Academic literature on the topic 'Logisitic regression'

This study determines the factors which correlate with attitudes towards mother-to-child transmission of HIV in pregnant women. Using a structured questionnaire, 527 pregnant women who visited a hospital to have prenatal checkups were interviewed. The survey items were: sociodemographic characteristics, experiences of pre-test counselling, knowledge of mother-to-child transmission, and attitude towards termination of pregnancy. Results showed that many pregnant women (80%) did not have proper knowledge of the possibility of mother-to-child transmission. Logisitic regression analysis also indicates that age and knowledge of the possibility of mother-to-child transmission were the significant determinants of attitudes towards termination of pregnancy. Older women who believe that all the babies of pregnant women with HIV will be infected are most likely to terminate their pregnancy when they are diagnosed as HIV positive. Considering the importance of informed decisions regarding pregnancy, this study must have important implications for future support progammes for HIV-positive pregnant women.

Background: This study is performed to evaluate vitamin D levels and metabolic parameters in patients with prediabetes, compared to healthy controls.Methods: This study was conducted between October and December 2013 in İstanbul Haseki Training and Research Hospital, internal medicine department. We enrolled total 247 individuals, 122 prediabetic (PreDM) patients (79 female, 43 male) and 125 control healthy individuals (94 female, 31 male) between 20-65 ages who admitted randomizely to the outpatient clinic with non spesific complaints. FPG, urea, creatinine, calcium, phosphate, albumin, alkaline phosphatase, thyriod stimulan hormon (TSH), 25 hydroxy vitamin D (25[OH]D), parathormon (PTH), c-peptide, insulin were analyzed.Results: Pre DM patients’ mean plasma 25[OH]D level (25.7±14.9 nmol/l) was statistically lower than the control group (31.4±17.8 nmol/l). Pre DM patients’ mean plasma insulin, c-peptide, calcium, PTH, HOMA-IR (10.8±8.7 IU/ml, 3.3±2.0 ng/ml, 9.7±0.4 mg/dl, 56.5±22.5 pg/ml, 3.0±2.68, respectively) levels were statistically higher than the control group’s (6.3±3.8 IU/ml, 2.4±1.0 ng/ml, 9.5±0.5 mg/dl, 44.0±16.0 pg/ml, 1.4±0.8, respectively) mean levels. There were negative correlations between 25[OH]D and BMI (r:- 0.13, p:0.03), FBG (r:- 0.14, p:0.02) and plasma insulin (r:-0.16, p:0.01) values. A multivariate logisitic regression model for prediabetes was performed and variables as female gender, age, HOMA-IR and lower 25[OH]D values were risk factors for pre DM.Conclusions: Serum low 25[OH]D level correlated with insulin resistance and metabolic parameters in prediabetic patients. Also, it may play an important role in the development of type 2 diabetes.

Abstract Background More than 10% of people aged 50 years and older report dizziness. Despite available treatments, dizziness remains unresolved for many people due in part to sub-optimal assessment. Aims to identify factors associated with dizziness handicap in middle-aged and older people to identify targets for intervention to address this debilitating problem. Methods This secondary analysis of baseline and prospective data from a randomised-controlled trial involved 305 individuals aged 50-92 years reporting significant dizziness in the past year. Participants were classified as having either mild or no dizziness handicap (score < 31) or moderate / severe dizziness handicap (score: 31-100) based on the dizziness handicap inventory (DHI). Participants completed health questionnaires and underwent assessments of psychological well-being, lying and standing blood pressure, vestibular function, strength, vision, proprioception, processing speed, balance, stepping and gait. Participants also reported dizziness episodes in monthly diaries for 6 months following baseline assessment. Participants reported dizziness episodes in monthly diaries for six months following baseline assessment. Results DHI scores ranged from 0-86 with 95 participants (31%) reporting moderate/severe dizziness handicap. Binary logisitic regression identified a postive Dix-Hallpike test for benign paroxysmal positional vertigo, cardiovascular medicationuse, high postural sway when standing on the floor with eyes closed and severe anxiety, as significant and independent predictors of moderate.severe dizziness handicap. Poor balance (high sway) also significantly identified participants who reported frequent (> or = weekly, n=134) dizziness episodes. Conclusion Assessment of cardiovascular medication use, benign paroxysmal positional vertigo, anxiety and postural sway identify middle aged and older people with significant dizziness handicap. A multifactorial assessment including these facotrs may assist in tailoring evidence-based therapies to alleviate dizziness handicap in this group.

