Discover the most talked about and latest scientific content & concepts.

Concept: Multivariate adaptive regression splines


OBJECTIVESThe effect of the lunar cycle and seasonal variation on ascending aortic dissection surgery outcomes is unknown. We investigated these temporal effects on risk-adjusted hospital mortality and then on the length of stay (LOS) following surgery for survivors.METHODSWe examined prospectively collected data from cardiac operations at two major centres within a single state between January 1996 and December 2011. We first examined the relationship between the lunar cycle and seasonal variation, along with demographic and risk profile covariates, with mortality using univariate analyses, followed by multiple logistic regression modelling that controlled for demographic and patient risk variables including age, gender, risk profile (diabetes, hypertension, dyslipidaemia and renal failure), and two surgical groups: Group A, consisting of patients having repair of ascending aorta dissection repair only, and Group B, with those having ascending aorta repair plus aortic valve surgery or coronary bypass surgery or both. We further examined the relationship with LOS using both univariate and multiple regression analyses.RESULTSThere were 210 patients who had repair of dissection in the study period, with 109 patients in Group A and 101 in Group B. The average age of this sample was 59.5 (standard deviation = 16.0), 65.7% were male and 18.1% died prior to discharge following repair. The greatest percentage of deaths occurred in winter (31.6%, n = 12), while the least were in summer (21.1%, n = 8) and fall (21.1%, n = 8). An overall χ(2) test found there was no difference in mortality for season (P = 0.55). Univariate analyses also found the age of patients who died vs lived was significantly higher (65.9 vs 58.1 years; P = 0.001), and a significantly greater (P = 0.029) percentage of patients with diabetes vs without diabetes died (41.7 vs 16.7%). Univariate analyses found all other covariates were not significantly related to mortality. In the multiple logistic regression model, there was no significant effect for season, while the odds of dying increased with age (odds ratio [OR] = 1.04, 95% confidence interval [95% CI] = 1.01-1.07, P = 0.012), and the odds of dying in the full-moon cycle vs the new moon cycle was significantly reduced (OR = 0.21, 95% CI = 0.05-0.81, P = 0.024). No other covariate significantly increased or decreased the odds of death, including diabetes risk, which had been significantly related to death in the univariate analysis. Within a linear regression model that examined the relationship with LOS, Group B (P = 0.020), male sex (P = 0.036) and the full-moon lunar phase (P = 0.001) were significantly related to shorter LOS.CONCLUSIONSSeason had no effect on mortality or LOS following aortic dissection repair, while patient age significantly increased the odds of death. The full-moon cycle appeared to reduce the odds of death, and the full-moon cycle, along with being male and requiring a concomitant cardiac procedure, was associated with shorter LOS.

Concepts: Regression analysis, Linear regression, Death, Econometrics, Multivariate adaptive regression splines, Aortic dissection, Aorta, Errors and residuals in statistics


Increased golf club head speed (CHS) has been shown to result in greater driving distances and is also correlated with golf handicap. The purpose of this study was to investigate the relationships between field-based measures of strength and power, and golf CHS, with a secondary aim to determine the reliability of the selected tests. A correlation design was used to assess the following variables; anthropometrics, squat jump height (SJ) and squat jump peak power (SJPP), unilateral countermovement jump heights (RLCMJ and LLCMJ), bilateral countermovement jump heights (CMJ), countermovement jump peak power (CMJPP), and seated (MBST) and rotational (MBRT) medicine ball throws. 48 male subjects participated in the study (age 20.1 ± 3.2 years, height 1.76m ± 0.07m, mass 72.8 kg ± 7.8, handicap 5.8 ± 2.2). Moderate significant correlations were reported between CHS and MBRT (r = 0.67; p < 0.01), MBST (r = 0.63; p < 0.01), CMJPP (r = 0.54; p < 0.01) and SJPP (r = 0.53; p < 0.01). Weak significant correlations (r = 0.3-0.5) were identified between club head speed and the other remaining variables excluding LLCMJ. Stepwise multiple regression analysis identified that the MBST and SJ were the greatest predictors of CHS, explaining 49% of the variance. Additionally the ICCs reported for tests of CHS and all performance variables were deemed acceptable (r = 0.7-0.9). The results of this study suggest that the strength and conditioning coach can accurately assess and monitor the physical abilities of golf athletes using the proposed battery of field tests. Additionally, movements that are more concentrically dominant in nature may display stronger relationships with CHS due to MBST and SJ displaying the highest explained variance following a stepwise linear regression.

