Lorena Guadalupe Barberia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Lorena Guadalupe Barberia has a total of 183 co-authors. Here are the top co-authors by number of shared publications:
- Tatiane Cristina Moraes Sousa
Affiliation Department of Epidemiology, Institute of Social Medicine, University of State of the Rio de Janeiro, Rio de Janeiro, Brazil
6 shared publications - Alexandra Crispim Boing
Universidade Federal de Santa Catarina, Brazil
4 shared publications - Isabel Seelaender Costa Rosa
Universidade de São Paulo, Brazil
4 shared publications - Marcia Caldas Castro
Harvard T.H. Chan School of Public Health, Boston, MA, USA
4 shared publications - Natália de Paula Moreira
Universidade de São Paulo, Brazil
4 shared publications - Antonio Fernando Boing
Universidade Federal de Santa Catarina
3 shared publications
Latest Publications
- Open Access
- 4 July 2026
- Published by Springer Nature in Journal of Computational Social Science
- Vol. 9 (3)
Sentiment and stance detection in COVID-19 vaccine discourse: a systematic review
Advances in machine learning have enhanced researchers’ ability to detect vaccine hesitancy on social media using Natural Language Processing. Our objective in this study was to evaluate, through a systematic review of studies that employed machine learning to analyze sentiment and stance regarding COVID-19 vaccines on Twitter in assessing their methodological quality and consistency. We searched for papers published between 1 January 2020 and 31 December 2023 in PubMed, Web of Science, and Scopus. The inclusion criteria were the use of supervised machine learning to assess COVID-19 vaccine hesitancy through stance detection or sentiment analysis on Twitter/X. We categorized the studies according to a taxonomy of five dimensions: tweet sample selection approach, self-reported study type, classification typology, annotation codebook definitions, and interpretation of results. We analyzed the risk of bias in the included studies, examining whether stance detection was used to report different hesitancy trends compared to those using sentiment analysis. We identified 51 papers that were published in 36 journals. Our review found that measurement bias is widely prevalent in studies employing supervised machine learning to analyze sentiment and stance toward COVID-19 vaccines and vaccination. The reporting errors are sufficiently serious to hinder the generalizability and interpretation of these studies, making it difficult to determine whether discursive content communicates reluctance to vaccinate against SARS-CoV-2. Our findings underscore the importance of more transparent reporting of NLP methods in vaccine discourse studies. Addressing methodological shortcomings is essential to improving our understanding of vaccine hesitancy on social media.
- 4 July 2026
- Published by Springer Nature in Journal of Computational Social Science
- Vol. 9 (3)
Open AccessSentiment and stance detection in COVID-19 vaccine discourse: a systematic review
Advances in machine learning have enhanced researchers’ ability to detect vaccine hesitancy on social media using Natural Language Processing. Our objective in this study was to evaluate, through a systematic review of studies that employed machine learning to analyze sentiment and stance regarding COVID-19 vaccines on Twitter in assessing their methodological quality and consistency. We searched for papers published between 1 January 2020 and 31 December 2023 in PubMed, Web of Science, and Scopus. The inclusion criteria were the use of supervised machine learning to assess COVID-19 vaccine hesitancy through stance detection or sentiment analysis on Twitter/X. We categorized the studies according to a taxonomy of five dimensions: tweet sample selection approach, self-reported study type, classification typology, annotation codebook definitions, and interpretation of results. We analyzed the risk of bias in the included studies, examining whether stance detection was used to report different hesitancy trends compared to those using sentiment analysis. We identified 51 papers that were published in 36 journals. Our review found that measurement bias is widely prevalent in studies employing supervised machine learning to analyze sentiment and stance toward COVID-19 vaccines and vaccination. The reporting errors are sufficiently serious to hinder the generalizability and interpretation of these studies, making it difficult to determine whether discursive content communicates reluctance to vaccinate against SARS-CoV-2. Our findings underscore the importance of more transparent reporting of NLP methods in vaccine discourse studies. Addressing methodological shortcomings is essential to improving our understanding of vaccine hesitancy on social media.
