Graphs are commonly selected as a model of scientific information: graphs can successfully represent imprecise, uncertain, noisy data; and graph theory has a well-developed mathematical apparatus forming a solid and sound foundation for graph research. Design and experimental confirmation of new, scalable, and practical analytics for massive graphs have been actively researched for decades. Our work concentrates on developing new accurate and efficient algorithms that calculate the most influential nodes and communities in an arbitrary graph. Our algorithms for graph decomposition into families of most influential communities compute influential communities faster and using smaller memory footprint than existing algorithms for the problem. Our algorithms solving the problem of influence maximization in large graphs use much smaller memory than the existing state-of-the-art algorithms while providing solutions with equal accuracy. Our main contribution is designing data structures and algorithms that drastically cut the memory footprint and scale up the computation of influential communities and nodes to massive modern graphs. The algorithms and their implementations can efficiently handle networks of billions of edges using a single consumer-grade machine. These claims are supported by extensive experiments on large real-world graphs of different types. / Graduate

abstract: The present study applied latent class analysis to a family-centered prevention trial in early childhood to identify subgroups of families with differential responsiveness to the Family Check-up (FCU) intervention. The sample included 731 families of 2-year- olds randomized to the FCU or control and followed through age five with yearly follow up assessments (Dishion et al., 2014; Shaw et al., 2015). A two-step mixture model was used to examine whether specific constellations of family characteristics at age 2 (baseline) were related to intervention response at age 3, 4, and 5. The first step empirically identified latent classes of families based on a variety of demographic and adjustment variables selected on the basis of previous research on predictors of response to the FCU and parent training in general, as well as on the clinical observations of FCU implementers. The second step modeled the effect of the FCU on longitudinal change in children's problem behavior in each of the empirically derived latent classes. Results suggested a five-class solution, where a significant intervention effect of moderate-to- large size was observed in one of the five classes. The families within the responsive class were characterized by child neglect, legal problems, and mental health issues. Pairwise comparisons revealed that the intervention effect was significantly greater in this class of families than in two other classes that were generally less at risk for the development of disruptive behavior problems, and post hoc analyses partially supported these results. Thus, results indicated that the FCU was most successful in reducing child problem behavior in the highly distressed group of families. We conclude by discussing the potential practical utility of these results and emphasizing the need for future research to evaluate this approach's predictive accuracy. / Dissertation/Thesis / Masters Thesis Psychology 2016

