A hepatite C é um processo inflamatório do tecido hepático que acomete milhões de pessoas em todo mundo. A evolução crônica da doença pode levar a transformação da fibrose hepática em cirrose e carcinoma hepatocelular em um grande número de casos. A avaliação dos graus de fibrose hepática é importante para o estudo da gravidade e progressão da doença, bem como decisão terapêutica e avaliação de sua eficácia. Os métodos não invasivos de avaliação da fibrose hepática têm sido uma opção aos possíveis riscos da biópsia hepática, ainda considerada o melhor método de avaliação de fibrose hepática. No presente trabalho, 107 pacientes portadores de hepatite C foram submetidos à biópsia hepática e Elastografia transitória hepática, 106 pacientes APRI e FIB4, ELF em 68 pacientes e ARFI em 51 pacientes. Usando a área abaixo da Curva ROC, AUROC, para obter a acurácia da fibrose hepática, encontramos na fibrose significativa ( >= F2): Elastografia transitória hepática: 0,83; FIB4: 0,76; ELF: 0,70; APRI: 0,69; ARFI: 0,67; na fibrose avançada ( >= F3): Elastografia transitória hepática: 0,85; ELF: 0,82; FIB4: 0,77; ARFI: 0,74; APRI: 0,71 e na cirrose ( >= F4) APRI: 1; FIB4: 1; Elastografia transitória hepática: 0,99; ARFI: 0,96; ELF: 0,94. Podemos dizer que em todos os graus de fibrose avaliados, a Elastografia transitória hepática foi o método que apresentou a melhor acurácia. Em se de tratando de cirrose (>=F4), todos os métodos não invasivos apresentam excelente acurácia. Utilizando o método Obuchowski, encontramos em cada grau de fibrose hepática classificado pelo escore METAVIR as acurácias: F1= Elastografia transitória hepática: 0,81; ARFI: 0,78; APRI: 0,72; FIB4: 0,67; ELF: 0,44; F2= Elastografia transitória hepática: 0,73; FIB4: 0,68; ELF: 0,63; APRI: 0,60; ARFI: 0,53; F3= ELF: 0,77; Elastografia transitória hepática: 0,70; FIB4: 0,67; ARFI: 0,64; APRI: 0,60 e F4: APRI e FIB4: 1; Elastografia transitória hepática: 0,98; ARFI: 0,96; ELF: 0,82. A Elastografia transitória hepática se mantém como um método eficiente para todos os graus de fibrose, sendo que nos extremos apresenta discreta superioridade em relação aos graus intermediários. A acurácia de todos os métodos é superior em F4 / Hepatitis C is an inflammatory condition of the hepatic tissue that affects millions worldwide. The chronic stages of the disease turns from hepatic fibrosis into cirrhosis and hepatocellular carcinoma in many cases. The evaluation of fibrosis staging is important for prognosis, as well as understanding the progression of the disease, choice of treatment options and assessing their effectiveness. Non-invasive methods of fibrosis assessment have increasingly become alternatives to liver biopsy, which is still considered the best method of fibrosis assessment. In this study, 107 consecutive patients with hepatitis C virus were submitted to liver biopsy and transient elastography, 106 underwent APRI and FIB-4, 68 underwent ELF and 51 underwent ARFI. Using the area under ROC curve (AUROC) to obtain the degree of accuracy of each test, the following cutoffs were found for significant fibrosis (F >= 2): transient elastography: 0,83; FIB4: 0,76; ELF: 0,70; APRI: 0,69; ARFI: 0,67; For advanced fibrosis (F>=3): transient elastography: 0,85; ELF: 0,82; FIB4: 0,77; ARFI: 0,74; APRI: 0,71; For cirrhosis (F >= 4): APRI: 1; FIB4: 1; transient elastography: 0,99; ARFI: 0,96; ELF: 0,94. Of the methods assessed, transient elastography presented the greatest diagnostic accuracy across all levels of fibrosis. When assessing cirrhosis (F>= 4), all of these non-invasive methods showed excellent diagnostic accuracy. Using the Obuchowski method, the accuracy of each degree of fibrosis categorised by the METAVIR score was determined: F1= transient elastography: 0,81; ARFI: 0,78; APRI: 0,72; FIB4: 0,67; ELF: 0,44; F2=transient elastography: 0,73; FIB4: 0,68; ELF: 0,63; APRI: 0,60; ARFI: 0,53; F3= ELF: 0,77; transient elastography: 0,70; FIB4: 0,67; ARFI: 0,64; APRI: 0,60; F4: APRI and FIB4: 1; transient elastography: 0,98; ARFI: 0,96; ELF: 0,82. Transient elastography remained the most effective method for all degrees of fibrosis, although at the higher levels this superiority was less than at the intermediate levels. The accuracy of all methodologies was best at F >= 4