Background Ethiopia has over 3,200 new cases of leprosy diagnosed every year. Prevention remains a challenge as transmission pathways are poorly understood. Susceptibility and disease manifestations are highly dependent on individual host-immune response. Nutritional deficiencies, such as protein-energy malnutrition, have been linked to reduced cell-mediated immunity, which in the case of leprosy, could lead to a higher chance of active leprosy and thus an increased reservoir of transmissible infection. Methodology/Principal findings Between June and August 2018, recently diagnosed patients with leprosy and individuals without known contact with cases were enrolled as controls in North Gondar regional health centers. Participants answered survey questions on biometric data, demographics, socioeconomic situation, and dietary habits. Descriptive statistics, univariate, and multivariate logisitic regression examined associations between undernutrition, specifically body mass index (BMI), middle upper arm circumference (MUAC), and leprosy. Eighty-one participants (40 cases of leprosy, 41 controls) were enrolled (75% male) with an average age of 38.6 years (SD 18.3). The majority of cases were multibacillary (MB) (90%). There was a high prevalence of undernutrition with 24 (29.6%) participants underweight (BMI <18.5) and 17 (21%) having a low MUAC. On multivariate analysis, underweight was significantly associated with leprosy (aOR = 9.25, 95% CI 2.77, 30.81). Also found to be associated with leprosy was cutting the size of meals/skipping meals (OR = 2.9, 95% CI 1.0, 8.32) or not having enough money for food (OR = 10, 95% CI 3.44 29.06). Conclusions/Significance The results suggest a strong association between leprosy and undernutrition, while also supporting the framework that food insecurity may lead to undernutrition that then could increase susceptibility to leprosy. In conclusion, this study highlights the need to study the interplay of undernutrition, food insecurity, and the manifestations of leprosy.

Abstract Abstract 553 First National Level Estimate of Emergency Department Charges for Sickle Cell Disease It is estimated that there are 100,000 people with sickle cell disease (SCD) living in the US. Little data is currently availabel on the number of emergency room visits made per year by this patient population. Given prior data demonstrating the cost savings associated with comprehensive care delivery and specialized day hospital units that have decreased ED visits and hospitalizations, our goal was to estimate the national burden of SCD care in the ED setting. The purpose of this study was to estimate the number of ED visits made by SCD patients in 2006, the total and average charges for those visits and the subsequent hospital admissions, and to identify factors that predicted if patients were admitted to the hospital. In addition, we used the online Healthcare Cost and Utilization Project program (HCUP) to compare the number of SCD visits with the number of ED visits and subsequent admissions for other chronic illnesses. Methods: The Nationwide Emergency Department Sample (NEDS) contains information from 950 hospitals and approximates a 20% stratified sample of US hospital based EDs. Data is currently only availabel for 2006. Using this dataset we identified all ED visits with the following SCD ICD9 codes 28260,28261,28262,28263,28264,28268,28269,28241,28242 listed anywhere in the diagnosis list. Bivariate and multivariate analyses were conducted using t-tests, oneway ANOVAs, and multiple linear or logisitic regressions as appropriate. Sampling weights are included in the NEDS to allow for the generation of national estimates. In addition to the full NEDS dataset the Agency for Healthcare Research and Quality makes availabel on-line a query system based on HCUP data free on its website. We used this system to collect data on total ED visits for asthma, HIV and congestive heart failure (CHF) that ended in a hospital admission using the clinical classification software codes 128, 5, and 108 as defined in the HCUP system. Results: There were a total of 58993 ED visits for patients with SCD sampled in the NEDS in 2006, representing a national estimate of 232,381 ED visits. Using the weighted data there were an estimated 190,000 adult ED visits and 44,000 pediatric visits. The total charges for all of the ED visits were estimated to be $266 million of which 191 million were for patients not subsequently admitted to the hospital. Being female and adult resulted in significantly higher ED charges. 50% of the visits were covered by Medicaid, 22% by Medicare, 18% were private pay, 7.5% self-pay and the remainder were no charge or other. 60% of the visits occurred in the South, 18% in the Northeast, 15% in the Midwest and 7% in the West. 5% of the visits had a code for pneumonia, 0.25% had a code for stroke and 7.3% had a code for asthma. These ED visits resulted in an estimated 91,000 hospitalizations. Pediatric patients were more likely to be admitted than adult patients (44% v 38% p=0.03). The combined ED and inpatient charges for those admitted were estimated to be $2.7 billion with a mean charge of $23,669 (range $591-871,111). 64 individuals are estimated to have died in the ED and another 522 died during their hospital stay. The average age of those that died in the ED was 33 and in the hospital was 42. In adjusted analyses being female, in the pediatric age group, or having a code for sickle crisis, asthma, stroke or pneumonia all were associated with a significant increase in odds of being admitted to the hospital. Table 1 shows the comparison number of admissions from the ED for SCD, asthma, HIV and CHF. Conclusions: Given the size of the US sickle population this data suggests that the ED is a major source of care for this patient population. Given that there are over 2 billion dollars in charges for those admitted from the ED, a focus of improving quality of care and decreasing costs of care should be eliminating ED use by the SCD population by providing alternative venues for treatment of painful crises that is more likely to result in a discharge home. Disclosures: No relevant conflicts of interest to declare.

