Graphical abstract
Keywords: Sepsis, Knowledge, Awareness, Lay people
Highlights
-
•
In this population-based survey among 6,312 Brazilian participants, we found an increase in public knowledge of sepsis over time.
-
•
In 2026, a high proportion of the participants demonstrated understanding of the importance of early diagnosis and treatment.
-
•
Educational attainment emerged as the strongest determinant of sepsis knowledge across all domains evaluated.
-
•
Despite the overall improvement in awareness, socioeconomic inequalities and gaps persisted.
-
•
Future efforts should focus on improving recognition of sepsis and reducing inequalities in access to health information.
Abstract
Background
Public awareness of sepsis is essential for early recognition of warning signs and timely healthcare seeking behavior, but data from resource-limited settings remain scarce.
Methods
We evaluated sepsis knowledge in Brazil using cross-sectional surveys conducted in 2014, 2017, and 2026. Standardized questionnaires were administered through face-to-face interviews using random sampling stratified by age, sex, and geographic region. Adequate basic knowledge was defined as having heard the term sepsis and correctly identifying it as a severe host response to infection. In 2026, we additionally assessed knowledge of sepsis time-dependent nature and long-term disability after sepsis. Multivariable logistic regression identified factors associated with knowledge.
Results
A total of 6,312 participants were interviewed (2014: n = 2,126; 2017: n = 2,100; 2026: n = 2,086). Adequate basic knowledge increased from 2.6% in 2014 to 5.8% in 2017 (OR 2.25, 95% CI 1.63–3.16; p < 0.001) and 10.3% in 2026 (OR 3.45, 95% CI 2.54–4.75; p < 0.001). In 2026, 70.0% of participants acknowledged that earlier diagnosis and treatment improve survival. Awareness of long-term disabilities was lower (24.4%). Higher educational attainment was the strongest determinant of knowledge across all domains. Younger age, male sex, residence outside state capitals, and lower household income were independently associated with lower knowledge.
Conclusion
Public awareness of sepsis in Brazil has improved substantially over the past decade, suggesting a positive impact of national and international awareness initiatives. Nevertheless, important knowledge gaps persist, and awareness continues to be strongly influenced by socioeconomic factors, highlighting the need for continued efforts to improve sepsis awareness and promote equitable access to health information.
Background
Sepsis is a major global health problem and remains one of the leading causes of death among hospitalized patients [1]. According to the Global Burden of Disease Study, more than 40 million cases and 11 million sepsis-related deaths occur annually worldwide [2] with more recent estimates suggesting approximately 21.4 million sepsis-related deaths in 2021 when COVID-19-associated deaths are included. Brazilian epidemiological studies also demonstrate high prevalence and mortality, with an estimated 240,000 deaths annually among adults admitted to intensive care units [3,4] and approximately 8,000 among children [5].
Despite advances in care, sepsis mortality remains high and is influenced not only by pathogen, host, and healthcare system factors, but also by delays in recognition and inadequate treatment [6,7]. Public awareness is particularly important in community-acquired sepsis, where early recognition of warning signs and prompt healthcare-seeking behavior may contribute to timely diagnosis and treatment. However, public awareness of sepsis remains limited worldwide and varies substantially across countries and healthcare settings [8,9]. In a recent scoping review, the proportion of individuals who had heard the term “sepsis” ranged from 2% in Japan to 88.6% in Germany, while the proportion able to correctly define sepsis ranged from 4.2% in Singapore to 92% in Sweden [9]. Although awareness appears to have improved over time, rates above 50% were reported in only a few countries. Younger age, female sex, higher educational attainment, healthcare-related employment, and previous personal or family experience with sepsis are associated with greater awareness [10]. Importantly, data from low- and middle-income countries remain scarce.
In Brazil, the Latin America Sepsis Institute (LASI) conducted nationwide population-based surveys evaluating public awareness of sepsis in 2014 and 2017, demonstrating low levels of awareness [11,12]. Since then, multiple educational initiatives directed at both healthcare professionals and the general public may have modified this scenario [4]. Therefore, the aim of this study was to evaluate the temporal trends on public awareness and knowledge of sepsis among the Brazilian population, comparing it with acute myocardial infarction awareness. We also sought to evaluate the current public perception regarding the time-dependent nature of sepsis treatment and long-term disabilities after sepsis. We hypothesized that, despite increasing health initiatives, knowledge regarding sepsis would remain inadequate and below that reported in high-income countries.
Methods
Study design
This was a nationally representative, cross-sectional, face-to-face survey conceived and coordinated by the Latin American Sepsis Institute. Fieldwork and interviews were conducted by the Datafolha Institute, an independent national research agency specialized in population-based surveys. Interviewers are required to be at least 18 years old, have completed secondary or higher education, and have prior research experience. All interviewers receive standardized training in research ethics, good research practices, and the Institute’s data collection procedures. For each survey wave, detailed written instructions regarding questionnaire administration were provided to field coordinators. These standardized guidelines ensured uniformity in the application of procedures across survey waves. Interviewers were instructed to administer all questions exactly as worded and in the predefined order specified in the survey instrument. The survey was performed through direct interviews in high-flow public areas. The study was approved by the Ethics Committee of the Federal University of São Paulo (CAAE 54357815.7.0000.5505), and all participants provided informed consent. Reporting followed the Checklist for Reporting of Survey Studies (CROSS) guidelines.
Sampling
Participants were randomly approached in high-traffic public locations, including squares, street intersections, avenues, and commercial streets, geographically distributed across the study areas, using the “flow point” methodology developed by Datafolha in collaboration with the University of São Paulo in 1983 [13]. Adults aged ≥16 years were eligible, and participation was voluntary and not incentivized.
