The research team, led by Jaime R. Sevillano Giraldez of Panama’s National Ministry of Health together with collaborators from the University of South Florida, Tulane University, and several Panamanian institutions, applied the Health Belief Model to understand why household source reduction succeeds or stalls. This framework, developed in the 1970s, organizes health decisions around perceived susceptibility, severity, benefits, barriers, self-efficacy, and cues to action. In dengue-endemic settings across Asia, Latin America, and the Caribbean, surveys built on this model have repeatedly shown that favorable attitudes do not guarantee routine preventive practice. A recent meta-analysis found substantially higher prevalence of positive attitudes than of good preventive behaviors, highlighting the persistent gap between what people believe and what they actually do.

Panama reported more than 32,000 dengue cases during the 2024 outbreak, part of unprecedented transmission across the Americas during 2024 and 2025. Twenty-five years of national surveillance demonstrate marked spatial heterogeneity in incidence, reporting, hospitalization, and inferred transmission intensity across the country. These patterns make local context central to prevention, yet Panama lacked a large, district-comparative analysis of perceived barriers to household larval-habitat control. The new study fills that gap with an epidemiological rather than merely psychometric contribution: it quantifies contextual heterogeneity in perceived barriers among surveyed residents despite near-universal intention to remove observed breeding sites.

The analysis drew on the Panama-adapted MosquitoWise survey, administered in Spanish between October 2024 and February 2025 across five districts: Panamá, San Miguelito, Chepo, Arraiján, and Chitré. Because no comprehensive sampling frame was available, participants were recruited through non-probability convenience sampling in community settings. Eligibility required that each respondent be the household member responsible for checking for or eliminating larval habitats, ensuring that the person answering questions about barriers was the same person expected to carry out the task. Of 2,115 records, 2,091 met the analytical criteria after excluding 21 respondents who declined data use and three residents with less than one year in the community.

The questionnaire measured six Health Belief Model constructs using three-category agreement statements, along with breeding-site knowledge, care-seeking knowledge, access to essential services, and behavioral intention. Each construct was scored as an unweighted sum, with the barrier items negatively framed so that higher scores indicated more perceived obstacles. During manuscript revision, after inspecting outcome distributions, the researchers designated the three-item perceived-barriers score as the primary analytical outcome because it retained sufficient variability for adjusted analysis. The two-item self-efficacy score was demoted to exploratory status after revealing a severe ceiling effect, with 88.4 percent of respondents at the maximum. Behavioral intention remained descriptive, and the survey did not measure whether or how often participants actually eliminated breeding sites.

The distribution of responses tells a clear story about measurement challenges in dengue prevention research. Perceived susceptibility averaged 7.96 on a 3-to-9 scale, with 62.9 percent of participants at the maximum. Perceived severity averaged 8.56, with 81.2 percent at the ceiling. Benefits averaged 11.71 on a 4-to-12 scale, with 88.4 percent at the maximum. Cues to action averaged 5.87, with 92.6 percent at the ceiling. Overall, 96.7 percent of participants agreed that an observed breeding site should be eliminated. Against this backdrop of near-universal favorable beliefs, the barriers score averaged 5.24 with a standard deviation of 2.34, and 42 percent of participants sat at its theoretical minimum while 19.6 percent reached its maximum. This variability made barriers the one construct capable of supporting meaningful adjusted analysis.

The psychometric evaluation, conducted using a reproducible random split of the sample, supported cautious within-sample use of the instrument. Parallel analysis in the exploratory half retained six factors, and the confirmatory six-factor model in the other half showed excellent fit: CFI 0.997, TLI 0.996, RMSEA 0.032, and SRMR 0.079. Ordinal reliability ranged from 0.84 to 0.91 across constructs, although conventional Cronbach alpha was modest for ceiling-concentrated two-item scales. Two severity items were nearly redundant, with a polychoric correlation of 0.976. These findings characterize instrument performance in this sample but do not constitute evidence of full external validation, particularly given that the archived materials did not document a formal forward and back translation or cognitive testing.

The primary multivariable linear regression, fitted to 2,079 complete cases with HC3 heteroscedasticity-robust standard errors, explained 14.8 percent of the barriers-score variance and produced the study’s central finding: substantial district-level heterogeneity. Compared with Panamá district, the adjusted mean barrier score was 0.98 points lower in San Miguelito, 1.10 points higher in Chepo, and 1.52 points higher in Arraiján, while Chitré did not differ clearly. Relative to the barriers-score standard deviation, these contrasts corresponded to approximately 0.65, 0.47, and negative 0.42 standard deviations respectively. Each one-point increase in breeding-site knowledge was associated with 0.17 fewer barrier points, and each one-point increase in care-seeking knowledge with 0.51 fewer points. The district contrasts were therefore larger than the per-point breeding-knowledge association and comparable to or larger than the care-seeking association, making geography the dominant contextual signal.

