Introdução: O estreptococo do grupo B (EGB) tem sido motivo de preocupação para obstetras e neonatologistas pela possibilidade de sepse neonatal, que apresenta um risco importante de mortalidade. O rastreamento do EGB entre 35-37 semanas associado com a profilaxia intra-parto com penicilina cristalina tem propiciado uma diminuição de sepse neonatal precoce por EGB. O índice de colonização por EGB é muito variável de acordo com região e condição social das gestantes. A freqüência de clones virulentos é desconhecida em nosso meio. Objetivos: 1 - Determinar a prevalência de EGB no trato genital-anal em gestantes que internam no Centro Obstétrico de um Hospital Universitário na Cidade de Porto Alegre, verificando também os fatores associados a esta colonização. 2- Verificar a presença de cepas invasivas do EGB nos casos positivos Métodos: O estudo envolveu 319 gestantes, com idade gestacional que variou entre24 e 41 semanas, que internaram na Maternidade do Hospital São Lucas da Pontifícia Universidade Católica do Rio Grande do Sul. Nestas pacientes foi feito um swab vaginal-anal e semeado em meio de cultura Todd-Hewitt. As amostras positivas para EGB foram enviadas para realização de reação em cadeia da polimerase (PCR) para extração do DNA e tipagem de cepas altamente virulentas do EGB. Resultados: A prevalência de EGB encontrada foi de 23,4% (IC 95%: 17,8- 27,2). Em relação aos possíveis fatores associados a colonização a variável idade mostrou uma tendência de maior colonização em pacientes com mais de 30 anos, quando comparada com idades mais jovens. Isso também aconteceu em relação a comparação com a cor da paciente onde apareceu uma tendência de maior colonização em pacientes de cor preta ou mista. As pacientes com idade gestacional abaixo de 35 semanas apresentaram um índice de colonização inferior às gestantes entre 35-37 semanas e também entre 37-39 semanas. A presença de ruptura prematura das membranas ovulares não apareceu como fator associado a índice maior de colonização pelo EGB. Em relação a freqüência de cepas virulentas a análise de 60 casos revelou 10 casos positivos (16,6%) para clone "sequence typing" - 17 (ST-17). Conclusões: A prevalência de EGB encontrado em nosso estudo é similar aos índices de países desenvolvidos, que realizam o rastreamento rotineiro no prénatal e fazem a intervenção com profilaxia intra-parto nos casos positivos. Houve um maior índice de colonização em gestações à termo comparadas com gestações de pré-termo. A freqüência de cepas "altamente virulentas" encontrada foi 16,6%. / Background: Group B Streptococcus (GBS) has been of concern to obstetrician and neonatologists to the possibility of neonatal sepsis, which presents a significant risk of mortality. Screening for GBS between 35-37 weeks associated with intrapartum prophylaxis with crystalline penicillin has resulted in reduction of neonatal sepsis by GBS. The rate of GBS colonization varies widely according to region and social status of women. The frequency of virulent clones is unknown in our contry. Objectives: 1. To determine the prevalence of GBS in the genital-anal tract of pregnant patients that are hospitalized at the obstetric center of a University Hospital of Porto Alegre, also check the factors associated with this colonization. 2. Verify the presence of invasive strains of GBS positive cases Methods: The study involved 319 pregnant women with gestational ages ranging from 24 and 41 weeks, who were hospitalized at the São Lucas Hospital of Catholic University of Rio Grande do Sul. These patients was done a vaginalanal swab and seeded un culture medium Todd-Hewitt. Samples positive for GBS were sent to carry out polymerase chain reaction (PCR) for DNA extraction and typing of higly virulent strains of GBS. Results: GBS prevalence found was 23.4% (CI 95%: 17.8-27.2). Regarding the possible factors associated with colonization of the age variable showed a trend of increased colonization in patients over 30 years compared with younger ages. This also happened for comparison with the color of the patient where it appeared a tend of increased colonization in patients of gestacional age below 35 weeks among pregnant women at 35-37 weeks and between 37-39 weeks. The presence of premature rupture of ovular membranes did not appear as a factor associated with higer rate of GBS colonization. Regarding the frequency of virulent strains analysis of 60 cases revealed 10 positive cases (16.6%) to clone sequence typing -17 (ST-17). Conclusions: GBS prevalence found in our study is similar to rates in developed countries, carrying out routine screening during prenatal care and make a intervention with intrapartum prophylaxis in positive cases. There was a higer rate of colonization in term pregnacies compared with preterm. The frequency of strains "highly virulent" found was 16.6%.