- Open Access
- 3 March 2026
- Published by Springer Nature in Archives of Public Health
- Vol. 84 (1)
Implementing a multi-centric cohort study during the SARS-CoV-2 pandemic in a resource-constrained setting: challenges, adaptive solutions, and lessons from Brazil
Conducting multicenter, longitudinal epidemiological studies on emerging pathogens in real time during health emergencies presents significant challenges,
- 3 March 2026
- Published by Springer Nature in Archives of Public Health
- Vol. 84 (1)
Open AccessImplementing a multi-centric cohort study during the SARS-CoV-2 pandemic in a resource-constrained setting: challenges, adaptive solutions, and lessons from Brazil
Conducting multicenter, longitudinal epidemiological studies on emerging pathogens in real time during health emergencies presents significant challenges,
- Open Access
- 1 January 2026
- Published by FapUNIFESP (SciELO) in Revista da Sociedade Brasileira de Medicina Tropical
- Vol. 59, e0151
Prevalence of SARS-CoV-2 Infection Among Vulnerable Populations Relying on Public Health Services: Findings from the AVISA Study in Brazil
Background: Understanding SARS-CoV-2 seroprevalence among vulnerable populations is crucial, especially in countries with universal but resource-constrained healthcare systems, such as Brazil. Methods: This was a cross-sectional, population-based household survey within the AVISA study, which aimed to assess seroprevalence among households enrolled in the Family Health Program (ESF). Two serological assays (rapid test and ECLIA) were used to detect anti-SARS-CoV-2 antibodies. Descriptive statistics were calculated and survey-adjusted Poisson regression models with robust variance were estimated, yielding prevalence ratios (PRs) and 95% confidence intervals (CIs). Results: The overall weighted seroprevalence was 35.66% (95% CI: 30.55-40.77) in 2,986 participants. Compared with individuals in the Southeast region, participants in the Northeast and North regions had a 1.48- and 1.38-times higher prevalence of SARS-CoV-2 testing positivity, respectively. Relative to individuals who self-identified as having Black or Brown skin color using IBGE-defined categories, White had a 22% lower prevalence of seropositivity (PR = 0.78, 95% CI 0.63-0.97). Compared to those living in households with four to five family members, individuals living in households with six or more family members had a higher prevalence of seropositivity (PR = 1.34, 95% CI 1.05-1.71). Individuals in the lowest per capita household income quintile were more likely to have SARS-CoV-2 infection (PR = 1.89, 95% CI 1.21-2.95) than those in the highest. Conclusion: This study is the first to assess SARS-CoV-2 seroprevalence among ESF participants, a socioeconomically vulnerable population, served by the primary care public health system. Lower socioeconomic status was associated with higher seroprevalence.
- 1 January 2026
- Published by FapUNIFESP (SciELO) in Revista da Sociedade Brasileira de Medicina Tropical
- Vol. 59, e0151
Open AccessPrevalence of SARS-CoV-2 Infection Among Vulnerable Populations Relying on Public Health Services: Findings from the AVISA Study in Brazil
Background: Understanding SARS-CoV-2 seroprevalence among vulnerable populations is crucial, especially in countries with universal but resource-constrained healthcare systems, such as Brazil. Methods: This was a cross-sectional, population-based household survey within the AVISA study, which aimed to assess seroprevalence among households enrolled in the Family Health Program (ESF). Two serological assays (rapid test and ECLIA) were used to detect anti-SARS-CoV-2 antibodies. Descriptive statistics were calculated and survey-adjusted Poisson regression models with robust variance were estimated, yielding prevalence ratios (PRs) and 95% confidence intervals (CIs). Results: The overall weighted seroprevalence was 35.66% (95% CI: 30.55-40.77) in 2,986 participants. Compared with individuals in the Southeast region, participants in the Northeast and North regions had a 1.48- and 1.38-times higher prevalence of SARS-CoV-2 testing positivity, respectively. Relative to individuals who self-identified as having Black or Brown skin color using IBGE-defined categories, White had a 22% lower prevalence of seropositivity (PR = 0.78, 95% CI 0.63-0.97). Compared to those living in households with four to five family members, individuals living in households with six or more family members had a higher prevalence of seropositivity (PR = 1.34, 95% CI 1.05-1.71). Individuals in the lowest per capita household income quintile were more likely to have SARS-CoV-2 infection (PR = 1.89, 95% CI 1.21-2.95) than those in the highest. Conclusion: This study is the first to assess SARS-CoV-2 seroprevalence among ESF participants, a socioeconomically vulnerable population, served by the primary care public health system. Lower socioeconomic status was associated with higher seroprevalence.