Orientador: Aarão Mendes Pinto-Neto / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Ciencias Medicas / Made available in DSpace on 2018-08-08T02:23:48Z (GMT). No. of bitstreams: 1 Costa_AurelioAntonioRibeiroda_D.pdf: 582824 bytes, checksum: b130d3e6304e9d014c75b64794caae29 (MD5) Previous issue date: 2006 / Resumo: a histeroscopia é um dos procedimentos mais realizados em ginecologia. Muitos ensaios controlados sugerem vantagens no uso prévio de misoprostol para diminuir a resistência da cérvix uterina; entretanto este tipo de estudo ainda é escasso quando se compara o uso do misoprostol prévio à histeroscopia diagnóstica em mulheres na pós-menopausa. avaliar os efeitos intra e pós-operatórios em mulheres na pós-menopausa, submetidas à histeroscopia diagnóstica, sem anestesia, com uso prévio de misoprostol para amadurecimento do colo uterino. Sujeitos e métodos: foi realizado um estudo tipo ensaio clínico, duplamente mascarado, randomizado, em pós-menopausadas que se submeteram à histeroscopia diagnóstica sem anestesia no IMIP e no hospital Barão de Lucena, em Recife, Pernambuco. Foram incluídas 120 pacientes alocadas aleatoriamente em dois grupos, sendo 60 com uso de 200µg de misoprostol via vaginal e 60 com uso de 200µg de placebo. Foram estudadas as seguintes variáveis: tempo do procedimento, freqüência e intensidade da dor durante o exame, necessidade de dilatação cervical adicional, efeitos colaterais (sangramento genital, náuseas, vômitos, diarréia, hipertermia) e complicações (perfuração uterina, falso pertuito, laceração cervical, infecções, dor no pós-operatório imediato). Para análise estatística, utilizaram-se os testes de qui-quadrado de associação, testes exato de Fisher e Mann-Whitney para comparação dos grupos, considerando-se significativo um erro alfa menor que 5%. Resultados: os grupos foram semelhantes em relação à média de idade (61,3 e 59,2, p=0,09), paridade (04 e 04, p=0,88), IMC (27 e 27,5, p=0,55) e tempo de menopausa (12,5 versus 9,6, p=0,52), intervalo entre medicação e exame (8,4 e 8,2, p=0,4), meio distensor com CO2 (39% versus 32%, p=0,44) e indicações por mioma, espessamento ou sangramento genital (p=0,52; 0,37; 0,62, respectivamente). O grupo do misoprostol apresentou menor intensidade da dor durante o procedimento, tanto na comparação das medianas (05 versus 07, p=0,02) quanto nos escores acima de cinco (44,6% versus 66,7%, p=0,01). Não houve diferenças entre os grupos com relação à duração do procedimento (2,4 versus 2,0 min, p=0,3), presença de dor ao exame ou na biópsia (p=0,74 e p=0,19, respectivamente). Não ocorreu diferença na necessidade de dilatação (17,2 versus 20,3, p=0,66) nem na presença de efeitos colaterais ou complicações. Também não se observaram diferenças significativas na presença e intensidade da dor após o procedimento. Conclusão: o uso prévio de misoprostol parece diminuir a intensidade da dor durante a histeroscopia diagnóstica em pós-menopausadas; entretanto não demonstrou diferenças significativas em relação à diminuição do tempo para o procedimento e nem da necessidade de dilatação cervical entre os grupos, sugerindo não utilizá-lo rotineiramente e sim em alguns casos selecionados / Abstract: hysteroscopy is one of the most useful procedure to study uterine cavity and widely used in gynecologic clinic. Although many controlled clinical trials have shown advantages in the use of misoprostol in realeasing cérvix uterine resistanse, there are few evidence regarding its use before office hysteroscopy in postmenopausal women. Objective: to compare the trans and postoperative results of postmenopausal women underwent office hysteroscopy without anesthesia that used misoprostol or placebo previouslly to the procedure to mature the uterine cérvix in two teaching hospitals in Recife. Patients and methods: a randomized, double masked, clinical trial was conducted enrolling 120 postmenopausal women who had been submitted to office hysteroscopy without anesthesia at IMIP and Barão de Lucena teaching hospital. Among those patients 60 were randomized allocated in the study group using 200µg of vaginal misoprostol and 60 took part in the placebo group using 200µg of a vaginal placebo.The following variable were studied: duration of the procedure, frequency and intensity of pain during the exame, necessity of additional dilatation, side effect as (genital bleeding, nauseas, vomits, diarrhea, and hyperthermia) and complications (uterine perforation, false way, cervical laceration, infections and pain in the imediate postoperatory period). For analysis statistics we used the tests of association qui-square, accurate test of Fisher and Mann-Whitney for comparison of the groups, considering itself significant with lesser alpha error that 5%. Results: We found no significant statistic difference between the groups regarding age (61,3 e 59,2, p=0,09), parity (04 e 04, p=0,88), IMC (27 e 27,5, p=0,55) ,time since menopause (12,5 versus 9,6, p=0,52), time between medication and hysteroscopy (8,4 e 8,2, p=0,4), distensor fluid with CO2 (39% versus 32%, p=0,44) and the indication of the exam (myoma, endometrial thickness and postmenopausal bleeding (p=0,52; 0,37; 0,62, respectively).There was no differences between the groups with regard to duration of the procedure, presence of pain to the examination or hystopathologic findings. No differences in the necessity of extra cervical dilatation nor in the presence of side effects or complications was observed. We also didn't find any statistic difference regarding the presence and intensity of pain after the procedure. The misoprostol group had less pain during hysteroscopy than placebo group (p= 0,02). Conclusions: Although the previous use of misoprostol seems to decrease the intensity of pain during the hysteroscopy in postmenopausal women when compared with those who had used placebo, there were no significant difference in duration of procedure and necessity of extra cervical dilataion. Therefore we do not suggest its routinely use, but only in selected cases / Doutorado / Doutor em Tocoginecologia

Background: A pilot study can be an important step in the assessment of an intervention by providing information to design the future definitive trial. Pilot studies can be used to estimate the recruitment and retention rates and population variance and to provide preliminary evidence of efficacy potential. However, estimation is poor because pilot studies are small, so sensitivity analyses for the main trial's sample size calculations should be undertaken. Methods: We demonstrate how to carry out easy-to-perform sensitivity analysis for designing trials based on pilot data using an example. Furthermore, we introduce rules of thumb for the size of the pilot study so that the overall sample size, for both pilot and main trials, is minimized. Results: The example illustrates how sample size estimates for the main trial can alter dramatically by plausibly varying assumptions. Required sample size for 90% power varied from 392 to 692 depending on assumptions. Some scenarios were not feasible based on the pilot study recruitment and retention rates. Conclusion: Pilot studies can be used to help design the main trial, but caution should be exercised. We recommend the use of sensitivity analyses to assess the robustness of the design assumptions for a main trial.