INTRODUÇÃO: O câncer de pulmão é a terceira neoplasia maligna mais frequentemente diagnosticada em todo o mundo e a primeira em termos de mortalidade. O tratamento cirúrgico é a melhor abordagem nos estágios iniciais, contudo, está associado a morbimortalidade considerável. Para que o impacto do tratamento cirúrgico na diminuição global da mortalidade pelo câncer de pulmão no estado de São Paulo seja maior, precisamos conhecer os indicadores de qualidade das instituições envolvidas no tratamento desta neoplasia, através da criação de uma base de dados abrangente, confiável e transparente. Este estudo envolveu a implementação do Registro Paulista de Tratamento Cirúrgico do Câncer de Pulmão (RPCP). O desfecho principal foi a análise da qualidade dos dados capturados através de um sistema de auditoria direta e indireta, com o intuito de identificar as variáveis com menor padrão de qualidade. MÉTODOS: Estudo prospectivo, multicêntrico, com participação de 10 instituições no estado de São Paulo. A auditoria dos dados foi realizada de forma direta por revisão dos prontuários, para análise da taxa de discordância, Coeficiente Kappa e Intraclass correlation e de forma indireta para análise dos índices de completude, acurácia e consistência. RESULTADOS: Dos 536 casos disponíveis, 511 foram incluídos para a auditoria indireta. O índice total de completude por questionário variou de 0,82 a 1, sendo que as seguintes variáveis obtiveram valor individual abaixo da meta estabelecida de 0,8: ECOG, MRC, hematócrito, potássio, uréia, creatinina, DHL, albumina, cálcio e FA, tempo de cirurgia e data da recidiva. O índice total de acurácia e consistência foi 0,99 e 0,96, respectivamente. Para auditoria direta foram randomizados 100 casos entre os 511 iniciais, sendo 4 excluídos, restando 96 para análise. As variáveis com maiores taxas de discordância ( > 20%), estavam no questionário de avaliação pré-operatória (ECOG, MRC, carga tabágica, DPOC, PFP, peso, altura, IMC e exames laboratoriais). Variáveis relacionadas ao estadiamento (tamanho da neoplasia, invasão de estruturas adjacentes, status linfonodal não invasivo) e dados cirúrgicos (tempo de cirurgia) também apresentaram taxas > 20%. CONCLUSÕES: A auditoria indireta dos dados mostrou índices de completude, acurácia e consistência aceitáveis para o padrão estabelecido e comparáveis a bancos de dados internacionais. Por outro lado, a auditoria direta, revelou algumas variáveis com altos índices de discordância, dados que serão analisados futuramente para aprimoramento do RPCP e que poderão contribuir para o desenvolvimento de outras bases de dados semelhantes / BACKGROUND: Lung cancer is the third malignant neoplasm most frequently diagnosed worldwide and the first in terms of mortality. Surgical treatment is the best approach in the initial stages; however, it\'s associated with considerable morbidity and mortality. In order to improve the surgical treatment global impact on lung cancer mortality in the state of Sao Paulo, we need to know the quality indicators of the institutions involved in the treatment of this neoplasm through the creation of a extensive, reliable and transparent database. The study involved the implementation of the Paulista Lung Cancer Registry (PLCR). The main outcome was the quality analysis of the data captured through a direct and indirect audit system, in order to identify the variables with the lowest quality standard. METHODS: A prospective, multicenter study with the participation of 10 institutions in the state of São Paulo. The data audit was performed directly, through the revision of medical registries, with the intention to analyze the discordance rate; and indirectly, with the intention to analyze the completeness, accuracy and consistency indexes. RESULTS: Of the 536 cases available, 511 were included for the indirect audit. The total completeness index per questionnaire ranged from 0.82 to 1, and the following variables had a in individual value bellow the established target of 0,8: ECOG, MRC, hematocrit, potassium, urea, creatinine, LDH, albumin, calcium, AF, surgical time, date of recurrence. The total accuracy and consistency index was 0.99 and 0.96, respectively. For direct audit, 100 cases were randomized among the initial 511, of which 4 were excluded, remaining 96 for analysis. The variables with the highest discordance rates ( > 20%) were in the preoperative evaluation questionnaire (ECOG, MRC, smoking rate, COPD, PFT, weight, high, BMI and lab tests). Variables related to staging (size of neoplasm, invasion of adjacent structures, noninvasive lymph node status) and surgical data (time of surgery) also presented rates > 20%. CONCLUSIONS: Regarding the established standards, the Indirect audit showed acceptable completeness, accuracy and consistency indices, comparable to international databases. On the other hand, the direct audit revealed some variables with high discordance indices, data that will be analyzed in the future for the improvement of the PLCR and that may contribute to the development of other similar databases

