Multilevel barriers to clinical and nutritional research in Latin America: a socioeconomic comparative analysis

Abstract

Introduction: Clinical and nutritional research in Latin America faces significant challenges that limit scientific development and evidence-based healthcare. Understanding these barriers is essential for developing effective strategies to enhance research capacity in the region. This study aimed to identify multilevel barriers to clinical and nutritional research in Latin America and compare them between countries of different socioeconomic levels.

Methods: A cross-sectional study was conducted with 327 healthcare professionals involved in clinical and nutritional research across Latin America. Data collection occurred via an online survey in which participants rated the importance of 16 potential barriers on a 3-point Likert scale. Analysis included descriptive statistics, chi-square tests to compare barriers between upper-middle and lower-middle-income countries, logistic regression to identify predictors of research participation, and k-means cluster analysis to identify researcher profiles.

Results: Funding (84.4%), research materials (71.6%), and time constraints (70.9%) emerged as the most significant barriers across all countries. Three barriers showed statistically significant differences between income levels: participant commitment (73.6% vs. 42.6%, p < 0.001), frequent appointments (56.6% vs. 37.8%, p = 0.02), and language barriers (39.6% vs. 22.9%, p = 0.02), all of which were higher in lower-middle-income countries. Logistic regression identified the importance of research materials (OR = 0.36, p = 0.002) and telemedicine (OR = 1.74, p = 0.044) as significant predictors of research participation. Cluster analysis revealed three distinct researcher profiles based on barrier perception patterns.

Conclusion: Multilevel barriers to research in Latin America are dominated by universal resource constraints (funding, materials, time), with lower-middle-income countries facing additional challenges in participant engagement and study logistics. The relative homogeneity of most barriers across income groups suggests that regional and institutional factors may be more influential than national income levels. These findings provide a foundation for developing targeted strategies to strengthen research capacity and infrastructure across Latin America.

Keywords: Latin America; clinical research barriers; global health equity; health research capacity; knowledge partnership; nutrition science; research inequality; scientific infrastructure.

Figure 1

High-importance research barriers by country income group. Percentage of respondents rating each barrier as high importance, stratified by country income group. Data represent the proportion of participants from lower-middle income (n = 53), upper-middle income (n = 249), and the overall combined sample (n = 302) who classified each barrier as highly important. Barriers are ordered from highest to lowest overall importance.

Figure 2

Barriers showing significant differences between upper-middle and lower-middle income countries. Differences in the proportion of respondents rating each barrier as highly important, comparing upper-middle income (n = 249) with lower-middle income (n = 53) countries. Bars represent the absolute difference in percentage points (upper-middle minus lower-middle). Negative values indicate a higher proportion of lower-middle income respondents rating the barrier as highly important. Only statistically significant differences (p < 0.05) are shown: participant commitment (Δ = −31.0 pp, p < 0.001), frequent appointments (Δ = −18.9 pp, p = 0.017), and language barriers (Δ = −16.7 pp, p = 0.019).

Figure 3

Predictors of research participation in a multivariable logistic regression model. Forest plot showing predictors of research participation in the multivariable logistic regression model (n = 302). Points represent adjusted odds ratios (ORs), and horizontal lines indicate 95% confidence intervals. The vertical dashed line denotes the null value (OR = 1). Predictors with p < 0.05 are shown in blue, while those with 0.05 ≤ p < 0.10 are shown in orange. Significant predictors included high-importance ratings for research materials (OR = 0.36, p = 0.002) and telemedicine (OR = 1.74, p = 0.044). Marginal associations were observed for profession (medicine), female sex, and education level.

Figure 4

Cluster analysis of mean importance scores for research barriers. Heatmap of mean importance scores (1 = high importance, 2 = medium, 3 = low) for 16 research barriers across three researcher clusters (n = 327). Each cell represents the average perceived importance of a barrier within each cluster, with darker shades indicating higher perceived importance (closer to 1). Cluster 1 (n = 49) shows overall moderate barrier perception; Cluster 2 (n = 145) displays consistently high importance across operational and resource-related barriers; and Cluster 3 (n = 133) exhibits uniformly high importance ratings across nearly all barriers. This visualization highlights distinct profiles of barrier perception within the research community.