- Open Access
- 1 October 2025
- Published by Elsevier in The Lancet Regional Health - Americas
- Vol. 51, 101252
Regional and national estimates of children affected by all-cause and COVID-19-associated orphanhood and caregiver death in Brazil, by age and family circumstance: a modeling study
- 1 October 2025
- Published by Elsevier in The Lancet Regional Health - Americas
- Vol. 51, 101252
Open AccessRegional and national estimates of children affected by all-cause and COVID-19-associated orphanhood and caregiver death in Brazil, by age and family circumstance: a modeling study
- Open Access
- 2 July 2025
- Published by Elsevier in Social Science & Medicine
- Vol. 383, 118352
How to incorporate social vulnerability into epidemic mathematical modelling: recommendations from an international Delphi
- 2 July 2025
- Published by Elsevier in Social Science & Medicine
- Vol. 383, 118352
Open AccessHow to incorporate social vulnerability into epidemic mathematical modelling: recommendations from an international Delphi
- Open Access
Clarifying Misconceptions in COVID-19 Vaccine Sentiment and Stance Analysis and Their Implications for Vaccine Hesitancy Mitigation: A Systematic Review
Background Advances in machine learning (ML) models have increased the capability of researchers to detect vaccine hesitancy in social media using Natural Language Processing (NLP). A considerable volume of research has identified the persistence of COVID-19 vaccine hesitancy in discourse shared on various social media platforms. Methods Our objective in this study was to conduct a systematic review of research employing sentiment analysis or stance detection to study discourse towards COVID-19 vaccines and vaccination spread on Twitter (officially known as X since 2023). Following registration in the PROSPERO international registry of systematic reviews, we searched papers published from 1 January 2020 to 31 December 2023 that used supervised machine learning to assess COVID-19 vaccine hesitancy through stance detection or sentiment analysis on Twitter. We categorized the studies according to a taxonomy of five dimensions: tweet sample selection approach, self-reported study type, classification typology, annotation codebook definitions, and interpretation of results. We analyzed if studies using stance detection report different hesitancy trends than those using sentiment analysis by examining how COVID-19 vaccine hesitancy is measured, and whether efforts were made to avoid measurement bias. Results Our review found that measurement bias is widely prevalent in studies employing supervised machine learning to analyze sentiment and stance toward COVID-19 vaccines and vaccination. The reporting errors are sufficiently serious that they hinder the generalisability and interpretation of these studies to understanding whether individual opinions communicate reluctance to vaccinate against SARS-CoV-2. Conclusion Improving the reporting of NLP methods is crucial to addressing knowledge gaps in vaccine hesitancy discourse.
