ABSTRACT
Background: The World Health Organization (WHO) recommends systematic tuberculosis (TB) screening in prisons. Evidence is lacking for accurate and scalable screening approaches in this setting. We aimed to assess the accuracy of artificial intelligence-based chest x-ray interpretation algorithms for TB screening in prisons. Methods: We performed prospective TB screening in three male prisons in Brazil from October 2017 to December 2019. We administered a standardized questionnaire, performed a chest x-ray in a mobile unit, and collected sputum for confirmatory testing using Xpert MTB/RIF and culture. We evaluated x-ray images using three algorithms (CAD4TB version 6, Lunit version 3.1.0.0 and qXR version 3) and compared their accuracy. We utilized multivariable logistic regression to assess the effect of demographic and clinical characteristics on algorithm accuracy. Finally, we investigated the relationship between abnormality scores and Xpert semi-quantitative results. Findings: Among 2075 incarcerated individuals, 259 (12.5%) had confirmed TB. All three algorithms performed similarly overall with area under the receiver operating characteristic curve (AUC) of 0.88-0.91. At 90% sensitivity, only LunitTB and qXR met the WHO Target Product Profile requirements for a triage test, with specificity of 84% and 74%, respectively. All algorithms had variable performance by age, prior TB, smoking, and presence of TB symptoms. LunitTB was the most robust to this heterogeneity but nonetheless failed to meet the TPP for individuals with previous TB. Abnormality scores of all three algorithms were significantly correlated with sputum bacillary load. Interpretation: Automated x-ray interpretation algorithms can be an effective triage tool for TB screening in prisons. However, their specificity is insufficient in individuals with previous TB. Funding: This study was supported by the US National Institutes of Health (grant numbers R01 AI130058 and R01 AI149620) and the State Secretary of Health of Mato Grosso do Sul.
ABSTRACT
Tuberculosis (TB) remains a leading cause of infectious mortality globally, yet most cases cannot be epidemiologically linked even with extensive contact investigations and whole genome sequencing. Consequently, there remain major gaps in our understanding of where and when M. tuberculosis (Mtb) exposures occur. We aimed to investigate whether Mtb can be detected in environments where TB patients were recently present, which could serve as a tool for characterizing exposure risk. We collected 389 environment surface (ES) swabs from two high TB burden prisons in Brazil, sampling 41 (n = 340) cells occupied by individuals with active TB and 7 (n = 49) cells from individuals without TB. In a subset of pooled swabs (n = 6) and a swab from a cigarette lighter from the cell with active TB patients, we enriched Mtb DNA using RNA-bait hybrid capture assays and performed whole genome sequencing. In prison cells, Mtb DNA was detected in 55/340 (16 %) of ES swabs from cells occupied by active TB patients and none (0/49) from cells in which no active TB patients were present. Mtb was detected in 13/16 (81 %) prison cells occupied by the individuals with high/medium sputum Xpert Mtb load and 8/25 (32 %) with low/very low sputum Mtb load (p = 0.003). Seven hybrid capture samples had a median genomic coverage of 140×. rpoB mutations conferring high-level rifampin resistance were detected in 3/7 ES swabs. Mtb was frequently detectable in environments recently occupied by individuals with active TB. This approach could be applied in congregate environments to identify and characterize high-risk settings for Mtb exposure.
Subject(s)
Mycobacterium tuberculosis , Tuberculosis , Humans , Mycobacterium tuberculosis/genetics , Rifampin , Sensitivity and Specificity , Sputum , Tuberculosis/epidemiologyABSTRACT
BACKGROUND: Although systematic tuberculosis screening in high-risk groups is recommended by the World Health Organization (WHO), implementation in prisons has been limited due to resource constraints. Whether Xpert Ultra sputum pooling could be a sensitive and efficient approach to mass screening in prisons is unknown. METHODS: In total, 1280 sputum samples were collected from incarcerated individuals in Brazil during mass screening and tested using Xpert G4. We selected samples for mixing in pools of 4, 8, 12, and 16, which were then tested using Ultra. In each pool, a single positive sample of differing Xpert mycobacterial loads was used. Additionally, 10 pools of 16 negative samples each were analyzed as controls. We then simulated tuberculosis screening at prevalences of 0.5-5% and calculated the cost per tuberculosis case detected at different sputum pooling sizes. RESULTS: The sensitivity and specificity of sputum pooling were high (sensitivity: 94%; 95% confidence interval [CI]: 88-98; specificity: 100%, 95% CI: 84-100). Sensitivity was greater in pools in which the positive sample had a high mycobacterial load compared to those that were very low (100% vs 88%). In settings with a higher tuberculosis prevalence, pools of 4 and 8 were more efficient than larger pool sizes. Larger pools decreased the costs by 87% at low prevalences, whereas smaller pools led to greater cost savings at higher prevalence at higher prevalences (57%). CONCLUSIONS: Sputum pooling using Ultra was a sensitive strategy for tuberculosis screening. This approach was more efficient than individual testing across a broad range of simulated tuberculosis prevalence settings and could enable active case finding to be scaled while containing costs.