Estudar os efeitos da estimulação tátil e cinestésica materna durante a hospitalização de recém-nascidos de muito baixo peso. Métodos: Trata-se de um ensaio clínico randomizado e cego onde foram incluídos recémnascidos pré-termo com peso de nascimento entre 750 g e 1500 g e idade gestacional ≤ 32 semanas. Os critérios de exclusão foram: óbito antes de completar 48 horas de vida, presença de malformações maiores e recusa dos pais em participar do estudo. Os recém-nascidos foram divididos em dois grupos: o grupo intervenção (GI), que recebeu o cuidado padrão e a estimulação tátil e cinestésica, até quatro sessões por dia de 15 minutos de duração cada uma, e o grupo controle (GC), que recebeu exclusivamente o cuidado padrão.Participaram do estudo 104 recém-nascidos, 52 em cada grupo. Os grupos foram similares quanto à idade gestacional (GI: 29,7 1,6 sem.; GC:29,4 1,6 sem.), ao peso de nascimento (GI:1186 194 g; GC:1156 198 g), ao sexo, à incidência de PIG, à mediana do SNAPPE-II e ao número de óbitos. Os ganhos de peso, de comprimento e de perímetro cefálico foram semelhantes nos grupos. A ocorrência de sepse tardia foi significativamente superior no grupo controle (GI: 10,8%, n=5; GC:38,3%, n=18, p=0.005). O grupo intervenção recebeu alta hospitalar 7 dias antes do grupo controle (GI: 42 dias; IC 95%: 38-46 dias; GC: 49 dias; IC 95%: 42-56 dias). Pela regressão de Cox, o grupo intervenção teve a Hazard Rate de 1,85 (IC95%:1,09-3,13; p=0,023) para alta precoce. Conclusão: A estimulação tátil e cinestésica realizada pela mãe diminui o tempo de hospitalização e a freqüência de sepse neonatal tardia em recém-nascidos de muito baixo peso. / Aim: To study the effect of maternal tactile-kinesthetic stimulation on hospital stay and incidence of sepsis in very low birth weight infants. Methods: A masked randomized study was performed including infants of birth weight ≥ 750 and ≤ 1500 grams and gestational age ≤ 32 weeks were included. Exclusion criteria were: death before completing 48 hours after birth, presence of major malformations and parents’ refusing to participate in the study. Neonates were divided into: intervention group (IG) (standard care plus maternal tactile-kinesthetic stimulation, up to four daily sessions of 15 minutes each) and control group (CG). Results: In total, 104 very low birth weight infants were included, 52 in each group. Both groups were similar in gestational age (IG: 29.7 1.6; CG: 29.4 1.6 weeks), birth weight (IG: 1186 194; CG: 1156 198 grams), gender, number of small for gestational age infants, SNAPPE-II median score, and number of deaths. Gains in weight, length and head circumference during hospital stay were similar when comparing both groups. The incidence of late-onset sepsis was significantly lower in the intervention group (IG: 10.8%, n=5; CG: 38.3%, n=18, p=0.005). The intervention group was discharged from hospital 7 days before the control group (GI: 42; CI 95%: 38 – 46; CG: 49; CI 95%: 42 – 56). Using Cox regression model, the intervention group presented 1.85 HR (CI 95%:1.09-3.13; p=0.023) for early discharge. Conclusions: Maternal tactile-kinesthetic stimulation in very low birth weight infants decreases the length of hospital stay and the incidence of late-onset neonatal sepsis.

