October 13, 2026
Predictors of death in the elderly with severe acute respiratory syndrome confirmed by Influenza A(H1N1)pdm09
Predictors of Death in Elderly Patients with Severe Acute Respiratory Syndrome Confirmed by Influenza A(h1n1)pdm09
Mônica Tilli Reis Pessoa Conde; Nícolas Volgarine Scaraboto
DOI: 10.22167/2675-6528-202603298
Article derived from a Course Conclusion Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.
Summary
Infection by the influenza A(H1N1)pdm09 virus can progress to Severe Acute Respiratory Syndrome (SARS), with a higher risk of death in the elderly. The objective was to analyze predictive factors of death in cases of SARS confirmed for influenza A(H1N1)pdm09, among the elderly aged 60 years or older, considering individual and contextual factors of the Administrative Regions (RA) of the State of São Paulo (ESP), and to compare the performance of regression models. A descriptive and analytical study was conducted in ESP, from 2023 to 2024, using data from SIVEP-Gripe and Fundação SEADE. 714 cases confirmed by RT-PCR for influenza A(H1N1)pdm09 in residents of ESP aged 60 years or older were included. Sociodemographic, clinical, hospitalization, treatment, vaccination, and residence location variables were analyzed. Binary logistic regression and hierarchical logistic regression were used, with death as the outcome. Of the 714 cases, 202 resulted in death. The final binary logistic regression model associated a higher chance of death with the male sex, advanced age, oxygen saturation below 95%, ICU admission, and use of invasive ventilatory support. The presence of cough and the use of antiviral were associated with a lower chance of death. The hierarchical logistic model showed similar performance, with a small residual contextual contribution from the RAs. It was concluded that the binary logistic regression model was sufficient and simpler to represent the data, being adopted as the final model.
Keywords: Influenza A(H1N1)pdm09; Hierarchical logistic modeling; Death by Influenza A(H1N1)pdm09; Logistic Regression; Severe Acute Respiratory Syndrome.
1. Introduction
The influenza virus, belonging to the Orthomyxoviridae family, is a single-stranded RNA virus classified into four types: A, B, C, and D (Brazil, 2024). Among these, type A stands out for its ability to infect both humans and animals, being subdivided into subtypes based on the hemagglutinin (H) and neuraminidase (N) proteins present on its surface (CDC, 2021). A notable example is the Influenza A (H1N1) virus, which has hemagglutinin H1 and neuraminidase N1. The genetic mutation capacity of influenza viruses, through processes such as drift and shift, is a crucial factor for their epidemiology. While drift involves small genetic alterations that justify the annual reformulation of vaccines, shift refers to major genetic changes, often resulting from recombination between viruses, which can lead to pandemics. A historical example is the emergence of the influenza A(H1N1)pdm09 virus in 2009, which rapidly spread to over 200 countries, resulting in a high number of severe cases and deaths (WHO, 2025b).
The clinical presentation of Influenza A(H1N1)pdm09 infection is highly variable, ranging from mild symptoms, such as fever and cough, to the development of Severe Acute Respiratory Syndrome (SARS). SARS is characterized by hypotension, dyspnea, and oxygen saturation below 95%, frequently requiring hospitalization and, in more severe cases, ventilatory support and intubation. The most common complications include pneumonia and bacterial infections, being particularly prevalent and severe in the elderly (Talbot, 2017). Risk factors for the development of severe cases and death include advanced age (60 years or older), children under two years of age, pregnant women, and individuals with chronic diseases such as heart disease, lung disease, kidney disease, obesity, and immunosuppression. The elderly population presents a high risk due to the higher prevalence of comorbidities and immunosenescence, which is the aging of the immune system (Braga, 2024; Talbot, 2017).
Public health measures for influenza management include specific treatment with antivirals, such as oseltamivir, recommended by the Ministry of Health for risk groups, ideally within the first 48 hours after symptom onset (Brazil, 2024). Prevention is reinforced by annual vaccination campaigns with the trivalent influenza vaccine, which includes the influenza A(H1N1)pdm09 virus, the influenza A H3N2 virus, and one of the influenza B types, targeted at priority groups such as the elderly aged 60 years or older and individuals with chronic diseases (Brazil, 2024). In Brazil, epidemiological surveillance of influenza is carried out through mandatory reporting of severe acute respiratory infection (SARI) cases and suspected deaths in the Influenza Epidemiological Surveillance System (SIVEP-Gripe). In the State of São Paulo (ESP), in 2024, 58,811 hospitalized SARI cases were reported, of which 6,494 (11.0%) were confirmed for influenza virus, with a case fatality rate of 11.0% (744 deaths) for influenza A (ESP, 2025a).
