Researchers have developed a set of locally derived transition probabilities for the EQ-5D health-related quality of life instrument tailored specifically to the Brazilian population.
While a validated EQ-5D-3L value set exists for Brazil, a nationally representative EQ-5D-5L value set is not yet available, necessitating the use of nonparametric crosswalk algorithms. These algorithms typically rely on empirical response-transition probabilities to map the more granular 5L descriptive system onto the 3L tariff.
The study was conducted as a cross-sectional assessment involving 413 Brazilian participants. The sample was drawn from two distinct groups: 310 individuals diagnosed with myasthenia gravis and 103 participants recruited from a social intervention program. This methodology allowed for the direct observation of empirical conditional response probabilities across each of the five health dimensions.
To estimate utilities for all 3,125 possible EQ-5D-5L health states, the research team utilized the validated Brazilian EQ-5D-3L tariff as a weight for the empirical probabilities derived from the sample. These locally generated results were subsequently compared to the estimates produced by the conventional van Hout algorithm, which applied the same Brazilian tariff but utilized European transition probabilities.
The comparison revealed notable differences between the two modeling approaches. The Brazilian-derived model produced a mean utility of 0.375 with a standard deviation of 0.162 across the health states, whereas the European-based model yielded a mean utility of 0.385 with a standard deviation of 0.147. Furthermore, the two models differed in their classification of extreme health states; the Brazilian model classified 1.73 percent of states, or 54 total, as being worse than death, compared to only 0.45 percent, or 14 states, categorized similarly by the European-based algorithm.
Despite these differences in specific utility scores and extreme state classification, the two models showed strong overall agreement in their rankings of health severity. Statistical analysis reported a Lin’s concordance correlation coefficient of 0.984 and a Spearman’s rho of 0.991, indicating that the local model effectively maintained the broad relative ordering established by the standard international crosswalk.
The most significant divergence between the two models was observed at the extreme end of the health severity spectrum. In these severe health states, the Brazilian model consistently assigned utilities up to 0.079 points lower than those predicted by the European model.
The study authors conclude that transition probabilities derived from the Brazilian sample differ substantively from those estimated in European populations. However, they emphasize that it remains unclear whether these variations are driven by population-specific health response patterns or by the specific composition of the study sample. While the locally derived crosswalk offers a valuable, sample-specific option for cost-utility analyses in similar Brazilian populations, the researchers note that these results require further validation in independent datasets and underscore the ongoing need for a definitive, nationally representative EQ-5D-5L value set for Brazil.
Subject of Research: Development and validation of local EQ-5D transition probabilities for health utility estimation in Brazil
Article Title: Locally derived transition probabilities for EQ-5D crosswalk in a Brazilian population: an alternative to European-based mapping algorithms
Article References: Correia, M. G., Von Doellinger, V. R., Tura, B. R., & Santos, M. (2026). Locally derived transition probabilities for EQ-5D crosswalk in a Brazilian population: an alternative to European-based mapping algorithms. Population Health Metrics. https://doi.org/10.1186/s12963-026-00512-2
Image Credits: AI Generated
DOI: 10.1186/s12963-026-00512-2
Keywords: EQ-5D, Health-Related Quality of Life, QALY, Brazil, Crosswalk, Transition Probabilities, Health Technology Assessment, Utility Estimation, Myasthenia Gravis, Mapping Algorithms, Population Health Metrics, Non-parametric Bootstrap, Health Economics, Clinical Assessment, Locally
News Source: Ophelia Keating. (November 17, 2026). Brazilian EQ-5D Mapping Algorithm Offers Alternative to European-Based Standards. Scienmag.



