrepresented in the literature, especially those related to coordination, governance, and community engagement. Notably, 7 out of 8 newly identified barriers were at the health system level, suggesting that the published literature adequately captures individual-level factors but underrepresents structural constraints highly salient to those delivering and receiving TB care.
Prioritization matrices provide a structured approach to differentiate between high-impact barriers that programs can address now, from those requiring longer-term investment. Four barriers emerged as consensus priorities for immediate action across both regions: patient/community knowledge, HCW knowledge, HCW attitudes, and lack of community engagement mechanisms. Because these span all three levels, they reinforce the need for coordinated, multilevel implementation strategies rather than isolated ones. Beyond these, regional variation emerged: Africa’s immediate priority quadrant additionally included systems-level drugs and supplies, holistic prevention and care, and intervention acceptability, patient/community-level stigma, and symptom perception, and HCW communication skills, while Asia’s included patient/community-level trust in healthcare. Such variation likely reflects differences in cultural norms, available resources, perceptions of government support, and prior exposure to successful care-improvement strategies. Nonetheless, the inclusion of both individual- and systems-level priorities across both regions demonstrates that communities and frontline HCWs are keenly aware of structural constraints, not solely of interpersonal interactions.
The barrier prioritization matrices should not be interpreted rigidly. While barriers in the immediate priority quadrant represent clear targets, several high-impact barriers in the strategic opportunities quadrant also warrant attention, despite being perceived as less modifiable — particularly in Asia, where 7 of the 10 highest-impact barriers fell into the strategic opportunities quadrant versus 4 in Africa. Patient financial factors, costs of care, financing and resource allocation, and TB commitment and policy fell into this quadrant in both regions; these are barriers where change is harder but the potential impact is substantial. Holistic prevention and care and intervention acceptability illustrate regional variation, perceived as changeable in Africa (immediate priorities) quadrant) but less modifiable in Asia (strategic opportunities). The structural and policy changes these barriers require, along with shifts in how interventions are developed and delivered, typically take longer than standard project cycles allow and are difficult to decouple from deeply entrenched practice,27 so sustained commitment is likely necessary to achieve meaningful improvements.1,3 Notably, even these barriers were rated as at least somewhat modifiable, suggesting stakeholders saw change as achievable. Programs must therefore weigh not only where change is easiest, but where it matters most.
TB programs have long emphasized individual-level behavior change strategies, such as health education, adherence support, and provider training, and stakeholders confirmed the importance of knowledge and attitudes. Most programs, however, already conduct TB sensitization and outreach to improve community knowledge and address symptom perception, and conduct training to address HCW knowledge and attitudes; this suggests the issue is not whether such activities are happening but how well they are operationalized. 5 Programs may need to reconceptualize how these strategies are delivered (e.g., who delivers them, in what format, how often, and with what content) rather than simply doing more of the same, to improve their reach, resonance, and impact.
This study should not be interpreted as defining the “true” priorities for any specific country or program. Rather, it provides a comprehensive, stakeholder-validated set of 39 barriers that can serve as a global reference for programs to systematically consider in their own settings, alongside a structured approach for prioritization based on local circumstances, resources, and policy windows. These findings serve as a foundation for TB-DASH,28 a global framework and tool to enhance TB demand, access, and care-seeking by modifying priority barriers by linking them to curated and highly promising evidence-based implementation strategies (i.e., “change strategies”).29,30 The tool aims to help programs move from identifying priority barriers to actionable, feasible, and context-specific strategies that can accelerate progress in high-burden settings.
This study has several strengths. Participants from 28 countries across the two highest-burden regions (Asia and Africa) contributed perspectives that reshaped the barrier landscape in ways the literature alone could not reveal. The in-person workshop format provided dedicated time and space for meaningful deliberation and refinement, and the community technical advisors and UCSF TRAC Community Advisory Board played key roles throughout, strengthening accessibility, trust, and engagement. The study also has limitations. We deliberately prioritized the voices of TB survivors, community advocates, and frontline HCWs, but policymakers and senior program managers, who were engaged later, may have different perceptions of barrier importance and modifiability. The sequential workshop design introduced asymmetry in the barrier sets evaluated across regions, but substantial overlap in top-ranked barriers suggests that this did not meaningfully bias regional comparisons. English-language requirements, while pragmatically necessary, may have hindered full engagement for some participants. BWS and modifiability assessments capture perceptions rather than objective effects or actual feasibility; these perceptions remain essential for designing realistic strategies but may not fully align with epidemiologic evidence or structural constraints. While the number of participants is modest, the depth of engagement, including two-day hybrid workshops with dedicated time for deliberation, refinement, and prioritization, yielded richer insights than larger, more superficial surveys could. Finally, while participants represented many diverse high-TB burden countries across Asia and Africa, findings may not generalize to Eastern Europe, Latin America, or the Western Pacific.