Open AccessClarifying Misconceptions in COVID-19 Vaccine Sentiment and Stance Analysis and Their Implications for Vaccine Hesitancy Mitigation: A Systematic Review
Background Advances in machine learning (ML) models have increased the capability of researchers to detect vaccine hesitancy in social media using Natural Language Processing (NLP). A considerable volume of research has identified the persistence of COVID-19 vaccine hesitancy in discourse shared on various social media platforms. Methods Our objective in this study was to conduct a systematic review of research employing sentiment analysis or stance detection to study discourse towards COVID-19 vaccines and vaccination spread on Twitter (officially known as X since 2023). Following registration in the PROSPERO international registry of systematic reviews, we searched papers published from 1 January 2020 to 31 December 2023 that used supervised machine learning to assess COVID-19 vaccine hesitancy through stance detection or sentiment analysis on Twitter. We categorized the studies according to a taxonomy of five dimensions: tweet sample selection approach, self-reported study type, classification typology, annotation codebook definitions, and interpretation of results. We analyzed if studies using stance detection report different hesitancy trends than those using sentiment analysis by examining how COVID-19 vaccine hesitancy is measured, and whether efforts were made to avoid measurement bias. Results Our review found that measurement bias is widely prevalent in studies employing supervised machine learning to analyze sentiment and stance toward COVID-19 vaccines and vaccination. The reporting errors are sufficiently serious that they hinder the generalisability and interpretation of these studies to understanding whether individual opinions communicate reluctance to vaccinate against SARS-CoV-2. Conclusion Improving the reporting of NLP methods is crucial to addressing knowledge gaps in vaccine hesitancy discourse.
- Open Access
Regional and national estimates of children affected by all-cause and COVID-19-associated orphanhood and caregiver death in Brazil, by age and family circumstance
Children affected by orphanhood of any cause may benefit from assessment and referral to appropriate services. Timely and accurate data can guide policy. We leveraged new data sources expanding previous reports on national COVID-19-associated orphanhood to estimate national and sub-national numbers of children newly affected by death of parents and co-residing elderly caregivers due to all-causes and to COVID-19-associated causes in 2020-2021 in Brazil. We estimated that 1,300,000 (95% uncertainty interval 1,190,000, 1,430,000) children in Brazil experienced loss of one or multiple parents and/or co-residing caregivers 60+. 673,000 (652,000, 690,000) children were estimated to have lost one or both parents, of which 149,000 (144,000, 154,000) were COVID-19-associated; 635,000 (534,000, 758,000) children were estimated to have lost a co-residing grandparent or other kin, of which 135,000 (85,900, 199,000) were COVID-19-associated. Orphanhood estimates varied across states. The highest all-cause rate of parental orphanhood was in Roraima, at 17.5 (95% uncertainty interval 15.6, 20.6) per 1000 children, and the lowest was in Santa Catarina, at 9.5 (8.7, 10.4) per 1000 children. COVID-19-associated orphanhood was also unevenly distributed, with Mato Grosso experiencing the greatest rate, at 4.4 (3.9, 5.3) per 1000 children, while Pará experienced the lowest rate of 1.4 (1.2, 1.8) per 1000 children. We compared our estimates with administrative data for COVID-19-associated orphanhood (from Brazil’s civil registry offices and manually reviewed death certificates in Campinas) and found that a similar demographic distribution of orphanhood. However, our estimates suggested that administrative sources undercount orphanhood, suggesting that approximately 32% and 56% of total orphanhood was captured in the two datasets, respectively. Our findings highlight the extent of orphanhood in Brazil and the large inequalities between states. Our comparisons with administrative data both validate our model and suggest that strengthening vital registration systems can put children at the center of public health responses globally.