Introdução: O estreptococo do grupo B (EGB) tem sido motivo de preocupação para obstetras e neonatologistas pela possibilidade de sepse neonatal, que apresenta um risco importante de mortalidade. O rastreamento do EGB entre 35-37 semanas associado com a profilaxia intra-parto com penicilina cristalina tem propiciado uma diminuição de sepse neonatal precoce por EGB. O índice de colonização por EGB é muito variável de acordo com região e condição social das gestantes. A freqüência de clones virulentos é desconhecida em nosso meio. Objetivos: 1 - Determinar a prevalência de EGB no trato genital-anal em gestantes que internam no Centro Obstétrico de um Hospital Universitário na Cidade de Porto Alegre, verificando também os fatores associados a esta colonização. 2- Verificar a presença de cepas invasivas do EGB nos casos positivos Métodos: O estudo envolveu 319 gestantes, com idade gestacional que variou entre24 e 41 semanas, que internaram na Maternidade do Hospital São Lucas da Pontifícia Universidade Católica do Rio Grande do Sul. Nestas pacientes foi feito um swab vaginal-anal e semeado em meio de cultura Todd-Hewitt. As amostras positivas para EGB foram enviadas para realização de reação em cadeia da polimerase (PCR) para extração do DNA e tipagem de cepas altamente virulentas do EGB. Resultados: A prevalência de EGB encontrada foi de 23,4% (IC 95%: 17,8- 27,2). Em relação aos possíveis fatores associados a colonização a variável idade mostrou uma tendência de maior colonização em pacientes com mais de 30 anos, quando comparada com idades mais jovens. Isso também aconteceu em relação a comparação com a cor da paciente onde apareceu uma tendência de maior colonização em pacientes de cor preta ou mista. As pacientes com idade gestacional abaixo de 35 semanas apresentaram um índice de colonização inferior às gestantes entre 35-37 semanas e também entre 37-39 semanas. A presença de ruptura prematura das membranas ovulares não apareceu como fator associado a índice maior de colonização pelo EGB. Em relação a freqüência de cepas virulentas a análise de 60 casos revelou 10 casos positivos (16,6%) para clone "sequence typing" - 17 (ST-17). Conclusões: A prevalência de EGB encontrado em nosso estudo é similar aos índices de países desenvolvidos, que realizam o rastreamento rotineiro no prénatal e fazem a intervenção com profilaxia intra-parto nos casos positivos. Houve um maior índice de colonização em gestações à termo comparadas com gestações de pré-termo. A freqüência de cepas "altamente virulentas" encontrada foi 16,6%. / Background: Group B Streptococcus (GBS) has been of concern to obstetrician and neonatologists to the possibility of neonatal sepsis, which presents a significant risk of mortality. Screening for GBS between 35-37 weeks associated with intrapartum prophylaxis with crystalline penicillin has resulted in reduction of neonatal sepsis by GBS. The rate of GBS colonization varies widely according to region and social status of women. The frequency of virulent clones is unknown in our contry. Objectives: 1. To determine the prevalence of GBS in the genital-anal tract of pregnant patients that are hospitalized at the obstetric center of a University Hospital of Porto Alegre, also check the factors associated with this colonization. 2. Verify the presence of invasive strains of GBS positive cases Methods: The study involved 319 pregnant women with gestational ages ranging from 24 and 41 weeks, who were hospitalized at the São Lucas Hospital of Catholic University of Rio Grande do Sul. These patients was done a vaginalanal swab and seeded un culture medium Todd-Hewitt. Samples positive for GBS were sent to carry out polymerase chain reaction (PCR) for DNA extraction and typing of higly virulent strains of GBS. Results: GBS prevalence found was 23.4% (CI 95%: 17.8-27.2). Regarding the possible factors associated with colonization of the age variable showed a trend of increased colonization in patients over 30 years compared with younger ages. This also happened for comparison with the color of the patient where it appeared a tend of increased colonization in patients of gestacional age below 35 weeks among pregnant women at 35-37 weeks and between 37-39 weeks. The presence of premature rupture of ovular membranes did not appear as a factor associated with higer rate of GBS colonization. Regarding the frequency of virulent strains analysis of 60 cases revealed 10 positive cases (16.6%) to clone sequence typing -17 (ST-17). Conclusions: GBS prevalence found in our study is similar to rates in developed countries, carrying out routine screening during prenatal care and make a intervention with intrapartum prophylaxis in positive cases. There was a higer rate of colonization in term pregnacies compared with preterm. The frequency of strains "highly virulent" found was 16.6%.