In conclusion, engaging TB survivors, community members, and frontline HCWs produced a more complete and actionable understanding of the barriers undermining TB prevention and care in Asia and Africa, demonstrating that centering these voices is not merely an ethical imperative but also essential for comprehensively identifying and prioritizing barriers that matter most. The priority barriers identified span patient/community, HCW, and health system levels, highlighting both immediate opportunities for impact, where change is perceived as highly feasible, and system-level constraints that require sustained investment but could offer substantial rewards. As TB programs navigate difficult decisions about resource allocation amid shifting global health funding, community-centered, stakeholder-driven barrier prioritization offers a principled approach to focusing implementation efforts where they are most needed and most likely to succeed.
Supporting information
Data Availability
All data produced in the present work are contained in the manuscript
Acknowledgements
We thank UC TRAC TB Community Advisory Board members and our stakeholder participants who made crucial contributions to this work by sharing their expertise and lived experience.
Funding
The parent project, Developing a Global Framework for Enhancing Demand, Access and Seeking of TB-related Healthcare Services (TB-DASH) is funded by Johnson & Johnson through an educational grant. The funder had no role in study design, data collection, analysis, interpretation of findings, or the decision to publish.
Competing interests
None declared.
Untitled section
Boxed Text
Stakeholder workshop design and procedures
Workshop preparation
Before the regional workshops, we refined barrier language to ensure accessibility to diverse stakeholders, including those without technical or medical backgrounds. First, an artificial intelligence (AI) model (Claude, Anthropic) provided a plain-language review (at an 8th-grade reading level) to identify overly technical terms or concepts and suggest accessible alternatives; all AI-suggested revisions were reviewed by the research team before adoption. The University of California, San Francisco Tuberculosis Research Advancement Center (UCSF TRAC) TB Community Advisory Board, largely comprised of TB survivors and community advocates from high-burden countries, then reviewed the barriers for resonance and appropriateness through a virtual-guided exercise on Zoom.
Participant recruitment and selection
We deliberately prioritized participation from groups typically underrepresented in TB policy discussions: TB survivors, community advocates, and frontline HCWs. This intentional focus reflected our aim to capture diverse perspectives that may be missing from the published literature, which is often shaped by the viewpoints of researchers. Participants were recruited purposively through TB advocacy networks, community-based organizations, national TB programs, and the research team’s professional networks. Eligibility criteria included direct experience with TB (as a survivor, caregiver, community advocate, or healthcare provider) or expertise in TB program implementation; residence or primary work in a high TB burden country in the relevant region; and ability to participate in English-language discussions. We carefully curated the list of participant invitations to ensure geographic representation across multiple countries within each region, balance across stakeholder types (TB survivors, community advocates, frontline HCWs, program implementers), relevant expertise distributed across the three barrier levels, and a mix of perspectives from different healthcare settings, including public and private sectors and urban and rural areas. Enrollment was capped at approximately 30 in-person participants per workshop to enable meaningful participation and discussion. The UCSF TRAC TB CAB emphasized the importance of including frontline HCWs and community members from rural areas, while final participant selection was guided by the research team and embedded community technical advisors (TB survivors and experienced advocates, residing in high-burden countries). To ensure meaningful participation and reduce financial barriers to engagement, we covered all travel, accommodation, and meal costs for in-person participants. A virtual participation option was offered for individuals unable to travel.
Workshop structure
Day 1 for each workshop comprised three parts: presentation of the barriers, participatory refinement and expansion, and prioritization. First, the study team community technical advisors presented the barriers one by one, grouped by theme and level, with opportunities for clarification and questions throughout. To accompany this process, participants received a barrier reference guide describing each of the 31 literature-derived barriers, including its definition in lay language, the TB care cascade step(s) it most directly affects, and its influence on TB prevention and care outcomes (Supplementary File 1). To refine and expand the barrier list, in-person and online participants were divided into three groups corresponding to the barrier levels (patient/community, HCW, health system), each assigned to the level most aligned with their lived or professional experience. During a 30-minute session, groups reviewed barriers assigned to their level and documented proposed refinements, additions, or consolidations on large poster paper. This was followed by a 30-minute crosswalk in which each group rotated to review, discuss, and further refine the barriers of the other two levels. Facilitators compiled proposed additions and refinements, and each group was then given dedicated time to report back their high-level findings to the larger group. New barriers meeting group consensus and deemed distinct from the literature-derived set were added to the final barrier set for further prioritization.
Barrier assessment and real-time results sharing
Following the refinement process, participants completed a Best-Worst Scaling exercise and modifiability assessment of the revised barrier set. A short presentation explained both assessment formats and emphasized the importance of honest feedback. Participants had one hour to complete the assessments on their personal devices or on computers provided by the research team. Technical support was available throughout (for both in-person and online participants), allowing for the clarification of any questions about barrier definitions. BWS and modifiability data were analyzed immediately following survey completion, and preliminary findings were presented to workshop participants before the end of Day 1. Real-time results sharing enabled stakeholder verification that the findings aligned with their collective understanding, facilitated discussion and co-interpretation of the results with the community that generated the data, and increased transparency and ownership of the findings.