Home About us Contact
|
Home About us Contact | ||
|
|
||
Logistic Regression Modelling (logistic + regression_modelling)
Selected AbstractsRisk of harm among gamblers in the general population as a function of level of participation in gambling activitiesADDICTION, Issue 4 2006Shawn R. Currie ABSTRACT Aims To examine the relationship between gambling behaviours and risk of gambling-related harm in a nationally representative population sample. Design Risk curves of gambling frequency and expenditure (total amount and percentage of income) were plotted against harm from gambling. Setting Data derived from 19 012 individuals participating in the Canadian Community Health Survey,Mental Health and Well-being cycle, a comprehensive interview-based survey conducted by Statistics Canada in 2002. Measurement Gambling behaviours and related harms were assessed with the Canadian Problem Gambling Index. Findings Risk curves indicated the chances of experiencing gambling-related harm increased steadily the more often one gambles and the more money one invests in gambling. Receiver operating characteristic analysis identified the optimal limits for low-risk participation as gambling no more than two to three times per month, spending no more than $501,1000CAN per year on gambling and investing no more than 1% of gross family income on gambling activities. Logistic regression modelling confirmed a significant increase in the risk of gambling-related harm (odds ratios ranging from 2.0 to 7.7) when these limits were exceeded. Conclusions Risk curves are a promising methodology for examining the relationship between gambling participation and risk of harm. The development of low-risk gambling limits based on risk curve analysis appears to be feasible. [source] Reporting of adverse drug reactions: predictors of under-reporting in Malaysia,PHARMACOEPIDEMIOLOGY AND DRUG SAFETY, Issue 2 2007Zoriah Aziz PhD Abstract Background Malaysia like many other countries worldwide uses spontaneous reporting systems as a mean of collecting data on suspected adverse drug reaction (ADR). However, compared to other countries, which use the system, the reporting rate in Malaysia is very low. Why some physicians do not report ADRs is not well understood. Objective To identify factors, which would predict physicians' failure to send ADR reports. Design and Setting Face-to-face interview using a structured questionnaire involving physicians working at the University of Malaya Medical Centre, Malaysia. Results About a third of the physicians in the Centre participated. Sixty-five of the 415 approached refused to participate. A high proportion of the respondents (81.4%) indicated that they had suspected an ADR but did not report it, while about 40% of the respondents were not aware of the existence of the national reporting system in Malaysia. Logistic regression modelling identified the variable ,ADR considered to be too trivial or too well known to report' as the strongest predictor of not reporting, followed by physicians' category and uncertainty that the reaction had been definitely caused by a drug. Conclusion Important predictor variables, which limit physicians from reporting ADR in Malaysia, were related to uncertainty of types of reaction to report, lack of awareness about the existence, function and purpose of national ADR reporting. The findings could be useful for planning strategies to improve the reporting rate. Copyright © 2006 John Wiley & Sons, Ltd. [source] Investigating the relationship between affluence and elective caesarean sectionsBJOG : AN INTERNATIONAL JOURNAL OF OBSTETRICS & GYNAECOLOGY, Issue 7 2005Bernadette Alves The proportion of women delivering by caesarean section has increased dramatically in England and many westernised countries. It has been suggested that one important reason for this increase is the growing proportion of women opting for elective caesareans for lifestyle reasons, a trend that is, it is argued, most common among the affluent. We investigated the hypothesis that affluent women are more likely to deliver by elective caesarean section. Logistic regression modelling was used to analyse data from half a million women who delivered in English NHS hospitals between 1996 and 2000. We found that women living in the most affluent areas of England were significantly more likely to have an elective caesarean section than their deprived counterparts. [source] Lipids, lipoproteins and the risk of benign prostatic hyperplasia in community-dwelling menBJU INTERNATIONAL, Issue 3 2008J. Kellogg Parsons OBJECTIVE To examine the associations of serum lipids and lipoproteins with benign prostatic hyperplasia (BPH) in community-dwelling men. SUBJECTS AND METHODS This analysis was conducted within the Rancho Bernardo Study, a prospective, community-based cohort study. BPH was defined as a history of prostate surgery for other than cancer, or a medical diagnosis of BPH. Logistic regression modelling, with adjustments for age and stratification by diabetes diagnosis, was used to estimate the odds ratio (OR) of BPH associated with fasting serum concentrations of total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, triglycerides, and the triglyceride to HDL ratio. RESULTS Among 531 eligible participants, 259 (48%) reported BPH and 272 (52%) reported no BPH. Men with BPH, with a mean (range) age of 75.8 (76.1,80.1) years, were older than men without BPH , at 72.7 (72.4,74.0) years. There were no significant associations of total cholesterol (P trend, 0.52), HDL cholesterol (0.56), triglycerides (0.30), or triglyceride to HDL ratio (0.13) with the risk of BPH. In a subset analysis in men with diabetes, those in the highest tertile (>133 mg/dL) of LDL cholesterol, compared with those in the lowest tertile (<110 mg/dL), were four times more likely to have BPH (odds ratio 4.00, 95% confidence interval 1.27,12.63, P trend 0.02). These results were not explained by the use of statins. CONCLUSIONS In these community-dwelling men, higher serum LDL was associated with a greater risk of BPH among diabetics. These data suggest that diabetic men with increased LDL cholesterol are at greater risk of BPH. This observation is consistent with the concept that cardiac risk factors are involved with the pathogenesis of BPH. [source] Socio-demographic and psychopathologic correlates of enuresis in urban Ethiopian childrenACTA PAEDIATRICA, Issue 4 2007Menelik Desta Abstract Aim: To examine the association between enuresis and psychopathology in urban Ethiopian children. Methods: A two-stage mental health survey of 5000 urban children found enuresis to be by far the most common disorder. Logistic regression modelling was carried out to determine the independent associations of a number of socio-demographic and psychopathological characteristics with enuresis. Results: Male sex, younger age and lower achieved educational grade of the child were all independently associated with childhood enuresis. The odds of having enuresis were significantly higher for children in families with significant financial worries and in children from homes where parents were separated. Children with DSMIII-R anxiety disorders, especially simple phobia, or disruptive behaviour disorders were found to have significantly higher odds of having enuresis. Conclusion: Psychopathology, both anxiety and behavioural disorders, as well as family stressors in urban Ethiopian children were found to be risk factors for enuresis. Although a cause-effect relationship could not be ascertained, the findings of higher association of psychopathology with enuresis in this and other studies indicate that there is a need for evaluating children with enuresis for the presence of concurrent psychopathology, especially in traditional societies where undetected psychopathology may be more common. [source] Logistic regression approach to modelling the variability of recombination rateJOURNAL OF ANIMAL BREEDING AND GENETICS, Issue 1 2000By J. Szyda The objective of the paper is to quantify the relationship between recombination rate and factors such as ,sex of sperm', paternal halfsib family, and individual, using logistic regression modelling. The analysed data set consists of 2214 single bovine sperm cell samples. Haplotypes at each single sperm were determined at eleven marker loci forming eight intervals located on chromosomes 6, 23 and X/Y. The experimental design comprises six paternal halfsib families. Six logistic regression models are fitted to the data from each interval. Departure from commonly assumed homogenous recombination is detected for one marker interval on chromosome 23 , influence of individual and ,sex of sperm' by family interaction, and for both intervals mapped to sex chromosomes , influence of ,sex of sperm' and paternal halfsib family. [source] Who needs colonoscopy to identify colorectal cancer?ALIMENTARY PHARMACOLOGY & THERAPEUTICS, Issue 2 2010Bowel symptoms do not add substantially to age, other medical history Aliment Pharmacol Ther 2010; 32: 270,281 Summary Background, Many bodies advise that people with bowel symptoms undergo colonoscopy to detect colorectal cancer. Aim, To determine which bowel symptoms predict cancer on colonoscopy. Methods, Information was collected on symptoms, demographics and medical history from patients subsequently undergoing colonoscopy. Multiple logistic regression modelling was used to identify predictors of colorectal cancer. An ROC curve was estimated for each model, and the area under the curve (AUC) was computed. Results, Cancer was found in 159 patients and no cancer or adenoma in 7577 patients. Bowel symptoms that predicted cancer were rectal bleeding, change in bowel habit and rectal mucus. Prediction was the strongest in patients who had symptoms at least weekly and commencing within the previous 12 months; abdominal pain was predictive only in such patients. The odds ratios never exceeded 4.27. A model based on age, gender, and medical history was highly predictive (AUC = 0.79). Adding symptoms to this model increased the AUC to 0.85. Conclusions, This model predicts patients in whom colonoscopy will have the highest yield. Conversely, colonoscopy can be avoided in people at low risk: in our study, 95% of cancers could have been detected by doing only 60% of the colonoscopies. [source] Predictors of persistence with 5-aminosalicylic acid therapy for ulcerative colitisALIMENTARY PHARMACOLOGY & THERAPEUTICS, Issue 8 2009S. V. KANE Summary Background, Individuals with ulcerative colitis (UC) are at risk for poor persistence with therapy. Aim, To identify factors predicting persistence with 5-aminosalicylic acid (5-ASA) therapy after 3 and 12 months in subjects with UC. Methods, In this retrospective cohort study, persistence with 5-ASA therapy was determined from prescription refill data from a commercial health insurance claims database. The analysis included subjects with UC who filled a prescription for any oral 5-ASA between October 2002 and September 2004. Persistence was defined as prescription refill at 3 and/or 12 months. Multivariate logistic regression modelling identified variables independently associated with persistence at 3 and 12 months. Results, In