This methodology improves efficiency and facilitates access to population groups often underrepresented in household surveys, including residents of high-income condominiums, informal settlements (favelas), and areas with high levels of violence. Sampling points were randomly selected and proportionally distributed according to the resident population, ensuring representation across all Brazilian geographic regions. Recruitment was controlled through quotas based on sex, age, geographic region, and socioeconomic distribution to ensure representativeness of the Brazilian population.
To ensure methodological reliability, Datafolha maintains an updated database containing demographic and socioeconomic characteristics of all survey locations. In-person or telephone quality checks were conducted in at least 20% of interviews performed by each interviewer.
Data collection
A preliminary questionnaire was developed in 2014 based on theoretical constructs and previous international studies evaluating sepsis awareness [10]. Content validity and comprehensibility were assessed by sepsis experts, resulting in minor modifications before the final version.
The surveys applied in 2014, 2017 and 2026, included both spontaneous and prompted questions (Electronic Supplement 1). Although the core questions remained consistent across the three surveys, minor differences existed over time. Participants were initially asked whether they had ever heard the term “sepsis.” In 2014 and 2017, only participants who answered “yes” to the question “Have you ever heard the term sepsis?” were asked additional questions assessing whether they correctly recognized sepsis as a severe host response to infection. In 2026, this question was applied to all participants regardless of their previous familiarity with the term. However, to ensure comparability across survey waves, the temporal trend analyses considered only participants on the 2026 survey who answered “yes” to the initial question. Accordingly, adequate basic knowledge of sepsis was defined identically in all survey waves as (1) heard the term sepsis and (2) correctly identified it as a severe host response to infection.
In 2017, recognition of the term “septicemia” was additionally evaluated to assess possible underestimation of sepsis awareness. Similar questions regarding myocardial infarction and recognition of its major symptoms were also included. To further assess participants’ understanding of the healthcare impact of sepsis, additional prompted questions explored whether sepsis was perceived as a rare disease, a hospital-only disease, a medical emergency, or a leading cause of death. In 2026, an additional option describing sepsis as “the body attacking its own tissues and organs in response to infection” was included and considered correct.
The 2026 survey additionally evaluated two additional domains of sepsis knowledge: (1) understanding of time-dependent nature of sepsis treatment (hereafter referred to as time-dependency knowledge) and (2) awareness of long-term disability after sepsis.
Adequate knowledge on time-dependency was defined as correctly recognizing that earlier diagnosis and treatment increase survival chances. Adequate knowledge on long-term disability was defined as recognizing that sepsis survivors may experience fatigue, cognitive impairment, depression, or functional limitations after recovery. Incorrect and “I do not know” responses were categorized as inadequate knowledge.
Demographic and socioeconomic variables included age, sex, educational attainment, household income, and geographic region. Educational attainment was categorized according to the Brazilian Institute of Geography and Statistics (IBGE) classification [15]. In addition, the surveys included a socioeconomic status (SES) score, a measure commonly used in Brazilian population surveys that is based on household characteristics and assets and reflects a construct closely related to household income [14].
All survey versions (2014, 2017, and 2026) are available in the electronic supplementary material.
Sample size calculation
The estimated minimum sample size was 2,100 participants, based on the Brazilian population aged ≥16 years, assuming a margin of error of 2 percentage points and a 95% confidence level.
Statistical analysis
The sampling strategy was designed to reflect the Brazilian population distribution according to age, sex, geographic region, type of city (capital versus countryside), household income, and socioeconomic class, based on data from the Brazilian Institute of Geography and Statistics (IBGE) and the National Household Sample Survey (PNAD) corresponding to each survey year.
Categorical variables were presented as frequencies and percentages. Continuous variables were assessed for normality using graphical methods and the Kolmogorov–Smirnov test and are presented as mean ± standard deviation or median (interquartile range), as appropriate. Comparisons of response rates across survey years were performed using the chi-square test, followed by partitioned chi-square analyses when appropriate.
Prior to multivariable modelling, unadjusted logistic regression models were fitted for each outcome (basic knowledge, time-dependency and long-term disability) with each sociodemographic covariate entered individually. Predicted probabilities of each outcome across categories of each covariate were estimated from these models.
To identify factors independently associated with basic sepsis knowledge in the overall study population, a multivariable logistic regression model was fitted using the pooled dataset from the 2014, 2017, and 2026 surveys, with survey year included as a categorical covariate to account for temporal differences across waves. Covariates were selected a priori based on epidemiological plausibility and previous literature and included survey year, age group, sex, educational attainment, household income, geographic region, and type of city. The presence of interactions between survey year and the remaining covariates was assessed using a likelihood ratio test comparing a model including all two-way interactions with survey year to a main-effects model. Post hoc, we fitted another logistic regression model including survey year, educational attainment, and their interaction to assess whether the magnitude of the increase in knowledge along the years differed according to educational attainment. Results were presented as odds ratios (ORs) with 95% confidence intervals (CIs).
For the 2026 survey, separate multivariable logistic regression models were fitted to evaluate factors associated with knowledge of the time-dependent nature of sepsis treatment and awareness of long-term disability after sepsis. These outcomes were analyzed separately because correct responses to these questions were not necessarily contingent upon fulfilling the definition of basic sepsis knowledge. Results are presented as odds ratios with corresponding 95% confidence intervals The relative importance of predictors was evaluated using permutation-based variable importance analysis, complementing the odds ratios derived from the logistic regression models. For each outcome, the area under the receiver operating characteristic curve (AUROC) of the full multivariable model was first calculated as the reference performance. Subsequently, each predictor was randomly permuted 1,000 times, while all remaining variables were kept unchanged, thereby disrupting its association with the outcome while preserving its distribution and the structure of the remaining predictors. The AUROC was recalculated after each permutation, and variable importance was quantified as the mean reduction in AUROC relative to the full model. Larger reductions indicate greater contribution of the predictor to model discrimination, and 95% confidence intervals were derived from the distribution of AUROC reductions across permutations. Because the objective of the multivariable logistic regression models was to identify factors independently associated with sepsis knowledge rather than to develop or validate prediction models, calibration and goodness-of-fit statistics were not formally assessed.