Backgrounds: Emerging evidence suggests that stress hyperglycemia ratio (SHR), an index of relative stress hyperglycemia, is of great prognostic value in acute myocardial infarction (AMI), but current evidence is limited in elderly patients. In this study, we aimed to assess whether SHR is associated with in-hospital outcomes in elderly patients with AMI.Methods: In this retrospective study, patients who were aged over 75 years and diagnosed with AMI were consecutively enrolled from 2015, January 1st to 2019, December 31th. Admission blood glucose and glycosylated hemoglobin (HbA1C) during the index hospitalization were used to calculate SHR. Restricted quadratic splines, receiver-operating curves, and logistic regression were performed to evaluate the association between SHR and in-hospital outcomes, including in-hospital all-cause death and in-hospital major adverse cardiac and cerebrovascular events (MACCEs) defined as a composite of all-cause death, cardiogenic shock, reinfarction, mechanical complications of MI, stroke, and major bleeding.Results: A total of 341 subjects were included in this study. Higher SHR levels were observed in patients who had MACCEs (n = 69) or death (n = 44) during hospitalization. Compared with a SHR value below 1.25, a high SHR was independently associated with in-hospital MACCEs (odds ratio [OR]: 2.945, 95% confidence interval [CI]: 1.626–5.334, P &lt; 0.001) and all-cause death (OR: 2.871 95% CI: 1.428–5.772, P = 0.003) in univariate and multivariate logisitic analysis. This relationship increased with SHR levels based on a non-linear dose-response curve. In contrast, admission glucose was only associated with clinical outcomes in univariate analysis. In subgroup analysis, high SHR was significantly predictive of worse in-hospital clinical outcomes in non-diabetic patients (MACCEs: 2.716 [1.281–5.762], P = 0.009; all-cause death: 2.394 [1.040–5.507], P = 0.040), but the association was not significant in diabetic patients.Conclusion: SHR might serve as a simple and independent indicator of adverse in-hospital outcomes in elderly patients with AMI, especially in non-diabetic population.

Introduction: In different societies, paying attention to spiritual dimension of health has been emphasized as one of the health dimensions according to the WHO. The study aimed to determine the relation between spiritual health and use of virtual social networks among high school students in Firouzabad, Fars province in 2017. Methods: In the cross-sectional study, 380 students (16-18 years) were selected using the multistage stratified sampling. Data collection instruments included demographic information questionnaire, virtual social network use questionnaire, and Poltzea and Ellisons questionnaire of spiritual health assessment. To analyze the data, analysis of variance, Pearson correlation coefficient, and logistic regression tests were used. Results: The mean score of spiritual health was 83.98 ± 17.01 and had a significant correlation with education level (p<0.001) and internet use time (p=0.02). A significant correlation was found between virtual social networks use and monthly income, (r=0.17) religions health (r= -0.64), individual health (r= -0.11), and total score of spiritual health (r= -0.09). Most of the students (65.77%) had internet access and the most frequently used virtual social network was WathsApp. Logisitc regression revealed that the mean score of desirable spiritual health among students who used virtual social networks for 11-20 , 21-30, and >30 hours per week were 0.83 , 0.72, and 0.27 times lower than the students who used these networks less than 10 hours a week. Conclusion: Internet access was higher in male students than female and the spiritual health was at the average level in both genders. Regarding the effect of virtual social networks on students and its increasing expansion, the issue must be considered by authorities.