Open AccessRegional and national estimates of children affected by all-cause and COVID-19-associated orphanhood and caregiver death in Brazil, by age and family circumstance
Children affected by orphanhood of any cause may benefit from assessment and referral to appropriate services. Timely and accurate data can guide policy. We leveraged new data sources expanding previous reports on national COVID-19-associated orphanhood to estimate national and sub-national numbers of children newly affected by death of parents and co-residing elderly caregivers due to all-causes and to COVID-19-associated causes in 2020-2021 in Brazil. We estimated that 1,300,000 (95% uncertainty interval 1,190,000, 1,430,000) children in Brazil experienced loss of one or multiple parents and/or co-residing caregivers 60+. 673,000 (652,000, 690,000) children were estimated to have lost one or both parents, of which 149,000 (144,000, 154,000) were COVID-19-associated; 635,000 (534,000, 758,000) children were estimated to have lost a co-residing grandparent or other kin, of which 135,000 (85,900, 199,000) were COVID-19-associated. Orphanhood estimates varied across states. The highest all-cause rate of parental orphanhood was in Roraima, at 17.5 (95% uncertainty interval 15.6, 20.6) per 1000 children, and the lowest was in Santa Catarina, at 9.5 (8.7, 10.4) per 1000 children. COVID-19-associated orphanhood was also unevenly distributed, with Mato Grosso experiencing the greatest rate, at 4.4 (3.9, 5.3) per 1000 children, while Pará experienced the lowest rate of 1.4 (1.2, 1.8) per 1000 children. We compared our estimates with administrative data for COVID-19-associated orphanhood (from Brazil’s civil registry offices and manually reviewed death certificates in Campinas) and found that a similar demographic distribution of orphanhood. However, our estimates suggested that administrative sources undercount orphanhood, suggesting that approximately 32% and 56% of total orphanhood was captured in the two datasets, respectively. Our findings highlight the extent of orphanhood in Brazil and the large inequalities between states. Our comparisons with administrative data both validate our model and suggest that strengthening vital registration systems can put children at the center of public health responses globally.
- Open Access
- 24 January 2025
- Published by SAGE Publications in Journal of Politics in Latin America
- Vol. 17 (2) , 157-186
The Cycle Continues: An Assessment of Democratic Transitions in Latin America with New Data and Indicators
This study examines how theoretically driven indicators modify our understanding of transition dynamics in Latin America. Our indicators show that a noticeable wave of reversion to authoritarian rule in Latin America occurred during the last two decades. Moreover, unlike the past, where authoritarian regimes were established with military coups, the Democracy-Dictatorship Reprise for Latin America (DDRLA) confirms that the region's recent experience with democratic reversion has been brought about by presidents who secured office through elections and who manipulated rules to remain in office for more than two terms. For the period from 1946 to 2022, we discuss how the minimalist conception of democratic regimes can be further advanced with our reformed political regime indicator, as well as measures of initial elections and political regime transitions. We then demonstrate the utility of the DDRLA by replicating studies recently published in leading journals.
- 24 January 2025
- Published by SAGE Publications in Journal of Politics in Latin America
- Vol. 17 (2) , 157-186
Open AccessThe Cycle Continues: An Assessment of Democratic Transitions in Latin America with New Data and Indicators
This study examines how theoretically driven indicators modify our understanding of transition dynamics in Latin America. Our indicators show that a noticeable wave of reversion to authoritarian rule in Latin America occurred during the last two decades. Moreover, unlike the past, where authoritarian regimes were established with military coups, the Democracy-Dictatorship Reprise for Latin America (DDRLA) confirms that the region's recent experience with democratic reversion has been brought about by presidents who secured office through elections and who manipulated rules to remain in office for more than two terms. For the period from 1946 to 2022, we discuss how the minimalist conception of democratic regimes can be further advanced with our reformed political regime indicator, as well as measures of initial elections and political regime transitions. We then demonstrate the utility of the DDRLA by replicating studies recently published in leading journals.
- Open Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 365-376
It’s about What and How you say it: A Corpus with Stance and Sentiment Annotation for COVID-19 Vaccines Posts on X/Twitter by Brazilian Political Elites
Lorena Barberia, Pedro Schmalz, Norton Trevisan Roman, Belinda Lombard, Tatiane Moraes de Sousa. Proceedings of the 5th International Conference on Natural Language Processing for Digital Humanities. 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 365-376
Open AccessIt’s about What and How you say it: A Corpus with Stance and Sentiment Annotation for COVID-19 Vaccines Posts on X/Twitter by Brazilian Political Elites
Lorena Barberia, Pedro Schmalz, Norton Trevisan Roman, Belinda Lombard, Tatiane Moraes de Sousa. Proceedings of the 5th International Conference on Natural Language Processing for Digital Humanities. 2025.