Missing data were minimal and were limited to participants who answered "I do not know" or "Prefer not to answer" to specific survey questions. These responses were treated as missing for the corresponding analyses, and complete-case analyses were performed. Because missingness resulted exclusively from these non-substantive survey responses rather than incomplete data collection, multiple imputation was not considered appropriate.
All analyses were performed using R software version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as a two-sided p-value < 0.05.
Results
The 2014 survey included 2,126 participants from 134 Brazilian cities and was conducted between March 7 and 9, 2014. The 2017 survey included 2,100 participants from 130 cities and was conducted between June 6 and 10, 2017. The 2026 survey included 2,086 participants from 130 cities and was conducted between January 12 and 13, 2026. Survey duration ranged from 6−10 min across the study years. Demographic and socioeconomic characteristics of the respondents are presented in Table 1. As expected, higher educational attainment was generally concentrated among higher household income categories (Figure S1).
Table 1.
General characteristics of the population across the 2014, 2017, and 2026 surveys.
| Variáveis | 2014 n = 2,126 | 2017 n = 2,100 | 2026 n = 2,086 |
|---|---|---|---|
| Age, years, mean ± SD | 39.6 ± 16.89 | 39.2 ± 16.4 | 42.6 ±16.7 |
| Age, years, n (%) | |||
| 16 – 24 | 446 (21.9) | 453 (21.6) | 348 (16.6) |
| 25 – 34 | 496 (23.3) | 534 (25.4) | 410 (19.6) |
| 35 – 44 | 407 (19.1) | 366 (17.4) | 407 (19.5) |
| 45 – 59 | 446 (21.0) | 441 (21.0) | 499 (23.9) |
| 60 ou mais | 311 (14.6) | 306 (14.6) | 422 (20.2) |
| Sex at birth, female, n (%) | 1101 (51.8) | 1066(50.7) | 1074 (51.5) |
| Education level, n (%) | |||
| Illiterate/Incomplete elementary I | 251 (11.8) | 215 (10.2) | 139 (6.7) |
| Incomplete/complete elementary II | 577 (27.1) | 509 (24.2) | 398 (19.1) |
| Complete/incomplete high school | 929 (43.7) | 972 (46.3) | 962 (46.1) |
| Complete/incomplete university level | 323 (15.2) | 345 (16.4) | 488 (23.4) |
| Postgraduate level | 46 (2.2) | 59 (2.8) | 99 (4.7) |
| Socioeconomic classification, n (%) | |||
| A | 38 (1.8) | -* | 75 (3.6) |
| B | 642 (30.2) | – | 535 (25.6) |
| C | 1094 (51.4) | – | 1007 (48.3) |
| D/E | 352 (16.6) | – | 469 (22.5) |
| Household income (minimal wage) | |||
| <2 | 929 (43.7) | 938 (44.7) | 886 (42.5) |
| 2 – 3 | 435 (20.5) | 444 (21.1) | 367 (17.6) |
| 3 – 5 | 338 (15.9) | 350 (16.7) | 363 (17.4) |
| 5 – 10 | 232 (10.9) | 165 (7.9) | 258 (12.4) |
| > 10 | 90 (4.2) | 62 (3.0) | 96 (4.6) |
| Don’t know/didn’t answer | 102 (4.8) | 141 (6.7) | 102 (4.9) |
| Employment status, n (%) | |||
| Active | |||
| Inactive | |||
| Socioeconomical score | 1.0 (0.97–1.0) | 0.96 (0.79–1.19) | 1.04 (0.68–1.23) |
| Region, n (%) | |||
| Southest | 914 (42.3) | 891 (42.4) | 896 (42.9) |
| Northeast | 573 (26.9) | 546 (26.0) | 546 (26.2) |
| South | 307 (14.4) | 308 (14.6) | 308 (14.8) |
| Middle west | 165 (7.7) | 168 (8.0) | 168 (8.0) |
| North | 167 (7.9) | 187 (8.9) | 168 (8.0) |
| City classification, n (%) | |||
| Capitals | 534 (25.1) | 539 (25.7) | 744 (35.7) |
| Countryside | 1243 (58.5) | 1209 (57.6) | 1022 (49.0) |
| Other municipalities in metropolitan region | 349 (16.4) | 352 (16.7) | 320 (15.3) |
*Data on socialeconomical classification was not available for the 2017 poll.
Public awareness of sepsis remained low throughout the study period. In 2014, only 140 of 2,126 participants (6.6%) reported having previously heard the term sepsis, and only 56 (2.6%) demonstrated adequate basic knowledge by correctly identifying sepsis as a severe host response to infection (Table 2). Considering only participants who answered “yes” to the initial question in all survey waves, adequate basic knowledge increased over time, reaching 121 participants (5.8%) in 2017 and 214 participants (10.3%) in 2026 (p < 0.001) (Fig. 1). Of note, in 2026, when all respondents regardless of their answer on the knowledge of the term sepsis were subsequently presented with the recognition question, 764 (36.6%) participants also correctly recognized sepsis as a severe host response to infection. In 2017, a higher percentage of participants recognized the term septicemia (941; 44.8%).
Table 2.