This thesis reports on research that examines the early stage invasion process of Asparagus asparagoides (L.) W. Wight (bridal creeper), primarily a bird-dispersed weed, in a remnant vegetation patch. The study site is on Phillip Island, approximately 100 kilometres south east of Melbourne, Victoria. Asparagus asparagoides invasion of the remnant vegetation reserve is a relatively recent phenomenon. Landscape elements that affect bird dispersal and vegetation types that affect seedling establishment may be important factors that limit or enhance the spread of A. asparagoides. A systematic sampling strategy was adopted and data collected for a variety of landscape and vegetative variables including cover and abundance of A. asparagoides and the data were presented in a Geographic Information System (GIS). Preliminary results show that the distribution of A. asparagoides within a remnant vegetation patch is not random. It appears to have entered the reserve from two boundaries, spreading toward the centre, which to date remains sparsely colonised despite the capacity of this weed to spread rapidly over long distances by birds. A number of other outcomes are noted. Asparagus asparagoides establishment is prevented in pasture where sheep and cattle graze, and paddocks subjected to tillage practices. The exclusion of grazing in fenced off vegetation in pastures demonstrates rapid weed establishment and colonisation several hundred metres from main infestation. Field observation and visual inspection of temporal progress of invasion (using above ground weed density with tuber appearance to infer age) appear to suggest that invasion into remnant is associated with the track network. This age/density assumption is strengthened when spatial distribution is examined using a data set where low-density values for A. asparagoides are removed and compared with a data set using all A. asparagoides density values. The mapping of A. asparagoides in fenced off farm remnants suggests that velocity of spread at 191m/yr is a considerable underestimate. Subsequent analysis shows that the spatial distribution of A. asparagoides is not completely spatially random while intensity surface analysis highlights regions of low and high intensity located near track network. Mapping a density surface within GIS provided confirmatory evidence for the establishment of satellite clusters along the track network. The change in the intensity surface observed using the two data sets (lowdensity values and all density values) is also consistent with an expanding invasion occurring between two time periods. Spatial point pattern analysis using K-function statistics shows that xxii the clustering observed using GIS appears to be occurring at two scales or distances (130m- 160m and 195m-205m). The association between tracks and the invasion process observed in the initial stages of the study is examined. There is a change in density as a function of distance from a track where the density of A. asparagoides appears to reduce the further away from the track a site is and this relationship holds regardless of track width. The final stages of the study look at the development of a predictive model. Visual exploration of the data through mapping in a GIS and field observation made during data collection provide the starting point for the development of logistic models to estimate the probability of A. asparagoides presence. Finally the best overall logistic model is applied to a second independent site to determine the general applicability of the model. A number of variables that impact on the presence of A. asparagoides, particularly during the initial stages of the invasion process, are identified. While all the identified variables and the overall model are statistically significant, the model is found to correctly predict presence/absence in only 67% of cases overall. The model however could be expected to correctly predict the presence of A. asparagoides in 74% of cases and has a false positive rate of 40%. The model is applied at a second independent site and found to have an overall percent correct rate of 80% and correctly predicted A. asparagoides presence in 94% of cases. The variables identified as influential in the early stage of invasion are relatively easy to acquire by simple field survey that does not require specialist skills. When considering the model as a tool for the management of remnant vegetation communities, high false positive rates may lead to limited resources being spent on searching sites where there is no weed. However, a high false negative rate would have a larger impact on the management of the weed since the undetected infestations would form sources for new propagules. The model performs well from this point of view in that it provided low false negative rates at both sites. The value of the predictive model is its ability to provide managers with information regarding specific areas to target for weed eradication and management can use the model to assess the effectiveness of any control measures by going back to obtain new cover density data, then using the model to examine the changes over time. The model also provides a starting point for the development of a generic model of A. asparagoides invasion at sites outside of Phillip Island and could also provide the starting point for developing models that could be used for other bird-dispersed fleshy-fruited weed species.