Considering the importance and severity of influenza A(H1N1)pdm09 infection in the elderly aged 60 years or older, especially in cases of hospitalized severe acute respiratory infection (SARI), and the existence of antiviral treatments, preventive measures, and care protocols in the State Health Department (ESP), it becomes relevant to deepen the analysis of predictors of death in this age group. This analysis should go beyond individual risk factors, also incorporating regional contextual factors of the Administrative Regions (AR) of the ESP, through a hierarchical modeling study. Hierarchical models are advantageous because they capture differences at both the individual level and between groups, allowing the inclusion of random effects at each layer of the analysis (Fávero and Belfiore, 2024). Thus, the objective of this study was to analyze the predictors of death in people aged 60 years or older among confirmed cases of severe acute respiratory infection (SARI) by influenza A(H1N1)pdm09, considering individual factors and group contextual factors related to the Administrative Regions [AR] of the ESP, as well as to compare the performance of the binary logistic regression model and the hierarchical logistic model in analyzing this outcome.
2. Material and Methods
The present study was characterized as descriptive and analytical, employing binary logistic regression analyses and hierarchical logistic modeling. The objective was to predict factors associated with the evolution to death in specific cases. The research was conducted in the period from January 1, 2023, to December 31, 2024, covering the State of São Paulo (ESP).
The study population consisted of individuals aged 60 years or older, residing in the State of São Paulo (ESP). The ESP, located in the Southeast region of Brazil, has 645 municipalities, which are grouped into sixteen Administrative Regions (AR) (SEADE, 2025a). The elderly population represents 17.8% of the state’s total population (SEADE, 2023).
All cases of Severe Acute Respiratory Syndrome (SARS) and deaths confirmed by influenza A(H1N1)pdm09, notified in the Influenza Epidemiological Surveillance System (SIVEP-Gripe) during the study period, were included. Participants should be sixty years of age or older, regardless of sex, and reside in the ESP. Cases younger than sixty years or who did not reside in the state were excluded.
SARI was defined as an individual with flu-like illness (fever, cough or sore throat, and symptoms such as myalgia, arthralgia and headache) who presented with dyspnea/respiratory distress, persistent chest pressure, oxygen saturation below 95% in ambient air or cyanosis (Brazil, 2024). A case was confirmed for influenza A(H1N1)pdm09 when the virus was detected by RT-PCR in respiratory secretion or other biological material, this being the gold standard technique (Brazil, 2024).
The data sources for this research were the open databases of SRAG Surveillance, available on the OpenDATASUS portal (OpenDATASUS, 2025), referring to the years 2023 and 2024. Additionally, the database from the SEADE Foundation repository (SEADE, 2025b) was used to obtain information on the Administrative Regions of ESP.
The explanatory variables analyzed for the individuals included sociodemographic data (sex, age, education level, and place of residence), clinical data (comorbidities, influenza vaccination history, intensive care unit [ICU] admission, use of invasive and non-invasive ventilatory support), and care-related data (time to hospitalization and antiviral use). The study’s dependent variable was progression to death.
The data were obtained from secondary, anonymized datasets available for download. After downloading, the 2023 and 2024 SRAG databases were prepared for merging, considering the structure and common variables. The new database underwent transformations for suitability for analysis and selection of variables of interest. A cleaning step included the exclusion of missing data and the application of inclusion and exclusion criteria.
The integration of SRAG dataset data was performed in a relational database management system (SQL), consolidating individual notifications into a single analytical table. This table was enriched with the contextual variable (RA of ESP), linked through the municipality code from the Brazilian Institute of Geography and Statistics (IBGE), structuring the final database for exploratory analyses.
The preprocessing and exploratory data analysis were conducted in the R programming language, with the aid of the tidyverse package. The selection of variables of interest and the treatment of missing data, coded as “9” in the original database, were performed. Categorical variables were converted to the factor type, and numerical variables were adjusted for the calculation of the time interval in days.