Concepts: Regression analysis, Linear regression, Econometrics, Multivariate adaptive regression splines, Ordinary least squares, Errors and residuals in statistics, Stepwise regression, Golf


This study examined the relationship between ratings of perceived exertion (RPE, CR-10), heart rate (HR), peak blood lactate (Lapeak), and immersion (IM) parameters in 17 elite synchronized swimmers performing 30 solo and duet routines during competition. All were video recorded (50 Hz) and an observational instrument was used to time the IM phases. Differences in the measured variables were tested using a linear mixed-effects model. RPE was 7.7±1.1 and did not differ among routines, and neither did any of the HR parameters. There were differences among routines in Lapeak (F3,7=16.5; P=0.002), number of IM (F3,15=14.0; P<0.001), total time immersed (F3,16=26.6; P<0.001), percentage of time immersed (F3,13=6.5; P=0.007) and number of IM longer than 10 s (F3,19=3.0; P=0.04). RPE correlated positively to HR pre-activation, range of variation and recovery, IM parameters and Lapeak, and inversely to minimum and mean HR. A hierarchical multiple linear regression (MLR) model (number of IM > 10 s, HR recovery, minimum HR, and Lapeak) explained 62% RPE variance (adj. Rm 2=0.62; P<0.001). A stepwise MLR model (Lapeak, mean IM time and pre-exercise HR) explained 46% of performance variance (adj. Rm 2=0.46; P<0.001). Findings highlight the psycho-physical stress imposed by the combination of intense dynamic exercise with repeated and prolonged apnea intervals during SS events.

Concepts: Regression analysis, Linear regression, Variance, Econometrics, Multivariate adaptive regression splines, Parametric statistics, Embedding, Synchronized swimming


Outdoor decontamination efforts have been ongoing since the Fukushima Daiichi nuclear power plant (FDNPP) accident; however, little is known about indoor contamination. Therefore, house dust was sampled based on particle size in 21 wooden buildings (19 residential houses and 2 community centers) within the evacuation area close to the FDNPP, Fukushima Prefecture, Japan. Activities of radiocesium (137Cs) per gram of house dust increased with decreasing particle size (mean: 6.1 × 103, 2.6 × 103, 1.6 × 103, 7.5 × 102, 5.0 × 102, and 4.6 × 102 Bq/g for <4-20, 20-63, 63-180, 180-500, 500-1000, and 1000-2000 μm, respectively). The137Cs concentrations in house dust were inversely related to the square of distance from the FDNPP for <4-1000 μm particles. From the results of the multiple linear regression analysis, distance from the FDNPP and direction from the FDNPP (northwest) were significantly related to the radioactivity of house dust. It was found that 19%, 33%, and 48% of137Cs in house dust were extracted in water, 1 M HCl, and not extracted, respectively. Considering the bioaccessibility and assuming a 20 mg/day daily intake of house dust, the daily doses would be 7.2 Bq/day (mean) and 18 Bq/day (95th percent quantile). These results provide valuable insight into indoor radioactive Cs contamination in the area around the FDNPP and possible oral exposure to indoor radioactive Cs after returning home.