Children in care (e.g., foster care) are at risk of a variety of negative developmental outcomes. Of particular concern are their often poor academic outcomes. Indeed, children in care often have below grade-level performance on math and reading, increased rates of school drop out, and learning disabilities. Despite these difficulties, relatively little research has been conducted to try to ameliorate the problem. The present thesis, consisting of three independent randomized controlled trials (RCTs), sought to address the need for improved educational outcomes for children in care. The first RCT was an evaluation of a one-on-one tutoring program, TutorBright. Children in care were randomized to either a tutoring group or a waitlist control group and assessed on their math and reading skills, as well as on other educationally relevant domains (executive functioning, behaviour, and caregiver involvement in school-related activities), pre and post test. ANCOVA via multiple regression revealed that the children in care that had received tutoring, compared to the waitlist control group, made significantly greater gains in reading comprehension (Hedges’ g = 0.34), reading fluency (g = 0.16), and math calculation (g = 0.39). Moreover, executive functioning and self-reported symptoms of PTSD (for older children) were significant moderators of the effects of tutoring. No significant spill-over effects of tutoring were found. The second RCT, building upon previous RCTs by Flynn et al. (2012) and Harper and Schmidt (2016), evaluated whether a shorter version (i.e., 15 weeks) of a one-on-one Direct Instruction tutoring program, Teach Your Children Well (TYCW), was as effective as a longer version of TYCW (i.e., 25 weeks) on improving the math and reading skills for children in care. ANCOVA via multiple regression did not reveal any significant group differences in academic performance for any of the WJ-III subtests. Collapsing across the two intervention conditions, follow-up paired samples t-tests revealed significant improvement in Letter-Word Identification (Cohen’s d = 0.22), Reading Fluency (d = 0.36), Calculation (d = 0.38), Math Fluency (d = 0.47), Applied Problems (d = 0.30), and Broad Reading (d = 0.30) and Broad Math (d = 0.40) composite scores, suggesting that the 15-week and 25-week TYCW programs were equally effective in improving math and reading skills. Moreover, an attribute-treatment interaction analysis revealed that children with higher executive functioning benefited more from the shorter tutoring dosage for Calculation. The third RCT was a small pilot study that aimed to assess whether working-memory training (WMT) can enhance: 1) working memory capacity, in the short and longer term (i.e., immediately following the completion of WMT and 6-months later), 2) symptoms of Attention Deficit Hyperactivity Disorder (i.e., inattention, hyperactivity, and executive functioning) in the short and longer term, and, 3) math and reading skills at 6-months post WMT. The findings from this study suggested that WMT can improve verbal working memory (g = 0.35) as well as visuo-spatial short-term memory (g = 1.10) in the shorter term but not in the longer term. Moreover, WMT did not have a significant impact on improving symptoms of ADHD or math and reading skills. Together, the results of this dissertation indicate that the math and reading skills of children in care can be improved via tutoring. The findings highlight the importance of providing children in care with effective academic supports in order to help them reach their full potential.

Abstract The aim of this study was to examine lifestyle risk factors for low bone mass, falls and fractures, and to determine the effect of 30-month exercise trial on bone mass, balance, muscle strength and gait in elderly women. Reliability of an inclinometric method for assessing postural sway was evaluated. Data on risk factors, falls and fractures were collected by questionnaires, and calcaneus and radius bone mass were measured from 1,222 women. Lifetime physical activity, low occupational physical activity, type 2 diabetes, hypertension, hormone replacement, thyroid hormone and thiazide use were associated with increased bone mass, while low current physical activity, high coffee intake and late menarche were associated with low bone mass in lean women. Factors associated with fractures were: low lifetime habitual physical activity, diabetes, living alone and calcaneum bone mass. One hundred and sixty women with low femoral neck bone mass were randomly assigned to the exercise group (n = 84) or to the control group (n = 76). The outcomes included radius, proximal femur and calcaneus bone mass, postural sway, muscle strength, gait speed and endurance. Bone mineral density (BMD) at proximal femur decreased in the control group, while no change occurred in the exercise group. Mean trochanter bone mineral content (BMC) decreased more in the control group. The women in the exercise group improved their performance in walking speed and endurance, body sway and leg strength compared to the control group. There were six falls that resulted in fractures in the exercise group and 16 in the control group. The inclinometric method proved to be reliable. In conclusion, lifestyle factors are determinants of bone mass in lean elderly women. Long-term exercise has a site-specific effect on BMC but not on BMD in elderly women. Weight-bearing exercise can modify risk factors for fractures, and may even prevent fall-related fractures in elderly women.