Public Knowledge of Sepsis and Myocardial Infarction Across the 2014, 2017, and 2026 Surveys.
| Variables | 2014 n = 2,126 | 2017 n = 2100 | 2026 n = 2086 | P value |
|---|---|---|---|---|
| Have your ever heard of sepsis? (yes)1 | 140 (6.6) | 304 (14.5) | 434 (20.8) | < 0.001 |
| What does it mean?1,2 | < 0.001 | |||
| A severe body response to infection | 56 (2.6) | 121(5.8) | 764 (36.6) | |
| Blood infection | 37(1.7) | 85 (4.0) | 700 (33.6) | |
| A rock band | 6 (0.3) | 8 (0.4) | 105 (5.0) | |
| A step of “lava-jato” operation | – | 16 (0.8) | – | |
| None of the above/ don’t know | 41 (1.9) | 74 (3.6) | 517 (24.8) | |
| Don’t know | ||||
| Adequate basic knowledge | 56 (2.6) | 121 (5.8) | 214 (10.3) | < 0.001 |
| Have you ever heard of septicemia (yes)1,3 | – | 941 (44.8) | – | |
| What does it mean1 | ||||
| A severe body response to infection | – | 475 (22.6) | – | |
| Blood infection | – | 375 (17.9) | – | |
| A rock band | – | 2 (0.1) | – | |
| A step of “lava-jato” operation | 12 (0.6) | |||
| None of the above/ I don’t know | – | 77 (3.7) | – | |
| In your opinion, sepsis is4: | < 0.001 | |||
| Something that happens when your body attacks its own tissues and organs in response to infection | – | 503 (23.9) | 708 (33.9) | |
| A rare disease | 226 (10.6) | 305 (14.5) | 308 (14.8) | |
| A medical emergency | 392 (18.4) | 322 (15.3) | 369 (17.7) | |
| Something you only catch at hospitals | 187 (8.8) | 268 (12.8) | 198 (9.5) | |
| Something that kills more than myocardial infarction and cancer | 286 (13.4) | 116 (5.5) | 122 (5.8) | |
| None of the above/don’t know | 1035 (48.6) | 586 (27.9) | 381 (18.3) | |
| Have you ever heard of myocardial infarction? (yes)1 | 2086 (98.1) | 2081 (99.1) | 2042 (97.9) | 0.005 |
| What are the symptoms of myocardial infarction?3 | ||||
| Chest pain radiating to the arm, cold sweats, and nausea | 1878 (90.0) | 1837 (87.5) | 1873 (89.8) | |
| Fever, cough, and wheezing | 25 (1.2) | 27 (1.3) | 33 (1.6) | |
| Abdominal pain and vomitting | 15 (0.7) | 16 (0.7) | 12 (0.6) | |
| Paralysis on one side of the body | 110 (5.3) | 114 (5.4) | 112 (5.3) | |
| None of the above/ don’t know | 58 (2.8) | 106 (5.1) | 56 (2.7) | |
| 2026 survey additional questions on sepsis knowledge | ||||
| Time-dependency nature1 | ||||
| The faster the diagnosis and start of sepsis treatment, the greater the chances of survival. | – | – | 1460 (70.0) | |
| Time has some influence in the outcomes, but the most important factor is treatment choices. | – | – | 224 (10.7) | |
| I've never heard that time influences sepsis treatment. | – | – | 193 (9.3) | |
| Time doesn't make that much difference; what's important is to be in a good hospital. | – | – | 129 (6.2) | |
| None of the above/ don’t know | – | – | 80 (3.8) | |
| Long-term disabilities1 | ||||
| They may have some long-term effects, but without affecting the person's quality of life. | – | – | 789 (37.8) | |
| Many people may experience fatigue, memory loss, depression, or other limitations after sepsis. | – | – | 509 (24.4) | |
| I have never heard that sepsis can have lasting effects. | – | – | 421 (20.2) | |
| Sepsis can be fatal while the person is in the hospital, but afterwards the person recovers completely. | – | – | 181 (8.7) | |
| None of the above/ don’t know | – | – | 186 (8.9) | |
1Single spontaneous answer. 2 In 2026, this question was available also to those who answered “no” in the first question. 3This question was available only for 2017 survey. 4Single prompted answer.
Fig. 1.
Predicted probability of adequate basic sepsis knowledge according to survey year. Probabilities were estimated from an unadjusted logistic regression model including survey year as the only independent variable. Points represent predicted probabilities and horizontal bars represent 95% confidence intervals. The figure illustrates the progressive increase in basic sepsis knowledge between 2014 and 2026.
Awareness of the clinical impact of sepsis remained limited. Recognition of sepsis as a medical emergency was low and changed little over time (2014: 392 [18.4%]; 2017: 322 [15.3%]; 2026: 369 [17.7%]). Similarly, only a small proportion of respondents identified sepsis as a major cause of death (2014: 286 [13.4%]; 2017: 116 [5.5%]; 2026: 122 [5.8%]). In contrast, awareness of myocardial infarction remained consistently high across all surveys (2014: 2,086 [98.1%]; 2017: 2,081 [99.1%]; 2026: 2,042 [97.9%]), and most participants correctly identified its major symptoms (2014: 1,878 [90.0%]; 2017: 1,837 [87.5%]; 2026: 1,873 [89.8%]). Detailed responses are presented in Table 2.
In the 2026 survey, knowledge regarding specific aspects of sepsis remained heterogeneous (Table 2). Most participants recognized that earlier diagnosis and treatment improved sepsis survival [1,460, 70.0%]. Awareness of long-term consequences was considerably lower. Only 509/2086 participants (24.4%) recognized that sepsis survivors may experience long-term physical, cognitive, or psychological impairments, whereas 970 (46.2%) underestimated these consequences and 421 (20.2%) reported not knowing whether long-term disabilities could occur.