all, 3574 subjects were identified. Fifty-seven per cent (2044) were persistent at 3 months. Glucocorticoid use before the index prescription predicted improved persistence at 3 months. Psychiatric diagnosis, mail order of the index prescription, female gender and co-pay predicted decreased persistence. At 12 months, 1124 (55%) remained persistent. Rectal 5-ASA use, older age and switching to a different 5-ASA predicted improved persistence at 12 months. Hospitalization for a gastrointestinal condition, mail order of the 3-month prescription and number of co-morbid illnesses predicted lower persistence. Conclusion, Persistence with 5-ASA treatment in UC is complex and multifactorial, and differs by time period. [source] Modelling the distribution of palsas in Finnish Lapland with logistic regression and GISPERMAFROST AND PERIGLACIAL PROCESSES, Issue 1 2002Miska Luoto Abstract The location of palsas (peat mounds with a perennially frozen core) was mapped in an area of 3370 km2 in Finnish Lapland by interpreting aerial digital photographs. Using environmental variables derived from digital land cover data and an elevation model, the distribution of palsas was modelled using geographic information system (GIS) techniques and multiple logistic regression. The relative roles of eight environmental variables potentially affecting the distribution of the palsas were studied in a spatial grid system with 3370 grid squares of 1.0 km2, of which 172 were found to contain palsas. The altitudes of the palsas varied from 180 m to 390 m. In the logistic regression model the probability of the presence of a palsa in a 1.0 km2 square increased with 1) the area of mire, 2) the proportion of flat topography, 3) water cover and 4) elevation of the lowest point in the 1.0 km2 analysis square. The palsa distribution model was validated by fitting it to an independent test area of 300 squares bordering on the main study area: the overall classification rate was 97.67%. The application of GIS data and techniques with logistic regression modelling has potential for wide use in studies on distribution patterns in periglacial processes and landforms. Copyright © 2002 John Wiley & Sons, Ltd. [source] Mental health and other clinical correlates of euthanasia attitudes in an Australian outpatient cancer populationPSYCHO-ONCOLOGY, Issue 4 2007G. L. Carter Abstract A majority of patients with cancer have been reported to endorse euthanasia and physician assisted suicide (PAS) in general and a substantial proportion endorse these for themselves. However, the potential influence of mental health and other clinical variables on these decisions is not well understood. This study of 228 outpatients attending an oncology clinic in Newcastle, Australia used a cross-sectional design and logistic regression modelling to examine the relationship of demographic, disease status, mental health and quality of life variables to attitudes toward euthanasia and PAS. The majority reported support for euthanasia (79%, n=179), for PAS (69%, n=158) and personal support for euthanasia/PAS (68%, n=156). However, few reported having asked their doctor for euthanasia (2%, n=5) or PAS (2%, n=5). Three outcomes were modelled: support for euthanasia was associated with active religious belief (adjusted odds ratio (AOR) 0.21, 95% CI: 0.10,0.46); support for PAS was associated with active religious belief (AOR 0.35, 95% CI: 18,0.70) and recent pain (AOR 0.87, 95% CI: 0.0.76,0.99); and personal support for euthanasia/PAS was associated with active religious belief (AOR 0.26, 95% CI: 0.14,0.48). Depression, anxiety, recent suicidal ideation, and lifetime suicide attempt were not independently associated with any of the three outcomes modelled. Copyright © 2006 John Wiley & Sons, Ltd. [source] Parents underestimate their child`s overweightACTA PAEDIATRICA, Issue 9 2010Nina Vuorela Abstract Aim:, The aim of this study was to evaluate parents' ability to perceive the weight status of their children. Methods:, This cross-sectional study was performed on 5 (n = 310) and 11-year-old (n = 296) children. The height, weight and waist circumference were measured. Body mass index (BMI, kg/m²) was calculated. The International Obesity Task Force criteria and the British cut-off points were used to classify BMI and waist circumference. Parents filled out a questionnaire concerning their perception of the weight class of their child. The parents and the 11-year-old children estimated their own weight class. For analysis, the measured and perceived weight classes were divided into two categories; normal weight (including underweight) and overweight or obese. To measure the agreement cross-tabulation with Cohen's Kappa was used. Explanatory variables associated with misclassification of overweight children as normal weight were examined by logistic regression modelling. Results:, The prevalence of overweight (including obese) was 17.4% and 21.6% in 5- and 11-year-old children in this study. Only a few parents misclassified their normal weight children as overweight. By contrast, a majority of parents to the 5-year-old children and about half of the parents to the 11-year-old children misclassified them as normal weight. Using waist circumference for body size classification did not improve parents' performance. Mothers performed best when estimating own weight class. Conclusion:, A majority of parents fail to recognize overweight or obesity in their 5- and 11-year-old children. The underestimation of overweight may impair the motivation of the parents to adopt weight control. [source] | ||||||||||