One result defied expectations. Access to essential services was positively associated with perceived barriers, with each additional service corresponding to 0.36 more barrier points. When the researchers disaggregated the four-service index into its components in a mutually adjusted exploratory model, sewer access carried most of this association, corresponding to a 0.67-point higher barriers score, while potable water, electricity, and garbage collection showed imprecise or near-zero estimates. The authors caution that sewer access may be a marker of urban form, population density, drainage, tenure, or other unmeasured contextual features rather than a causal infrastructure effect. Service presence also does not measure continuity or quality. Residual confounding, district composition, sparse exposure categories, and selection into recruitment sites remain plausible explanations, and the result should not be read as evidence that infrastructure increases barriers.

Sensitivity analyses reinforced the robustness of the central pattern. Cumulative-logit models produced the same directions for all principal district, service, and knowledge associations, with ordinal odds ratios of 0.44 for San Miguelito, 2.45 for Chepo, 3.31 for Arraiján, 1.32 for essential services, 0.85 for breeding-site knowledge, and 0.67 for care-seeking knowledge in predicting a higher barriers category. However, nominal-effect tests indicated possible proportional-odds violations for education, district, and breeding-site knowledge, so the ordinal estimates serve as robustness summaries rather than definitive effect estimates. Omitting education, substituting province for district, and excluding 66 influential observations did not overturn the central pattern of district and knowledge associations.

The exploratory self-efficacy model, interpreted cautiously given the severe ceiling effect, revealed small adjusted differences. Female participants had a 0.08-point lower adjusted mean than male participants. Chepo was 0.10 points higher than Panamá, and Arraiján was 0.19 points lower. Breeding-site and care-seeking knowledge were associated with 0.12 and 0.06 higher points per index point respectively. The authors emphasize that these magnitudes, while statistically detectable in a large sample, are behaviorally trivial and poorly discriminated by the scale. Acquiescence and social-desirability processes may also inflate agreement across the favorable-belief constructs, and future instruments should use documented translation and cultural adaptation, cognitive testing, more discriminating response options, and items that separate confidence from actual behavior.

The study’s principal contribution lies in its contextual signal rather than a causal prescription. The district coefficients must not be interpreted as effects of residence itself; they may reflect infrastructure, environmental conditions, service reliability, water storage, vector-control operations, local communication, recruitment venue, interviewer administration, or other unmeasured characteristics. The observed differences justify district-specific investigation rather than ranking districts or generalizing to their populations. Local programs could combine district-specific qualitative work, service-reliability assessment, community mapping, and observed household practices before tailoring communication or source-reduction support. Knowledge messages should be paired with practical opportunity, reliable services, and sustained community engagement.

The authors also acknowledge important limitations. Convenience sampling and an unknown approach denominator create selection bias and prevent population prevalence estimates. Exact recruitment venues, recruiters, schedules, and site identifiers were unavailable, so recruitment-site clustering could not be modeled. Cross-sectional data cannot establish temporality or causality. All substantive measures were self-reported, actual preventive behavior was not measured, and several scales had floor or ceiling concentration. Knowledge indices cover distinct content and are better interpreted as correct-answer counts than reflective scales. District coefficients may combine genuine context with differential recruitment or administration.

Despite these constraints, the findings carry clear implications for dengue control in Panama and beyond. Nearly all participants reported an intention to eliminate observed breeding sites, yet intention was not a measure of action. Perceived barriers varied substantially by district and were lower with greater dengue-related knowledge, suggesting that willingness alone does not guarantee feasibility. The results identify contextual hypotheses for local investigation and tailored prevention rather than supporting representative, behavioral, or causal conclusions. As dengue transmission intensifies across the Americas, understanding where and why residents perceive obstacles to eliminating mosquito breeding sites may prove as important as knowing that they intend to act.

Subject of Research: Perceived barriers to Aedes aegypti larval habitat control among surveyed residents in five Panamanian districts: a cross-sectional health belief model analysis

Article Title: Perceived barriers to Aedes aegypti larval habitat control among surveyed residents in five Panamanian districts: a cross-sectional health belief model analysis

Article References: Sevillano Giraldez, J. R., Oliveira, C., Rodriguez, N., Duncan, C., Shinde, S., Thompson, G., Quintero, W., Acevedo, J., Kirpalani, R., Rodríguez, J., Ortiz Gonzalez, K., Jessany, J., Salcedo, T., Núñez-Samudio, V., Moses, L., & Calvo, A. (2026). Perceived barriers to Aedes aegypti larval habitat control among surveyed residents in five Panamanian districts: a cross-sectional health belief model analysis. Frontiers in Tropical Diseases, 7, Article 1864749. https://doi.org/10.3389/fitd.2026.1864749

Image Credits: AI Generated

DOI: 10.3389/fitd.2026.1864749

Keywords: Perceived, barriers, Aedes, aegypti, larval, habitat, control, surveyed, residents, five, Panamanian, districts