Participant characteristics according to basic knowledge status are shown in Table S1. Predicted probabilities of adequate basic knowledge across categories of each demographic and socioeconomical covariate are displayed in Figures S2–S7. In the multivariate model, survey year was independently associated with higher knowledge, with respondents in 2017 (OR 2.25, 95% CI 1.63–3.16) and 2026 (OR 3.45, 95% CI 2.54–4.75) demonstrating substantially greater odds of adequate knowledge compared with those surveyed in 2014 (Table 3). Higher educational attainment was the strongest sociodemographic determinant of basic knowledge, showing a marked dose-response relationship across educational categories. Younger age, male sex, residence outside state capitals, and lower household income were independently associated with lower levels of knowledge (Table 3). No significant interactions were observed between survey year and the sociodemographic variables included in the model (p = 0.628). In our post-hoc model, while visual inspection suggests a larger increase in knowledge among participants with higher educational attainment, formal testing did not demonstrate a significant interaction between survey year and educational level (Fig. 2, p = 0.42). Variable importance assessed through permutation analysis confirmed educational attainment and survey year as the most influential predictors of basic sepsis knowledge (Fig. 3).
Table 3.
Factors associated with adequate basic knowledge in the pooled 2014–2026 survey population.
| Variables | Odds ratio (95%CI) | p-value |
|---|---|---|
| Year | ||
| 2014 | Reference | – |
| 2017 | 2.25 (1.63, 3.16) | < 0.001 |
| 2026 | 3.45 (2.54, 4.75) | < 0.001 |
| Sex | ||
| Female | Reference | – |
| Male | 0.68 (0.54, 0.85) | 0.001 |
| Age group | ||
| 16−19 | Reference | – |
| 20−29 | 2.10 (1.27, 3.65) | 0.006 |
| 30−39 | 2.48 (1.51, 4.31) | 0.001 |
| 40−49 | 2.09 (1.23, 3.71) | 0.008 |
| 50−59 | 2.20 (1.27, 3.95) | 0.006 |
| 60−69 | 2.25 (1.26, 4.16) | 0.008 |
| 70+ | 3.88 (1.66, 8.74) | 0.001 |
| Educational level | ||
| Illiterate/Incomplete elementary I | Reference | – |
| Incomplete/complete elementary II | 1.53 (0.70, 3.84) | 0.323 |
| Complete/incomplete high school | 3.84 (1.86, 9.31) | 0.001 |
| Complete/incomplete university level | 7.48 (3.56, 18.37) | < 0.001 |
| Postgraduate level | 9.14 (4.00, 23.73) | < 0.001 |
| Household income (minimal wage) | ||
| Don’t know/none/didn’t answer | 0.62 (0.31, 1.13) | 0.145 |
| Up to 2 | Reference | – |
| 2 to 3 | 1.31 (0.96, 1.79) | 0.090 |
| 3 to 5 | 1.31 (0.94, 1.81) | 0.110 |
| 5 to 10 | 1.54 (1.07, 2.20) | 0.019 |
| More than 10 | 1.88 (1.17, 2.97) | 0.008 |
| City type | ||
| Capital | Reference | – |
| Countryside | 0.60 (0.47, 0.76) | < 0.001 |
| Other municipalities in the metropolitan region | 0.72 (0.52, 0.99) | 0.045 |
| Region | ||
| Middle-West | Reference | – |
| Northeast | 1.27 (0.84, 1.99) | 0.272 |
| North | 1.02 (0.59, 1.76) | 0.934 |
| Southeast | 1.23 (0.83, 1.87) | 0.312 |
| South | 0.69 (0.41, 1.17) | 0.166 |
CI: confidence interval. Results are presented as adjusted odds ratios (ORs) with 95% confidence intervals (CIs) obtained from a multivariable logistic regression model including survey year, age group, sex, educational attainment, household income, geographic region, and type of city.
Fig. 2.
Predicted probability of adequate basic sepsis knowledge according to educational attainment across survey years. Probabilities were estimated from a logistic regression model including survey year, educational attainment, and their interaction. Points represent predicted probabilities and error bars indicate 95% confidence intervals. While visual inspection suggests a larger increase in knowledge among participants with higher educational attainment, formal testing did not demonstrate a significant interaction between survey year and educational level (p = 0.42).
Fig. 3.
Relative importance of sociodemographic variables for predicting basic sepsis, time-dependency and long-term disability knowledge. Variable importance was assessed using permutation analysis based on the reduction in the area under the receiver operating characteristic curve (AUROC) following random permutation of each predictor in the final multivariable logistic regression models. Larger reductions in AUROC indicate greater contribution of the predictor to model discrimination. Error bars represent 95% confidence intervals. Educational attainment emerged as the most important predictor across all knowledge domains.
Separate multivariable logistic regression models were fitted for knowledge regarding the time-dependent nature of sepsis treatment and awareness of long-term disability after sepsis. Participant characteristics according to these outcomes are shown in Tables S2 and S3, and the predicted probabilities of knowledge across categories of each demographic and socioeconomical covariate are displayed in the Figures S8–S13 (time-dependency) and Figures S14–S19 (long-term disability). For knowledge regarding time-dependency, higher educational attainment remains as the strongest sociodemographic determinant. Higher household income, and female sex were independently associated with correct responses (Table S4). For awareness of long-term disability, associations were generally weaker (Table S5). Despite differences in the regression models, permutation analyses consistently identified educational attainment as the most important predictor across all knowledge domains evaluated. In contrast, geographic region and city type contributed minimally to model discrimination (Fig. 3)
Discussion
In this large population-based analysis spanning more than a decade, public awareness of sepsis in Brazil increased substantially between 2014 and 2026, suggesting a positive impact of sustained national and international efforts to improve sepsis recognition. Although only a minority of participants fulfilled our strict definition of adequate basic knowledge, a considerably larger proportion demonstrated some understanding of sepsis and its time-dependency nature. Educational attainment emerged as the strongest determinant of sepsis knowledge across all domains evaluated. Importantly, despite the overall improvement in awareness over time, socioeconomic inequalities in knowledge persisted, suggesting that gains in awareness were not driven by reductions in disparities.