Descriptive analyses were performed, presenting continuous variables through measures of central tendency and dispersion (mean, standard deviation, median, minimum and maximum values). Categorical variables were described in absolute and relative frequencies. The association between the variables of interest, the group variable, and the clinical evolution of the patients was evaluated using the chi-squared (X²) test or Fisher’s exact test.
For the analysis of factors associated with death, a binary logistic regression model (Fávero and Belfiore, 2024) was initially proposed, where the dependent variable was qualitative and binary. The null model, without explanatory variables, was evaluated using the glm function from the stats package in R, with a binomial distribution. Subsequently, the explanatory variables were included sequentially, evaluating the statistical significance of the coefficients through Wald’s z-statistic, with a selection criterion of p less than 0.20.
The model was estimated by maximum likelihood. For the choice of the final model, McFadden’s pseudo-R², Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), and maximum likelihood, calculated with the stats package, were evaluated. Sensitivity and specificity, and the Receiver Operating Characteristic (ROC) curve, obtained with the pROC package, were also analyzed. The area under the ROC curve (AUC) was used to evaluate the discriminatory capacity and compare the models (Fávero and Belfiore, 2024).
After defining the binary logistic regression model, hierarchical logistic modeling was performed. This analysis allowed evaluating the relationship of the dependent variable with predictive factors at the individual level, considering that individuals were inserted into groups (RA of residence) with their own characteristics (Fávero and Belfiore, 2024). The hierarchical logistic model was tested based on two levels: level one for the individual and level two for the Administrative Region of residence.
The null logistic hierarchical model was initially estimated, using the gimer function from the lme4 package, to verify the existence of variance between groups and the relevance of the model. Subsequently, the coefficients of the level-one explanatory variables were tested, using the same methods as binary logistic regression. The logistic hierarchical model was then estimated with the insertion of the group variable, also with the lme4 package.
The choice of the best final model was made in a manner analogous to binary logistic regression, considering statistical significance and biological plausibility. For comparison between models, ROC curves were constructed and the Intraclass Correlation Coefficient (ICC) was calculated using the performance package, in order to estimate the percentage of variance attributed to differences between groups.
The analyses were performed in the software MySQL Workbench 8.0.41 Community, RStudio® 2026.1.0.392, and R® version 4.5.2.
3. Results and Discussion
The analysis of Severe Acute Respiratory Syndrome (SARS) data for the years 2023 and 2024, from OpenDATASUS, revealed a significant panorama of influenza A(H1N1)pdm09 infection in the State of São Paulo (ESP). In Brazil, 518,738 suspected cases of SARS were reported, of which 142,122 (27.4%) were residents of ESP. Among the confirmed cases for influenza A(H1N1)pdm09, the country registered 7,413 occurrences and 1,149 deaths. In ESP, confirmed cases totaled 1,710, with 304 deaths, resulting in a lethality of 17.8%. Specifically for the age group of 60 years or older in ESP, which was the focus of this study, 714 confirmed cases were identified, of which 202 progressed to death, representing a lethality of 28.3%. The annual distribution of these cases in ESP showed 297 records and 74 deaths in 2023, and 417 cases and 128 deaths in 2024, indicating a persistence of the disease’s severity in this group.
The sociodemographic characteristics of the 714 cases of SARI confirmed by influenza A(H1N1)pdm09 in the elderly at ESP revealed that 56.7% were female. However, in the univariate logistic analysis, males showed an Odds Ratio (OR) of 1.46 (95% confidence interval [CI95%]: 1.05-2.03; p=0.023) for death, compared to females. This finding, although the female predominance is consistent with previous studies (Braga, 2024; Mata-Marin et al., 2025; Sumner et al., 2023), points to a greater male vulnerability to the fatal outcome. The average age of the patients was 74.5 years (standard deviation=9.3), and the univariate analysis indicated an OR of 1.01 (CI95%: 0.99-1.02; p=0.542) for death, suggesting that advanced age is a relevant factor.
Regarding race, the majority of cases (79.6%) were white individuals, followed by mixed-race (14.9%). Mixed-race individuals, compared to white individuals, showed a higher chance of death in the univariate analysis, with an OR of 1.66 (95% CI: 1.05-2.61; p=0.029). As for schooling, elementary education (1st and 2nd cycles) predominated in 63.0% of patients, with no statistically significant difference between the cure and death groups in the univariate logistic analysis. It is important to note that the schooling variable had data filled in for only 43% of cases, which may have influenced the absence of statistical significance, although the profile of the cases did not differ from the elderly population of ESP, which has an average of 8.3 years of study (ESP, 2023).