Doctor of Philosophy
Department of Physics
Dean Zollman
As part of a collaborative study of the science preparation of elementary school teachers, we investigated the quality of students’ reasoning and explored the relationship between sophistication of reasoning and the degree to which the courses were considered inquiry oriented. To probe students’ reasoning, we developed open-ended written content questions with the distinguishing feature of applying recently learned concepts in a new context. We devised a protocol for developing written content questions that provided a common structure for probing and classifying students’ sophistication level of reasoning. In designing our protocol, we considered several distinct criteria, and classified students’ responses based on their performance for each criterion. First, we classified concepts into three types: Descriptive, Hypothetical, and Theoretical and categorized the abstraction levels of the responses in terms of the types of concepts and the inter-relationship between the concepts. Second, we devised a rubric based on Bloom’s revised taxonomy with seven traits (both knowledge types and cognitive processes) and a defined set of criteria to evaluate each trait. Along with analyzing students’ reasoning, we visited universities and observed the courses in which the students were enrolled. We used the Reformed Teaching Observation Protocol (RTOP) to rank the courses with respect to characteristics that are valued for the inquiry courses. We conducted logistic regression for a sample of 18 courses with about 900 students and reported the results for performing logistic regression to estimate the relationship between traits of reasoning and RTOP score. In addition, we analyzed conceptual structure of students’ responses, based on conceptual classification schemes, and clustered students’ responses into six categories. We derived regression model, to estimate the relationship between the sophistication of the categories of conceptual structure and RTOP scores. However, the outcome variable with six categories required a more complicated regression model, known as multinomial logistic regression, generalized from binary logistic regression. With the large amount of collected data, we found that the likelihood of the higher cognitive processes were in favor of classes with higher measures on inquiry. However, the usage of more abstract concepts with higher order conceptual structures was less prevalent in higher RTOP courses.

In this study, we compare the two statistical techniques logistic regression and discriminant analysis to see how well they classify companies based on clusters – made from the solvency ratio ­– using principal components as independent variables. The principal components are made with different financial ratios. We use cluster analysis to find groups with low, medium and high solvency ratio of 1200 different companies found on the NASDAQ stock market and use this as an apriori definition of risk. The results shows that the logistic regression outperforms the discriminant analysis in classifying all of the groups except for the middle one. We conclude that this is in line with previous studies.

This thesis presents a robust solution to the classification and variable selection problem when the dimension of the data, or number of predictor variables, may greatly exceed the number of observations. When faced with the problem of classifying objects given many measured attributes of the objects, the goal is to build a model that makes the most accurate predictions using only the most meaningful subset of the available measurements. The introduction of [cursive l] 1 regularized model titling has inspired many approaches that simultaneously do model fitting and variable selection. If parametric models are employed, the standard approach is some form of regularized maximum likelihood estimation. While this is an asymptotically efficient procedure under very general conditions, it is not robust. Outliers can negatively impact both estimation and variable selection. Moreover, outliers can be very difficult to identify as the number of predictor variables becomes large. Minimizing the integrated squared error, or L 2 error, while less efficient, has been shown to generate parametric estimators that are robust to a fair amount of contamination in several contexts. In this thesis, we present a novel robust parametric regression model for the binary classification problem based on L 2 distance, the logistic L 2 estimator (L 2 E). To perform simultaneous model fitting and variable selection among correlated predictors in the high dimensional setting, an elastic net penalty is introduced. A fast computational algorithm for minimizing the elastic net penalized logistic L 2 E loss is derived and results on the algorithm's global convergence properties are given. Through simulations we demonstrate the utility of the penalized logistic L 2 E at robustly recovering sparse models from high dimensional data in the presence of outliers and inliers. Results on real genomic data are also presented.