Our findings are also consistent with previous studies demonstrating substantial heterogeneity in public sepsis awareness across countries and healthcare settings [8,9,[16], [17], [18], [19], [20]]. A scoping review showed large disparities in awareness, with higher knowledge in high-resource settings and among populations with greater access to education and healthcare information [8]. In this study, the awareness increased over time, particularly after major international sepsis campaigns and the establishment of World Sepsis Day. Our results extend these observations by demonstrating that, even within a middle-income country, sepsis awareness is strongly socially patterned. Together, these findings suggest that disparities in knowledge occur not only between countries, but also within populations, reinforcing the importance of locally adapted and socially targeted strategies. This is particularly relevant for middle-income countries, where educational inequalities and heterogeneous access to health information may limit the reach of broad public awareness campaigns [21,22].
Importantly, public awareness of sepsis increased substantially over the study period. Although only 10.3% of participants in 2026 fulfilled the standardized definition of adequate basic knowledge used for temporal comparisons, the broader picture was more encouraging. When presented with a stimulated question, regardless of whether they had previously heard the term sepsis, a considerably larger proportion of respondents correctly recognized sepsis either as an abnormal response of the body to infection or, albeit imprecisely, as a “blood infection.” This finding was analyzed separately because incorporating it into the temporal trend analysis would have compromised comparability with the 2014 and 2017 surveys. Nonetheless, our findings suggest that public understanding of sepsis extends beyond the strict definition used in our primary outcome and likely reflects the impact of national and international efforts over the past decade to increase awareness of the condition. Interestingly, knowledge regarding the time-dependent nature of sepsis treatment was substantially higher than basic awareness of sepsis itself. This finding reinforces that individuals may understand the concept of severe infection requiring urgent treatment even when they are unfamiliar with the term 'sepsis'. Future public awareness strategies may therefore benefit from focusing not only on the term itself, but also on recognition of severe infection and the need for urgent medical evaluation.
Nevertheless, important gaps remain. The striking contrast between awareness of myocardial infarction and awareness of sepsis suggests that low sepsis awareness may not simply reflect limited health literacy. Rather, it may also reflect decades of sustained public education efforts for cardiovascular diseases that have no equivalent to sepsis. The reasons for persistently low adequate awareness of sepsis are likely multifactorial. First, sepsis itself remains a complex and difficult syndrome to define and recognize, even among healthcare professionals [[23], [24], [25]]. This complexity may impair communication with patients and families, leading healthcare professionals to describe the underlying infection rather than explicitly using the term “sepsis”. Of note, a higher percentage of the participants in 2017 recognized the term “septicemia”. Another challenge may be related to the nonspecific nature of sepsis manifestations. In contrast to acute myocardial infarction or stroke, sepsis lacks a highly recognizable symptom that can be easily disseminated through public campaigns and readily identified by laypersons. Most manifestations are related to the underlying infection or become apparent only after organ dysfunction develops, which may limit public recognition and delay healthcare-seeking behavior.
Improving public awareness of sepsis may have important implications at multiple levels. In other acute conditions like stroke and acute myocardial infarction, educational campaigns have been associated with reductions in delays in seeking emergency care [[26], [27], [28]]. In sepsis, earlier recognition and timely treatment are strongly associated with improved outcomes [29]. Increased awareness among the public and policymakers may also contribute to greater prioritization of sepsis as a healthcare problem, potentially influencing resource allocation, implementation of quality improvement initiatives, and research funding.
An important finding of our study is that different dimensions of sepsis knowledge appear to be shaped by partially distinct social determinants. Basic awareness of sepsis was independently associated with both educational attainment and socioeconomic conditions. In contrast, the contribution of socioeconomic variables became less consistent when evaluating knowledge regarding the time-dependent nature of sepsis treatment and its long-term consequences. This pattern suggests that different forms of health knowledge may rely on distinct mechanisms of information acquisition and processing. Basic awareness may depend on broader exposure to healthcare systems, media, and social networks. Complex concepts, such as time-dependent treatment and post-sepsis morbidity, likely require higher health literacy and may therefore depend more directly on formal education.
Importantly, although awareness increased substantially over the study period, inequalities did not significantly narrow. We did not identify statistically significant interactions between survey year and key socioeconomic variables, including educational attainment. Nevertheless, the observed temporal patterns are noteworthy. While awareness improved across educational groups, the absolute gains appeared more pronounced among individuals with higher educational attainment, whereas improvements among the least educated groups were comparatively modest. Although these differences should be interpreted cautiously given the absence of a formal interaction, they raise the possibility that gains in sepsis awareness have not been evenly distributed across the population. These findings have important public health implications. Broad awareness campaigns may be sufficient to increase general recognition of sepsis, but improving deeper understanding likely requires educational strategies adapted to different levels of health literacy. This is particularly relevant because previous studies have shown that public awareness of sepsis is generally higher in high-resource settings and tends to improve following awareness campaigns and educational initiatives. Efforts targeting vulnerable populations should therefore move beyond dissemination of the term “sepsis” and incorporate accessible communication regarding warning signs, the urgency of treatment, and the potential long-term consequences of sepsis.