Orientador: Sergio Tadeu Martins Marba / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Ciências Médicas / Made available in DSpace on 2018-08-23T19:06:43Z (GMT). No. of bitstreams: 1 Pavan_TychaBiancaSabaini_M.pdf: 1135745 bytes, checksum: 87193a679a977d0df4230db53eab4297 (MD5) Previous issue date: 2013 / Resumo: O objetivo do trabalho foi descrever e avaliar a capacidade da "RFLP-PCR" para detectar patógenos bacterianos em recém-nascidos (RN) e determinar seu uso diagnóstico na sepse neonatal precoce (SNP). Foram avaliados RNs assintomáticos, filhos de gestantes com fatores de risco para infecção ovular e RNs com apresentação de sintomas clínicos sugestivos para SNP, nas primeiras 48 horas de vida, atendidos na unidade neonatal do CAISM/UNICAMP/SP. Realizou-se extração genética das amostras de 200?L de sangue total através do kit QIAamp DNA Mini kit column (Qiagen), seguida da detecção do gene universal bacteriano 16 S rRNA e identificação microbiana pelas sucessivas digestões com enzimas de restrições - HaeIII, AluI, DdeI, MnlI. Foram avaliados 65 RN assintomáticos e 178 RN com apresentação de sinais clínicos sugestivo de SNP. Os patógenos detectados foram K. pneumoniae, S. pneumoniae, L. monocytogenes, P. aeruginosa, E.coli, S.pyognes, S. agalactiae, S. epidermidis e S. aureus. Resultados duvidosos ou bactérias não identificadas ocorram em 56 amostras. A positividade da "RFLP-PCR" foi de 55,3% no grupo assintomático e 51,6% no grupo sintomático, sem diferença estatística (p=0,583). Houve comprovação de sete RN com sepse por cultura e 24 com sepse clínica e foram encontrados valores de sensibilidade de 87,8%, especificidade de 46,6%, valores preditivo positivo de 23,3% e preditivo negativo de 95,4%. Em conclusão, o uso da técnica mostrou uma taxa de acurácia baixa para o diagnóstico para a SNP satisfatória / Abstract: Study objective was to describe and evaluate the ability of RFLP-PCR for detection of bacterial pathogens in newborn and to determine its diagnostic utility in early neonatal sepsis. There were evaluated asymptomatic newborns, children of mothers with risk factors for infection ovulate and newborns with presentation of clinical symptoms suggestive of early neonatal sepsis, occurring before 48 hours of life, evaluated at neonatal unit at CAISM-UNICAMP. Whole blood samples were taken from newborns - 200 ?L each. A genetic extraction were performed using QIAamp DNA Mini kit column (Qiagen) and also the detection of bacterial universal 16 S rRNA gene, subsequent microbial identification using restriction digestion enzymes - HaeIII, AluI, DdeI, MnlI. There were 65 asymptomatic infants and 178 symptomatic newborns. The identified bacteria were K. pneumoniae, S. pneumoniae, L. monocytogenes, P. aeruginosa, E.coli, S.pyognes, S. agalactiae, S. epidermidis and S. aureus. Uncertain or unidentified pathogens occurred in 56 samples. RFLP-PCR positivity was 55,3% in the asymptomatic group and 51,6% in the symptomatic group, with no statistical difference (p=0,583). There were culture comproved sepsis in 7 infants and in 24 newborns were made a diagnostic of clinical sepsis. Diagnostic indexes were: sensibility 87.8%, specificity 46.6%, positive predictive value 23.3% e negative predictive value 95.4%. In conclusion, RFLP-PCR had an unsatisfactory accuracy index early neonatal sepsis / Mestrado / Saude da Criança e do Adolescente / Mestra em Ciências

The vital signs monitoring data of an infant receiving intensive care are a rich source of information about its health condition. One major concern about the state of health of such patients is the onset of neonatal sepsis, a life-threatening bloodstream infection. As early signs are subtle and current diagnosis procedures involve slow laboratory testing, sepsis detection based on the monitored physiological dynamics is a clinically significant task. This challenging problem can be thoroughly modelled as real-time inference within a machine learning framework. In this thesis, we develop probabilistic dynamical models centred around the goal of providing useful predictions about the onset of neonatal sepsis. This research is characterised by the careful incorporation of domain knowledge for the purpose of extracting the infant’s true physiology from the monitoring data. We make two main contributions. The first one is the formulation of sepsis