The occurrence of nosocomial infection was low, with only 18 (2.8%) of the 641 cases with available data acquiring the infection within the hospital. Of these, 3 (16.7%) progressed to death, while 15 (83.3%) progressed to recovery. Although the number is small, this data reinforces the importance of influenza vaccination in healthcare professionals, especially in hospital settings, as a protective measure for patients, as highlighted by Talbot (2017) in a study that revealed higher lethality in institutions with unvaccinated healthcare professionals. The absence of statistical significance for nosocomial infection in the univariate analysis may be attributed to the low number of events.
Regarding the clinical aspects, cough was the most prevalent symptom, present in 81.2% of all cases, and more frequent among patients who progressed to recovery (87.5%). Univariate analysis showed that the presence of cough was a protective factor, associated with a lower chance of death, with an OR of 0.26 (95% CI: 0.17-0.39; p<0.001). In contrast, respiratory distress was significantly more common among cases that progressed to death (86.3%) than among those who recovered (69.4%), with an OR of 2.78 (95% CI: 1.75-4.41; p<0.001) in univariate analysis. Oxygen saturation below 95% was also a significant predictor of death, with an OR of 2.53 (95% CI: 1.56-4.11; p<0.001) in univariate analysis, being observed in 87.6% of deaths.
Other symptoms such as fever, sore throat, dyspnea, diarrhea, vomiting, abdominal pain, fatigue, loss of smell, and loss of taste were evaluated. Fever was observed in 56.7% of cases, with an OR of 0.83 (95% CI: 0.58-1.18; p=0.291). Dyspnea was reported in 80.8% of patients, with an OR of 1.55 (95% CI: 0.97-2.47; p=0.065). These findings corroborate the literature that points to fever, cough, and respiratory discomfort as frequent symptoms of influenza (Braga, 2024; Braga and Ribeiro, 2024). However, statistical significance varied among symptoms, with cough and respiratory discomfort showing the most robust associations with the outcome of death in the univariate analysis.
The evaluation of risk factors and comorbidities did not reveal statistically significant differences between patients who progressed to cure or death in the univariate logistic regression analysis. Comorbidities such as chronic cardiovascular disease (OR=0.82; p=0.366), diabetes (OR=0.67; p=0.060), asthma (OR=0.48; p=0.070), and others did not show a significant association with death. This divergence from the literature, which frequently associates comorbidities with a higher risk of mortality from influenza A(H1N1)pdm09 (Braga, 2024; Braga and Ribeiro, 2024), may be explained by the high basal prevalence of comorbidities in the elderly population studied, which tends to reduce the differences between the cure and death groups. Furthermore, the presence of a considerable number of cases without information on these variables may have limited the ability to detect associations.
Aspects related to hospital admission were also crucial. It was observed that 12 (6.0%) of the deaths occurred without hospitalization, indicating the severity of the disease even outside the hospital environment. Intensive Care Unit (ICU) admission was significantly more frequent among cases that progressed to death (61.7%) compared to those who recovered (21.9%), resulting in an OR of 5.74 (95% CI: 3.97-8.30; p<0.001) in univariate analysis. The use of invasive ventilatory support also stood out as a strong predictor of death, present in 53.7% of fatal cases and only 7.8% of recovered cases, with an OR of 16.43 (95% CI: 9.30-29.05; p<0.001) in univariate analysis. Non-invasive ventilatory support, in turn, showed an OR of 1.31 (95% CI: 0.79-2.18; p=0.289) in univariate analysis. These results are in line with national and international studies that identify ICU admission and the use of mechanical ventilation as risk factors for mortality from influenza A(H1N1)pdm09 (Braga and Ribeiro, 2024; Minney-Simmth et al., 2019; Sumner et al., 2023).
Regarding preventive and treatment measures, influenza vaccination did not show a significant difference between cases that progressed to recovery or death, with an OR of 0.79 (95% CI: 0.51-1.25; p=0.324). Only 62.5% of cases had this variable filled in, and low vaccine adherence among the elderly, in both recovery and death groups, may have influenced these results, reflecting the low vaccination coverage observed in the population aged 60 years or older in 2023 and 2024. Vaccine effectiveness in the elderly can be affected by immunosenescence and waning immunity over time (Talbot, 2017).