Our study has several strengths, including its population-based design, large sample size, standardized methodology, and the ability to evaluate temporal trends over more than a decade. However, some limitations should be acknowledged. First, certain survey questions differed slightly across years, which may have affected temporal comparisons. Second, as in most survey-based studies, residual selection bias and limitations in response rate assessment cannot be completely excluded. Third, we evaluated awareness and knowledge rather than actual healthcare-seeking behavior. Therefore, we cannot determine whether greater awareness would necessarily translate into earlier presentation, faster diagnosis, or improved clinical outcomes. Finally, although the Datafolha Institute employs a rigorous sampling strategy designed to obtain a sample representative of the Brazilian population, no post-survey weighting was applied in the statistical analyses. Consequently, small differences between the study sample and the target population may have remained.
Conclusion
In conclusion, public awareness of sepsis in Brazil has improved substantially over the past decade, reflecting the success of sustained national and international efforts to increase recognition of this major public health problem. Nevertheless, important gaps remain, and awareness continues to be strongly influenced by educational attainment and socioeconomic factors. These findings highlight both the progress achieved and the need for continued efforts involving public health authorities, healthcare professionals, educators, researchers, and policymakers to further improve sepsis awareness and promote equitable access to health information.
Author contributions
Conceptualization: FRM, LCPA, DCS.
Formal analysis: TMS, FRM.
Funding acquisition: FRM, DCS.
Investigation: FRM, LCPA, LRP, JFF, JLDS, TL, RBM, RS, RR, DCS.
Methodology: FRM, LCPA.
Project administration: LRP, JFF, JLDS.
Writing – original draft: FRM, LCPA.
Writing – review and editing designed the study: FRM, LCPA, LRP, JFF, JLDS, TL, RBM, RS, TMS, RR, DCS.
Ethical approval and consent to participate
The study was approved by the Ethics Committee of the Federal University of São Paulo (CAAE 54357815.7.0000.5505), and all participants provided informed consent.
Declaration of Generative AI and AI-assisted technologies in the writing process
During the preparation of this work, the authors used ChatGPT only for English editing. The authors reviewed and edited the output as needed and take full responsibility for the content of the published article.
Funding
This study received financial support from the National Research Council through the Research Productivity Grant for the first and second authors. BioMérieux supported the direct cost of Datafolha Institute for the 2026 poll. The funding source has no influence on study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.
Consent for publication
Not applicable
Availability of data
All the data sets and data dictionary generated for this study will be publicly available immediately following publication (no end date) upon reasonable from researchers who provide a methodologically sound proposal request to the Latin America Sepsis Institute (LASI, secretaria@ilas.org.br), subject to prior assessment and approval of the proposal by LASI council. To gain access, data requestors will need to sign a data access agreement.
Data availability
The data that has been used is confidential.
Data will be made available on request.
Declaration of competing interest
The authors declare no conflicts of interest.
Acknowledgements
Not applicable
Footnotes
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.aicoj.2026.100128.
Appendix A. Supplementary data
The following are Supplementary data to this article:
References
- 1.Rudd K.E., Johnson S.C., Agesa K.M., Shackelford K.A., Tsoi D., Kievlan D.R., et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet. 2020;395(10219):200–211. doi: 10.1016/S0140-6736(19)32989-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.GBD 2021 Global Sepsis Collaborators Global, regional, and national sepsis incidence and mortality, 1990-2021: a systematic analysis. Lancet Glob Health. 2025;13(12):e2013–e2026. doi: 10.1016/S2214-109X(25)00356-0. [DOI] [PubMed] [Google Scholar]
- 3.Machado F.R., Cavalcanti A.B., Bozza F.A., Ferreira E.M., Angotti Carrara F.S., Sousa J.L., et al. The epidemiology of Sepsis in Brazilian intensive care units (the Sepsis PREvalence Assessment Database, SPREAD): an observational study. Lancet Infect Dis. 2017;17(11):1180–1189. doi: 10.1016/S1473-3099(17)30322-5. [DOI] [PubMed] [Google Scholar]
- 4.Souza D.C., Rosa R.G., Salomao R., Machado F.R. Improving the outcomes of sepsis in Brazil: strategies and initiatives. Crit Care Sci. 2025;37 doi: 10.62675/2965-2774.20250313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.de Souza D.C., Goncalves Martin J., Soares Lanziotti V., de Oliveira C.F., Tonial C.T., de Carvalho W.B., et al. The epidemiology of Sepsis in paediatric intensive care units in Brazil (the Sepsis PREvalence Assessment Database in Pediatric population, SPREAD PED): an observational study. Lancet Child Adolesc Health. 2021;5(12):873–881. doi: 10.1016/S2352-4642(21)00286-8. [DOI] [PubMed] [Google Scholar]
- 6.Prescott H.C., Antonelli M., Alhazzani W., Møller M.H., Alshamsi F., Azevedo L.C.P., et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2026. Crit Care Med. 2026 Apr 1;54(4):725–812. doi: 10.1097/CCM.0000000000007075. [DOI] [PubMed] [Google Scholar]
- 7.Seymour C.W., Gesten F., Prescott H.C., Friedrich M.E., Iwashyna T.J., Phillips G.S., et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med. 2017;376(23):2235–2244. doi: 10.1056/NEJMoa1703058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Surviving Sepsis Campaign Steering Committee. Rubulotta F.M., Ramsay G., Parker M.M., Dellinger R.P., Levy M.M., Poeze M. An international survey: public awareness and perception of Sepsis. Crit Care Med. 2009;37(1):167–170. doi: 10.1097/ccm.0b013e3181926883. [DOI] [PubMed] [Google Scholar]