Concepts: Regression analysis, Linear regression, Econometrics, Uranium, Nuclear fission, Multivariate adaptive regression splines, Errors and residuals in statistics, Nuclear power


Appointment nonadherence is a health behavior that represents a burden to health care systems. On March 1, 2015, a new negative reinforcement intervention involving “service fees” for a visit without appointment was implemented at King Abdullah University Hospital in Jordan. To evaluate the effect of this intervention in improving patient adherence to medical appointment, a retrospective preintervention and postintervention analysis was used, including all patients (n = 65 535) who had scheduled appointments at 39 outpatient clinics. A repeated-measures analysis of variance was first performed. Then, a multivariate linear regression model was used to identify factors that might predict individuals who are likely to attend or miss their appointments and those who have a greater tendency to visit the hospital with or without appointments. Although the average percentage of appointments attended was more than missed preintervention and postintervention, the decrease in percentage of missed appointments was more pronounced postintervention. Also, the average percentage of visits without appointments was less than visits with appointments in both times, but the decrease in the percentage of visits without appointments was more prominent after. The regression analysis revealed that younger, married and male patients were more likely to miss their appointment before and after the intervention. Also, younger patients had a tendency to attend without appointments. Conversely, patients with the lower copayment rate had a tendency to adhere to appointment times. In conclusion, negative reinforcement interventions could improve patient appointment adherence rates. Accordingly, interventions designed that consider evidence and are theory-based are needed to change patient behavior.

Concepts: Regression analysis, Linear regression, Hospital, Econometrics, Multivariate adaptive regression splines, Ordinary least squares, Forecasting, Errors and residuals in statistics


Targeting Proviral integration-site of murine Moloney leukemia virus 1 kinase, hereafter called Pim-1 kinase, is a promising strategy for treating different kinds of human cancer. Headed for this a total list of 328 formerly reported Pim-1 kinase inhibitors has been explored and divided based on the pharmacophoric features of the most active molecules into 10 subsets projected to represent potential active binding manners accessible to ligands within the binding pocket of Pim-1 kinase. Discovery Studio 4.1 (DS 4.1) was employed to detect potential pharmacophoric active binding manners anticipated by Pim-1 Kinase inhibitors. The pharmacophoric models were then allowed to compete within Quantitative Structure Activity Relationship (QSAR) framework with other 2D descriptors. Accordingly Genetic algorithm and multiple linear regression investigation were engaged to find the finest QSAR equation that has the best predictive power r 262 (2)  = 0.70, F = 119.14, r LOO (2)  = 0.693, r PRESS (2) against 66 external test inhibitors = 0.71 q(2) = 0.55. Three different pharmacophores appeared in the successful QSAR equation this represents three different binding modes for inhibitors within the Pim-1 kinase binding pocket. Pharmacophoric models were later used to screen compounds within the National Cancer Institute database. Several low micromolar Pim-1 Kinase inhibitors were captured. The most potent hits show IC50 values of 0.77 and 1.03 µM. Also, upon analyzing the successful QSAR Equation we found that some polycyclic aromatic electron-rich structures namely 6-Chloro-2-methoxy-acridine can be considered as putative hits for Pim-1 kinase inhibition.

Concepts: Regression analysis, Linear regression, Econometrics, Inhibitor, Multivariate adaptive regression splines, National Cancer Institute, Medicinal chemistry, Murine leukemia virus


Rheumatoid arthritis (RA) is a bone destructive autoimmune disease. Many patients with RA recognize fluctuations of their joint synovitis according to changes of air pressure, but the correlations between them have never been addressed in large-scale association studies. To address this point we recruited large-scale assessments of RA activity in a Japanese population, and performed an association analysis. Here, a total of 23,064 assessments of RA activity from 2,131 patients were obtained from the KURAMA (Kyoto University Rheumatoid Arthritis Management Alliance) database. Detailed correlations between air pressure and joint swelling or tenderness were analyzed separately for each of the 326 patients with more than 20 assessments to regulate intra-patient correlations. Association studies were also performed for seven consecutive days to identify the strongest correlations. Standardized multiple linear regression analysis was performed to evaluate independent influences from other meteorological factors. As a result, components of composite measures for RA disease activity revealed suggestive negative associations with air pressure. The 326 patients displayed significant negative mean correlations between air pressure and swellings or the sum of swellings and tenderness (p = 0.00068 and 0.00011, respectively). Among the seven consecutive days, the most significant mean negative correlations were observed for air pressure three days before evaluations of RA synovitis (p = 1.7×10(-7), 0.00027, and 8.3×10(-8), for swellings, tenderness and the sum of them, respectively). Standardized multiple linear regression analysis revealed these associations were independent from humidity and temperature. Our findings suggest that air pressure is inversely associated with synovitis in patients with RA.