On the other hand, the use of antiviral (oseltamivir) showed a statistically significant difference, being a protective factor. It was used in 33.2% of cases that evolved to recovery, but in only 16.5% of those that evolved to death. Univariate analysis revealed an OR of 0.40 (95% CI: 0.25-0.63; p<0.001), indicating a lower chance of death for those who received the treatment. This finding is in agreement with meta-analyses that point to a reduction in mortality associated with the use of neuraminidase inhibitor antivirals for influenza (Muthuri et al., 2014). The absence of a significant difference in the time between the onset of symptoms and the use of antiviral between the recovery and death groups may be explained by the lack of information in 20.9% of patients.
The characteristics related to the place of residence of SARI cases by influenza A(H1N1)pdm09 in ESP showed a predominance in the Administrative Regions (AR) of São José do Rio Preto (28.6%) and the Metropolitan Region of São Paulo (24.9%). The lethality varied among the ARs, being 33.1% in the MR of São Paulo and 19.6% in the AR of São José do Rio Preto. In the univariate logistic regression analysis, some ARs presented higher chances of death compared to the Central AR (São Carlos). The AR of Bauru had an OR of 5.18 (95% CI: 1.15-23.29; p=0.032), associated with agro-industrial characteristics and schooling indicators below the ESP average (Desenvolve SP, 2026).
The Campinas AR also presented an OR of 4.22 (95% CI: 1.15-15.45; p=0.030), reflecting a diversified economic strength, but with pressures on the health infrastructure due to population growth (Desenvolve SP, 2026). The Santos AR stood out with a very high OR of 38.00 (95% CI: 3.30-436.94; p=0.004), possibly due to the small number of cases and the high proportion of deaths, in addition to being a port complex with lower social indicators and needs for improvement in the hospital network (Desenvolve SP, 2026). The Itapeva AR presented a very high OR, probably overestimated by the presence of a single case that evolved to death, while the Registro AR did not report cases.
For the modeling of predictor variables of death, binary logistic regression models were fitted. The final chosen model, Model 2, showed good fit and biological plausibility, with an area under the ROC curve (AUC) of 0.837 and a GINI index of 0.67. In this model, the variables associated with a higher chance of death were male sex (adjusted OR=1.52; 95% CI: 0.90-2.61; p=0.121), advanced age (adjusted OR=1.03; 95% CI: 1.00-1.06; p=0.037), oxygen saturation below 95% (adjusted OR=2.11; 95% CI: 1.04-4.48; p=0.042), ICU admission (adjusted OR=2.05; 95% CI: 1.10-3.80; p=0.023), and use of invasive ventilatory support (adjusted OR=11.35; 95% CI: 4.91-27.65; p<0.001). Although male sex and non-invasive ventilatory support (adjusted OR=1.65; 95% CI: 0.81-3.51; p=0.177) did not reach statistical significance in the adjusted model, the original TCC maintained them as factors with a higher chance of death.
In contrast, the presence of cough (OR adjusted=0.35; 95% CI: 0.19-0.64; p<0.001) and the use of antiviral (OR adjusted=0.50; 95% CI: 0.26-0.91; p=0.023) were identified as protective factors, associated with a lower chance of death. These results from the final binary logistic regression model are aligned with the literature, which recognizes the importance of early antiviral therapy and the severity of severe respiratory symptoms (Braga, 2024; Braga and Ribeiro, 2024; Minney-Simmth et al., 2019; Sumner et al., 2023). The inclusion of the contextual variable of Administrative Regions in the binary logistic regression was not associated with the outcome after multivariate adjustment, and therefore, was not included in the final model.
Hierarchical logistic modeling was performed to assess the contribution of contextual factors of the Administrative Regions. The null hierarchical model indicated that 5.8% of the outcome variability could be attributed to the ARs (adjusted ICC=0.058), with a variance of random effects of 0.2035. However, in the full hierarchical logistic model, which included individual variables and ARs as level 2, the variance of random effects decreased to 0.03413, and the AUC (0.836) and GINI index (0.672) were very similar to those of the binary logistic regression model. This suggests that the residual contextual contribution of the ARs was small after adjusting for individual and care variables.