- 9.Fiest K.M., Krewulak K.D., Brundin-Mather R., Leia M.P., Fox-Robichaud A., Lamontagne F., et al. Patient, public, and healthcare professionals’ sepsis awareness, knowledge, and information seeking behaviors: a scoping review. Crit Care Med. 2022;50(8):1187–1197. doi: 10.1097/CCM.0000000000005564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Parsons Leigh J., Krewulak K.D., Brundin-Mather R., Stelfox H.T., Righy C., Rattray J., et al. Public awareness and knowledge of sepsis: a cross-sectional survey of adults in Canada. Crit Care. 2022;26(1):337. doi: 10.1186/s13054-022-04215-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Azevedo L.C., Carrara F., Machado F.R., Lubarino J., Salomao R., Costa-Filho R.C. Conhecimento do público leigo sobre sepse no Brasil: uma comparação com infarto agudo do miocárdio (in portuguese) Rev Bras Terapia Intensiva. 2014;S1:S7–S43. [Google Scholar]
- 12.Azevedo L.C., Souza J.L.A., Bossa A., Monteiro M.A., Lisboa T.L., dal Pizzol F., et al. Conhecimento do público leigo sobre sepse no Brasil: uma comparação 2014-2017 (in portuguese) Rev Bras Terapia Intensiva. 2018;S1:S1–S31. [Google Scholar]
- 13.Bittencourt P.L., Codes L., Cesar H.F., Gomes Ferraz M.L. Public knowledge and attitudes toward liver diseases and liver cancer in the Brazilian population: a cross-sectional study. Lancet Reg Health Am. 2023 Jun 8;23 doi: 10.1016/j.lana.2023.100531. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Associação Brasileira de Empresas de Pesquisa . ABEP; Sao Paulo: 2024. Critério de Classificação Econzômica Brasil: CCEB 2024. [Google Scholar]
- 15.Instituto Brasileiro de Geografia e Estatística (IBGE) IBGE; Rio de Janeiro: 2022. Censo Demográfico 2022: População e Domicílios - primeiros resultados. [Google Scholar]
- 16.Waalders N.J.B., Broens B., de Keizer N., et al. A nationwide survey to explore sepsis awareness among the Dutch general public and former patients. Neth J Crit Care. 2021;29(5):231–237. [Google Scholar]
- 17.vaccination60+ study group. Eitze S., Fleischmann-Struzek C., Betsch C., Reinhart K. Determinants of sepsis knowledge: a representative survey of the elderly population in Germany. Crit Care. 2018;22(1):273. doi: 10.1186/s13054-018-2208-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Phua J., Lim H.F., Tay C.K., Aung N.W. Public awareness of sepsis and stroke in Singapore: a population-based survey. Ann Acad Med Singap. 2013;42(6):269–277. [PubMed] [Google Scholar]
- 19.Mellhammar L., Christensson B., Linder A. Public awareness of sepsis is low in Sweden. Open Forum Infect Dis. 2015;2(4) doi: 10.1093/ofid/ofv161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Born S., Rose N., Piedmont S., Schwarzkopf D., Pletz M.W., Abels W., et al. The role of perceived urgency and barriers in using emergency medical care for sepsis: a German-wide, representative survey using case vignettes. Crit Care Med. 2026 Feb 1;54(2):291–300. doi: 10.1097/CCM.0000000000006967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Machado F.R., Azevedo L.C. Sepsis: a threat that needs a global solution. Crit Care Med. 2018;46(3):454–459. doi: 10.1097/CCM.0000000000002899. [DOI] [PubMed] [Google Scholar]
- 22.Pontes Azevedo L.C. Surviving sepsis: don’t break the chain. Crit Care Med. 2025;53(10):e2086–8. doi: 10.1097/CCM.0000000000006832. [DOI] [PubMed] [Google Scholar]
- 23.Poeze M., Ramsay G., Gerlach H., Rubulotta F., Levy M. An international sepsis survey: a study of doctors’ knowledge and perception about sepsis. Crit Care. 2004;8(6):R409–R413. doi: 10.1186/cc2959. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Rhee C., Kadri S.S., Danner R.L., Suffredini A.F., Massaro A.F., Kitch B.T., et al. Diagnosing sepsis is subjective and highly variable: a survey of intensivists using case vignettes. Crit Care. 2016;20:89. doi: 10.1186/s13054-016-1266-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Assuncao M., Akamine N., Cardoso G.S., Mello P.V.C., Teles J.M.M., Nunes A.L.B., et al. Survey on physicians’ knowledge of sepsis: do they recognize it promptly? J Crit Care. 2010;25(4):545–552. doi: 10.1016/j.jcrc.2010.03.012. [DOI] [PubMed] [Google Scholar]
- 26.Caminiti C., Schulz P., Marcomini B., Iezzi E., Riva S., Scoditti U., et al. Development of an education campaign to reduce delays in pre-hospital response to stroke. BMC Neurol. 2017;17(1):20. doi: 10.1186/s12873-017-0130-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mellon L., Doyle F., Williams D., Brewer L., Hall P., Hickey A. Stroke warning campaigns: delivering better patient outcomes? A systematic review. Patient Relat Outcome Meas. 2015;6:61–73. doi: 10.2147/PROM.S54087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hoschar S., Albarqouni L., Ladwig K.H. A systematic review of educational interventions aiming to reduce prehospital delay in patients with acute coronary syndrome. Open Heart. 2020;7(1) doi: 10.1136/openhrt-2019-001175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Noritomi D.T., Ranzani O.T., Monteiro M.B., Ferreira E.M., Santos S.R., Leibel F., et al. Implementation of a multifaceted sepsis education program in an emerging country setting: clinical outcomes and cost-effectiveness in a long-term follow-up study. Intensive Care Med. 2014;40(2):182–191. doi: 10.1007/s00134-013-3131-5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All the data sets and data dictionary generated for this study will be publicly available immediately following publication (no end date) upon reasonable from researchers who provide a methodologically sound proposal request to the Latin America Sepsis Institute (LASI, secretaria@ilas.org.br), subject to prior assessment and approval of the proposal by LASI council. To gain access, data requestors will need to sign a data access agreement.
The data that has been used is confidential.
Data will be made available on request.