Abstract Though not a life threatening condition, involuntary loss of urine is a miserable situation. It has a multidimensional effect on the afflicted individuals, both men and women – and for the society. The purpose of this study was to evaluate two modern, minimally invasive surgical techniques for the treatment of female stress urinary incontinence (SUI). The techniques evaluated were laparoscopic colposuspension with mesh and staples (LCM) and the tension-free vaginal tape procedure (TVT). The study consisted of four parts. The first part (Study I) was an observational retrospective follow-up study. Data on the first forty patients operated on with LCM at Oulu University Hospital were collected. Patients had SUI or mixed urinary incontinence (MUI) with predominantly stress incontinence. The Studies II–IV were parts of a randomized, multicenter clinical trial, where LCM was compared with TVT. According to the predefined inclusion criteria 128 SUI women were randomly allocated into two treatment groups: 70 patients received TVT treatment as allocated and 51 LCM. There were seven drop-outs after randomization. After one year of follow-up the cure and improvement rate of the patients operated with LCM were ~ 90%. Also a significant improvement was found in Urinary Incontinence Severity Scores (UISS). At base line the score was 12.1 and after one year follow-up 2.7 (p < 0.001). The bladder perforation rate was 15%. In Study II immediate cure rates and complications of LCM and TVT were studied. After six weeks of follow-up there was no difference in cure rates (~ 90%) between the procedures. There was no difference in complication rates. A significant difference was found in the use of anti-inflammatory / opioid drugs in the immediate post-operative period to relief the pain in favour for TVT. Hospital care was also significantly shorter after TVT than LCM. After one year of follow-up (Study III) TVT was found to give better result both objectively and subjectively. Negative stress test result was recorded in 85.7% in the TVT group and 56.9% in the LCM group. A significant difference was also found, when Visual Analoque Scale (VAS), King's College Health Questionnaire (KHQ) and UISS were used as outcome measures, in the favour of TVT. When 48-hour pad test was used as outcome measure there was no statistically significant difference between the groups. The cost-effectiveness (Study IV) of TVT was found to be better than that of LCM after one year of follow-up. In conclusion, the results of this study suggest, that TVT procedure is on the whole a cost-effective alternative for LCM in the treatment of female SUI.

Background: With usual supplementation practices, most children are Vitamin D Deficient (VDD) following Congenital Heart Disease (CHD) surgery and alternative regimens need consideration. Methods/Results: i) A systematic review identified 88 pediatric trials of high dose vitamin D. Studies evaluating the Institute of Medicine (IOM) Tolerable Upper Intake Level (UL) did not rapidly normalize levels, while loading therapy (≥ 40000 IU) did so within 3 days. Hypercalcemia occurred more often with doses above 400000 IU. ii) A double blind RCT was designed to determine whether pre-operative administration of the IOM UL can prevent post-operative VDD. Results after the first 30 participants completed study procedures demonstrated it was possible to recruit (1.8 patients per month) and complete study procedures (i.e. blood collection). Unfortunately few participants (45%) received more than 30 doses of study drug. Conclusion: Prevention of post-operative VDD in the majority of CHD patients will require alternatives to the IOM recommendations.

Background: In September 2005, scientific journals began requiring trial protocol registration to increase transparency and accountability. Objective: My primary objectives were: develop a database of linked protocols and publications for surgical randomized control trials (RCTs); estimate the proportion published; and determine the proportion exhibiting selective outcome reporting. Methods: A systematic search of the clinicaltrials.gov database was conducted identifying surgical RCTs, completed between 2006 and 2012. Protocols were linked with publications. Primary outcomes were compared. Results: We identified a cohort of 743 surgical RCT protocols. The proportion of registered trials which published their primary results was 0.49 (n=364). The proportion of selective outcome reporting was estimated to be 0.244, significantly lower than the previous estimate (p<0.001). Conclusion: More than half of the completed surgical RCTs were unpublished, and one quarter of those published selectively reported their primary outcome. This supports the notion that significant bias is present in the surgical literature.

Class of 2008 Abstract / Objectives: 1) To determine if data on weight change in the placebo arm of RCT's for weight loss were available, and 2) to conduct a pilot meta-analysis to estimate the average weight change in the placebo arm. Methods: Four randomized placebo controlled trials of rimonabant for weight loss were retrieved. A draft data extraction form was developed to record weight loss and demographic data. Potential for bias was assessed on design issues related to withdrawals, blinding, allocation procedure, adherence, and manufacturer influence. Based on available data, a forest plot was constructed and heterogeneity was assessed. The a priori alpha level was 0.05. Results: The placebo groups from all studies were similar. The pooled data indicated that individuals in the placebo arm lost an average of 3.3 kg, p < 0.001. One study had a significantly greater completion rate than the other studies. Participants were prescribed a hypocaloric diet and were instructed to increase physical activity but no data were reported on calories consumed or amount of physical activity. Weight loss of 5% ranged from 15-20% of participants. There was potential for bias relating to reported adherence, allocation concealment process, and manufacturer funding. Conclusions: Participants in the placebo arm of rimonabant trials lost an average of 3.3 kg, which was statistically significant. Little can be learned about weight loss in the placebo arm because no data on calories consumed, amount of exercise, or hunger were reported. Information from other RCT's is needed to provide additional data and to confirm the findings.