<p dir="ltr">As a new spaceborne laser remote sensing system, the Global Ecosystem Dynamics Investigation, or GEDI, is being widely used for monitoring forest ecosystems. However, its measurements are subject to uncertainties that will affect the calculation of ground elevation and vegetation height. This research intends to investigate the quality of the GEDI elevation data and its relevance to topography and land cover.</p><p dir="ltr">In this study, the elevation of the GEDI data is compared to 3DEP DEM, which has a higher resolution and accuracy. All the experiments in this study are conducted for two locations with vastly different terrain and land cover conditions, namely Tippecanoe County in Indiana and Mendocino County in California. Through this investigation we expect to gain a comprehensive understanding of GEDI’s elevation quality in various terrain and land cover conditions.</p><p dir="ltr">The results show that GEDI data in Tippecanoe County has better elevation accuracy than the GEDI data in Mendocino County. GEDI in Tippecanoe County is almost four times more accurate than in Mendocino County. Regarding land cover, GEDI have better accuracy in low vegetation areas than in forest areas. The ratio can be around three times better in Tippecanoe County and around one and half times better in Mendocino County. In terms of slope, GEDI data shows a clear positive correlation between RMSE and slope. The trend indicates as slope increases, the RMSE increases concurrently. In other words, slope and GEDI elevation accuracy are inversely related. In the experiment involving slope and land cover, the results show that slope is the most influential factor to GEDI elevation accuracy.</p><p dir="ltr">This study informs GEDI users of the factors they must consider for forest biomass calculation and topographic mapping applications. When high terrain slope and/or high vegetation is present, the GEDI data should be checked with other data sources like 3DEP DEM or any ground truth measurements to assure its quality. We expect these findings can help worldwide users understand that the quality of GEDI data is variable and dependent on terrain relief and land cover.</p>

This bachelor's thesis explores the integration of machine learning (ML) with sensor fusion techniques to enhance spatial data accuracy in Extended Reality (XR) environments. With XR's revolutionary impact across various sectors, accurate localization in virtual environments becomes imperative. The thesis conducts a comprehensive literature review, highlighting advancements in indoor positioning technologies and the pivotal role of machine learning in refining sensor fusion for precise localization. It underscores the challenges in the XR field, such as signal interference, device heterogeneity, and data processing complexities. Through critical analysis, this study aims to bridge the gap in practical application of ML, offering insights into developing scalable solutions for immersive virtual productions. It offers insights into the practical integration of advanced machine learning techniques in XR applications, thereby providing valuable implications for technology development and user experience in XR. This contribution is not merely theoretical; it showcases practical applications and advancements in real-time processing and adaptability in complex environments, aligning well with existing research and extending it by addressing scalability and practical implementation challenges in XR environments. This study identifies key themes in the integration of ML with sensor fusion for XR, such as the enhancement of spatial data accuracy, challenges in real-time processing, and the need for scalable solutions. It concludes that the fusion of ML and sensor technologies not only enhances the accuracy of XR environments but also paves the way for more immersive and realistic virtual experiences.