In the comparison between the models, it was observed that, although the adjusted Odds Ratios of the hierarchical logistic model were similar to those of the binary logistic regression model for most variables, some, such as male sex (OR=1.63; 95% CI: 0.99-2.68; p=0.053), oxygen saturation below 95% (OR=1.89; 95% CI: 0.91-3.92; p=0.085), and use of non-invasive ventilatory support (OR=1.61; 95% CI: 0.80-3.26; p=0.182), lost statistical significance. This loss of significance, with confidence intervals close to one, reinforces the coherence between the two models and the conclusion that the binary logistic regression model was sufficient and simpler to represent the data, being adopted as the final model.
The study presented some limitations that deserve consideration. The two-year time frame (2023-2024) may have introduced a bias, limiting the capture of broader territorial patterns and the necessary heterogeneity of the Administrative Regions to influence the outcome more pronouncedly. Furthermore, the quality of the analyzed data and the large number of variables with missing information may have restricted a more complete and in-depth analysis. The presence of missing data, especially in variables such as schooling and comorbidities, may have impacted the ability to identify statistically significant associations for some factors.
In summary, the research identified that individual factors such as male sex, advanced age, oxygen saturation below 95%, ICU admission, and use of invasive ventilatory support were predictors of a higher chance of death in the elderly with SARI due to influenza A(H1N1)pdm09 at ESP. On the other hand, the presence of cough and the use of antiviral were associated with a lower chance of death, acting as protective factors. Although hierarchical modeling allowed for the evaluation of the contextual influence of the Administrative Regions, its residual contribution to the explanation of death was small after adjustment for individual and care variables, which led to the adoption of the binary logistic regression model as the most adequate and parsimonious to represent the findings.
4. Conclusion
This study analyzed the predictor factors of death in elderly individuals aged 60 years or older, affected by Severe Acute Respiratory Syndrome (SARS) confirmed for influenza A(H1N1)pdm09 in the State of São Paulo, considering both individual and contextual aspects of the Administrative Regions. It was found that, of the 714 cases included, 202 progressed to death, resulting in a lethality of 28.3% in this age group. The findings indicated that male sex, advanced age, oxygen saturation below 95%, admission to the Intensive Care Unit, and the use of invasive ventilatory support were associated with a higher chance of death. In contrast, the presence of cough and the use of antiviral were identified as protective factors, associated with a lower chance of death. The study’s main contribution lies in the identification of these mortality predictors in a vulnerable population, offering a simpler and more parsimonious binary logistic regression model that proved sufficient to represent the data, with predictive performance equivalent to the hierarchical logistic model.
Despite the relevance of the results, the study presented some limitations, such as the two-year time frame (2023-2024), which may have restricted the capture of broader territorial patterns and the necessary heterogeneity of the Administrative Regions to influence the outcome more pronouncedly. Additionally, the quality of the data and the presence of missing information in several variables may have limited a more in-depth analysis. For future studies, it is recommended to extend the observation period, include other grouping variables, and evaluate different levels of territorial aggregation, in order to deepen the knowledge about the predictor variables of death by influenza A(H1N1)pdm09 in the elderly in the State of São Paulo.
Bibliographic References
Braga, J.U. 2024. Clinical predictors of severe forms of influenza A (H1N1)pdm09 in adults and children during the 2009 epidemic in Brazil. PLoS ONE 19(2): e0291843.
Braga, J.U.; Ribeiro, A.F. 2024. Biological, social, and healthcare factors for death due to influenza A(H1N1) during the 2009 epidemic in Brazil. Revista de Saúde Pública 58(32): 1-11.
Brasil. Ministério da Saúde [MS]. 2024. Guia de vigilância em saúde. v.1. 6ed. rev. Ministério da saúde, Brasília, DF, Brasil. Disponível em:< https://www.gov.br/saude/pt-br/assuntos/saude-de-a-a-z/t/tetano-acidental/publicacoes/guia-de-vigilancia-em-saude-6a-edicao.pdf/view>. Acesso em: 1 mar. 2025.
Centers for Disease Control and Prevention [CDC]. 2021. Influenza. In: Hall E.; Wodi A.P.; Hamborsky J.; et al. In: Epidemiology and Prevention of Vaccine-Preventable Diseases. 2021. Public Health Foundation.14th ed. Washington, D.C., EUA. Disponível em: <https://www.cdc.gov/pinkbook/hcp/table-of-contents/chapter-12-influenza.html>. Acesso em: 20 fev. 2025.