Concepts: Regression analysis, Linear regression, Econometrics, Rheumatoid arthritis, Multivariate adaptive regression splines, Ordinary least squares, Errors and residuals in statistics, Segmented regression


Misconceptions about the assumptions behind the standard linear regression model are widespread and dangerous. These lead to using linear regression when inappropriate, and to employing alternative procedures with less statistical power when unnecessary. Our systematic literature review investigated employment and reporting of assumption checks in twelve clinical psychology journals. Findings indicate that normality of the variables themselves, rather than of the errors, was wrongfully held for a necessary assumption in 4% of papers that use regression. Furthermore, 92% of all papers using linear regression were unclear about their assumption checks, violating APA-recommendations. This paper appeals for a heightened awareness for and increased transparency in the reporting of statistical assumption checking.

Concepts: Psychology, Regression analysis, Linear regression, Econometrics, Multivariate adaptive regression splines, Ordinary least squares, Errors and residuals in statistics, Censored regression model


We investigated the association between breastfeeding and cognitive development in infants during their first 3 years. The present study was a part of the Mothers' and Children’s Environmental Health (MOCEH) study, which was a multi-center birth cohort project in Korea that began in 2006. A total of 697 infants were tested at age 12, 24, and 36 months using the Korean version of the Bayley Scales of Infant Development II (K-BSID-II). The use and duration of breastfeeding and formula feeding were measured. The relationship between breastfeeding and the mental development index (MDI) score was analyzed by multiple linear regression analysis. The results indicated a positive correlation between breastfeeding duration and MDI score. After adjusting for covariates, infants who were breastfed for ≥ 9 months had significantly better cognitive development than those who had not been breastfed. These results suggest that the longer duration of breastfeeding improves cognitive development in infants.

Concepts: Regression analysis, Linear regression, Infant, Econometrics, Breastfeeding, Multivariate adaptive regression splines, Ordinary least squares, Child development


An understanding of past hydroclimatic variability is critical to resolving the significance of recent recorded trends in Australian precipitation and informing climate models. Our aim was to reconstruct past hydroclimatic variability in semi-arid northwest Australia to provide a longer context within which to examine a recent period of unusually high summer-autumn precipitation. We developed a 210-year ring-width chronology from Callitris columellaris, which was highly correlated with summer-autumn (Dec-May) precipitation (r = 0.81; 1910-2011; p < 0.0001) and autumn (Mar-May) self-calibrating Palmer drought severity index (scPDSI, r = 0.73; 1910-2011; p < 0.0001) across semi-arid northwest Australia. A linear regression model was used to reconstruct precipitation and explained 66% of the variance in observed summer-autumn precipitation. Our reconstruction reveals inter-annual to multi-decadal scale variation in hydroclimate of the region during the last 210 years, typically showing periods of below average precipitation extending from one to three decades and periods of above average precipitation, which were often less than a decade. Our results demonstrate that the last two decades (1995-2012) have been unusually wet (average summer-autumn precipitation of 310 mm) compared to the previous two centuries (average summer-autumn precipitation of 229 mm), coinciding with both an anomalously high frequency and intensity of tropical cyclones in northwest Australia and the dominance of the positive phase of the Southern Annular Mode.

Concepts: Regression analysis, Linear regression, Precipitation, Climate, Econometrics, Multivariate adaptive regression splines, Tropical cyclone, Decade