Thesis (PhD)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: The prevention of mother-to-child transmission (PMTCT) of HIV is a key maternal and child health intervention in the context of the HIV/AIDS pandemic in South Africa. Accordingly, the PMTCT programme has been incorporated in the District Health Management Information System (DHMIS) that collects monthly facility-based data to support the management of public health services. To date, there has not been a comprehensive evaluation of the PMTCT information system. By comparing the experiences in two health districts, using the Performance of Routine Information System Management (PRISM) framework and tools, this study seeks to evaluate the availability, quality and use of process and output indicators for monitoring PMTCT interventions. A comparative analytical and observational study was undertaken using a multi-method approach which included: a self-administered survey of health information personnel to assess confidence and competence levels for routine health information system (RHIS) tasks, an assessment of the routine PMTCT data for quality, completeness, accuracy, and data use; and a facility survey of RHIS processes and resources. In addition, in-depth interviews with 22 key informants and observations in health facilities were conducted. Data were collected from 57 health facilities in a convenience sample of two health districts, and also from 182 health information personnel in the 57 health facilities, three sub-districts, and two district offices. Descriptive statistics, χ2-test, correlation and multiple regression analyses were conducted using STATA® Version 13. A general inductive approach was also used to analyse the qualitative data, which was used for triangulation. The study revealed considerable data quality concerns for the PMTCT information with an average accuracy between the register and routine monthly report of 51% and between the routine monthly reports and DHMIS database of 84% suggesting that the primary point of departure for accurate transfer of data is during the collation process. The importance of human factors was emphasised by the observation that the average confidence level for performing RHIS-related tasks (69%) was not commensurate with the average competence levels (30%). Education was found to be associated with competence, implying that levels of education may be associated with the level at which RHIS competencies are acquired; and that three years or more of post-matriculation education is necessary. Motivation, on the other hand was not associated Stellenbosch University https://scholar.sun.ac.za iv with competence. The study observed the absence of processes such as data-quality checks and data-analysis in place in facilities. There was a general absence of a culture of information use, as a result of lack of trust in the data, and the inability of programme and facility managers to analyse, interpret and use information. We observed differences in the data accuracy by organisational authority, and multivariate analysis and qualitative information suggested that feedback may be an essential process to ensure quality. Although the PRISM framework has been developed from a multi-disciplinary evidence base, this study has been able to validate some of the internal assumptions but has also found some aspects that were not supported such as motivation and data display. Data collected from a larger number of facilities will be required to investigate this further. Institutional capacity to improve RHIS processes, ensure core competencies for RHIS-related tasks are needed, and in the longer term, measures to tackle problems associated with low pass rates in numeracy subjects among high school learners are needed. Further exploration of the possible factors that may influence data accuracy, such as supervision, training and leadership are needed as well as investigating the relationships between human and institutional agency-related aspects, in particular, how individual actions can bring about changes in institutional routines. Further study is needed to determine how decision for planning and evaluating key programmes such as PMTCT are made, and what informs such decisions if not routine data. / AFRIKAANSE OPSOMMING: In die lig van Suid Afrika se MIV/VIGS-pandemie kan ’n ingryping op gesondheidsvlak ’n belangrike rol speel om moeder-na-kind-oordrag (beter bekend as PMTCT) van MIV te voorkom. ’n Inligtingstelsel vir distriksgesondheidsbestuur – die DHMIS – was ontwerp vir die invordering van maandelikse fasiliteitsdata, wat gebruik kan word om die bestuur van openbare gesondheidsdienste en -programme te ondersteun. Die inligtingstelsel self was nog nie omvattend evalueer nie. Hierdie studie het die ervarings van twee gesondheidsdistrikte vergelyk met behulp van die PRISM- (Performance of Routine Information System) raamwerk en -instrumente. Derhalwe het hierdie studie die beskikbaarheid, gehalte en gebruik van proses- en uitsetaanwysers probeer bepaal om die PMTCT-ingrypings te monitor. ’n Vergelykende analitiese en waarnemingstudie is onderneem met behulp van ’n veelvuldige benadering. Die verskillende metodes het ’n selfopname onder gesondheidsinligtingspersoneel ingesluit om hul selfvertroue en bevoegdheid in roetinegesondheidsinligtingstelsel (RHIS)-take te evalueer. Daar was ook ’n assessering van die PMTCT-roetinedata om datagehalte, -volledigheid, -akkuraatheid en -gebruik