Desenvolve SP. 2026. Mapa da Economia Paulista. Disponível em: <https://www.desenvolvesp.com.br/mapadaeconomiapaulista/>. Acesso 26 mar. 2026.
Estado de São Paulo [ESP]. 2023. Fundação para o Desenvolvimento da Educação – FDE. Educação e Escolaridade. PNAD Contínua 2023. Disponível em: <https://arquivo.fde.sp.gov.br/fde.portal/PermanentFile/File/Escolaridade%20e%20Escolariz a%C3%A7%C3%A3o%20-%20Pnad%20Cont%C3%ADnua%202023.pdf >. Acesso em: 05 abr. 2026.
Estado de São Paulo [ESP]. 2025a. Centro de Vigilância Epidemiológica. Sazonalidade Influenza- Semana epidemiológica (SE) 1 a 52/2024. Disponível em: <https://www.saude.sp.gov.br/resources/cve-centro-de-vigilancia-epidemiologica/areas-de-vigilancia/doencas-de-transmissao-respiratoria/influenza/2024/infoflu24_se0152.pdf>. Acesso em: 21 fev. 2025.
Fundação Sistema Estadual de Análise de Dados [SEADE]. 2023. Seade Municípios. Disponível em:< https://municipios.seade.gov.br/>. Acesso em: 1 de mar. 2025.
Fundação Sistema Estadual de Análise de Dados [SEADE]. 2025a. Códigos e nomes dos municípios e suas respectivas regiões do Estado de São Paulo. Disponível em: < https://repositorio.seade.gov.br/dataset/nomes-codigos-municipios-esp/resource/1871ac05-7b7f-4c13-9b4a-23c7d41fd988>. Acesso em: 1 mar. 2025.
Fundação Sistema Estadual de Análise de Dados [SEADE]. 2025b. Tema: demografia. Conjunto de dados. Disponível em:< https://repositorio.seade.gov.br/dataset/>. Acesso em: 22 de fev. 2025.
Fávero, L.P.; Belfiore, P. 2024. Manual de análise de dados: estatística e machine learning com Excel ®, SPSS ®, Stata ®, R® e Phyton®. 2ed. LTC, Rio de Janeiro, RJ, Brasil.
Mata-Marín L.A., Mata-Marín J.A., Vásquez-Mota V.C., Arroyo-Anduiza C.I., Gaytán-Martínez J.E., Manjarrez-Téllez B. et al. 2015.Risk factors associated with mortality in patients infected with influenza A/H1N1 in Mexico. BMC Res Notes 11; 8:432.
Minney-Smith C.A., Selvey L.A., Levy A., Smith D.W. 2019.Post-pandemic influenza A/H1N1pdm09 is associated with more severe outcomes than A/H3N2 and other respiratory viruses in adult hospitalisations. Epidemiol Infect 147:e310.
Muthuri S.G.; Venkatesan, S.; Myles, P.R.; Leonardi-Bee, J.; Al Khuwaitir, T.S.; Al Mamun, A. et al. 2014. Effectiveness of neuraminidase inhibitors in reducing mortality in patients admitted to hospital with influenza A H1N1pdm09 virus infection: a meta-analysis of individual participant data. Lancet Respir Med 2(5): 395-404.
OpenDATASUS. 2025.Conjunto de Dados. Síndrome Respiratória Aguda Grave (SRAG). Disponível em:< https://opendatasus.saude.gov.br/dataset/?groups=dados-sobre-srag>. Acesso em: 19 maio 2025.
Sumner, K.M.; Masalovich, S.; O’Halloran, A; Holstein, R.; Reingold. A.; Kirley, P.D. et al. 2023.Severity of influenza-associated hospitalizations by influenza virus type and subtype in the USA, 2010-19: a repeated cross-sectional study. Lancet Microbe (11): e903-e912.
Talbot, H.K. 2017. Influenza in Older Adults. Infect Dis Clin North Am 31(4): 757-766.
World Health Organization [WHO]. 2025b. The burden of Influenza. Disponível em: < https://www.who.int/news-room/feature-stories/detail/the-burden-of-influenza>. Acesso em: 1 Mar. 2025.
Article originating from the Final Course Project of the Specialization in Data Science and Analytics of the MBA USP/Esalq
To learn more about the course, click here and access the MBX Academy platform