detection as learning and inference in an Auto-Regressive Hidden Markov Model (AR-HMM). The model investigates the extent to which physiological events observed in the patient’s monitoring traces could be used for the early detection of neonatal sepsis. In addition, the proposed approach involves exact marginalisation over missing data at inference time. When applying the ARHMM on a real-world dataset, we found that it can produce effective predictions about the onset of sepsis. Second, both sepsis and clinical event detection are formulated as learning and inference in a Hierarchical Switching Linear Dynamical System (HSLDS). The HSLDS models dynamical systems where complex interactions between modes of operation can be represented as a twolevel hidden discrete hierarchical structure. For neonatal condition monitoring, the lower layer models clinical events and is controlled by upper layer variables with semantics sepsis/nonsepsis. The model parameterisation and estimation procedures are adapted to the specifics of physiological monitoring data. We demonstrate that the performance of the HSLDS for the detection of sepsis is not statistically different from the AR-HMM, despite the fact that the latter model is given “ground truth” annotations of the patient’s physiology.

>Magister Scientiae - MSc / The aim of this study was to compare the trough plasma levels of Aspen-Vancomycin® (AV); and Sandoz-Vancocin CP® (SV) in premature infants with suspected Methicillin Resistant Staphylococcus aureus (MRSA) infection. The study was designed as a prospective, double blind, randomised trial involving male and female premature infants admitted in the Neonatal Intensive care Unit (NICU) at Netcare Blaauwberg and N1-city Hospitals for treatment of suspected MRSA-infection between April 2012 and June 2013. The inclusion criteria were: 29-35 weeks postmenstrual age (PMA), informed and written consent from parents of each premature infant enrolled in the study. Blood samples (0.3-0.4ml) were collected for renal function test and vancomycin trough levels determination. Blood samples for vancomycin trough level assay were collected thirty minutes prior to the administration of the third dose of vancomycin. Statistical analysis was performed and estimation was made giving an indication of how many infants will be needed to make the study statistically significant. Wilcoxon Two-Sample test was performed to determine the p-values and Spearman correlation coefficients were used to determine the correlation between trough levels and variables. P-values < 0.05 were considered significant. A total of 19 premature infants met with study criteria, 10 (5 females and 5 males) received AV and 9 (6 females and 3 males) receive d SV. There was no statistical significant difference between the demographic (GA, BW, PMA, PNA, weight at trial entry, height at trial entry) and biological (albumin, serum creatinine concentration and glomerular filtration rate) parameters of the premature infants in the AV and SV group. There were no statistical significant difference between trough level 1 of AV and SV, although trough level 1 had a lower trend in the SV group (p=0.118). No AV trough level 1 was below the minimum effective concentration (<5μg/ml). It was found that 30% of AV trough level 1 was within the therapeutic range (5-10μg/ml) and 70% of AV trough level 1, were above minimum toxic concentration (>10mg/l). It was found that 22.2% of SV trough level 1 was below minimum effective concentration, 44.4% of SV trough level 1 was within therapeutic range and 33.3% of trough level 1 was above minimum toxic concentration. No correlation was found between trough level 1 and the demographic and biological parameters of the premature infants in the AV group. SV had a positive correlation with GA, BBW, PMA and a negative correlation with PNA