te beoordeel.’n Fasiliteitsopname oor RHIS-prosesse en –hulpbronne was ook gedoen. Ander navorsingsmetodes het diepte-onderhoude met 22 sleutelpersone ingesluit, sowel as waarnemings in gesondheidsfasiliteite. Data is van 182 gesondheidsinligtingpersoneel van die 57 gesondheidsfasiliteite in ’n geriefsteekproef van twee gesondheidsdistrikte ingesamel. Deskriptiewe statistiek, χ2-toetsing, korrelasie en veelvoudige regressie is met behulp van STATA® weergawe 13 ontleed. ʼn Algemene induktiewe benadering is ook gevolg om die kwalitatiewe data te ontleed. Die studie toon dat menslike faktore ’n impak op datagehalte en -inligting kan hê, met ’n gemiddelde akkuraatheidsyfer van 51% van beide die register en roetine maandelikse verslae. Die akkuraatheid van die maandelikse verslae en RHIS databasis is 84%, wat aandui dat akkuraatheid slegs toegepas word indien inligting uit die staanspoor korrek aangeteken word. Die impak van menslike hulpbronafaktore was beklemtoon toe daar bevind was dat hoewel 69% van RHIS-dataverwerkers vertroue getoon het in die gebruik van RHIS-verwante take, slegs 30% wel bevoeg was om die werk te doen. Opvoeding was grootliks geassosieer met bevoegdheid, wat moontlik voorstel dat sekere vlakke van opvoeding benodig word vir spesifieke RHIS-bevoegdhede. Minsten drie jaar tersiêre opleiding word aanbebeel. Motivering was nie met Stellenbosch University https://scholar.sun.ac.za vi bevoegdheid geklassifeer nie. Die studie het bevind dat daar te min aandag aan datagehalte en –analise gegee word in fasiliteite. Oor die algemeen was daar nie ’n ordentlike kultuur van inligtinggebruik nie, a.g.v. die feit dat daar nie vertroue in die data was nie. Terselftertyd was program- en fasiliteitbestuurders nie bevoeg om inligting te analiseer en ontleed nie. Ons het verskille in die akkuraatheid van data opgetel wat deur organisasie-hoofde gedoen was. Meervoudige analise en kwalitatiewe informasie stel voor dat terugvoering ’n belangrike deel van die proses moet wees om kwaliteit te verseker. Hoewel die PRISM-raamwerk saamgestel was uit ’n multi-dissiplinêre bewyslewering, kon hierdie studie sommige van die interne voorneme valideer, maar daar was aspekte wat nie gestaaf kon word nie. Inligting van ’n groter aantal fasiliteite sal benodig word om verder hierna ondersoek in te stel. Institusionele kapasiteit word benodig om RHIS-prosesses te verbeter en basiese vaardighede vir RHIS-verwante take te verseker. Op langtermynvlak moet daar ook gekyk word na probleme wat lei tot laë slaagsyfers in syfervaardighede in hoërskoolleerders. Verdere ondersoek moet ingestel word om vas te stel watter faktore moontlik akkurate data teweeg kan bring. Dit sluit toesig, opleiding en leierskap, asook die verhoudings tussen menslike en agentskap-verwante aspekte in. Die feit dat optrede op individuele vlak veranderings in institusionele roetines kan aanbring, moet spesifiek na gekyk word. Verdere studies kan help om vas te stel hoe besluite vir beplanning en evaluaring vir hoofprogramme soos PMTCT gemaak word – asook hoe die besluite gemaak word indien hulle nie roetine voorafgaan nie.

The fossiliferous sediments of the Hueso and Tapiado formations exposed in the Vallecito Creek-Fish Creek section (VCFC) of the Anza-Borrego Desert (ABD) span the boundary between Blancan and Irvingtonian North American Land Mammal 'Ages.' Historically, the determination of the Blancan-Irvingtonian (B-I) boundary in the VCFC proved problematic. A new study to determine the B-I boundary in the VCFC revealed data inaccuracies in both published works and curatorial records of the mammal and other vertebrate fossils from the ABD. Most individual inaccuracies were minor, but an accumulation of over 50 years of multiple inaccuracies had significant effects on local and regional biostratigraphic and biochronologic correlations. A detailed investigation of the inaccurate data resulted in recognition of 17 types of error, with at least five sources. The two most prominent sources of error are those derived from publication and curation protocols. Examination of over 150 publications and in-house faunal lists produced over 830 taxonomic names and format variants, for 110 mammalian taxa identified from the ABD and VCFC. Approximately 50% of the taxonomic identifications were previously published without voucher catalogue numbers or fossil descriptions. A critical review of the taxonomic assignments resulted in an updated faunal list of ABD terrestrial Mammalia, including 110 taxonomic names, 66 unqualified genera, and 46 unqualified species. A supplemental list of 'retired' taxa includes 178 previously published or listed taxonomic names and format variants. The 4.5 km sequence of originally superposed sediments within VCFC is now tilted at 23 degrees and exposed in plan view. This exposure was captured in both aerial photos and satellite imagery. GIS layers of the Geomagnetic Polarity Time Scale and individual collecting localities are readily superposed onto the stratigraphic images. When united with the faunal database, GIS maps of biostratigraphic data facilitate detection and correction of data errors. Resulting corrected maps show highest and lowest stratigraphic occurrences of taxa, as well as geographic clustering of taxonomic groups, outlining possible paleohabitats. As a result of data improvement and GIS display, the local B-I boundary currently is best estimated by the presence of Ovibovini at >1.42 Ma. / text