Subject(s)
Mycobacterium tuberculosis , Tuberculosis , Humans , Mass Screening , Mycobacterium tuberculosis/genetics , Prisons , Sensitivity and Specificity , Sputum/microbiology , Tuberculosis/diagnosis , Tuberculosis/epidemiologyABSTRACT
Objetivo avaliar a infraestrutura de unidades básicas de saúde, quanto à adequação às normas preconizadas pelo Ministério da Saúde. Métodos estudo descritivo, com abordagem quantitativa, realizado em 18 unidades básicas de saúde. Para coleta de dados, utilizou-se checklist construído com base na Portaria Ministerial 2.226/09. A análise dos dados foi realizada por meio do Statistical Package for the Social Sciences e empregou-se escala de pontuação (0-10) para classificação das unidades de saúde. Resultados das unidades analisadas, 16,6% ficaram com a nota igual a 3,5; 16,6% com 3,0; 5,5% com 2,5; 16,6% com nota 2,0; 11,1% com 1,5; 11,1% com 1,0; e 22,2% ficaram com 0,5, estando, portanto, todas com estrutura física inadequada. Conclusão o município não oferece ao público um serviço que contemple os padrões ideais em sua estrutura, na maioria das unidades inspecionadas.
Objetivo evaluar la infraestructura de unidades básicas de salud, cuanto a la adaptación a las normas recomendadas por el Ministerio de la Salud. Métodos estudio descriptivo, con abordaje cuantitativa, realizado en 18 unidades básicas de salud. Para recolección de datos, se utilizó lista construida con base en el Decreto Ministerial 2.226/09. Se realizó el análisis de datos utilizando el Statistical Package for the Social Sciences y empleó a escala de puntuación (0-10) para clasificación de los unidades de salud. Resultados de las unidades analizadas, 16,6% estaban con puntuación media de 3,5; 16,6% con 3,0; 5,5% con 2,5; 16,6% con nota 2,0; 11,1% con 1,5; 11,1% con 1,0; y 22,2% alcanzaron 0,5, por lo tanto, todas con estructura física inadecuada. Conclusión la ciudad no ofrece al público servicio que contemple las normas ideales en su estructura, en la mayoría de las unidades inspeccionadas.
Objective to evaluate the infrastructure of basic health units, as the adaptation to the standards recommended by the Ministry of Health. Methods descriptive study with a quantitative approach, carried in 18 basic health units. For data collection, we used a checklist built based on Ministerial Decree 2,226/09. Data analysis was performed using the Statistical Package for Social Sciences and a scoring scale (0-10) was used for the classification of health facilities. Results 16.6% of the analyzed units received the average score of 3.5; 16.6% received 3.0; 5.5%, 2.5; 16.6% received grade 2.0; 11.1%, 1.5; 11.1%, 1.0; and 22.2% received 0.5 and, therefore, all units had inadequate physical structure. Conclusion the city does not offer the public a service that addresses the ideals standards in its structure in most of the inspected units.
Subject(s)
Primary Health Care , Health Infrastructure , Health Services ResearchABSTRACT
Objetivou-se investigar a vulnerabilidade dos usuários de crack cadastrados nos Centro de Atenção Psicossocial Álcool e Drogas (CAPSad) do Piauí, com relação à infecção pelo Human Imunodeficiency Vírus (HIV). Trata-se de um inquérito epidemiológico, realizado nos CAPSad do Piauí, com 343 usuários. Predominaram o sexo masculino, com média de idade de 29,2 anos. Os fatores de riscos predominantes foram múltiplos parceiros sexuais, uso de álcool e drogas antes do sexo, história de prisão e tatuagem. Conclui-se que os usuários de crack constituem um grupo com a susceptibilidade aumentada à infecção pelo vírus HIV, devido comportamentos de risco por eles assumidos.
The aim of this work was to investigate the vulnerability of crack users enrolled in Psychosocial Care Center Alcohol and Drugs (CAPSad) Piauí, with respect to infection Imunodeficiency Human Virus (HIV). This is a survey carried out in CAPSad the state of Piaui, with 343 users. A predominance of male sex with mean age of 29.2 years. The risk factors were predominant multiple sexual partners, use of alcohol and drugs before sex, and history of prison tattooing. It is concluded that crack users are a group with increased susceptibility to HIV infection, risk behaviors because they assumed.
El objetivo era investigar la vulnerabilidad de los consumidores de crack inscrito en Alcohol Centro de Atención Psicosocial y Drogas (CAPSad) Piauí, con respecto a la infección del Virus Imunodeficiency Humana (VIH). Se trata de un estudio llevado a cabo en CAPSad el estado de Piauí, con 343 usuarios. Se identificó predominio de hombres del sexo masculino, con una edad media de 29,2 años. Los factores de riesgo predominantes fueron múltiples parejas sexuales, uso de alcohol y drogas antes de tener sexo, y la historia de los tatuajes prisión. Se concluye que los usuarios de crack son un grupo con una mayor susceptibilidad a la infección por VIH, las conductas de riesgo, ya que se supone.