Introduction : La variabilité du rythme cardiaque est étudiée à partir des variations de durée des cycles cardiaques (intervalle R-R de l’électrocardiogramme). Ces variations peuvent être analysées par des méthodes linéaires (temporelles et fréquentielles) et non linéaires (théorie de l’information ou des fractales) de quantifications mathématiques et statistiques qui donnent des informations innovantes sur les signaux analysés. L’application de ces méthodes d’étude en néonatologie a démontré un intérêt pour le diagnostique précoce de l’infection néonatale tardive du prématuré mais n’avait pas été étudié dans l’infection néonatale précoce du nouveau-né à terme, dans le contexte des évènements cardio-respiratoires suivant la primo-vaccination des prématurés ou pour évaluer un effet neurologique de l’hyperbilirubinémie dans l’ictère néonatal. Notre hypothèse dans ce travail était qu’il était possible de : (i) caractériser la variabilité du rythme cardiaque en cas d’infection materno fœtale ou de méningite néonatale, (ii) mettre en évidence des facteurs prédisposant à la survenue d’évènements cardio-respiratoires post-vaccinaux, (iii) Identifier un éventuel retentissement neurologique de l’ictère néonatal par étude de la variabilité du rythme cardiaque. / The heart rate variability measures permitted to evaluate equilibrium state and perturbation in the regulation of cardio-vascular system.  These tools, based on heart rate variability analysis, helped to recognize associated disease state as early onset neonatal sepsis and non-infectious inflammatory response induced to immunization. An increase in global variability (SD), long term variability (SD, LF) and low approximated entropy (ApEn) were observed in the proven-sepsis full term infants. Importance of decrease in ApEn was correlated to the severity of sepsis assessed by blood markers. These suggest an association of sepsis with uncoordinated sympatho-vagal coactivation together with loss of adaptability. In premature infants, the risk of increase in cardio-respiratory events after the first immunization was associated with a specific pre-immunization profile: sympathetic predominance in heart rate control (high LF/HF ratio), abnormal oversimplification of heart rate variability and persistence rhythm control immaturity. Increased ApEn after immunization reflects a marginal result from adaptability of the heart rate to environmental changes without possibility to reserve in case of severe infection.

Neonatal sepsis is a potentially fatal medical conditiondue to an infection and is attributed to about 200 000annual deaths globally. With healthcare systems that are facingconstant challenges, there exists a potential for introducingmachine learning models as a diagnostic tool that can beautomatized within existing workflows and would not entail morework for healthcare personnel. The Herlenius Research Teamat Karolinska Institutet has collected neonatal sepsis data thathas been used for the development of many machine learningmodels across several papers. However, none have tried to studydecision tree ensemble methods. In this paper, random forestand XGBoost models are developed and evaluated in order toassess their feasibility for clinical practice. The data contained24 features of vital parameters that are easily collected througha patient monitoring system. The validation and evaluationprocedure needed special consideration due to the data beinggrouped based on patient level and being imbalanced. Theproposed methods developed in this paper have the potentialto be generalized to other similar applications. Finally, usingthe measure receiver-operating-characteristic area-under-curve(ROC AUC), both models achieved around ROC AUC= 0.84.Such results suggest that the random forest and XGBoost modelsare potentially feasible for clinical practice. Another gainedinsight was that both models seemed to perform better withsimpler models, suggesting that future work could create a moreexplainable model. / Nenatal sepsis är ett potentiellt dödligt‌‌‌ medicinskt tillstånd till följd av en infektion och uppges globalt orsaka 200 000 dödsfall årligen. Med sjukvårdssystem som konstant utsätts för utmaningar existerar det en potential för maskininlärningsmodeller som diagnostiska verktyg automatiserade inom existerande arbetsflöden utan att innebära mer arbete för sjukvårdsanställda. Herelenius forskarteam på Karolinska Institet har samlat ihop neonatal sepsis data som har använts för att utveckla många maskininlärningsmodeller över flera studier. Emellertid har ingen prövat att undersöka beslutsträds ensemble metoder. Syftet med denna studie är att utveckla och utvärdera random forest och XGBoost modeller för att bedöma deras möjligheter i klinisk praxis. Datan innehör 24 attribut av vitalparameterar som enkelt samlas in genom patientövervakningssystem. Förfarandet