Search

Machine learning algorithms are being increasingly used in healthcare settings but their generalizability between different regions is still unknown. This study aims to identify the strategy that maximizes the predictive performance of identifying the risk of death by COVID-19 in different regions of a large and unequal country. This is a multicenter cohort study with data collected from patients with a positive RT-PCR test for COVID-19 from March to August 2020 (n = 8477) in 18 hospitals, covering all five Brazilian regions. Of all patients with a positive RT-PCR test during the period, 2356 (28%) died. Eight different strategies were used for training and evaluating the performance of three popular machine learning algorithms (extreme gradient boosting, lightGBM, and catboost). The strategies ranged from only using training data from a single hospital, up to aggregating patients by their geographic regions. The predictive performance of the algorithms was evaluated by the area under the ROC curve (AUROC) on the test set of each hospital. We found that the best overall predictive performances were obtained when using training data from the same hospital, which was the winning strategy for 11 (61%) of the 18 participating hospitals. In this study, the use of more patient data from other regions slightly decreased predictive performance. However, models trained in other hospitals still had acceptable performances and could be a solution while data for a specific hospital is being collected. [ABSTRACT FROM AUTHOR]

Copyright of AIDS & Behavior is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Salud Colectiva is the property of Instituto de Salud Colectiva Universidad Nacional de Lanus and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Interface - Comunicação, Saúde, Educação is the property of Fundacao UNI and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Cadernos de Saude Publica is the property of Escola Nacional de Saude Publica Sergio Arouca and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Despite the known benefits of early treatment initiation for individual morbidity and mortality, as well as for reducing the risk of transmission, late presentation (LP) to HIV/AIDS services remains a major concern in many countries. There is little information on LP from middle- and low-income countries and studies that do evaluate LP commonly disaggregate data by sex. It is rare, however, for researchers to further disaggregate the data by pregnancy status so it remains unclear if pregnancy status modifies the effects associated with sex. The study was conducted at the only State Reference Center for HIV/AIDS in Salvador, Brazil's third largest city. LP was defined as a patient accessing services with a CD4 < 350 cells/mm3. Data were abstracted from the electronic medical records of 1421 patients presenting between 2007 and 2009. CD4 counts and viral load (VL) information was validated with data from the National CD4/VL Database. Descriptive and bivariate statistics were conducted to inform the multivariate analysis. Adjusted prevalence ratios (APR) were estimated using generalized linear models due to the high frequency of the outcome. Half of the sample (52.5%;n= 621) was classified as LP. Compared to the prevalence among pregnant women (21.1%), the prevalence of LP was more than twice as high among non-pregnant women (56.0%) and among men (55.4%). The multivariate analysis demonstrated no statistical difference between men and nonpregnant women (APR 1.04; 95%CI 0.92–1.19), but the APR of LP for nonpregnant women was 53% less than men (APR 0.47; 95%CI 0.33–0.68). These results highlight the importance of analyzing data disaggregated not only by sex but also by pregnancy status to accurately identify the risk factors associated with LP so that programs and policies can effectively and efficiently address LP in Brazil and beyond. [ABSTRACT FROM PUBLISHER]

Background: Brazil has many people living with HIV (PLWH) who are unaware of their serostatus. The public health system has recently added HIV self-testing (HIVST) for key populations such as men who have sex with men (MSM). This study estimates HIVST acceptability among Brazilian MSM and explores factors associated with acceptability among MSM who have never tested for HIV or who had a previous negative result.Methods: Respondent-driven sampling (RDS) was used to recruit 4176 MSM in 12 Brazilian cities in 2016 to this biological and behavioral surveillance study. We excluded from this analysis all MSM who were aware of their positive HIV serostatus. Descriptive, bivariate and multivariate analyses were conducted. Overall proportions were weighted with Gile's estimator in RDS Analyst software and 95% confidence intervals were calculated. The analyses of HIVST acceptability were stratified by prior HIV testing (never or one or more times).Results: For this analysis, 3605 MSM were included. The acceptability of HIVST was 49.1%, lower among those who had never tested for HIV (42.7%) compared to those who had a previous HIV negative test (50.1%). In the subgroup of MSM who had never tested for HIV, those who reported discrimination or who had a medical appointment in the last 12 months reported higher HIVST acceptability. Among MSM who had a previous negative HIV test, only those reporting condomless receptive anal sex reported higher HIVST acceptability. In addition, we observed that high levels of knowledge of HIV/AIDS, taking part in lesbian, gay, bisexual, and transgender nongovernmental organizations (LGBT-NGO), or complete secondary or incomplete higher undergraduate education reported higher acceptability.Conclusions: The acceptability of HIVST was low among MSM, especially among those who never tested for HIV. Given access to HIVST in Brazil, we point to the need for programs that enhance promotion of testing addressed to MSM. [ABSTRACT FROM AUTHOR]