Demographic Inaccuracies and Biases in the Depiction of Patients by Artificial Intelligence Text-to-Image Generators DOI Creative Commons
Tim L. T. Wiegand, Leonard Jung,

Luisa S. Schuhmacher

и другие.

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Июль 25, 2024

Abstract The wide usage of artificial intelligence (AI) text-to-image generators raises concerns about the role AI in amplifying misconceptions healthcare. This study therefore evaluated demographic accuracy and potential biases depiction patients by two commonly used generators. A total 4,580 images with 29 different diseases was generated using Bing Image Generator Meta Imagine. Eight independent raters determined sex, age, weight group, race ethnicity depicted. Comparison to real-world epidemiology showed that failed depict demographical characteristics such as accurately. In addition, we observed an over-representation White well normal individuals. Inaccuracies may stem from non-representative non-specific training data insufficient or misdirected bias mitigation strategies. consequence, new strategies counteract inaccuracies are needed.

Язык: Английский

Fear of COVID-19 among professional caregivers of the elderly in Central Alentejo, Portugal DOI Creative Commons
Felismina Rosa Parreira Mendes, Margarida Sim-Sim, Maria Laurência Gemito

и другие.

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Фев. 7, 2024

Abstract The coronavirus disease 2019 (COVID-19) has infected many institutionalised elderly people. In Portugal, the level of pandemic fear among professional caregivers is unknown, as are its predictive factors. This study aimed to investigate predictors COVID-19 workers caring for people in nursing homes. a cross-sectional using multiple linear regression applied population 652 located 14 municipalities Central Alentejo, at March 2021. criterion variable was COVID-19. Standardised coefficients showed that higher education, lower (β = − 0.158; t 4.134; p < .001). Other were gender, with women having levels 0.123; 3.203; 0.001), scores on COVID-19-like suspicious symptoms 3.219; 0.001) and received flu vaccine 0.086; 2.252; 0.025). model explains 6.7% variation (R 2 Adj 0.067). Health literacy can minimise impact physical mental health these workers. play fundamental role social balance. Further studies needed better understand factors improve their personal well-being.

Язык: Английский

Процитировано

1

Cytokine release syndrome after treatment with immune checkpoint inhibitors: an observational cohort study of 2672 patients from Karolinska University Hospital in Sweden DOI Creative Commons

Osama Hamida,

Frans Karlsson,

Andreas Lundqvist

и другие.

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Март 16, 2024

Abstract Immune checkpoint inhibitors (ICIs) are linked to diverse immune-related adverse events (irAEs). Rare irAEs surface first in clinical practice. Here, we systematically studied the rare irAE, cytokine-release syndrome (CRS), a cohort of 2672 patients treated with ICIs at Karolinska University Hospital Stockholm, Sweden. We find that risk ICI-induced CRS – defined as fever, negative microbiological findings and absence other probable causes within 30 days after ICI treatment is approximately 1%, higher than previously reported. was often mild rechallenge generally safe. Two out 28 experienced high-grade CRS, one fatal. While C-reactive protein procalcitonin were not discriminative fatal our data suggest quick Sequential Organ Failure Assessment (qSOFA) score might identify high-risk patients. These provide framework for assessment motivate multicenter studies improve early diagnosis. Highlights Cytokine release following immune inhibition mild. Risk using Assessment, but serum CRP, can potentially detect severe cytokine decisions. Rechallenge well tolerated.

Язык: Английский

Процитировано

1

Hybrid Immunity and the Incidence of SARS-CoV-2 Reinfections during the Omicron Era in Frontline Healthcare Workers DOI Creative Commons
Carmen-Daniela Chivu, Maria Crăciun, Daniela Pițigoi

и другие.

Vaccines, Год журнала: 2024, Номер 12(6), С. 682 - 682

Опубликована: Июнь 19, 2024

During the coronavirus disease (COVID-19) pandemic healthcare workers (HCWs) acquired immunity by vaccination or exposure to multiple variants of severe acute respiratory syndrome 2 (SARS-CoV-2). Our study is a comparative analysis between subgroups HCWs constructed based on number SARS-CoV-2 infections, vaccination, and dominant variant in population. We collected analyzed data using χ2 test density incidence reinfections Microsoft Excel for Mac, Version 16.84, MedCalc®, 22.026. Of 829 HCWs, 70.1% (581) had only one infection 29.9% (248) two infections. subjects with 77.4% (192) worked high-risk departments 93.2% (231) second infections were registered during Omicron dominance. The was higher vaccinated primary schedule than those first booster, ratio 2.8 (95% CI: 1.2; 6.7). probability reinfection five times lower 2.9; 9.2) if booster three 1.9; 5.8) Omicron. characteristics such as gender, age group, job category, department also significant differences incidence. history important when interpreting understanding public health results studies related vaccine efficacy hybrid subgroup populations.

Язык: Английский

Процитировано

1

Sex differences in COVID-19 mortality: A large US-based cohort study (2020–2022) DOI Creative Commons
Samer Kharroubi, Marwa Diab-El-Harake

AIMS Public Health, Год журнала: 2024, Номер 11(3), С. 886 - 904

Опубликована: Янв. 1, 2024

In the present study, we aim to assess trend in mortality COVID-19 by time and sex a large cohort using Datavant's Death Index database. The main objectives of this study are analyze cases over time, which categorized age, identify potential reasons for observed differences.

Язык: Английский

Процитировано

1

Demographic Inaccuracies and Biases in the Depiction of Patients by Artificial Intelligence Text-to-Image Generators DOI Creative Commons
Tim L. T. Wiegand, Leonard Jung,

Luisa S. Schuhmacher

и другие.

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Июль 25, 2024

Abstract The wide usage of artificial intelligence (AI) text-to-image generators raises concerns about the role AI in amplifying misconceptions healthcare. This study therefore evaluated demographic accuracy and potential biases depiction patients by two commonly used generators. A total 4,580 images with 29 different diseases was generated using Bing Image Generator Meta Imagine. Eight independent raters determined sex, age, weight group, race ethnicity depicted. Comparison to real-world epidemiology showed that failed depict demographical characteristics such as accurately. In addition, we observed an over-representation White well normal individuals. Inaccuracies may stem from non-representative non-specific training data insufficient or misdirected bias mitigation strategies. consequence, new strategies counteract inaccuracies are needed.

Язык: Английский

Процитировано

1