för validering och utvärdering krävde särskild hänsyn med tanke på att datan var grupperad på patientnivå och var obalanserad. Den föreslagna metoden har potential att generaliseras till andra liknande tillämpningar. Slutligen, genom att använda receiveroperating-characteristic area-under-curve (ROC AUC) måttet kunde vi uppvisa att båda modellerna presterade med ett resultat på ROC AUC= 0.84. Sådana resultat föreslår att både random forest och XGBoost modellerna kan potentiellt användas i klinisk praxis. En annan insikt var att båda modellerna verkade prestera bättre med enklare modeller vilket föreslår att ete skulle kunna vara att skapa en mer förklarlig skininlärningsmodell. / Kandidatexjobb i elektroteknik 2022, KTH, Stockholm

Neonatal sepsis is associated with most cases ofmortality in the neonatal intensive care unit. Major challengesin detecting sepsis using suitable biomarkers has lead people tolook for alternative approaches in the form of Machine Learningtechniques. In this project, Random Forest classification wasperformed on a sepsis data set provided by Karolinska Hospital.We particularly focused on tackling class imbalance in the datausing sampling and cost-sensitive techniques. We compare theclassification performances of Random Forests in six differentsetups; four using oversampling and undersampling techniques;one using cost-sensitive learning and one basic Random Forest.The performance with the oversampling techniques were betterand could identify more sepsis patients than the other setups.The overall performances were also good, making the methodspotentially useful in practice. / Neonatal sepsis är orsaken till majoriteten av mortaliteten i neonatal intensivvården. Svårigheten i att detektera sepsis med hjälp av biomarkörer har lett många att leta efter alternativa metoder. Maskininlärningstekniker är en sådan alternativ metod som har i senaste tider ökat i användning inom vård och andra sektorer. I detta project användes Random Forest klassifikations algoritmen på en sepsis datamängd given av Karolinska Sjukhuset. Vi fokuserade på att hantera klassimbalansen i datan genom att använda olika provtagningsoch kostnadskänsliga metoder. Vi jämförde klassificeringsprestanda för Random Forest med sex olika inställningar; fyra av de använde provtagingsmetoderna; en av de använde en kostnadskänslig metod och en var en vanlig Random Forest. Det visade sig att modellens prestanda ökade som mest med översamplings metoderna. Den generella klassificeringsprestandan var också bra, vilket gör Random Forests tillsammans med ingsmetoderna potentiellt användbar i praktiken. / Kandidatexjobb i elektroteknik 2022, KTH, Stockholm

Background: Neonatal sepsis is one of the most important causes of elevated morbidity and mortality rates in neonatal intensive care units worldwide. While the clinical manifestations of neonatal sepsis tend to be nonspecific, its rapid development and life-threatening potential call for reliable markers for early detection. Methods: We conducted a retrospective single-center study including all neonates suspected of having developed neonatal sepsis from 2013 to 2016. Perinatal and clinical characteristics as well as microbiological and laboratory findings were evaluated. Neonatal sepsis was defined as either culture-proven sepsis (positive blood culture) or clinical sepsis (at least one symptom and elevated C-reactive protein (CRP) concentrations within 72 h with negative blood culture). We further differentiated between early-onset (EOS) and late-onset (LOS) sepsis. Results: Microbiological colonization screening by throat and rectal swabs frequently did not detect the organism that subsequently caused the sepsis. Depending on the age of the newborn with sepsis (EOS or LOS), associations between different anamnestic and clinical factors (prenatal or postnatal ones) were found. In particular, the central–peripheral temperature difference showed a strong association with LOS. Laboratory results useful for the early detection of neonatal sepsis included interleukin-6 (IL-6) and CRP concentrations. Conclusions: Elevated IL-6 >100 ng/L was a strong marker for neonatal sepsis. When choosing the antibiotics for treatment, data from microbiological colonization screening should be considered but not solely relied on. Some indicators of infection also depended on postnatal age.